Selective Leviathans: Explaining State Strategies of Counterinsurgency and Consolidation MASSACHUSETTS INWTUME. by OF TECHNOLOGY OCT 0 e Sameer Prem Lalwani B.A. Political Science University of California, Berkeley, 2005 201 LIBRARIES SUBMITTED TO THE DEPARTMENT OF POLITICAL SCIENCE IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN POLITICAL SCIENCE AT THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY SEPTEMBER 2014 C 2014 Sameer P. Lalwani. All rights reserved. The author hereby grants to MIT permission to reproduce and to distribute publicly paper and electronic copies of this thesis document in whole or in part in any medium now known or hereafter created. Signature of Author: Signature redacted A epartment of Poiitical Science '-'September 5, 2014 Certified by: Signature redacted Roger D. Petersen Professor of Political Science Thesis Supervisor Signature redacted Accepted by Andrea Campbell Professor Political Science Chair, Graduate Program Committee Selective Leviathans: Explaining State Strategies of Counterinsurgency and Consolidation by Sameer Prem Lalwani Submitted to the Department of Political Science on September 5, 2014 in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Political Science ABSTRACT States respond to rebellion differently, often shaping the intensity, duration, and outcome of the conflict as well as prospects for stability and state consolidation. Puzzling cross-national variation in counterinsurgency-sometimes brutal, sometimes minimalist-has drawn some inquiry, but sub-national variation by the same state incumbent has largely been neglected. This study develops a core-periphery theory to explain why states choose different strategies to fight rebellions based on what is at stake and who is rebelling. The theory identifies the variation in state strategy along two dimensions of effort and violence and contends the sub-national variation can best be explained by the value of the contested territory and the identity of the rebelling group. First, the contested territory's value, in economic, strategic, and ideational terms, will shape the strategy of effort the state is willing to deploy. When a rebellion, regardless of its size, threatens core territory, the state is likely to employ a strategy involving significant effort to decisively defeat the rebellion and regain control. Peripheral territory offers comparatively lower incentives resulting in minimalist containment strategies. Second, the positional status of the identity group composing the rebel base will shape the state's incentives and constraints for the strategic use of violence. Core social identity groups with high group worth and embeddedness within the state will motivate greater restraint as coidentity serves to limit state violence through affective mechanisms of empathy and trust and strategic mechanisms of vulnerability and information. Peripheral rebel groups are less likely to trigger such restraint and are therefore met with strategies of greater violence. The theory is tested with 29 cases of rebellion from India and Pakistan to leverage within-country variation with macro-structures of state capacity and institutions fairly constant. The study draws on extensive fieldwork, including over 150 interviews, as well as primary and secondary sources to examine state counterinsurgency through a medium-N analysis for each country's set of campaigns, as well as structured, focused comparisons of a select number of cases along specific dimensions of these conflicts and strategies. The two approaches provide strong support for the theory relative to competing explanations. Thesis Supervisor: Title: Roger D. Petersen Professor of Political Science 3 ACKNOWLEDGEMENTS I first owe a debt of gratitude to my dissertation committee. Roger Petersen guided me through this process with excellent feedback, moral support, and encouragement to push to my theory into more creative directions. I have truly appreciated his good humor, humility, and commitment to his students. He is the kind of thoughtful scholar and generous mentor that I one day hope to be. It is because of Barry Posen that I first acquired an interest in military power, and have learned much from our conversations during his classes, Wednesday seminar lunches, and Security Studies Program (SSP) field trips. While I was drawn to the MIT program specifically because of his lucid and penetrating diagnosis of contemporary grand strategy challenges, I believe one of his greatest legacies may still be the SSP community he has fostered through his deep commitment to training the next generation of security scholars. For better or worse, Fotini Christia helped spark the initial idea for this project and was the person who encouraged me to develop it form a paper into a dissertation project. Along with this, she has provided an invaluable model for how to conduct good fieldwork and interviews. Vipin Narang was a mentor even before he became a member of my committee and was willing to stick with it even during his brief year away from the department. In addition to his constructive advice on this project, he has helped open doors for me to the South Asia research community and helped keep me honest in my empirical analysis. I have been truly blessed with a wonderful cohort of MIT colleagues, who made time to read drafts, listen to presentations, bounce around ideas, and offer constructive feedback on how to improve the project. I'd especially like to thank Dan Altman, Noel Anderson, Nathan Black, Mark Bell, Chris Clary, Brian Haggerty, Toby Harris, David Jae, Nick Miller, Miranda Priebe, Kai Quek, Andrew Radin, Josh Shifrinson, Joseph Torigian, David Weinberg, and Alec Worsnop for all their advice and support. Though they have not been subjected to my unpolished drafts or meandering presentations, I am also appreciative of past generations of MIT SSP scholars who have been both models of scholarship, readily volunteered their time to serve as chairs or discussants of conference panels, offered professional advice, and taken an active interest in mentoring students like myself. For that, my sincere thanks goes out to Greg Distlehorst, Kristin Fabbe, Keren Fraiman, Brendan Green, Phil Haun, Colin Jackson, Stephanie Kaplan, Peter Krause, Jon Lindsey, Austin Long, Josh Rovner, Paul Staniland, and Rachel Wellhausen. No matter how difficult the challenge, I always felt at home at MIT because of the support of people, who were generous with their time, always had the students' best interests in mind, and made my experience at MIT easier. I am especially grateful to Joli Divon Saraf, Lynne Levine, Harlene Miller, Valerie Thaddeus, Casey Johnson-Houlihan, Laurie Scheffler, Susan Twarog, Diana Gallagher, and Paula Kreutzer. I could not have asked for a better home away from home with the Institute for Security and Conflict Studies (ISCS) at George Washington University's Elliott School. Charlie Glaser was incredibly generous with his time and support and I profited immensely from the support, feedback, and discussions with many in the ISCS community including Rose Kelanic, Catherine Silvey, Stephen Biddle, Alex Downes, Harris Mylonas, Joanna Spear, Rachel Stein, Caitlin Talmadge, along with all the fellows whom I had the good fortune to share an office with for two 4 years. In addition to ISCS, I also received tremendous feedback from workshops hosted by the MIT Strategic Use of Force Working Group, the Tobin Project, the Harvard-MIT-Yale graduate student conference, the Triangle Institute for Security Studies, Syracuse University, and the University of Pennsylvania's Center for the Advanced Study of India, along with a number of ISA and APSA conference panels. The fieldwork I conducted in South Asia proved critical for my research and it would not have been possible without the help of a number of people. I benefitted from the support of Ramel Dahiya Rasul Baksh Rais, Sophiya Mirza, Surinder Oberoi, Basharat Peer, Anu Varghese, Krishnappa Venkatshamy as well as the Lahore University of Management Sciences and the Institute for Defence Studies and Analysis for hosting me. I am also grateful to colleagues and mentors who helped me prepare for fieldwork and set up meetings prior to my travel including Peter Bergen, Yelena Biberman, Christine Fair, Anatol Lieven, Sahar Khan, Haider Mullick, Dick Samuels, Shuja Nawaz, Josh White, and Moeed Yusuf. This fieldwork and dissertation research would also not have been possible without the generous financial support from the MIT Entrepreneurship Center's Carol L. Wilson Award, the MIT Center for International Studies, the Smith Richardson Foundation, the Tobin Project, the American Institute for Pakistan Studies, and the Institute for Security and Conflict Studies. While writing this dissertation, I have also had the good fortune of befriending a number of colleagues outside of the MIT community, who have been also been instrumental to my intellectual growth including Andrew Bell, Ahsan Butt, Austin Carson, Erica DeBruin, Jeff Friedman, Inwook Kim, Adam Lebovitz, Julia Macdonald, Sara Moller, Lindsey O'Rourke, Katherine Tiedemann, Jane Vaynman, Tristan Volpe, and Jon Wallace. Along the way, I also learned and drew inspiration from a number of other experiences and opportunities outside of academia and benefitted from the advice, guidance, and support of mentors including Steve Clemons, Dan Egel, Terry Kelly, Olya Oliker, Sherle Schwenninger, and Steve Watts. I'm particularly grateful to a set of people outside of the world of political science who were removed enough to provide doses of perspective, but close enough to empathize and feel invested in my endeavors. To Matt Cordell, David Henderson, Ben Katcher, Jenna Levine, Jeff Meyer, Mary and Pat Moore, Suzannah Phillips, Gaurav Reddy, Erin Renner, Aaron Scheinwald, Sam Sherraden, Karthik Srinivasan, Elliott Wimmer, and Yao Wu, I am sincerely grateful for your friendship and support, without which I would not have been able to make it this far. The greatest thanks belongs to my family. Alley, Jane, Patrick (and Desmond) Sawran generously opened their homes to me the past few years, supported my nomadic travels between Boston and DC with room and board and copious amounts of coffee in the morning, and celebrated my triumphs as if they were their own. I cannot even begin to thank my wife, Nicole, for all the years she has supported me. Since the beginning of my time in this program, she has been a true partner in this endeavor. She has borne the 5am drop-offs at the airport, the absence during months of field work, the sculpted piles of paper that have become fixtures on the dining table, the endless library renewal runs, the books that collect their weight in dust, and my moments of frustration and self-doubt, all with immense 5 patience, grace, and love. For all the years she has selflessly been my sounding board, copy editor, confidante, ally and advocate, I cannot hope to repay her, but will spend a lifetime trying. Finally, I owe everything to my original mentors, my parents, for every sacrifice they made and every opportunity they have given me. I would not be the person I am today without their love and gifts of intellectual curiosity, a strong work ethic, and desire to contribute something meaningful. I have always known that I am standing on the shoulders of giants, and I aspire to one day be worthy of that perch. 6 Table of Contents Abstract p. 3 Acknowledgements p. 4 Table of Contents p. 7 Chapter 1: Introduction p. 8 Chapter 2: Theorizing A Typology of Counterinsurgency Strategy p. 29 Chapter 3: Explaining Strategy with a Theory of Core-Periphery Relations p. 74 Chapter 4: A Medium-N Analysis of Pakistan's Counterinsurgency Campaigns p. 128 Chapter 5: Core-periphery Mechanisms in East Pakistan, FATA, and Malakand p. 248 Chapter 6: A Medium-N Analysis of India's Counterinsurgency Campaigns p. 327 Chapter 7: Core-periphery Mechanisms in Punjab, Kashmir, and the Naxal Belt p. 451 Chapter 8: Conclusion p. 572 Bibliography p. 622 7 Chapter 1 Introduction I. The Puzzle In the early 1990s, the flames of rebellion engulfed India. Facing multiple distinct rebellions, the Indian state battled for its life. Yet the same state, with a lengthy history of democratic institutions governing a country that was primed for takeoff, fought back the rebels in a selective, divergent manner. In Kashmir, the Indian military unleashed a massive and brutal campaign to violently crush a separatist insurgency and punish its civilian base. By contrast, the government overhauled its police force and intelligence capacity in Punjab to systematically target rebels while working to rehabilitate them and return separatist leaders back into the political fold. In Manipur, the state maintained a phantom presence to battle separatist insurgents through a dirty war of proxies and assassinations, intensifying violence and disorder. Meanwhile, the Indian state casually ignored a Maoist rebellion, offering meager central support to a struggling Andhra state government that cycled between appeasement and occasional police actions. Two decades later, India's archrival Pakistan was now on the hot seat facing multiple insurgencies even while bolstered with new injections of capacity from its American ally. Like its rival, the Pakistani state also displayed a penchant for selectivity in battling rebellion. The Pakistani Army pummeled insurgents and civilians indiscriminately with air power and artillery as it moved in multiple divisions to take over a section of South Waziristan, displacing hundreds of thousands of people who would never return. By contrast, in Malakand, the military had taken great pains to dislodge insurgents, return millions of people their homes, help police the region, de-radicalize former rebels, and implement judicial reforms. Further to the west, the state 8 interspersed aerial bombing of Balochistan rebels while prosecuting a dirty war from the shadows with para-statal agents, forced disappearances, and dumped bodies. And in much of North Waziristan, the state primarily coopted local warlords to its cause while reluctantly battling rebels sporadically. Baffled by Pakistani erratic behavior, one American national security advisor remarked, "there is appeasement one day, confrontation another day, and direction a third day."i Though quite different in their state capacity, degree of democratic institutions, and strategic culture, India and Pakistan both displayed a curious pattern: the state over the same period of time of deploying divergent approaches to combat rebellion. All at once, the state leviathan revealed four different shades of itself-brutal, sensitive, aggressive, and reluctant. This project fundamentally seeks to explain the puzzling variation in state strategies deployed to counter insurgency. II. Introduction India and Pakistan are not isolated cases. Around the world, states exhibit puzzling behavior in civil wars by choosing brutal or limited strategies (and sometimes both) to fight rebellion. One puzzling empirical pattern is how often states seem to defy normative, political, or strategic incentives and choose costly, heavy-handed strategies producing high levels of violence and civilian casualties. Recent work finds that only eleven percent of 20th century counterinsurgency campaigns sought to protect civilians 2 and even incumbents that provided the model for Western counterinsurgency doctrine regularly departed from strategies of minimal 1Bob Woodward, Obama's Wars, New York: Simon and Schuster, 2010, p. 286. 2Patrick Johnston, "The Treatment of Civilians in Effective Counterinsurgency Operations," PhD Dissertation, Northwestern University, December 2009. 9 force. 3 One need look no further than cases of the Sudanese, Sri Lankan, and Syrian states' recent tilts with insurgencies to recognize the frequency of this phenomena even in the 21st century as well as the high risks and costs of such ventures. A separate but related puzzling pattern of behavior is the seemingly casual, haphazard, and limited manner with which states sometimes go about fighting rebellion, evidenced by the high number of civil conflicts left as "draws" or with enduring "low-activity," particularly in Asia.4 While they have crushed others, countries such as the Philippines, Indonesia, and Myanmar have applied limited effort only to contain some insurgencies for decades rather than defeat them outright. Even strong states like Russia and China, have never summoned the requisite strength they possesses to root out decades-long insurgencies that persist on their frontiers. While outcomes are quite difficult to define, states will often persist with these costly, seemingly losing strategies. Why then do states choose brutal or minimalist strategiesto fight rebellion? Understanding the strategic logic of these incumbents in civil war-like when they escalate and when they under react-is essential for policymakers to gauge destabilizing strategies, stem spillovers, and restructure incentives to mitigate violence. The fact is we simply do not know much about what motivates state counterinsurgency strategies other than some general structural features endemic to most civil wars. Security scholars have studied the sources of conventional military strategy to understand the choices 3David French, The British Way in Counter-Insurgency, 1945-1967, New York: Oxford University Press, 2011. 4 Staniland observes this of Lyall and Wilson III. See Paul Staniland, "States, Insurgents, and Wartime Political Orders,"Perspectives on Politics, 10 (2), June 2012, pp. 243-264; Jason Lyall and Isaiah Wilson III, "Rage Against the Machines: Explaining Outcomes in Counterinsurgency Wars," InternationalOrganization63 (1), 2009, pp. 67106; James Fearon "Why Do Some Civil Wars Last So Much Longer Than Others," JournalofPeace Research, 41 (3), May 2004, pp. 275-302; David E. Cunningham, Kristian Skrede Gleditsch, and Idean Salehyan, "It Takes Two: A Dyadic Analysis of Civil War Duration and Outcome," Journalof Conflict Resolution, 53 (4), August 2009, pp. 570- 597. 10 states make in inter-state wars and their subsequent effects on stability and outcomes. Little work, however, has sought to explain domestic incumbents' strategies within civil war. While conflict is inherently a strategic interaction between at least two parties, in asymmetric conflicts, the state incumbent's strategy plays a particularly important, if not dominant, role in shaping conflict dynamics, processes, and outcomes. My research seeks to explain the variation in state strategies of fighting rebellion. I argue that strategy is shaped by a core-periphery relationship prior to rebellion-that is between the state's geopolitical core and the contested region, as well as between the identity of state elites and the identity group rebelling. This argument challenges conventional wisdom on the way states fight civil wars. We have come to view incumbent strategy as a function of structural settings-a state's capacity, its institutions, or its strategic culture. These functionalist explanations, while useful, can only predict super-aggregated phenomena, and are not designed to explain variation below this level. Consequently we might be missing explanatory accounts and causal mechanisms even for countries that do not exhibit the same type of stark sub-national variation as India and Pakistan. Civil wars, even more than inter-state wars, are politics by other means and pre-war political incentive structures shape the choices that states make within these conflicts. These choices cannot be reduced to those of a single individual (i.e. great man theories), which are often historically inaccurate, "maddeningly incomplete" and ignore the environmental limitations 5 John J. Mearsheimer, ConventionalDeterrence,Ithaca: Cornell University Press, 1985; Barry R. Posen, The Sources of Military Doctrine:France, Britain, and Germany Between the World Wars, Ithaca: Cornell University Press, 1984; Robert A. Pape, Bombing to Win: Air Power and Coercionin War, Ithaca: Cornell University Press, 1996; Stephen D. Biddle, MilitaryPower: Explaining Victory andDefeat in Modern Battle, (Princeton: Princeton University Press, 2004. 11 of the individual. 6Nor can we pretend that states are abstract, dispassionate, neutral entities operating behind a veil of ignorance or maximizing utility for the greatest good. Between these micro- and the macro- levels resides the meso-level of organizations and sub-groups, which have proven to be increasingly important in civil wars scholarship. 7 These groups are often bound by identity and inhabit and shape structures like states, sometimes in collaboration with other groups. In other words, meso-level sub-groups often make "the state" their meso-level organization. These identity groups have collective interests, preferences, biases, and capabilities, all of which they channel into the state. And two aspects of politics that feature prominently in these groups' decisions are what is at stake and who is involved. Translated to the realm of civil wars, the choices states make hinge on the material importance of the contested territory and the identity of the combatants. 111. The Argument What then might explain varied state strategies to different cases of rebellion? I develop a theory based on a country's core-periphery relations-that is a state's relative value of territory and the positionalstatus of the rebellingidentity group interact to produce distinct strategies. The territory shapes the state's estimation of what is at stake, and consequently how much effort it should deploy. The identity group shapes the state's perceptions of what type of force is required or permissible, and consequently how much violence (or substitutes for violence) the state should employ. 6 Colin Jackson, "Defeat in Victory: Organizational Learning Dysfunction in Counterinsurgency," PhD Dissertation, Massachusetts Institute of Technology, June 2008, p. 9; Joshua Rovner, "The Heroes of COIN," Orbis, 56 (2), Spring 2012, pp. 215-232. Paul Staniland, Networks ofRebellion: Explaining InsurgenctCohesion and Collapse, Ithaca: Cornell University Press, 2014; Harris Mylonas, The Politics ofNation-Building:Making Co-Nationals, Refugees, and Minorities, New York: Cambridge University Pres, 2013; Fotini Christia, Alliance Formation in Civil Wars, New York: Cambridge University Press, 2012. 12 Territory. Rebellions often begin by concentrating around a particularly territory or region. This may the source of the dispute for a separatist rebellion, but even if they are centerseeking rebels that expand their coverage, they almost always begin in one place and retain a base where they are most concentrated or operationally effective. This territory then becomes the site of violent contestation against the state incumbent. The type of territory-core or peripheral-will shape the material stakes and the incentives for the state's deployment of effort. A core territory possesses significant material, strategic, or ideational value, in addition to physical and social infrastructure that ultimately lowers the costs of maintaining incumbent control. Consequently a state incumbent facing a rebellion that threatens its core territory would be likely to employ a strategy involving a high degree of effort to decisively defeat the rebellion and regain control of the contested territory. A peripheral territory offers comparatively lower material, strategic, or ideational incentives and, lacking physical and social infrastructure, raises the costs of (re)establishing control. As a result, a state incumbent facing a rebellion that threatens its peripheral territory would be likely to employ a low degree of effort to contain the rebellion below a certain level, perhaps enough to access the territory. States not only have choices about whether to fight a rebel group but also how they fight, and specifically, how hard they fight in terms of effort. The state may possess plenty of incentives to choose to fight rebels, yet if the rebels contest or operate on peripheral territories, this does not provide sufficient incentive for the state to deploy a high-effort strategy of decisive control. The costliness of sustaining a lengthy, resource-intensive campaign and building up a dominant state presence to control peripheries can be prohibitive. A state may balk at such a costly venture with low returns, conserve its resources, and deploy a minimalist strategy that 13 merely limits conflict intensity below a certain threshold. On the other hand, high incentives and lower costs can compel a high effort strategy. Identity. Most rebellions are built from a particular identity group or social base within which insurgent entrepreneurs are embedded. These groups often share both grievances motivating them as well as social properties that enable them to organize and engage in collective action. Though they may vary in density and type, a rebel group generally has some shared identity, often ethnicity, as it is an extremely powerful organizing tool. For instance, even the ideological Maoist rebellions in Nepal and India are constructed mostly around particular ethnic groups rather than just a class.8 The identity of the rebels and their organizational base can play a pivotal role in the state's response. The positional status of an identity group-whether it is a core or peripheral group-will shape the state's incentives and constraints for its employment of violence. A group with high position and proximate to the state's core "in-group" possesses both moral worth in the eyes of the dominant group(s) and embeddedness within important state institutions, principally the security forces and bureaucracy. When a state incumbent faces a rebellion built from a core identity group, the incumbent will be more likely to restrain its use of violence and employ strategies that involve more selective targeting and tactical substitutes for violence. The rebel group's identity serves to restrain the state's use of violence through affective mechanisms of empathy and trust and strategic mechanisms of vulnerability and information. State officials advising or leading the counterinsurgency campaign are more likely to empathize with rebel group or base, even if they are minorities, rendering the state sensitive to the groups' grievances, concerned about indiscriminate violence, and willing to attributing actions to S. Mansoob Murshed and Scott Gates, "Spatial-Horizontal Inequality and the Maoist Insurgency in Nepal," Review ofDevelopment Economics, 9 (1), 2005, pp. 121-134. 8 14 circumstance rather than group character. If security forces are composed of high numbers from the rebels' identity group, the incumbent will be more vulnerable to backlashes-like desertion or mutiny-that pose strategic threats to cohesion. Group embeddedness within state institutions can also offer the state greater information for more discriminate targeting. Both worth and embeddedness can facilitate greater trust for a state to address the "demand-side" of rebellion and substitute credible bargaining, public goods, and redress of grievances for high violence within a military campaign. By contrast, a group with low position, a peripheral group, possesses neither moral worth nor embeddedness within state institutions. When a state incumbent faces a rebellion built from such a peripheral group, it will be unrestrained from choosing a strategy of high violence, which involves brute force and indiscriminate targeting. In this case the rebel group's identity is too distant or worthless to restrain the state through any of the four mechanisms described above. Instead, the state will be insensitive to the group's grievances or backlash, unconcerned about collateral damage, more willing to attribute rebel actions to disloyal character, lacking information for discriminate targeting, and harboring distrust that prevents any substitutes for violence. Strategies of Counterinsurgency The interaction of contested territory and rebel group identity leads the state to choose from four ideal-type counterinsurgency strategies with different combinations of effort and violence and distinct theories of victory: Attrition (I). States facing peripheral groups contesting core territory are likely to choose high-effort, high-violence strategies of attrition. This approach targets the supply-side of 15 rebellion and focuses on degrading, disarming, and punishing the rebels and their base to destroy their combat power and capacity to resist the state. The state's return on investment is linear and Figure 1.1: Explaining Variation in Counterinsurgency Strategy Contested Territory (Value) Periphery Core (Low) (High) Enfeeblement [III] Attrition [I] Peripheral (low effort, high (high effort, high Positional Status of (Low) violence) violence) Rebel Identity GrouD Population Control [II] Mitigation [IV] Core (high effort, low low effort, (low (High) violence) violence) requires continuous elimination of insurgents until the rebellion is broken. This strategy is often accompanied by the center's usurpation of political authority and rewiring the circuits of local politics for total control. Pakistan attempted this strategy in East Pakistan though it eventually failed due to external intervention, but India has successfully applied this strategy in Kashmir, and though it is not covered in this study, it appears Sri Lanka succeeded in doing the same in its region of Tamil Eelam. Population Control (II). States facing core groups contesting core territory are likely to adopt high-effort, low-violence strategies of population control. This strategy seeks victory through both denial and inducement by investing significant resources to control the population and gradually deplete the insurgency through more intelligence-driven, targeted violence, but also addressing the demand-side of rebellion with inducements for surrenders, political accommodations, and rehabilitation of the rebel base. The most notable signatures of this approach are small, dispersed units policing areas while gathering information for targeted strikes or arrests of insurgents. The returns on investment in this strategy take time to yield but 16 generate increasing returns over time. This strategy involves devolution and power sharing of political authority between the center and locals. Some examples of this strategy in the following chapters include Pakistan's second campaign in Malakand and India's campaign in Punjab, both of which have been considered fairly successful. Enfeeblement (III). States facing peripheral groups in peripheral territory are likely to field low-effort, high-violence strategies of enfeeblement. The strategy posits victory can be achieved with a light-footprint approach by periodically disrupting, interdicting, and deterring rebels so that they are never eliminated but restricted in their operational abilities. Common tactics for this strategy may include the use of periodic punitive airstrikes, leadership decapitation, and/or shadowy intelligence operatives or special forces to punish an insurgent's base, target its network, or cripple its operational capabilities for a period of time. While also a supply-side strategy focused on kinetic methods of dealing with insurgents, the strategy anticipates diminishing returns over time and focuses on containing insurgency below a certain threshold. This strategy is often accompanied by an aloof form of political authority termed administrative occupation, in which the center controls the region and its politics remotely, often through tools of coercion, subversion, and manipulation, and with little interest other than managing violence. Pakistan has generally employed this strategy in Balochistan while India has done something similar against the rebellions of its Northeast. Though not considered model approaches, as both states have been engaged in fighting rebellion in these regions on and off for almost sixty years each, both states have managed to retain a hold over these territories. Mitigation (IV). Finally, states facing core groups in peripheral territory are likely to select low-effort, low-violence strategies of mitigation. Similar to enfeeblement, this strategy involves an even lighter footprint to contain a rebellion through tactical denial and often with 17 some cooptation. In these cases, the state also tries to manage violence, which can never be eliminated but only kept below a certain threshold. Incumbents can mitigate violence by addressing some aspects of the demand-side of rebellion, though often the demands are those of a narrow set of local power brokers in the region. These warlords, tribal leaders, or militias may be part of the rebellion or possess the capacity to help contain it, and while the state periodically has to inflict violence, it generally can rely on proxies or non-kinetic tools. Given the aloofness of the approach, this strategy is accompanied by an effective abdication of the center's political authority that devolves defacto control to local, often informal authorities. Pakistan has done this with large parts of its tribal areas in the northwest part of its country, while India has done roughly the same in the tribal areas of Central India, sometimes described as its Naxal belt. IV. Methods and Data In many respects, strategy is inherently endogenous to the varied contingencies of conflict, but there is still significant value to identifying historically rooted, pre-war economic and political factors that shape strategy below the macro-level structure but above the choices of individuals. These theories can improve our understanding of conflict dynamics by understanding what filters and shapes the strategic choices of meso-level organizations in conflict. To tackle this, I primarily rely on qualitative comparative case study method that utilizes process tracing, longitudinal analysis and most importantly structured, focused comparisons, tightly matched to control for a number of confounding variables. This methodological approach allows for detailed, fine-grained historical research necessary to accurately measure a complex and multi-faceted dependent variable like strategy as well as the variables and mechanisms of my theory. 18 I employ two empirical strategies for studying and testing my theory of core periphery relations. First I conduct two within-country, medium-N analyses of counterinsurgency campaigns that generate enough variation to test my theory, control for other variables, and still maintain close attention to concept validity and causal mechanisms. I examine 14 cases of counterinsurgency conducted by the Indian state, and separately, 15 cases of counterinsurgency conducted by the Pakistani state to compare strategy across multiple campaigns by the same incumbent. These analyses rely on extensive historical research and quantitative data to first identify the cases (many of which are not well-known counterinsurgency campaigns) and measure the values of the independent and dependent variables. The medium-N analysis generates enough observations to identify discernable patterns, establish the theory's external validity beyond a narrow set of cases, and evaluate competing theories, particularly those only testable over time such as reputational interests and learning. Second, I select a set of cases-tightly overlapping in time-from each country for more structured, focused comparisons while holding constant regime type, state strength, organizational culture, historical institutions, the international system, and world time. These indepth case studies also allow me to process trace my theory's hypothesized mechanisms and eventually relax my controls to evaluate the relative importance of macro-structures in these cases. Data on the state's selection of counterinsurgency strategy (and its components of effort and violence), the value of territory, and position of the identity group is drawn from historical and journalistic accounts, published interviews, memoirs, government records, service journals, doctrinal works, and quantitative data collected from numerous sources. Furthermore, the arguments and evidence presented in this project are underpinned by more than five months of 19 field and archival research in Lahore, Islamabad, Delhi, Mumbai, Srinagar, and London contributing to over 140 interviews, roughly half with current or retired officials, military officers, intelligence analysts, civil servants, and politicians and the rest with knowledgeable scholars, journalists, and civil-society leaders. The tight and continuous interaction of the two countries offered synergistic value during fieldwork and historical research, where an analyst or official's account from one country could contribute insights and information on strategic choices of the other. This dissertation offers the most systematic and comprehensive comparative study of counterinsurgency campaigns within these two countries compiled to date. It threads a path between macro-structure and micro-choices that has typified most of the limited research on these countries' civil conflicts and examines historical cases to balance a recency bias, particularly endemic to current research on Pakistan. The study seeks to combine finely-detailed historical accounts, systematic quantitative data, rigorous comparisons, and a novel theory to contribute to the field's understanding of strategy and civil war. V. Contributions In addition to explaining the dependent variable of interest, the dissertation offers a number of novel theoretical contributions to the study of civil war. First, it distinguishes two dimensions of strategy often conflated as one in the civil conflict literature. Much of the literature combines the dimensions of effort and violence, wherein "fighting hard" and "fighting violently" are often considered the same thing or at least co-varying. The argument theoretically unpacks counterinsurgency strategy by drawing from the conventional strategy literature and demonstrates empirically that these dimensions are in fact distinct. By conceptualizing this novel 20 dimension of effort, the argument advances our understanding of the range of choices states may have. Measuring this dimension challenges a certain state strength determinism that assumes increasing state capacity necessarily translates into greater effort. In fact, the empirical sections of this project reveal time and again, that structure is not deterministic; states can select not to expend effort even if they possess the requisite capacity. War making need not involve state building. Instead states routinely husband "spare capacity" for other potential uses or contingencies. Second, by unpacking core and peripheral territory, this study also poses a theoretical challenge to the very concept of state strength or capacity. Contemporary states, many of which were hastily grafted on to territories carved out after the post-WWII/post-colonial era, cannot be internal security-maximizing but instead are internal-security satisficing. A weak state's (or strong state's) weakness may be unevenly distributed. These pockets of weakness are in the seams between the Westphalian state and the Weberian state, and when large enough, we call it a failed state. 9 But for the majority of states well short of failure, their strength or weakness depends on the region of interest. In international relations, this may matter less because strength can be aggregated at the center, but in the study of civil wars, this understanding is essential for state strength to have any meaning. Third, my theory explains how states make strategic choices within a world of bounded rationality. States exercise agency below the level of structure but they do not necessarily possess the capacity for perfect calibration to strategic dynamics. Theories of strategic interaction-such as insurgent strength and goals, reputational stakes, or external backers of rebellion-may be useful in explaining broader dynamics of a conflict (such as duration and 9 This insight drawn from work by Adnan A. Naseemullah, "Foundations of the Frontier: Norms, Resources and Limited Sovereignty in Pakistan's Tribal Areas," Unpublished Paper, 2011. 21 intensity) ex post, but they are not easily apprehended by a state prior to or in the midst of a conflict. Instead, states formulate strategy with what they have and rely heavily on pre-war assumptions and assessments of territory, as well as pre-war sources of knowledge and biases. These filters limit the seamless updating of information and render state incumbents less sensitive to these fine-grained details and shifts than otherwise thought. Fourth, my theory clarifies the mechanisms that link a state's domestic polity to its strategy, particularly its use of violence and restraint. Identity has made a recent comeback in the civil conflict literature after a serious of large-N studies that discounted its role. However, studies of ethnicity in conflict still underspecified the mechanisms by which it might impact violence. We have learned that attribution biases abound in interactions between foreign forces and local identity groups, and that co-ethnics somehow "improve" counterinsurgency outcomes and reduce violence, but it was unclear how these phenomena occurred. Combing insights from social identity theory and ethnic politics, this project articulates and empirically demonstrates how preexisting mechanisms of empathy, trust, vulnerability and information serve as the transmission belt between the incumbent's comprehension of the conflict, tactical choices, and levels of violence within a conflict. These mechanisms are robustly present even under very different regime types suggesting that they are not dependent on or by-products of institutions. Furthermore, all these mechanisms are powerful because they have purchase at different levels of analysis. The expenditure of effort and the restraint on violence has been studied primarily at the strategic level but it can also explain choices at the tactical level. Groups can, on average, harbor collective feelings of empathy and vulnerability, and access tools of information and trust, but these mechanisms can also be exhibited in interactions between local commanders or individual soldiers with individual rebels. 22 Fifth, the argument identifies to some limits to the relationship between violence and control at the meso-level. It first reveals how control is a choice endogenous to strategy, and different choices still possess both plausible theories of victory and are empirically valid. More importantly, it demonstrates that control alone does not determine the type of violence an incumbent will wield. Incumbents choosing low effort strategies of containment rather than control can do this more or less violently. Even with area dominance and a local preponderance of power, a state incumbent can still exercise more or less violence depending on its affective preferences and available tools to exercise selective violence or introduce substitutes. Finally, this study theorizes a dependent variable that is, along with its components, a major explanatory variable for other closely-scrutinized conflict phenomena including outcomes, termination, duration, settlement durability, and insurgent fragmentation. This study in particular distinguishes between counterinsurgency processes and outcomes. Much of the counterinsurgency literature conflates process with outcomes, or simply fixates on outcomes by ignoring broad strategic (as opposed to tactical) choices. Instead, this study shifts the emphasis back to what an incumbent chooses to do to achieve its theory of victory, mindful of the fact that it might result in what outside observers regard as a "loss" or a "draw." Theorizing some conflict processes as deliberate choices or strategies helps us to reevaluate other variables of interest such as outcomes. Draws may in fact be victories for a state-so long as it doesn't lose, it wins. In this sense, strategies of containment-whether enfeeblement or mitigation-may succeed in their endeavors when resulting in a "draw," while strategies for control, even ones abiding by counterinsurgency "best practices" may fail or fall short. Empirical Contributions. In addition, this study offers two unique empirical contributions. First, it contributes to our broader knowledge of domestic counterinsurgents. Most 23 of the literature (estimated at 75 percent of existing counterinsurgency studies') that examines the nuts and bolts of military strategy has been confined to Western and third-party counterinsurgents. Few works have bothered to explore the challenges and different approaches of domestic incumbents. This is particularly problematic given that policymakers prescribe and try to develop capacity for "best practices" amongst domestic incumbents based on this narrow set of studies. My research and argument helps to identify challenges that are not necessarily unique to domestic counterinsurgents, but uniquely revealed by their cases because they have far less opportunity to select out of a conflict. Second, this study offers a major empirical contribution to the study of South Asia by providing a systematic look at India and Pakistan's histories of counterinsurgency. While some studies and edited volumes have tried to produce political-military histories of various Indian cases or sets of them, few if any involve systematic measurement and comparison, let alone an inclusion of the full set of cases. The shortfall is even more pronounced in the study of Pakistan's counterinsurgency campaigns. Despite a common meme in South Asia studies to juxtapose India and Pakistan due to their divergent macro-structures, it is striking how similar patterns of coreperiphery relations have shaped their respective counterinsurgency strategies and patterns of conflict." It suggests that their shared colonial legacy of the British Raj-both its treatment of 1 An examination of the policy-oriented Journalof Strategic Studies illustrates this point. Between the early 1990s and the writing of this report, the journal published 82 articles on counterinsurgency and counterter- rorism; 93 percent of them contained case studies, but only 26 percent contained case studies of non-Western domestic incumbents. A similar pattern is found in International Security, where only 21 percent (nine articles) of the 43 articles since 1996 that mention counterinsurgency and deal with the subject of state response to violent non-state actors include a domestic-incumbent case study, while 65 percent (28 articles) focus on campaigns by foreign interveners. Excluding Israel and the Soviet Union/Russia, that figure falls to 12 percent (ten articles). The result is a limited understanding of the way much of the world strat- egizes for and fights insurgencies. "1 Paul Staniland, "Explaining Civil-Military Relations in Complex Political Environments: India and Pakistan in Comparative Perspective," Security Studies, 17 (2), 2008, pp. 322-362; Maya Tudor, The Promise ofPower: The OriginsofDmeocracy in India and Autocracy in Pakistan,New York: Cambridge University Press, 2013. This is also implicit in Aqil Shah, The Army and Democracy: Military Politics in Pakistan, Cambridge: Harvard University Press, 2014; T.V. Paul, The WarriorState: Pakistanin the ContemporaryWorld, New York: Oxford University 24 peripheral zones as well as favoring minority groups as martial races like the Pashtuns and Sikhs-remains a powerful contemporary political force in South Asia, which molds Indian and Pakistani state behavior to be far similar than generally acknowledged. Policy Contributions This project offers a number of policy contributions. First, the clear implication for scholars of counterinsurgency is that there are multiple counterinsurgency strategies that offer potential paths to victory, depending on how victory is defined. The strategy of population control, which is the closest match to the Western ideal type approach, is rarely practiced. Because the states most likely to experience rebellion are those that are weak and face aggrieved, often excluded, groups, this makes population control the least likely strategy for an incumbent. Therefore, international observers and policymakers should not expect emulation or adaptation to the "counterinsurgency ideal" is unlikely to happen anytime soon. Second, international assistance to improve an incumbent's capacity, professionalism, or institutions in an effort to expedite a domestic incumbents' convergence to counterinsurgency best practices is unlikely to succeed in the short to medium term, or frankly even the long-term. The United States has been at the forefront of trying to mold other states' capacity and political institutions, particularly with polices to support foreign internal defense.' 2 However, these are bound to face serious problems because these even if they manage to produce some structural Press, 2014. A notable exception to this is Anatol Lieven, Pakistan:A Hard Country, New York: Public Affairs, 2011, particularly chapter 1. 12 Robert M. Gates, "Helping Others Defend Themselves: The Future of U.S. Security Assistance," ForeignAffairs, 89 (3), May/June, 2010, pp. 2-6; Daniel L. Byman, "Friends Like These: Counterinsurgency and the War on Terrorism," InternationalSecurity, 31 (2), 2006, pp. 79-115; Stephen Watts, Jason H. Campbell, Patrick B. Johnston, Sameer Lalwani, Sarah H. Bana, CounteringOthers'Insurgencies:Undestanding U.S. Small-Footprint Intrvention in Local Context, Santa Monica, CA: RAND, 2014. 25 changes, a state will still be guided by the far less malleable incentives of its territory and identity, which take decades to change. Furthermore, rather than being exhausted to the point of reaching "hurting stalemates," in some undervalued regions, states are more likely to deploy a minimalist approach that merely contains rebellion below a certain threshold, sometimes yielding a steady-state of chronic, lowlevel warfare that policymakers term "ungoverned spaces." This novel, yet disturbing finding belies the commonly held assumption that civil war is the most costly enterprise states seek to avoid or quickly end, when in fact monopolization of violence is the far more daunting task. International actors may need to revise their approach to stabilization and institution-building and work around or within these "violence management" strategies.' 3 Finally, I find that democracy alone is simply not sufficient to restrain states from unleashing high levels of violence. India in 1990, Pakistan in 1973 and 2010, and even Sri Lanka in 2006 were all nominally democracies (and rated as such by the commonly used Polity index) when they unleashed brutal attrition campaigns involving high levels of violence in Kashmir, Balochistan, South Waziristan, and Tamil Eelam respectively. Policymakers interested in institutional design to install limits on brutal strategies like these need to pay more attention to the more enduring operational arms of the state that exercise considerable influence over strategy. As much as representative decision making institutions are important, more attention needs to be paid to the composition of military forces and civil and security bureaucracies. V. Roadmap of the Dissertation The rest of the dissertation provides a more detailed explanation of the theory and the empirical evidence to support the arguments I have previewed in this chapter. Chapter 2 13 Staniland, 2012 26 theorizes a typology of counterinsurgencystrategy to identify what values the dependent variable can take on and how to measure it. To do this, it draws from historical research on counterinsurgency to discern patterns of variation and then draws from theoretical work on counterinsurgency, conventional warfare, and military doctrine to conceptualize the dimensions and ideal types of counterinsurgency strategy. It then clarifies definitions, scope conditions, and my approach to measuring strategy in qualitative and quantitative terms. Chapter 3 offers a theory of core-periphery relations to explain why states vary their counterinsurgency strategy in different campaigns and/or against different adversaries. By developing a theory based on a state's core and peripheral territoryas well as its core and peripheral identity groups, I make observable, testable predictions on the dependent variable. I also specify mechanisms to be observed in the course of a counterinsurgency campaign that are associated with particular strategies. I then describe my research design in evaluating 29 campaigns from South Asia in two medium-N sub-national analyses and closely examining seven/eight of those cases to observe the presence of hypothesized mechanisms. Chapter 4 offers a detailed, medium-N analysis of 14 counterinsurgency campaigns conducted by the Government of Pakistan (GOP) from 1947 - 2013. It first measures the value that the Pakistani state placed on different territories and identity groups to generate predictions of their counterinsurgency strategies. It then describes the GOP's actualstrategy in each campaign and evaluates how the theory fares in explaining the range of cases relative to the alternative explanations. Chapter 5 examines three campaigns from Pakistan-two temporally overlapping cases of FATA (2002 - present) and Malakand (2007 - present) and one critical case of East Pakistan (1971)-for more in-depth structured, focused comparisons. The chapter seeks to identify the 27 presence of hypothesized territory and identity mechanisms as predicted in the core-periphery theory. It finds that territorial mechanisms incentivized significant incumbent effort in the East Pakistan and Malakand but not in FATA and Balochistan, and mechanisms of co-ethnicity restrained incumbent violence in both Malakand and FATA, but were distinctly absent in Balochistan and East Pakistan. Chapter 6 applies the same medium-N approach used in chapter 4 to analyze 14 counterinsurgency campaigns conducted by the Government of India (GOI) from 1947 - 2013. It first measures the value that the Indian state placed on different territories and identity groups to generate predictions of their counterinsurgency strategies. It then describes the GOI's actual strategy in each campaign and evaluates how the theory fares in explaining the range of cases relative to the alternative explanations. Chapter 7 follows the approach of chapter 5 and closely examines three cases from India-Punjab (1983 - 93), Kashmir (1990 - present), and the Naxal belt (2003 - present)-all overlapping in time but subject to very different strategies. Since the state incumbent deployed a different counterinsurgency strategy in each case, I again leverage structured, focused comparisons. It finds that territorial mechanisms incenting incumbent effort were strongly present in the Punjab and Kashmir while absent in the Naxal belt. It also finds that mechanisms of co-identity restrained incumbent violence in Punjab and in the Naxal belt, but remained distinctly absent in Kashmir as well as a shadow case of Telangana (1948 - 51). Chapter 8 concludes by briefly evaluating the evidence for my argument and weighing it against alternative explanations, examining the theory's external validity on a number of mini shadow cases, and expanding on the core theoretical and policy implications of the argument. 28 Chapter 2 Theorizing A Typology of Counterinsurgency Strategy I. Introduction Much of the contemporary discussion on counterinsurgency over the past decade has been refined through the experience of the U.S.-led war in Iraq and Afghanistan, leading scholars to appreciate a broader understanding of strategy with considerable variance on at least two dimensions: effort and violence. During the Iraq war (2003 - 2011), analysts observed that a strategy emphasizing politics and engaging the population rather than brute force offered an alternative approach to fighting a counterinsurgency, at least in theory.' Years later during a period of strategic reappraisal over the war in Afghanistan, a debate raged over the utility and effectiveness of "counter-terrorism" as an alternative to counterinsurgency, which demanded fewer inputs but operated with a lower threshold of success. 2 Implicitly, U.S. policymakers and strategic elites were debating distinct strategies but the critical dimensions that distinguished these strategies-the levels of effort and violence-have not been sufficiently theorized. The goal of this project is to understand why incumbents choose particular strategies of counterinsurgency but in order to begin theorizing, we need to first identify and organize the tactics that, in composite, define different counterinsurgency strategies. The formation of 1Kalev I. Sepp, "From 'Shock and Awe' to 'Hearts and Minds': The Fall and Rise of US Counterinsurgency Capability in Iraq," Third World Quarterly, 28 (2), 2007: 217-230; Colin H. Kahl, "How We Fight," Foreign Affairs, 85 (6), November-December 2006; Bradley Graham, "Calls for Shift in Iraq Strategy Growing; Lawmakers, Experts Urge Military to Focus More on Protecting Population Centers," Washington Post, November 11, 2005; Richard A. Oppel Jr., Eric Schmitt and Thom Shanker, "Baghdad Bombings Raise Anew Questions About US Strategy in Iraq," New York Times, Sept 17, 2005; Nathaniel Fick, "Fight Less, Win More," Washington Post, August 12, 2007; Bruce Hoffman, "We Cant Win if We Don't Know the Enemy," Washington Post, March 25, 2007; Thomas E. Ricks, "US Counterinsurgency Academy Giving Officers a New Mind-Set; Course in Iraq Stresses the Cultural, Challenges the Conventional," WashingtonPost, February 21, 2006; Michael R. Gordon, "Military Hones a New Strategy On Insurgency," New York Times, October 5, 2006; David Rhode, "Army Enlists Anthropology in War Zones," New York Times, October 5, 2007. 2 Fred Kaplan, "CT or COIN? Obama Must Choose This Week Between Two Radically Different Afghanistan Policies," Slate (War Stories), March 24, 2009; Austin Long, "Small is Beautiful: The Counterterrorism Option in Afghanistan," Orbis, Spring 2010: 199-214. 29 concepts and typologies is essential for theorization; we cannot explain events, processes and mechanisms without first identifying and organizing what they include. 3 This chapter conceptualizes the dependent variable of counterinsurgency strategy by identifying its range of values based on dimensions of effort and violence. The intersection of these dimensions produce four different "ideal type" counterinsurgency strategies. A dominant strain in the scholarly literature has identified only a single successful approach to counterinsurgency strategy-"the counterinsurgency ideal" 4 -that emphasizes the following tactics: "respect for noncombatants" 5 ; "protecting the population, not killing the enemy" because "political power [has] primacy over military power," 6; and most notoriously, winning "hearts and minds" 7 However, given the wide array of strategies deployed by various counterinsurgents in the post-WWII period (some quite successfully) as well as puzzling incumbent choices and outcomes described in the introduction chapter, this suggests more than one rational approach to combat insurgency, a variation found not merely between different incumbents but also between campaigns by the same incumbent. Whether explicitly acknowledged or not, at its core counterinsurgency is a form of competitive state-building intended to restore political order, but if strategies of political authority can vary regionally and spatially depending on local contexts and social patterns,, it also stands to reason that counterinsurgency strategies vary in the types of order they seek. 3 Giovanni Sartori, "Concept Misformation in Comparative Politics," American PoliticalScience Review 64 (4), 1970: 1033-1053. 4 Byman, 2006: 81 5 David J. Kilcullen, Counterinsurgency,New York: Oxford University Press, 2010. 6 Daniel Marston and Carter Malkasian, Counterinsurgencyin Modern Warfare, New York: Osprey Publishing, 2008. 7 The US. Army/Marine Corps CounterinsurgencyFieldManual, Chicago: University of Chicago Press, 2007, A5 8 Catherine Boone, PoliticalTopographiesof the African State: TerritorialAuthorityand InstitutionalChoice (New York: Cambridge University Press 2003; Jeffrey Herbst, States andPower in Africa: ComparativeLessons in Authority and Control, Princeton, N.J.: Princeton University Press, 2000; Joel S. Migdal, Strong Societies and Weak States: State-Society Relations andState Capabilitiesin the Third World (Princeton, N.J.: Princeton University 30 II. Literature Even as current policy debates have tended to treat counterinsurgency as homogenous and a single type rather than a range of types,9 variation has been observed empirically across numerous quantitative and qualitative studies suggesting that while there may be more or less effective strategies, there is no single "ideal type" of counterinsurgency.' 0 Little of this work has sought to theorize a typology for the complex and multidimensional empirical phenomenon of counterinsurgency strategy. Thus, there is a pressing need for sorting strategies into "taxonomic conceptual containers."" At least four distinct approaches to counterinsurgency strategy can be observed in the empirical literature (as opposed to the normative literature) to begin sorting this concept. The first approach has been to study discrete sub-sets of strategy, or tactics, as a dichotomous independent variable to explain patterns of conflict, battlefield effectiveness, or outcomes. This has included features of a counterinsurgency such as mechanization, 2 manpower, co-ethnic forces, 17 13 punishment tactics, service provisions, 18 1 civilian victimization, 15 forcible resettlement, 16 and leadership decapitation. 19 While these are certainly strategic Press, 1988; Robert H. Jackson, Quasi-states: Sovereignty, InternationalRelations, and the Third World (New York: Cambridge University Press, 1990; Michael Mann, "The Autonomous Power of the State: Its Origins, Mechanisms and Results," EuropeanJournalofSociology, 25 (2), 1984, pp. 185-213. 9 US. Army/Marine Corps CounterinsurgencyField Manual, 2007; Kilcullen, 2010; Marston and Malkasian, 2008 10 Byman, 2006. " Giovanni Sartori, 1970, p. 1052. 12 Lyall and Wilson, 2009. 13 Jeffrey A. Friedman, "Manpower and Counterinsurgency: Empirical Foundations for Theory nd Doctrine," Security Studies, 20 (4), 2011, pp. 1-36. 1 Monica Duffy Toft and Yuri M. Zhukov, "Denial and Punishment in the North Caucasus : Evaluating the Effectiveness of Coercive Counter-insurgency, JournalofPeaceResearch, 49 (6), 2012, pp. 785-800. 15 Johnston, 2009; Alexander B. Downes, "Draining the Sea by Filling the Graves: Investigating the Effectiveness of Indiscriminate Violence as a Counterinsurgency Strategy," Civil Wars, 9 (4), December 2007, pp. 420-444. 16 Yuri M. Zhukov, "Population Resettlement in War Theory and Evidence from Soviet Archives ," Journalof Conflict Resolution, (forthcoming). 17 Jason Lyall, "Are Coethnics More Effective Counterinsurgents? Evidence from the Second Chechen War," American PoliticalScience Review, 104 (1), February 2010, pp. 1-20. (Henceforth referred to as Lyall, 2010a). " Eli Berman, Jacob N. Shapiro, and Joseph Felter, "Can Hearts and Minds Be Bought? The Economics of Counterinsurgency in Iraq," JournalofPoliticalEconomy, 119 (4) , 2011, pp. 766-819. 31 ' components of a counterinsurgency campaign, alone they do not capture the full strategy or theory of victory. Different combinations of tactics could produce different strategies. For instance, very specific elements like manpower, decapitation, and service provision are theoretically and empirically central to two very different strategies.2 0 Broader conceptualizations like "direct" and "indirect" strategies used to explain battlefield outcomes suffer from the opposite problem of vagueness and under specification. 2 A second approach in the literature explores more nuanced strategy choices like degrees of accommodation for dealing with violent political mobilization, but these inevitably focus most of their attention on pre-conflict choices rather than strategy within conflict. A third set of literature centers on the type of force employment and treats the level of incumbent violence as synonymous with strategy, whether as selective or indiscriminate,2 mass killing, or restrained.25 These approaches tend to be much closer to the mark at capturing overall strategy within asymmetric conflicts but the singular focus on violence, though an important aspect of counterinsurgency, appears to be incomplete. Violence, while instrumental, does not imply a clear purpose. The fact that both selective and indiscriminate violence can be brutal, constitute terror, and in theory produce the similar amounts of fatalities can also pose Patrick B. Johnston, "Does Decapitation Work? Assessing the Effectiveness of Leadership Targeting in Counterinsurgency Campaigns," InternationalSecurity, 36 (4), Spring 2012, pp. 47-79. 20 Watts et al, 2014. 21 Ivan Arreguin-Toft, How the Weak Win Wars: A Theory ofAsymmetric Conflict, Cambridge: Cambridge University Press, 2005. 19 22 Monica D. Toft, The Geography of Ethnic Violence: Identity, Interests, and the Indivisibility of Territory, Princeton, NJ: Princeton University Press, 2003; Barbara F. Walter, Reputation and Civil War: Why Separatist Conflicts Are So Violent, New York: Cambridge University Press, 2009. 23 Stathis N. Kalyvas, The Logic of Violence in Civil War, Cambridge: Cambridge University Press, 2006. 24 Benjamin A. Valentino, FinalSolutions: Mass Killing and Genocide in the 20th Century. Ithaca: Cornell University Press, 2004; Benjamin Valentino, Paul Huth, and Dylan Balch-Lindsay, "'Draining the Sea': Mass Killing and Guerrilla Warfare," InternationalOrganization, 58 (2), 2004, pp. 375-407. 25 Jessica Stanton, "Strategies of Violence and Restraint in Civil War," PhD Dissertation, Columbia University, 2009. 32 challenges for relying on this distinction. 2 6 In fact, the conventional strategy literature provides a useful model for identifying empirical variation and distinct strategic models based not on violence, but rather a "theory of victory, a notion of the combination of human and material resources and tactics that it believes is most likely to produce success on the battlefield."2 7 A final set of literature has advanced the treatment of strategy beyond dichotomous measures of violence to the full package of tactics, targeting, training, resources, organization, and parallel non-violent economic and political activities combined in distinct ways and guided by a theory of victory.2 8 Based on deductive and inductive reasoning, this approach has identified distinct strategies for countering insurgency that an incumbent can draw upon during a civil conflict, even while operating under the same doctrine29 or within the same conflict.30 Because my argument moves beyond dichotomous or continuous treatments of violence, this work has organized strategy based on a number of overlapping typologies (See end of chapter, table A2. 1). While a major advance in the study of counterinsurgency strategy, these typologies have concentrated on third-party incumbents fighting rebels on foreign soil. Both the literature on violence and this newer typology literature still tend to miss an entire dimension of strategyforce extraction and deployment, either by treating counterinsurgents' bid for control as largely exogenous rather than a choice or by focusing on third-party incumbents that select into conflicts. Recognizing these limitations, my project builds on these works by developing a more Stathis N. Kalyvas, "Wanton and Senseless? The Logic of Massacres in Algeria," Rationality and Society, 11 (3), 1999, p. 251, 259. 27 Barry R. Posen, InadvertantEscalation: Conventional War and Nuclear Risks, Ithaca: Cornell University Press, 1991, pp. 71-72. This concept of strategy also draws from and builds on literatures in comparative conventional strategy including Mearsheimer, 1983; Posen, 1984; Pape, 1996; and Biddle, 2004. 2 Jackson, 2008; Erin Marie Simpson, "The Perils of Third-Party Counterinsurgency Campaigns," PhD Dissertation, Harvard University, June 2010, 2010; Jon Lindsay and Roger Petersen, Varieties ofInsurgency and Counterinsurgencyin Iraq, 2003-2009, Newport, RI: US Naval War College, 2011. 29 Jackson, 2008. For an approach that treats distinctions based on doctrine, see John Nagl, Learning to Eat Soup With a Knife: CounterinsurgencyLessons FromMalaya and Vietnam, Westport, Conn.: Praeger Publishers, 2002; Anthony James Joes, Resisting Rebellion: The History and Politics of Counterinsurgency.Lexington, KY: University Press of Kentucky, 2004. 3 Petersen and Lindsay, 2011 26 33 complete typology that also applies to domestic counterinsurgents by specifying the dimension of effort. 1II. Theory Building! through History To begin this project, it is useful to first scour the empirically rich and well-developed historical literature on great power and/or foreign incumbent counterinsurgency to generate concepts potentially useful for theorization. If the most sophisticated and capable militaries in history were observed to have conducted different counterinsurgency strategies, it would validate the notion that there are indeed distinct counterinsurgency strategies to be deliberately selected rather than better or worse imitations of the ideal model. Much of the historical research has still tended to focus on outcomes over dynamics and compare "national approaches to counterinsurgency." 31 However, some new research has begun to systematically examine conflict processes conducted by the same incumbent in different regions, generating fresh insights. For example, historian David French's recently published archival research on British counterinsurgency campaigns between 1945 -67 offers several insights on the incumbent regarded as the pioneering architect of modern counterinsurgency warfare, and his findings provide opportunities to identify systematic variation between campaigns. First, the British clearly varied the level of effort they exerted in different campaigns during this period. Above all, the British counterinsurgency model required leaders "to muster men and money"32 and the costliness of such endeavors led the British to settle for cheaper approaches, at times. Though the counterinsurgency model British commanders developed in Malaya was well regarded as a breakthrough, the British inconsistently applied the model to 31 32 Joes, 2004: 218. French, 2011: 219. 34 other campaigns that were ongoing or followed it.33 Many campaigns were poorly resourced and understaffed, which limited success and restrained the range of activities they could pursue.3 4 Innovations developed in one campaign were not necessarily employed in others.3 5 For instance, the Aden campaign in the Radfan desert involved considerably less effort and resources, and the British even declared their intent to leave more than a year before their defeat.3 6 Second, violence was far more central to British campaigns than previously assumed. Many historians have pointed out that despite the treatise offered by Sir Robert Thompson on "ideal-type" counterinsurgency, the British approach frequently diverged from this and "coercion.. .was the bedrock of British counter-insurgency policies."3 Counterinsurgency employing "minimum necessary force" was a welcome concept to bureaucrats seeking to sanitize a messy decolonization process.38 This is not a novel insight for most scholars but it is something to emphasize given the abundance of recent literature that has posited benevolent counterinsurgency strategy, whether as an ideal model or as a straw man to be knocked down. To put it bluntly, there are no cases of sanitized, minimal force counterinsurgency. Equally important however, the interpretation of this minimal force doctrine and degree of divergence from the ideal-type still involved considerable variation, especially in terms of violence. The British use of force varied due to their elastic definition of what an insurgent is, varying surrender and amnesty terms offered, and at times placing less emphasis on defeating political subversion than merely killing guerillas. 3 9 French points out that while campaigns in the 3 French, 2011: 250. 34 French, 2011: 30-31. French, 2011: 225. 36 Scott Smitson, The Road to Good Intentions: British Nation-building in Aden, 35 Complex Operations: Naval Postgraduate School, 2010. 37 French, 2011: 247. 3 Frnech, 2011: 249. 39 French, 2011: 72, 80, 92. 35 Case Study No. 12, Center for Canal Zone, Palestine, British Guiana, and Cyprus more closely adhered to this ideal type, Kenya, Nyasaland, Oman, and Borneo strayed from it. 40 Using data mined by French from the archives, additional data sources, and a couple additional British counterinsurgency campaigns, 4 1 figure A2.1 (see end of chapter) arrays British campaigns on two dimensions using two different measures for each dimension (thus yielding four combinations). The result is a striking correspondence in the relative locations of each of these campaigns suggesting four clusters of strategy (see figure 2.1). Figure 2.1: Distribution of British Counterinsurgency Cases Effort High Low I High Oman, Nyasaland Low Aden, Borneo, British Guiana K(a | Paletne Violence Canal Zone, Cyprus, Malaya, Northern Ireland, Palestine (Zionist) Plenty of other incumbents also displayed noticeable variation on at least one of these two dimensions. Nazi Germany, known for their reliance on terror and harsh countermeasures to defeat nationalist and communist partisan insurgents during their occupation of much of Europe, displayed some notable differences in their approach to different regional subgroups. "In the West," (i.e. France), the Germans proved to be "less indiscriminate and bloody than in the East," 40 41 French, 2011: 132-33. In addition, I added Northern Ireland and the Arab Revolt from 1936-39. 36 and embraced a "'moderate' form of anti-partisan warfare," 4 2 while in the Soviet theater they embraced a "harsh and inflexible" approach.43 Even within the East, they held more trust and empowered certain minority groups above others.44 Varying strategy despite strong military doctrine and organization was not confined to foreign incumbents. The Soviets, known for a generally repressive counterinsurgency model,45 clearly operated with different strategies of violence against different insurgencies, relying on indiscriminate violence against the Chechans and Ural Kulaks, versus more selective violence against the post-WWII insurgencies on the Western border.4 6 Furthermore, between two nearly simultaneous Western border campaigns, the Lithuanian campaign depended on "out-and-out brutality," 4 7 but the Ukrainian campaign soon exhibited rational learning and a "major reversal" to more selective violence.4 8 Soviet archival data appears to substantiate this. Relative to its Lithuanian campaign, Soviet counterinsurgency in Ukraine perpetrated fewer casualties and deportations, enjoyed a much higher composition of local militias and informants, and embraced political concessions and economic development as part of their campaign (Table 2. 1).49 42 Peter Lieb, "Few Carrots and a Lot of Sticks,: German Anti-Partisan Warfare in World War Two," Counterinsurgencyin Modern Warfare, Ed. Daniel Marston and Carter Malkasian, Oxford: Osprey Publishing, 2008, p. 75 43 Lieb, 2008, p. 70 4Lieb, 2008, p. 65 45 Joes, 2004. 46 Yuri M. Zhukov, "Counterinsurgency in a Non-Democratic State: The Russian Example," The Rutledge Handbook of Insurgency and Counterinsurgency, Ed. Paul B. Rich and Isabella Duyvesteyn, New York: Routledge, 2012, p. 293. 47 Roger D. Petersen, Resistence andRebellion: Lessonsfrom EasternEurope, New York: Cambridge University Press, 2001, p. 170 48 Serhiy Kudelia, "Choosing Violence in Irregular Wars: The Case of Anti-Soviet Insurgency in Western Ukraine," Eastern European Politics and Societies and Cultures, 27 (1), February 2013, p. 161 49 Based on data from Kudelia, 2013; Friedman, 2011; Alexander Statiev, The Soviet Counterinsurgency in the Western Borderlands,New York: Cambridge University Press, 2010; Yuri Zhukov, "Examining the Authoritarian Model of Counter-insurgency: The Soviet Campaign Against the Ukrainian Insurgent Army, Small Wars and Insurgencies, 18 (3), September 2007, pp. 439-466; V. Stanley Vardys, "The Partisan Movement in Postwar Lithuania," Lituanus (Lithuanian Quarterly Journal of Arts and Sciences), 15 (1), Spring 1969; Michael Clodfelter, Warfare and Armed Conflicts: A StatisticalEncyclopedia of Casualty and Other Figures, 1494-2007, Jefferson, NC: McFarland & Company, 2008. 37 Table 2.1: Comparative Intensity of Tactics in Two Soviet Counterinsurgencies (Per 100,000 people) Fatalities/Yr (1946-52) Deportations Surrenders Guerillas Amnestied Informers Local Forces (Source: Statiev, 2010) Lithuania 308 4334 155 286 153 531 Ukraine 250 1144 525 373 212 657 The United States during the civil war was a very different domestic incumbent in terms of regime type and state strength yet it too exhibited similar variations. While the civil war was principally a conventional conflict, there was a substantial amount of guerilla warfare alongside it. Large numbers of federal forces were required to fight guerillas in the Western theater, but commanders were under pressure to send their resources to the much-prioritized eastern theater where the "real war" was being fought. 50 Thus Union forces were poorly equipped and less determined, resulting in numerous live and let-live bargains in the western zones of conflict. 5 1 The use of force also varied during the U.S. civil war. Union forces in Virginia abided by a code of conduct with prisoners of war and worked well with civilians, but the Union applied much more brutal methods against the irregular militia in Missouri and treated all civilians as rebel sympathizers until proven otherwise.52 This brief historical survey confirms that counterinsurgency campaigns by the same incumbent could vary dramatically but the relevant dimensions of this variation still need to be better articulated and theorized. 50 51 52 Anthony James Joes, America and Guerilla Warfare, Lexington: University of Kentucky Press, 2000, p. 65. Joes, 2000, p. 58, 66. Joes, 2000, pp. 52-57, 61-65. 38 IV. Dimensions of the Dependent Variable Because the existing literature's conceptualization of counterinsurgency is incapable of analytically capturing the puzzling patterns and full variation observed between campaigns, the first goal of this project is to devise a typology of strategy in a counterinsurgency campaign. A typology would allow us to overcome the problems of "conceptual stretching" common in the counterinsurgency literature. 53 This approach offers richer theorization and measurement of the dependent variable beyond dichotomous measures like the use of individual fighting tactics, accommodation, or indiscriminate violence. The typology is devised through deductive logic based on concepts from the strategic and military studies literature-specifically the conventional military strategy that has long recognized empirical variation and theorized distinct strategic models5-as well as inductive detection of patterns in the empirical research and cases of counterinsurgency and asymmetric or irregular conflict. 55 A number of recent studies of counterinsurgency based on deductive and inductive reasoning have also identified multiple incumbent strategies in civil conflicts (see table 1), potentially salient for explaining outcomes. These are not simply stylistic differences by militaries with different doctrines or cultures; 56 this variation is exhibited by the same military under the same official doctrine, 57 or even during the same conflict. 58 Additionally, the identified variation is not simply based on a fatality threshold but other features of a broader military 53 Giovanni Sartori, "Concept Misformation in Comparative Politics," American PoliticalScience Review, 64 (4), 1970: 1033-1053. This concept of strategy draws from and build on literatures in comparative conventional strategy including Mearsheimer, 1983; Posen, 1984; Reiter and Meeks, 1999; and Biddle, 2004. For instance, Reiter and Meek use the levels of mechanized forces/assets to infer a conventional strategy of maneuver as opposed to attrition. See Dan Reiter and Curtis Meek, "The Determinants of Military Strategy," InternationalStudies Quarterly 43 (2), 1999, pp. 363-387. 5 Inductive pattern seeking is even used by quantitative-oriented scholars to discover new classes or types of observations generally lumped together. For instance, Fearon, 2004 is able to do this by disaggregating types of civil wars and noticing patterns in a particular type-"sons of the soil" conflicts. 56 Joes, 2004. 57 French, 2011; Jackson, 2008. 58 Lindsay and Petersen, 2011. 54 39 strategy that includes doctrine, tactics, targeting, training, resources, organization, and parallel non-kinetic activities (e.g. negotiations, side-deals, public goods) and political authority that accompanies it. The combination of deductive logic, inductive pattern recognition from empirical cases, and a review of these other typologies reveal two major dimensions of an incumbent's counterinsurgency strategy: effort and violence. Both are further developed in this chapter. Violence The more obvious and well-recognized dimension in the literature is the level of counterinsurgent's violence, or "brute force,"59 described in different concepts like "unconventional warfare,"60 "civilian victimization," 61 "firepower-intensive," 62 or even "mass killing," 63 all of which prove useful as dichotomous measures of violence for large-N research. However, most of these approaches tend to study violence as an end rather than integrated into a strategy with a clear "theory of victory" as has been done for conventional military strategy. Violence as a dimension of counterinsurgent strategy is best thought of as a set of sticks and carrots in varying proportions, distributed differently across a variety of targets or recipients in order to defeat the insurgents' political subversion.6 4 This balance of carrots and sticks falls along a continuum and an incumbent's distribution of its resources can err more towards securing support for the incumbent (a strategic hamlet strategy) or towards killing insurgents (a 59 6 0 Thomas Schelling, Arms and Influence, New Haven, CT: Yale University Press, 1966, pp. 2-6. Kalyvas, 2006. 61 62 Johnston, 2009; Alexander B. Downes, Targeting Civilians in War, Ithaca: Cornell University Press, 2008. Jonathan D. Caverley, "The Myth of Military Myopia: Democracy, Small Wars, and Vietnam," International Security, 34 (3), Winter 2009/10: 119-157. 63 Valentino et al, 2004; Valentino, 2004. 6 Petersen, 2001: 226-28. 40 "partisan hunt" strategy).6 5 This choice of strategy is akin to conventional doctrine that privileges offense or defense. 66 Incumbents can choose the degree of violence simply based on its preferences, or based on its theory of what is driving the rebellion. In economic terms, a supply-side strategy concentrates more attention on eliminating insurgents who "pull" rebellion (as well as interdicting their organization and resources), while a demand-side strategy concentrates more attention addressing the motives that "push" an insurgency, whether they be economic deprivation, political exclusion, or social repression. 67 While supply-side strategies may focus more on the level of violence necessary to trigger the adversary's "breaking point,",6 a demandside strategy might focus on the package of incentives, reforms, or rehabilitation to precipitate an adversary's satiation point. Furthermore, demand-side strategies might seek to avoid actions (like heavy or indiscriminate violence) that exacerbate the factors "pushing" insurgency. Realistically, all counterinsurgency strategies involve the use of force, but in tandem with this violence, some strategies can substitute a greater proportion of non-kinetic tactics to satiate and demobilize some proportion of insurgents and their support base without having to kill them. Sticks and carrots can be deployed and measured by more than just guns and butter. Though the sticks frequently are measured or operationalized as levels of casualties, in fact they encompasses a whole range of repressive tactics that may not get captured by accessible data on violence including capture or imprisonment, internment, forced disappearances, torture, aerial bombing, and seizure or destruction of property. Carrots intended to offset or substitute for the Petersen, 2001, p. 227. Posen, 1984, p. 14. 67 This supply demand framework is made famous by Nathan Leites and Charles Wolf, Jr., Rebellion and Authority: An Analytic Essay on Insurgent Conflicts, Chicago: Markham, 1970, pp. 28-30. 68 One indicator of this perception is the language used by the incumbent. An incumbent that refers to enemy combatants as "miscreants," "thugs," or "criminals," versus "militants," "fighters," or "terrorists" may reveal the incumbent's different estimation of the enemy combatants' breaking point and/or the instruments necessary to defeat them. 65 66 41 use of pure force and address the demand-side of rebellion can extend beyond material (re)distribution or public goods-typically favored by Western counterinsurgency doctrine-to include two-way bargaining and accommodations (including amnesties and rehabilitation), 69 support for moderates,70 political or economic reforms, 7 ' and devolution of political authority.7 2 Effort The more novel contribution of this typology is conceptualizing the second dimension of the dependent variable, namely the scale of counterinsurgent effort mobilized, deployed, and expended. Since the incumbent, typically a state, has to be concerned with the "financialeconomic-industrial capacity element," its security strategy depends not only on the manner in which forces are mobilized and configured to fight, but also the extraction of material and political resources allotted to initiate and sustain the campaign.73 Furthermore, we know that states regularly under balance even against clear and direct threats, 74 therefore incumbents facing multiple threats or challenges that compete for resources, time, or focus, may need to economize efforts when fighting a particular rebellion. The literature has underappreciated this notion of effort. 69 Jackson, 2008; Watts et al, 2014. 70 Arjun Chowdhry and Ronald Krebs, "Making and Mobilizing Moderates: Rhetorical Strategy, Political Networks, and Counterterrorism," Security Studies, 18 (3), 2009, pp. 371-399. 71 Michael Albertus and Oliver Kaplan, "Land Reform as a Counterinsurgency Policy: Evidence from Columbia," of Conflict Resolution, 57 (2), 2012: 198-23 1. Journal 72 Boone, 2003. 73 Steven E. Lobell, "War is Politics: Offensive Realism, Domestic Politics, and Security Strategies," Security Studies, 12 (2), Winter 2002/03: 165-95. 7 Randall L. Schweller, "Unanswered Threats: A Neoclassical Realist Theory of Underbalancing," International Security, 29 (2), Fall 2004, pp. 159-201. 42 ' Scholars of grand strategy and conventional conflict sought to formulate a more costefficient "indirect approach" to strategy75 which has spread to counterinsurgency scholarship as well.76 While some conceptions have proved too elastic, there is a valid strategy, even in conventional fights, namely a "limited aims strategy," which contends that "minimizing contact with the [adversary],"77 may be less costly. Recent debates have positioned these lower scale and smaller footprint campaigns, labeled "counterterrorism," as an alternative strategic option to counterinsurgency though they are both strategies to fight insurgency and can involve equally high levels of incumbent violence and lethality. The presumption in much of the literature, including the rationalist bargaining literature 79 (conceiving strategy as dichotomous) is that all incumbents that choose to fight rebels in an asymmetric conflict embark on a "costly"8 0 campaign to "defeat" insurgency or "destroy" 8 guerilla forces and regain decisive sovereign control over a contested region. However, these are problematic assumptions given the previously introduced puzzles and the variety of wartime political orders, 82 assumptions made possible due to how infrequently domestic incumbent strategies are studied. 8 3 In subtle ways, recent counterinsurgency studies have begun to recognize the empirical validity of effort as a strategic dimension. Even moderately capable states can prefer containment B.H. Liddell Hart, Strategy: the IndirectApproach, London: Faber and Faber Limited, 1967: 333-346; John J. Mearsheimer, Lidell Hart and the Weight offHistory, Ithaca, NY: Cornell University Press, 1988: pp. 84-98. 76 David H. Ucko, "Counterinsurgency in El Salvador: The Lessons and Limits of the Indirect Approach," Small Wars andInsurgencies, 24 (4), 2013: 669-695; Thomas H. Henriksen, Afghanistan, Counterinsurgency, and the IndirectApproach, Hurlburt Field, Fl: Joint Special Operations University Press, 2010 77 Mearsheimer, 1983: 30. 78 Long, 2010; Stanley McChrystal, "Generation Kill," ForeignAffairs, 92 (2), March/April, 2013, pp. 2-8; Linda Robinson, "The Future of Special Operations, Foreign Affairs, 91 (6), Nov/Dec 2012, pp. 110-122. 79 Walter, 2009. 80 Walter, 2009: 4 81 David Galula, Robert Thompson, and Frank Kitson quoted in Martson and Malkasian, 2008, pp. 13-15. 82 Staniland, 2012. 8 It should be noted that both Toft (2003) and Walter (2009) do in fact study indigenous incumbents-namely Russia, Georgia, Indonesia, and the Philippines-but I would submit that these studies focus more on the political bargaining and onset of conflict rather than the dynamics of military strategy and operations within the conflict. 7' 43 to outright defeat of an insurgency to retain contested territory but hold down costs.8 4 Operational costs, "the expense of mobilizing, deploying, and maintaining forces,"8 5 are not incidental but in fact play a central role for strong and weak states alike in strategic decisions about the type and degree of force mobilization and deployment . Meanwhile, the costs to an incumbent of enduring violent political disorder are substantially lower in the post-1945 period compared to previous periods of history because state survival is less threatened due to international norms of border fixity and territorial sovereignty that prevent state-death.86 Cost control has even been central to some third-party counterinsurgency campaigns. Even at the height of their fabled counterinsurgency prowess, the British pioneered low-effort counterinsurgency campaigns during periods of fiscal austerity in the early twentieth century. The campaigns of "air policing"8 7 to combat uprisings sought what one scholar described as "control without occupation."8 8 As a strategy of limited aims and objectives, the goal was to regain some hold over territory but it was "not concerned with punishing an opponent by inflicting great damage on his army." 89 As a methodological note, the fact that most studies of foreign incumbents fail to observe limited aims counterinsurgency strategies is a result of a selection effect. "Modern nation states display a strong tendency to favor the pursuit of decisive victories, and decision makers 84 Watts et al, 2014; David R. Haines, COIN in the Real World," Parameters,Winter 2008-09, pp. 43-59. 85 Mearsheimer 1983: 23; also see Lobell 2002/03. 86 Boaz Atzili, Good Fences, Bad Neighbors:Border Fixity andInternationalConflict, Chicago: University of Chicago Press, 2011; Tanisha M. Fazal, State Death: The Politicsand Geographyof Conquest, Occupation, and Annexation, Princeton, NJ: Princeton University Press, 2007; Ryan D. Griffiths, Catch and Release: Expansion, Contraction, andthe Downsizing ofStates, PhD dissertation, Columbia University, 2010. 87 David Omissi, "Technology and Repression: Air Control in Palestine, 1922-36," JournalofStrategic Studies, 13 (4), 1990, pp. 41-63. 88 David French, "Have Options Really Changed? British Defence Policy in the Twentieth Century," Journalof Strategic Studies, 15 (1), 1992, pp. 50-72. 89 Mearsheimer, 1983, p. 54. 44 invariably seek a way to score one." 90 In the most frequently studied counterinsurgency campaigns, foreign incumbents chose to enter a conflict in the pursuit of these goals, but sustained power projection is still very difficult and expensive so when costs prove prohibitive, foreign incumbents are much more likely to refrain from entering (or exit completely) rather than adopt a limited aims strategy. However domestic incumbents facing rebellions-most likely because they are already weak or resource-poor states 91-have no option to select out of the conflict and are compelled to field some strategy, whether limited or ambitious. V. Four Ideal Type Strategies These two analytical dimensions of violence and effort intersect to create space for four possible ideal-type strategies-attrition,population control, enfeeblement, and mitigation-that are each held together by a set of operational tactics and objectives and correspond with approaches theorized in the counterinsurgency literature. (see figure 2.2 and table 2.2).92 High effort/low violence is coded as a strategy of population control; High effort/high violence is coded as a strategy of attrition; low effort/high violence is coded as a strategy of enfeeblement; and low effort/low violence is coded as a strategy of mitigation. If strategy is a means-ends chain, 9 3 both effort and violence are the means that interact to form four distinct ends, with distinct theories of victory. 90 Mearsheimer, 1983: 56. 91 James D. Fearon and David Laitin, "Ethnicity, Insurgency, and Civil War," American PoliticalScience Review, 97 (1), February 2003, pp. 75-90. 92 A detailed explanation of the measurement for each aspect of the dependent variable is included later in this chapter. 93 Posen, 1984: 13-14; Lidell-Hart, 1962: 335-36. 45 Figure 2.2: Typology of Counterinsurgency Strategy (Dependent Variable) Low (Containment) H High Violence Effort High (Control) L 0 U C 0 0 Enfeeblement Attrition .0 Low 0 tn .0 0 Population Mitigation MitigtionControl (Observable: Manpower) Table 2.2: Strategy Means-Ends Chain Goal Theory of Victory (Means) Attrition Population Control Enfeeblement Control (Rollback) through Disarm and Punishment Control (Rollback) through Denial and Inducement Containment through Disruption and Punishment Controlling the population to deny insurgents a target and base of support; gradually draining insurgent strength through violence as well as the capture, surrender, flipping, and reintegration of insurgents; harnessing extensive information for selective targeting to hold down violence and avoid excess collateral damage; and persuasion and bargaining with public goods, political concessions, devolution of authority, and rehabilitation of Interdiction and Deterrence of insurgent organization's ability to violently threaten state authority along with harassment and disruption of insurgent central nervous system (leadership decapitation), until violence is limited below a tolerable threshold. Direct annihilation and military destruction of adversary's combat power and capacity to resist; inflict such unrelenting pressure and pain on the insurgent and base that eventually, it becomes obvious that victory is impossible, and insurgents are wiped out or cease fighting Mitigation Containment through Tactical Denal and Cooptation Use untrained local forces and auxiliaries to manage violence until it wears down and sinks below a tolerable threshold. Side payments are used to co-opt local power brokers to back the state or "flip" sides and manage the region. the contested area Functional Linear Returns Form Orientation (Leites and Wolf) "Approach" (Liddell Hart) Political Authority (Boone) Increasing Returns Decreasing Returns Decreasing ________________________Returns Supply-side Demand-side Supply-side Demand-side Direct Direct Indirect Indirect Usurpation Powersharing Administrative Occupation Non-incorporation 46 Relationship to State Political Strategy. Since counterinsurgency involves more than war fighting, seeks to install or restore order to wield power in the long-term, and envisions some political end goal, the discussion of each of the distinct counterinsurgency strategies requires a vision of political authority. The parallels to Boone's work on political topography are striking as the political strategies identified closely track with each theorized counterinsurgency strategy (figure 2.3). Boone argues political authority can vary on two dimensions. Spatial dispersion (or concentration) estimates how physically extended and present the state is in the region and corresponds with the dimension of counterinsurgent effort. Meanwhile, processural centralization (or devolution) measures whether power is concentrated in the hands of a few state-authorized actors or devolved to more local nodes of power. 94 Since this second dimension is intimately tied to whether local actors can be trusted or require repression, 95 it corresponds with the counterinsurgent's level of violence. Figure 2.3: Overlap of Counterinsurgency and Political Authority Strategies Effort (Degree of State Penetrationinto region or "Spatial Deconcentration") Violence (Centralization ofAuthority) Low High (ConcentratedState (DeconcentratedStructure Suspended "Balloon-Like") - State Penetration) High Enfeeblement Attrition (Centralization) (AdministrativeOccupation) (Usurpation) Low (Devolution) Mitigation (Nonincorporation) Population Control (Powersharing) This model expects each counterinsurgency strategy to correspond with a strategy of political authority. The strategy of population control requires powersharinginvolving 94 Boone, 2003: 22. 95 Boone, 2003: 23, 136. 47 dispersion and devolution of authority. Attrition is not only repressive but transformative usurpationthat involves little devolution and requires "rewir[ing] the circuits of local authority." 9 6 Enfeeblement employs a form of administrativeoccupation, with the stategoverning with relative autonomy and a few elites-"suspended balloon-like" over the region. 97 Finally, mitigation, like non-incorporationis where "the regime seems to abdicate authority" and "does not seek to engage or impose," remaining spatially concentrated and aloof but allowing for local authority structures to prevail.9 8 Counterinsurgency Strategies in Detail Having analytically distinguished four counterinsurgency strategies within the two critical dimensions, each of which involves a concept of political authority, I now seek to ground each of these strategies in the specific theoretical and empirical literature on counterinsurgency to demonstrate that these are real world rather than just ideal-type categories,. Control Strategies Population Control. The tactical corollaries ofpopulation control (sometimes referred to as classic counterinsurgency or "Clear, Hold, Build") are well popularized. The central premise is that killing insurgents is necessary but not sufficient for victory and that resourceintensive, broadly targeted, parallel efforts are required to address the demand-side of rebellion including protecting and winning over the population, restoring the incumbent's legitimacy, and redressing grievances with political concessions and/or devolution of authority. Protection and persuasion of the population can be achieved through recruitment and training of local security 96 97 98 Boone, 2003: 32. Boone, 2003: 33. Boone, 2003: 33. 48 forces, cultivation of local information, economic and political concessions, and the discriminate application of force (with minimal collateral damage) against irreconcilable insurgents, with opportunities for rehabilitating reconcilable insurgents. 99 This would involve high amounts of effort and relatively less violence, or more restraint, resulting in a wider dispersion of forces to control the population, far less use of indiscriminate area weapons like artillery and airpower, and activities to win legitimacy through institutional reforms, resource transfers, development, social services, persuasion, and inclusion in decisionmaking.' 00 To be clear, this strategy still involves a nontrivial amount of violence. The "hearts and minds" myth of pure benign force is either a label that has never been put into practicel' or extremely infrequent. 0 2 Historical reevaluations of the paradigmatic models for such a "hearts and minds"-most notably the British counterinsurgency campaign in Malaya-have revealed them to actually be more brutal affairs.1 03 But dismissing the myth need not throw the baby out with the bathwater. A costly and sustained strategy of population control can "win" utilizing the following methods: controlling the population to deny insurgents a target and base of support; gradually draining insurgent strength through violence as well as the capture, surrender, flipping, and reintegration of insurgents; harnessing extensive information for selective targeting to hold down 99 Jackson 2008: 111-114; Lindsay and Petersen, 2011: 41-42, 46. 100 Nagl, 2002; The U.S. Army/Marine Corps CounterinsurgencyFieldManual, 2007; Kilcullen, 2010; Marston and Malkasian, 2008; Stephen D. Biddle, "The New U.S. Army/Marine Corps Counterinsurgency Field Manual as Political Science and Political Praxis," Perspectives on Politics, Vol. 6, No. 2, June 2008; David Galula, CounterInsurgency Warfare: Theory and Practice, London: Pall Mall Press, 1964. 101 Jacqueline L. Hazelton, "Compellence and Accommodation in Counterinsurgency Warfare," PhD Dissertation, Brandeis University, 2011: 3. 102 Johnston, 2009: 112. Johnston finds that 89% of the 158 cases of counterinsurgency from 1800-2006 involved civilian victimization through intentional or indiscriminate violence. The figure is about 88% when looking only at the 73 cases post 1945. 103 French, 2011; Huw Bennett, "The Other Side of the COIN: Minimum and Exemplary Force in British Army Counterinsurgency in Kenya," Small Wars and Insurgencies, 18 (4), December 2007, pp. 638-664; Bruno C. Reis, "The Myth of British Minimum Force in Counterinsurgency Campaigns during Decolonisation (1945-1970)," JournalofStrategicStudies, 34 (2), April 2011, pp. 245-279. 49 .& violence and avoid excess collateral damage; and persuasion and bargaining--"violent negotiation"10 4--that involves public goods, political concessions, devolution of authority, and rehabilitation of the contested area. Scholars agree the British campaign in Northern Ireland generally adhered to this population control strategy of relative restraint.1 05 The nearest conventional warfare strategy analog for population control might be something similar to a positional defense (potentially with elements of maneuver defense) 06 and the envisioned political authority would involve a form of power-sharing. The functional form of this strategy might best be captured by an exponential curve with high effort required early on for little immediate payoff but eventually producing increasing returns to scale (figure 2.4). Figure 2.4: Functional Form for Different Counterinsurgency Strategies Attrition: Linear Returns Y= X Population Control: Increasing Returns Y= e^(X) Containment (Enfeeblement Mktiation, DimInishinj Returns Y log(X); Y=log(X)/2 Attrition. Attrition is inherently a supply-side focused strategy intended to degrade adversary combat power. Rather than refraining from excess violence, the strategy of attrition seeks the "direct annihilation of the enemy's forces," which is why "the attrition strategy relies 104 Jackson, 2008, p. 1 1 105 Brian A Jackson, "Counterinsurgency in a 'Long War': The British Experience in Northern Ireland, Military Review, Jan/Feb 2007, pp. 74-85; Paul Dixon (2009) 'Hearts and Minds'? British Counter-Insurgency Strategy in Northern Ireland, JournalofStrategicStudies, 32:3, 445-474. 106 John G. Hines and Donald Mahoney, "Positional and Maneuver Forms in Defense," Defense and Counteroffensive Under the New Soviet Military Doctrine, Santa Monica: RAND, 1991, pp. 62-86 50 heavily on firepower,"'07 and extensive effort to destroy "the adversary's physical capacity to resist."10 8 This strategy often involves or is synonymous with brute force and repression'0 9 that, "... inflict[s] such unrelenting pressure and pain on the insurgent or terrorist that eventually, despite perhaps fanatical commitment to the cause for which he fights, it becomes obvious that victory is impossible, and he stops fighting."' 10 Though it commonly involves firepowerintensive area weapons and air power, regardless of the weaponry, a strategy of attrition is one that measures success in kills and body counts of the adversary. In this approach, incumbents assume that "the military defeat of their armed opponents will allow them to dictate the desired political outcome,"" and regain full control of contested space. Like population control, attrition often involves extensive resources, but relies heavily on the considerable use of firepower. This drives up the level of state inflicted violence in two ways - first through obvious collateral damage and second through deliberate targeting of the insurgents' material and popular base to "drain the sea" within which they maneuver." 2 Distinguishing insurgents from their support base and from neutral civilians is difficult and can easily result in seemingly indiscriminate violence." 3 This is why attrition strategies like those pursued by the French in Algeria, U.S. in Vietnam, or the Soviets in Afghanistan may at times seem indistinguishable from "collective punishment" and broach mass killing levels of violence, even if the objective is not cleansing or democide." 4 Mearsheimer, 1983: 33-34. Arreguin-Toft, 2005, pp. 30-31. 109 Leites and Wolf, Jr., 1970; Edward Luttwak, "Dead End: Counterinsurgency Warfare as Military Malpractice," Harper's,February, 2007, pp. 33-42. 110 J. Boone Bartholomees, Jr., "The Issue of Attrition," Parameters,Spring 2010; 16. 111 Jackson, 2008: 13. 11 2 Valentino et al, 2004; Valentino, 2004; Downes, 2007. 113 See Kalyvas, 2006: 87-91, 147-49. 114 French, 2011: 134; Valentino, 2004, pp. 217-227; Jonathan D. Caverly, "The Myth of Military Myopia: Democracy, Small Wars, and Vietnam," InternationalSecurity, 34 (3), Winter 2009/10, pp. 119-157; Matthew 107 51 In contrast to population control, which is based around defense and denial, attrition is rather similar to conventional attrition, offensive in nature, and focused on punishing and gradually degrading insurgent forces and their base, however broadly defined. 1 5 Since a strategy of attrition has little interest in two-way politics, a victory would most likely impose a political order that Boone describes as "usurpation" rather than involving any devolution of authority." 6 While population control has increasing returns to investment over time, the functional form of attrition looks more like a linear function with returns matching the level of inputs (see figure 2.4), but this still motivates high efforts to substantially degrade the insurgency. Containment Strategies In some cases, an incumbent may be willing to accept fragmented sovereignty to avoid the costly area domination operations that the two "control strategies" require and may opt for an alternative. Drawing on similar concepts from old cold war strategic debates, the cost of "containment" may be considered more bearable than the costs of aggressive "rollback" to achieve control." 7 While control is considered natural for sovereign states, in this case: Containment is a good option, sometimes the only option.... If the state wishes to retain control over the area being contained, doing so often requires saturating the region with security forces. Otherwise the national government can choose to isolate the region with security forces and let it 8 govern itself." Containment strategies may offer states a cost-efficient route to periodically suppress or hold an insurgency at bay while wrestling other problems that beset weak states. In contrast to the previous control strategies, an incumbent's expectation of the functional form of containment Adam Kocher, Thomas B. Pepinsky, and Stathis N. Kalyvas, "Aerial Bombing and Counterinsurgency in the Vietnam War," American JournalofPoliticalScience, 55 (2), 2011, pp. 201-218. 1 5 Pape, 1996; Posen, 1984. 116 Boone, 2003. 117 Laszlo Borhi, "Rollback, Liberation, Containment, or Inaction? U.S. Policy and Eastern Europe in the 1950s," Journalof Cold War Studies, 1 (3), 1999, pp. 67-110; 118 Haines, 2008-09, p. 54. 52 strategies would look more like log functions with decreasing returns to scale, providing minimal incentives to increase effort (see figure 2.4). There are at least two distinct containment strategies with different levels of violence. Table 2.3: Tradeoffs Of Counterinsurgency Strategie Tradeoffs Costs Attrition 1) Material (manpower and money) trade off with other regime priorities or threats; 2) Risk of Regime loss/collapse: excess force can benefit insurgents by a) solving collective action problem, b) attracting foreign support from foreign rival or diaspora, or c) drawing international intervention; 3) International Reputation/Normative (Pariah) costs Population Control 1) Material (manpower and money) trade off with other regime priorities or threats; 2) Political/Regime Costs: Counterinsurgent has to offer goods, concessions, or accommodations that at best alienate a political base and at worst threaten to undo critical institutions of the Enfeeblement 1) Extended Duration of Conflict (and risks of future escalation or contagion); 2) Preservation of dysfunctional, under-productive regions (Prolonged violence, lawlessness, criminal activity) Mitigation 1) Extended Duration of Conflict (and risks of future escalation or contagion); 2) Preservation of dysfunctional, underproductive regions (Prolonged violence, lawlessness, criminal activity); 3) Principal-Agent Problem and empowerment of new violent non-state actors Affordable, lowcost stability (durable disorder) with few concessions; Some Affordable, low-cost stability (durable disorder) regime Long-term stability; Forced Integration with no Benefits concessions; Internal deterrence Long Term Stability; Integration of embattled region(s) internal deterrence Enfeeblement. Enfeeblement is one containment strategy, derived in part from the concepts of limited war, punishment, and tactical interdiction.' 19 Though appearing similar to attrition in its heavy reliance on violence, targeting the "supply-side" of the insurgency,120 and "belligerent" approach, 12' enfeeblement differs in the level of material and political resources invested, sustainment of operations, and degree of adaptation. Whether described as a "shortcut 119 Mearsheimer, 1983; Posen, 1984; Pape, 1996. I owe thanks to Barry Posen for this term and its conceptual development. 120 Leites and Wolf, 1970. 121 Haines, 2008-09: 49 53 of skillfully applied blows," "dislocation," or "paralysis of the defense," 22 the objective is to minimize direct confrontation in favor of psychological attacks to wear down rebels and their base. Unlike control strategies, enfeeblement will not try to deplete insurgent combat power or fully resolve conflict, and therefore will not involve costly efforts to retain large deployments, to innovate and adapt to new tactics, or to win over a population. Instead of degrading, disarming, and defeating insurgents through attrition, enfeeblement is meant to disrupt, interdict, and deter insurgent capacity and limit violence below a "tolerable" threshold. This may include harassing tactics such as limited raids and leadership decapitation, to weaken, fracture, or deter insurgent organization.123 Some cases of enfeeblement include the U.S. strategy in Laos, 24 Israel's strategy against Hamas and Hezbollah,12 5 and the "very indelicate method of policing" employed by the British to manage peripheral conflicts during and after its colonial empire in a costefficient manner using aerial policing.1 2 6 Its signature is the military avoiding substantial or prolonged close-range confrontation by conserving troop deployments and relying on sporadic sweeps or raids, airstrikes, or artillery barrages. As with the other strategies, this approach has its own distinct measure of victory. For instance, despite its overwhelming advantage and capacity, Israel's minimalist strategy towards local violent non-state actors "is not aiming for victory or for ending the conflict." 7 And despite Mearsheimer describes these theories of Clausewitz and Lidell-Hart in ConventionalDeterrence, 1983: 37-41. Lindsey and Petersen, 2011: 42-43; Simpson, 2010: 89. Long, 2010. Efraim Inbar & Eitan Shamir, 'Mowing the Grass': Israel's Strategy for Protracted Intractable Conflict, Journal ofStrategicStudies, 37 (1), 2014: 65-90 1 6 Omissi, 1990. See also, Bruce Hoffman, BritishAir Power in PeripheralConflict, 1919-1976, Santa Monica: RAND, 1989. 12 7 Inbar and Shamir, 2014: p. 71 122 123 124 12' 54 its limited effect, the British Air campaign against a tribal uprising in Oman was lauded by the Times as " a brilliant example of economy in the use of force." 128 It might be argued that highly sophisticated militaries (e.g. the U.S. military) may have the requisite capabilities and technology (e.g. precision airpower, drones) to employ this minimalist strategy to not only contain, but also defeat an insurgency outright, rendering the distinction of goals moot. 129 There are three responses to this contention. First, at best this is arguably unique to the United States, as no other state possesses this level of capacity, so that the theory remains in tact for all other incumbents. Second, while the effort may not demand traditional (manpower) resources, such a strategy to achieve outright defeat would require such intensive use of other resources (money, R&D, ISR, political resources) that the strategy could cumulatively escalate to a maximalist, high effort strategy. Third, the limited empirical record actually shows U.S. manned and unmanned airpower to be capable of containment but not outright defeat or elimination of belligerents in asymmetric conflicts.13 0 Enfeeblement's theory of victory might be closest to the conventional strategy of blitzkrieg (especially with the analogies drawn between leadership decapitation and destroying the central nervous system 3 1), but the limited input of effort and expectation of the adversary's rapid collapse and capitulation might make enfeeblement closer to American "shock and awe" doctrine.12 Incumbents may prosecute limited and sporadic retaliation-sometimes derided as 1 From "Insurgents Surprised. Brilliant But Little Known British Desert Action," Times, April 9, 1959, quoted in Peter Dye, "The Jebel Akhdar War: The Royal Air Force in Oman 1952-1959, Air PowerReview, 11 (3), Winter 2008, pp. 18-32. 129 For recent work on this debate, see Richard B. Andres, Craig Wills, and Thomas E. Griffith Jr., "Winning with Allies: The Strategic Value of the Afghan Model," InternationalSecurity, 30 (3), Winter 2005/06, pp. 124-160; Stephen D. Biddle, "Allies, Airpower, and Modem Warfare: The Afghan Model in Afghanistan and Iraq," International Security, 30 (3), Winter 2005/06, pp. 161-176. 130 Mark Mazetti, The Way of the Knife, New York: Penguin Press, 2013; Biddle, 2005/06. 131 Mearsheimer, 1983: 36. 132 Sepp, 2007; John T. Correll, "What Happened to Shock and Awe?" Air ForceMagazine, 86 (11), November 2003, pp. 52-57; Harlan Ullman, James Wade Jr., L.A. Edney, Frederick Franks, Jr., Charles Homer, Jonathan 55 "whack a mole" or "mowing the grass" 1-if not to collapse the insurgency then at least to re- establish deterrence. Even with expectations of rapid gains, incumbents can stick to this strategy for a protracted period of time believing it to be effective and cost-efficient given other constraints.134 The type of political authority likely accompanying this strategy would be "administrative occupation" with the state suspended balloon-like over the disputed region with neither state penetration nor devolution of authority.1 35 Mitigation. Similar to enfeeblement, mitigation seeks containment with low effort as well as low violence (though a nonzero amount), often by bargaining with and co-opting the power of pre-existing social orders and power brokers such as tribes, militias, and warlords. In these cases, the weak state tries to leverage the strong society to its advantage, 13 6 anointing local forces like militias or warlords as a direct substitute rather than an auxiliaryto state forces and outsourcing violence management authority backstopped by coercion and periodic formal reinforcement. To create this "durable, low cost political order"1 37 where some state authority is accepted, the incumbent willfully accepts the principal-agent risk and forfeits a degree of sovereignty by relying on local irregulars, sometimes through insurgent defections and amnesty. 138 Though characterized by occasional periods of collusion or tacit coexistence,139 the incumbent does not necessarily align with local armed actors and more likely tolerates them out Howe, and Keith Brendly, Shock andAwe: Achieving Rapid Dominance, Washington: D.C., National Defense University Press, 2002 133 Inbar and Shamir, 2014; Sergio Catignani, "Israeli Counterinsurgency: The Never-ending 'Whack-A-Mole,'" The Rutledge Handbook of Insurgency and Counterinsurgency, Ed. Paul B. Rich and Isabella Duyvesteyn, New York: Routledge, 2012, pp. 263-275; Peter Spiegel, "'Oil Spot' Replaces 'Whack-a-Mole' Strategy," FinancialTimes, November 30, 2005. 13 Austin Long, Doctrine of EternalRecurrence: The US. Military and CounterinsurgencyDoctrine, 1960-1970 and 2003-2006, Santa Monica, CA: RAND, 2008; Inbar and Shamir, 2014. 1 Boone, 2003, p. 32-33. 136 Migdal, 1988. 137 Jackson, 2008: 116-119. 138 Lindsay and Petersen, 2011: 45-47; 139 Staniland, 2012, p. 249. 56 of inertia or a higher strategic objective-a form of rational appeasement. 140 Again, the appeal is resource conservation as Marten notes, "States choose to cooperate with warlords as a low-cost method for achieving immediate, short-term security benefits." 14 1 Mitigation seeks to address the demand side of rebellion but through informal, inconsistent, and incomprehensive methods (e.g. discretionary side-payments). Local proxies with better information are able to coerce complicity without high violence1 4 2 and perhaps with some distribution of inducements. While co-opting locals and "bolstering moderates," 43 mitigation primarily differs from population control due to the lower level of resources invested resulting in less control or influence. Rather than integrating and rehabilitating a contested area, with this strategy, incumbents' embrace the indirect political rule of "non-incorporation" 144 allowing a degree of informal autonomy so long as its agents remain relatively loyal and contain rebellion below a certain threshold. Historic examples of mitigation include British colonial divide-and-rule strategies, the Ottoman Empire's use of local notables to manage distant provinces,1 45 U.S. reliance on the tribal awakening in Iraq,1 46 and Russia's approach in Chechnya. For example, "Russian leaders effectively outsourced Chechnya to a private contractor because they no longer wished to expend 140 Daniel Treisman, "Rational Appeasement," International Organization, 58 (2), Spring 2004, pp. 345-373. 141 Kimberly Marten, Warlords: Strong-armBrokers in Weak States, Ithaca, NY: Cornell University Press, 2011: 105. 142 Simpson, 2010: 88. 143 Chowdhry and Krebs, 2009. 144 Nonincorporation is a political strategy theorized by Boone, 2003, pp. 32-33. 145 Albert Hourani, "Ottoman Reform and the Politics of Notables," The Modern Mddle East: A Reader, Ed. Albert Hourani, Philip S. Khoury, and Mary C. Wilson, Berkeley, CA: University of California Press, 1993, pp. 83-109. 146 Lindsey and Petersen, 2011; Austin Long, "The Anbar Awakening," Survival, 50 (2), 2008, pp. 67-94; Stephen Biddle, Jeffrey A. Friedman, and Jacob N. Shapiro, "Testing the Surge: Why Did Violence Decline in Iraq in 2007?" InternationalSecurity, 37 (1), Summer 2012, pp. 7-40. 57 the state resources and political capital that would be required for direct control over Chechen territory." 147 Containment tradeoffs. Containment may be cheaper than control strategies but it incurs other tradeoffs. At the same time containment is not purely a product of state capacity as it presents appeals and risks to both strong and weak states alike. The tradeoff with containment strategies is that while extensive efforts may not appear worthwhile at present, minimalist efforts incur some nonzero risk of future escalation. This dilemma confronted the United States just a few months after the 2003 invasion of Iraq was described by an open-source intelligence report: The classical counterinsurgency dilemma now confronts the United States. The quantity of forces needed to defeat the guerrillas is disproportionate to the military advantage gained by defeating them. Failure to engage the guerrilla force could result in a dramatic upsurge in their numbers, allowing them to become unmanageable. The ineffective engagement of guerrillas could result in both the squandering of resources and the failure to contain them. 148 Furthermore, control strategies may appear more costly for weak states in the near term, but future escalation costs that risk toppling states could also prove more consequential. In the summer of 2014, the upsurge in violence in Iraq demonstrated just this risk. While the Iraqi regime was "initially willing to contain ISIL within Fallujah," 149 after its advances in 2013 in order to conserve focus on the 2014 electoral process, this insurgent threat escalated into an existential threat that rapidly took over much of North-western Iraq within a week and threatened to overtake the capital of Baghdad." 0 147 Marten, 2011: 105. For more, see Marten, 2011: 105-113. 148 "US Counterinsurgency Strategies in Iraq," Stratfor Global Intelligence, July 7, 2003, (http://www.stratfor.com/weekly/u s counterinsurgencystrategies_iraq). Accessed June 13, 2014. 149 Michael Knights, The ISIL's Stand in the Ramadi-Falluja Corridor," CTC Sentinel, 7 (5), May 2014, pp. 8-12. "s Ali A. Nabhan and Matt Bradley, "Islamist Insurgents Advanced Toward Baghdad," Wall Street Journal,June 11,2014. 58 VI. Codingy and Measuring the Dependent Variable It is clear that in some of the recent literature there is increasing convergence on idealtype counterinsurgency strategies, but a key challenge is how to identify what "type" was used in a particular campaign. I propose a rigorous approach to accurately identify or "code" a strategy using available and consistent measures. Measurement. This project makes use of a wide range of qualitative and quantitative data to identify and code an incumbent's dominant counterinsurgency strategy in a campaign. While quantifiable measures are in theory more objective and replicable, they are still crude proxies. Qualitative measures are more sensitive to the nuances and details in observations of complex events like strategy. In measuring strategy, I try to use both types of data to compensate for each other's weaknesses. The first step is to identify strategy based on qualitative measures of effort and violence using rich descriptive analysis of the conflicts from numerous sources for each country."' Along each dimension of strategy (effort and violence), I assign a description (low, medium, or high) corresponding to a score (0, 1, or 2) for each of the four key factors cumulating to a maximum score of 8 on both the effort (X) and violence (Y) axes. Effort is scored on four factors: 1) material (manpower and fiscal) resources invested, 2) major political or administrative reforms/reorganizations undertaken, 3) sustainment of operations in terms of breadth and tempo, and 4) consistent attempts to monitor and measure progress for adaptive learning. The strategy's level of violence is also scored along four factors: 1) collateral damage to civilians152 (including non-fatal damage such as arrests, torture, These sources include numerous descriptions and explanations offered by state elites and officials through news accounts, memoirs, interviews, and other sources detailed in the data sources section. 152 This includes both intentional and unintentional (collateral) victimzation of civilians. Since there is such a strong norm against direct civilian victimizaiton, particularly in the cold-war era, looking only at intentional targeting 151 59 displacement, and property-damage), 2) the relative focus on offensive, enemy-centric targeting, or kinetic operations over tactical substitutes, 3) (un)willingness to offer strategic accommodations, grievance redress, and/or political concessions, and 4) (dis)regardi5 3 for improving quality of life and public welfare for the affected population (in terms of safety and development).1 54 For each of the campaigns evaluated and scored along these measures, extensive documentation is provided in the appendix to support the judgment of each metric. Table 2.4: Measures of Counterinsurgency Strategies Effort Material (Manpower/Fiscal) Resources Political/Administrative Investments Operational Sustainment Adaptive Learning Violence Collateral Damage/Civilian Victimization Offensive Orientation Unwilling to make Accommodations or Concessions Disregard for Public Welfare Attrition High High Population Control High High Enfeeblement Low Low Mitigation Low Low High High Low Low High High High High High High Low Low Low Low High Medium Low Low Low Low High High Medium Low High High Low Low High Low High Medium The qualitative coding is complemented with quantitative measures of effort and violence (figure 2.5). Peak manpower deployed is measured to estimate effort and state-inflicted civilian casualties to estimate violence. Other data used to validate the value of the effort dimension includes fiscal resources expended per capita, the ratio of security force operations to insurgent activity, the duration of federal control or presence, the proportion of top-tier forces deployed, and cumulative or annual security force casualties. Data that helps validate the value would underestimate how much deliberate violence has been used (since incumbents are likely to chalk it up to unintentional violence, particularly for international audiences). 153 Note - 3 and 4 are phrased in the negative because higher scores correspond with higher violence 154 These qualitative factors were originally conceived of as 12 qualitative measures of effort and another 12 for violence, for which I consulted numerous accounts from primary and secondary sources. Later I condensed these down to 4 overarching factors for each dimension. 60 of the violence dimension includes exchange ratios (of security forces to insurgents), offensive to defensive operations ratios, levels of population displacement (and or repatriation), and finally ratios of rebels killed, arrested, surrendered, and rehabilitated. Figure 2.5: Sample Distribution of Dependent Variable based on Qualitative Coding 1,7 E -,Ree (A trition) -,7,6 me i TY 6,3 (Pol ation -4--conl 0 1 2 3 4 5 roTI) 6 7 8 Both the quantitative and qualitative values locate the campaign in a quadrant on the XYplane corresponding to an ideal type counterinsurgency strategy: attrition (high, high); population control (high, low); enfeeblement (low, high); and mitigation (low, low). In order to confirm the campaign's strategy, other correlates of that strategy would need to be observed including mission sets, targets, force types, operations scale, basing, intelligence assets/usage, and the type of political authority during and after the campaign (table 2.5). Decision Rule. The dependent variable of strategy will ultimately be coded based on where the qualitative measures plot the campaign on an scatter plot (XY) graph and checked for robustness based on the other three factors: 1) the location of the strategy in the quantitative scatter plot (XY) graph based on a number of critical quantitative measures as detailed above; 2) the type of political authority the counterinsurgency strategy most resembles; and finally, 3) the 61 presence of other tactical correlates (or "signatures") of the strategy. If these measures point to different strategies, the first decision rule will rely on what strategy the majority of the indicators converge, and if split, the second will rely solely on the first qualitative measure. Table 2.5: Tactical Correlates of Counterinsurgency S rategies Enfeeblement Sporadic "mowing the grass" operations; some high-value target and counter-network missions that disrupt organizational Mitigation Enlisting militias, warlords to counter insurgents OR flipping insurgents and have them "selfpolice" Attrition cordon and search, search and destroy, large-scale raids and sweeps Population Control Discriminate use of force with enduring, dispersed force presence; specific use of economic tools (and perhaps even some political concessions) to address grievances Primary Target Dismantling and degrading all hardcore insurgents (mobile combatants), local insurgents, and unorganized resistance Periodically striking leadership, hardcore insurgents (mobile combatants) and some local insurgents Balancing local insurgents and policing neutrals as well as unorganized resistance and supporters Primary Incumbent Forces Primarily government security forces (military) with local auxiliary forces Winning/eliciting support from unorganized supporters, neutrals, and unorganized resistance; Flipping and rehabilitating local insurgents; selectively targeting hardcore insurgents (mobile combatants) Government security forces coordinated with local security forces (police), auxiliaries, and population intelligence Limited government security forces (paramilitary) with some local militias or auxiliary forces Operations Scale large unit operations; airpower dispersed, small unit (company or platoon) operations Basing based on large outposts based on smaller, dispersed outposts Intelligence some tactical intelligence through coercion and defectors but little strategic intelligence heavily reliance on local forces, police, and population for intelligence inputs special forces missions (small units); air raids for tactical interdiction forces generally on large bases or not in the area of operation some tactical intelligence through coercion and defectors but little strategic intelligence Primarily local security forces (police) and selfdefense forces (tribes, militias, warlords, etc) Limited regular forces, backstopping irregular militia units forces generally on large bases or not in the area of operation outsource intelligence/informa tion problem to locals Mission Sets cohesion on motives on motives Dominant Strategies. While it is important to acknowledge that the strategies identified are not totally mutually exclusive, this will not pose a problem for tests of the theory. The sharing of some overlapping properties is a challenge anticipated for any articulation of ideal 62 types, particularly those of strategy, which is why some scholars have chosen to treat them as strategic "models" or "archetypes."' 55 While some tactics or "means" may overlap between strategies, how an incumbent organizes the political-military means-ends chain into a theory of victory will remain distinct.1 56 Like in other major scholarship on strategy, a dominant strategy, even amidst a mix of different properties, can be discerned for a number of reasons.1 57 First, strategies frequently involve tiers of importance and are identifiable even in messy contemporary conflicts.1 58 Second, in order for large organizations like state security forces to operate effectively and in a coordinated fashion, they require overarching guidance with at least a dominant strategic objective and approach.1 59 Third, there are inherent tradeoffs, costs, and risks to every strategy, which create strong disincentives for more than one to be employed co-equally in the same campaign (see Table 2). Fourth, what might be depicted as a heterogeneous strategy may more likely be a variation of strategy over time and can be captured as separate observations. Concerns may be raised that the incumbent may be divided along factional or civilmilitary lines with a shadow state or deep state pursuing a different track, but ultimately, what the incumbent actually does is what deserves analysis. This challenge of non-unified actors and "5 Jackson, 2008; Simpson, 2010. 156 I am combining the language and insights from both Posen, 1984: 13 and Posen, 1991: 71-72. 157 Mearsheimer, 1984; Posen, 1984; Pape, 1986. For recent works identifying strategy, see Jeremy Weinstein, Inside Rebellion: The Politics ofInsurgent Violence, New York: Cambridge University Press, 2007; Roger D. Petersen, Western Intervention in the Balkans: The Strategic Use ofEmotion in Conflict, New York: Cambridge University Press, 2011; Vipin Narang, Nuclear Strategy in the Modern Era, Princeton, NJ: Princeton University Press, 2014. 158 One recently declassifid government document on US counterinsurgency strategy in Afghanistan states that in response to the emerging in insurgency in 2005, the US Govt of Afghanistan's strategy had three different components but in ranked order: "first based on military action; second, on the Taliban reconciliation process; and third, improving relations and security cooperation with Pakistan." See U.S. Department of State, Issue Paper for Vice President of the United States (VPOTUS), "Counterterrorism Activities (Neo-Taliban)," December 9, 2005, Secret, 4 pp. [Excised] (from National Security Archive Electronic Briefing Book 325, document 14 [http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB325/]). 159 Admittedly there are cases of the United States pursuing two near-simultaneous yet different strategies, like in Vietnam or Iraq, but it is the outlier in this regard as few if any states have the capacity to do this. After all, the U.S. has the luxury of multiple branches each with their own separate air, naval, intelligence, and logistics capacities. 63 principal-agent problems is inherent to any study of strategy above the micro-level, but in cases of material threat and war, factions within the state are more likely to align.1 60 For this study, the presumption is that the dominant incumbent elites (whether it be civilian or the military) will prevail, fielding their preferred strategy and plan. Additionally, to avoid the worst instances of this problem, the scope conditions exclude cases of symmetric conflicts in periods of state collapse where incumbents are absent and leadership structures are fluid. Non-correlation ofMeasures. To ensure the measures and conceptual dimensions of strategy were not correlated across a broader universe of cases, I examined data on state-directed civilian casualties and manpower for 81 observations of civil conflict drawn from two different datasets.161 No significant correlation was found between the two dimensions, even when controlling for intensity (civilian casualties per conflict year) and population size in the area of operation. Additionally, taking the log measure of these two dimensions reveals substantial spatial variation (see figure 2.6) to suggest that all four quadrants combining high and low levels of violence and effort were empirically valid and well populated. Table 2.6: Correlation of Observables Correlation Coefficient Correlation Coefficient Correlation Coefficient Correlation Coefficient (Manpower, Civ (Manpower, Civ (Manpower/AO, (Manpower/AO, Cas): Intensity): Civ Cas/AO-Pop): Civ Intensity/ AO-Pop): 0.0928 -0.0183 0.0378 0.2127 Schweller, 2004; Posen, 1984. Friedman, 2011; Jay Ulfelder and Benjamin Valentino, "Assessing Risks of State-Sponsored Mass Killing," Unpublished Paper, Political Instability Task Force, February 2008. 160 161 64 - Figure 2.6: Distributions of Effort and Violence (N=81) [X=Log(Incumbent Manpower/Area of Operation); Y=Log(Civilian Casualty Intensity/Area Population)] 0.0 3.1 -6 .2 - - - -- - -3.0 - - - - - - - - -- - 2.0 -0.5 Log(Manpower/AO) VII. Scope and Definitions At this point, it is important to define some of the terms and variables used in this project along with the scope conditions of the theory. Definitions In this study, the dependent variable-counterinsurgency strategy-is an incumbent's strategy within a political-military campaign against an insurgency. Strategy is defined as, "an actor's plan for using armed forces to achieve military or political objectives."1 62 The objectives of strategy are to counter the challenge of an insurgency, "a protracted violent struggle by nonstate actors to obtain their political objectives-often independence, greater autonomy, or subversion of existing authorities-against the current political authority, the incumbent."' 63 162 Arreguin-Toft, "How the Weak Win Wars: A Theory of Asymmetric Conflict," International Security, 26 (1), Summer, 2001, p. 99. 163 Lyall and Wilson, 2009, p. 70. 65 Finally, the campaign itself is "a phase of a war involving a series of operations related in time and space and aimed towards a single, specific, strategic objective or result in the war."1 64 Scope To address some of the weaknesses and gaps of the literature described above, I specifically focus on how incumbents contend with insurgencies, which are a narrower subset of civil wars and include most self-determination movements that turn violent. Therefore, the unit of analysis in this study is a counterinsurgency campaign against a non-state actor that at least partly relies on guerilla warfare.1 65 This scope primarily seeks to focus on asymmetric or irregular conflicts while excluding purely symmetric conflicts such as conventional civil wars (Nigeria (Biafra) 1967-70), coups (Argentina 1955) or cases when the state has entirely collapsed and a dominant incumbent is absent (Afghanistan 1996-01). 166 However, this would still include most (though not all) multi-party civil wars so long as there is a dominant incumbent.1 67 A campaign begins when the state incumbent activates or deploys some non-local forces against the violent non-state challenger. Campaigns within a country may be considered spatially distinct if they are conducted in different spaces-that is territories, regions, or polities-against distinct insurgent organizations. They may also be considered sequentially distinct if they are 6 T.N. Dupuy, Understandingwar: History and Theory of Combat, Leo Cooper, London, 1992. He goes on to write, "A campaign may include a single battle, but more often it comprises a number of battles over a protracted period of time or a considerable distance, but within a single theatre of operations or delimited area. A campaign may last only a few weeks, but usually lasts several months or even a year." 165 This is the definition used by similar to the unit of analysis used by Lyall and Wilson, 2009; Johnston, 2009; and Friedman, 2011, p. 568. Ivan Arreguin-Toft's dependent variable, the "strategic interaction," of which there is generally one but there may be multiple-sequential interactions in which one conflict could generate more than one observation with a strategy shift. I will employ the same approach of separating out campaigns within a single conflict dyad if there is a distinct strategy shift and hence a new campaign. 166 For the distinction between symmetric and asymmetric civil wars, see Stathis N. Kalyvas and Laia Balcells, "International system and Technologies of Rebellion: How the End of the Cold War Shaped Internal Conflict," American PoliticalScience Review, 104 (3), August 2010, pp. 415-429. When comparing Lyall and Wilson's (2009) dataset to Kalyvas and Balcells, Lyall and Wilson include 9 of 18 symmetric non-conventional conflicts and 37 of 50 conventional conflicts, presumably since they have irregular elements or periods to them. 167 Christia, 2012. 66 conducted in the same space but if an incumbent dramatically changes strategy over time, effectively initiating a new campaign.1 68 My argument's scope conditions and unit of analysis generally comport with those offered by Lyall, Friedman, and Johnston as well as others, but with one exception. In the event the conflict fails to reach the battle death threshold specified by Lyall and Wilson (1,000 battle deaths with at least 100 casualties suffered on each side), but the incumbent explicitly believes the conflict to be a counterinsurgency campaign (based on official publications or statements), and this judgment is supported by outside observers, it was counted within the universe of cases and include it amongst an incumbents' counterinsurgency campaigns. For instance, British counterinsurgency campaigns in Oman, the Canal Zone, and Nyasaland may not have crossed the battle death threshold but they are nevertheless important and formative cases in pantheon of British counterinsurgency. 169 There is good reason to do this because while the battle death threshold is useful, it is inherently endogenous to the dependent variable: counterinsurgency strategy. The effectiveness of a strategy could cause the insurgents to concede early before the conflict escalates past the threshold, or the nastiness required of a strategy to achieve victory could cause the state to exit early. These are both strategic choices and fully adhering to the battle death threshold would select out or "left censor" the universe of cases. There is good justification to censor this data in cross-national studies of counterinsurgency efficacy, but it is important they be included for studying subnational variation of counterinsurgency approaches. 168 169 This approach to distinguishing campaigns is drawn from Arreguin-Toft's approach, 2001; 2005. French, 2011. 67 VIII. Conclusion This chapter has presented the case for a richer theorization of counterinsurgency strategy given its absence in the empirical study of civil conflict and insurgency despite clear historical evidence of variations in the approach of the same capable incumbent across different campaigns and/or theaters. The two critical dimensions of the campaign-effort and violence-have been central, recurring concepts in the strategic theorization of grand strategy, military doctrine, conventional strategy, coercion, and even counterinsurgency. A typology of counterinsurgency is derived from these two dimensions yielding four ideal-type strategies that are frequently identified in the literature: attrition, population control, enfeeblement, and mitigation. Each of these strategies is based on a combination effort and violence, measurable in qualitative and quantitative terms, in conjunction with a distinct theory of victory, an approach to political authority, and several tactical correlates. The next chapter offers a core-periphery theory based on the independent variables of territory and identity to explain the varying counterinsurgency strategies of an incumbent. 68 Chapter 2 Appendix Table A2.1: Overlap of Different Strategy Typologies Author Typology Counterinsurizencv Lalwani, 2012 [Pure To Conceptualize Types, focus on... Demand Side Strate Mitigation ies Population Control Enfeeblement Suply Side Strategies Attrition Concessions/ [Collective Effort, &Violence Punishment]* Non-Military]* Observables to Measure Concept Money, Manpower, Material, Sustainment, Reforms, Monitoring; Civilian and enemy casualties, targeting, Watts et. al. 2014 Lindsay & Petersen, 2011 Homogenization Informal Classic Accommodation Counterinsurgency Tribal Mobilization Johnston, 2009 Containment Strong-State Repression Clear, Hold, Build Decapitation Persuasion Coercion Annihilation Political Accommodation; political concessions, public welfare Political Accommodation; Public Goods; Discriminate Public Goods; Discriminate Violence Violence Concentration of effort along certain nodes of role spectrum within conflict Individual Triggering and Sustaining Mechanisms Civilian Targeting total conflict casualties; dummy variables for civilian victimization and mass Jackson, 2008 Model 3 - "violent negotiation" Simpson, 2010 Model 2 population control" Local Security Model killing 1- "small war" Targeted Killing Raiding Mass Killing Assumptions about significance, agency, and motivation of insurgents and population "routines that symbolize these abstract foundations" (i.e. Lots of Tactics, Assessments, Statements of Leaders) Approach & Information GDP/capita, road density, mortality, relative political capacity Chowdhry and Krebs, 2009 Policy Concessions Bolster Moderates Stanton, 2009 Hearts and Minds Restrained Control Repression Alone Restrained Coercion (no clear theory) Cleanse Coercion Kalyvas, 2006 Selective Violence Indiscriminate Violence Selectivity of Violence Denunciations vs. Random Killings Conventional Strategv (Analogs) Posen, 1984 Mearsheimer, '83 Defense Limited War Offense Positional Defense? Stam, 1996; Reiter & Meek '99 Pape, 1996 Denial Arreguin-Toft '05 Indirect Blitzkrieg Attrition Maneuver Attrition Deterrence Punishment Punishment Direct Case Studies Case Studies Indirect Maneuver Forces/Capabilities Targeting Targeting percentage of forces that are mobile Target key assets or Values (Physical Capacity) or Target Population (Will) -* Figure A2.1: Distribution of British COIN Cases -1.9 Oman 4 Ken a I -19 *Oman Arab Revolt enya 4 Palestine (Arab Revolt) (Palestine}) *Aden * Born o ii Mai .5 * Aden ayas * MIaya Borneo, onits * Pa lestine) 'I I * -5.1 N. irelant Canal Zone Nyasaland -- - --1.6 lestine (Zionists) Canal Zone * - - -5.1 0 . reland -- Ntasaland '"--_-_-_------ - - -0.5 L6 0.75 Lodflops/Sq Km) 2 LgMmops/ftbe) (A) (B) N 1.6 L * Nyasaland Arab Revolt * (Palestine) .AW 0.5 W6 0.5 NvasaandJ Palestine I Arab Revolt) #Omati Aden O Adent Canal Zone Borneo Cyprus Cyprus I Palestine Zwionist reland, &Palestine) 4.6 (Zionists) N -1.6 -0.6 C L6 Canavone -0.5 0.75 Log(Troaps(Sqc m) tog(TroepsRebeQ (C) (D) 4N freland - . __- __ W Table A2.2: All Properties of Counterinsurgency Strategies (Means-Ends Chain, Tactical Correlates, Tradeoffs, and Measures) Goal (Ends) Theory of Victory (Means) Functional Form Supply-Demand Orientation (Leites and Attrition Population Control Enfeeblement Mitigation Control (Rollback) through Containment through Disruption and Punishment Interdiction and Deterrence of insurgent organization's ability to violently threaten state authority along with harassment and disruption of insurgent central nervous system (leadership decapitation), until violence is limited below a tolerable threshold. Containment through Tactical Denial and Co-optation Use untrained local forces and auxiliaries to manage violence until it wears down and sinks below a tolerable threshold. Side payments are used to co-opt local power brokers to back the state or "flip" sides and manage the region. Linear Returns Control (Rollback) through Denial and Inducement Controlling the population to deny insurgents a target and base of support; gradually draining insurgent strength through violence as well as the capture, surrender, flipping, and reintegration of insurgents; harnessing extensive information for selective targeting to hold down violence and avoid excess collateral damage; and persuasion and bargaining with public goods, political concessions, devolution of authority, and rehabilitation of the contested area Increasing Returns Decreasing Returns Decreasing Returns Supply-side Demand-side Supply-side Demand-side Direct Direct Indirect Indirect Usurpation Powersharing Administrative Occupation Non-incorporation Attrition Population Control Discriminate use of force with enduring, dispersed force presence; specific use of economic tools (and perhaps even some political concessions) to address grievances Winning/eliciting support from unorganized supporters, neutrals, and unorganized resistance; Flipping and rehabilitating local insurgents; selectively targeting hardcore insurgents (mobile combatants) Enfeeblement Mitigation Sporadic "mowing the grass" operations; some high-value target and counter-network missions that disrupt organizational cohesion Enlisting militias, warlords to counter insurgents OR flipping insurgents and have them "selfpolice" Periodically striking leadership, hardcore insurgents (mobile combatants) and some local insurgents Balancing local insurgents and policing neutrals as well as unorganized resistance and supporters Disarm and Punishment Direct annihilation and military destruction of adversary's combat power and capacity to resist; inflict such unrelenting pressure and pain on the insurgent and base that eventually, it becomes obvious that victory is impossible, and insurgents are wiped out or cease fighting Wolf) "Approach" (Liddell Hart) Political Authority (Boone) Tactical Correlates Mission Sets Cordon and search, search and destroy, large-scale raids and sweeps Primary Target Dismantling and degrading all hardcore insurgents (mobile combatants), local insurgents, and unorganized resistance Primary Incumbent Forces Primarily government security forces (military) with local auxiliary forces Operations Scale Large unit operations; airpower Government security forces coordinated with local security forces (police), auxiliaries, and population intelligence Dispersed, small unit (company or platoon) operations Limited government security forces (paramilitary) with some local militias or auxiliary forces Primarily local security forces (police) and self-defense forces (tribes, militias, warlords, etc) Special forces missions (small units); air raids for tactical interdiction Forces generally on large bases or not in the area of operation Some tactical intelligence through coercion and defectors but little Limited regular forces, backstopping irregular militia Basing Based on large outposts Intelligence Some tactical intelligence through coercion and defectors but little Based on smaller, dispersed outposts Heavily reliance on local forces, police, and population for strategic intelligence on motives intelligence inputs strategic intelligence on motives Attrition 1) Material (manpower and money) trade off with other regime priorities or threats; 2) Risk of Regime loss/collapse: excess force can benefit insurgents by a) solving collective action problem, b) attracting foreign support from foreign rival or diaspora, or c) drawing international intervention; 3) International Reputation/Normative (Pariah) costs Population Control Enfeeblement 1) Material (manpower and money) trade off with other regime priorities or threats; 2) Political/Regime Costs: Counterinsurgent has to offer goods, concessions, or accommodations that at best alienate a political base and at worst threaten to undo critical institutions of the regime 1) Extended Duration of Conflict (and risks of future escalation or contagion); 2) Preservation of dysfunctional, under-productive regions (Prolonged violence, lawlessness, criminal activity) units Forces generally on large bases or not in the area of operation Outsource intelligence/ information problem to locals Tradeoffs Costs Mitigation 1) Extended Duration of Conflict (and risks of future escalation or contagion); 2) Preservation of dysfunctional, under-productive regions (Prolonged violence, lawlessness, criminal activity); 3) Principal-Agent Problem and empowerment of new violent non-state actors Affordable, low-cost stability (durable disorder) with few concessions; Some internal deterrence Affordable, low-cost stability (durable disorder) Population Control Enfeeblement Mitigation High Low Low High High Low Low High High Low Low Adaptive Learning High High Low Low Low Low Violence High High High Low High Low Benefits Long-term stability; Forced Integration with no concessions; Internal deterrence Long Term Stability; Integration of embattled region(s) Attrition High Measures Effort Material (Manpower/Fiscal) Resources Political/Administrative Investments Operational Sustainment (Ni Collateral Damage/Civilian Victimization Offensive Orientation Unwilling to make Accommodations or Concessions Disregard for Public Welfare High Low Medium Low High Medium High Low High Low High Low High Low High Medium Chapter 3 Explainin2 Strategy with a Theory of Core-Periphery Relations I. Introduction This project advances a theory of core-peripheryrelationsin which an incumbent's prewar relationship with different contested territories and the incumbent's pre-war relationship to the rebellious sub-national group identity will shape counterinsurgency strategy. Specifically the value and accessibility of territoryas well as the positional status of the rebel identity group, based on its worth and embeddedness, all shape and constrain counterinsurgency strategy. The starting point of the core-periphery theory is that the geopolitical and human terrain of a country is uneven, differentiated, and prioritized by an incumbent into core and peripheral regions and groups. Meanwhile, counterinsurgency campaigns are costly endeavors-in distinct, rather than correlated, dimensions of material resources as well as violence meted out. Certain territories and peoples are more valuable to the incumbent than others, thus incentivizing or constraining the strategy adopted to counter a rebellion. To explain these varying strategies, the theory of coreperiphery relations offers a set of predictions with respect to levels of effort and violence in counterinsurgency, and therefore strategy. The theory first argues that state incumbents will exert more effort and commit greater resources when facing rebel challenges that threaten territories of high value or its "core," while they will exert lower levels of effort to fight challenges in territories that constitute its "periphery."' Second, the theory argues that state incumbents facing rebellion will be 1This distinction stems from Immanuel Wallerstein, The Modern World System: CapitalistAgriculture and the Origins ofthe European World Economy in the Sixteenth Century, New York: Academic Press, 1974. This paradigm has since been adopted by a number of scholars for the purpose of explaining spatial dynamics in political geography. For instance, see Michael Hechter, Internal Colonialism: The Celtic Fringein British National Development, 1536-1966, Berkeley, University of California Press, 1975; E. Spencer Wellhofer, "Core and Periphery: Territorial Dimensions in Politics," Urban Studies, 26 (3), June 1989, pp. 340-355. 74 unconstrained and willing to employ greater violence against peripheral identity groups that are of low worth and un-embedded within state institutions, but will be much more circumspect against identity groups part of or close to the "core group," possessing high worth and embedded in state institutions. 2 This chapter lays out the core-periphery theory's specific hypotheses on strategy and specific mechanism that link territory and identity to a counterinsurgent's use of effort and violence. Section two provides some initial background and context to the theory. Section three details the territory hypotheses, measurements, and mechanisms, while section four does the same for identity. Section five combines the two parts of the theory to offer predictions and then considers alternative explanations. Finally section six offers the research design to test this theory and guide the rest of this project. II. Background and Context This theory stands in contrast to much of the civil conflict literature that has generally discounted the role of territory and identity in conflict dynamics and state responses. In crossnational studies of reputation, the value of territory has been found to be empirically wanting,3 but these studies have focused on conflict onset rather than dynamics and employed a very narrow conceptualization of territory to its extractable or strategic value in absolute terms. What has not been seriously considered is how the relative importance of territory and the type of state political authority wielded over that region, its "political topography," might incent or constrain the type of state response to rebellion. The core group is "the members of the ruling political organization that has the military and administrative capacity to enforce its decisions within the borders of the state." See Mylonas, 2013, p. 23. 3 Toft, 2003; Walter, 2009. 2 75 Much of the cross-national empirical research on civil conflict has also dismissed the role of identity in conflict onset, dynamics, or state strategy. 4 But much of this might be a result of poor operationalization or under-specification of identity that is beginning to be re-evaluated by other scholarship.5 A more fine-grained examination not simply of general diversity but specific ethno-political hierarchies, sub-group status, and who is fighting whom offers evidence to the contrary. In direct contrast to this conventional wisdom, my core-periphery theory offers novel predictions and mechanisms for how both territory and identity, re-specified, can do much to explain sub-national variation. By re-conceptualizing and specifying these variables, I use them to construct a semi-exogenous theory of state counterinsurgency strategy. The Uneven-ness of Political Topography A central premise of this theory is that states do not have uniform relationships with its territorial units or identity sub-groups, and that significant subnational variation exists. A combination of literature fixated on third-party foreign incumbents and a focus on the macro and the micro at the expense of the meso-level have left scholars inattentive to the subnational variation a state's uneven relationship to its territory and its people, and the challenges that arise from it. As third-party counterinsurgents struggle with problems of commitment-since their interest is peripheral relative to the defense of the homeland-and with problems of information 4 Collier, Paul, and Anke Hoeffler, "Greed and Grievance in Civil War," World Bank Policy Research Working Paper 2355, World Bank, 2000; Fearon and Laitin, 2003; Fearon, 2004; Walter, 2009. s Roger Petersen, UnderstandingEthnic Violence: Fear, Hatred, and Resentment in Twentieth-CenturyEastern Europe, New York: Cambridge University Press, 2002; Evan S. Lieberman and Prerna Singh, "The Institutional Origins of Ethnic Violence," ComparativePolitics, 45 (1), October 2012: 1-24; Lars-Erik Cederman, Andreas Wimmer, and Brian Min, "Why Do Ethnic Groups Rebel? New Data and Analysis," World Politics, 62 (1) 2010, 87119; Kanchan Chandra and Steven Wilkinson, "Measuring the Effect of 'Ethnicity'," ComparativePoliticalStudies, 41(4-5), 2008: 515--63. 76 and legitimacy resulting from having a uniformly different identity (by definition, a third party),6 it is often assumed that domestic incumbents do not face these challenges. In fact, an often overlooked feature of domestic incumbent counterinsurgents is that sometimes they do encounter similar challenges as third-parties, and at other times don't, a product of their uneven relationships to sub-territories and sub-groups. States view and exercise authority over their territories unevenly, particularly over "frontier zones." 7 Historically, the British exerted uneven patterns of authority on its colonial holdings with some areas ruled directly and others ruled indirectly.8 In many areas, where the antebellum U.S. frontier remained ambiguous and entwined with overlapping sovereignties and quasi-sovereignties, the lines blurred between public and private coercion to police and enforce authority.9 There are even fewer disincentives to rule unevenly in the modem era. The landscape of post-1945 international politics includes norms of fixed borders and territorial integrity, sovereignty, and international state system, all propped up by institutions like the UN, which have nearly eliminated the risk of state death.' 0 One implication is that weak states are able to persist without having to do much, but this also allows for a capable state's weak grip over certain territories to persist as well. States can highly value some territories while others are deemed peripheral, lumped into their polity by artifact of colonial legacies and retained by Simpson, 2010; Hazelton, 2011; Keith Darden and Harris Mylonas, "The Promethean Dilemma: Third-party Statebuilding in Occupied Territories," Ethnopolitics, 11 (1), March 2012, pp. 85-93; James Lebovic, The Limits of US. Military Capability:Lessonsfrom Vietnam andIraq, Baltimore: The John Hopkins University Press, 2010. 7 Boone, 2003, p. 21. 8 Lakshmi Iyer, "Direct versus Indirect Colonial Rule in India: Long-term Consequences," Review ofEconomics and Statistics 92 (4), November 2010, pp. 693-713. 9 Ira Katznelson, "Flexible Capacity: The Military and Early American Statebuilding," Shaped by War and Trade: InternationalInfluences on American PoliticalDevelopment, Eds. Ira Katznelson and Martin Shefter, Princeton, N.J.: Princeton University Press, 2002, p. 101. 10 Herbst, 200; Fazal, 2007; Atzili, 2011; Griffiths, 2010. 6 77 inertia. Uneven patterns of value and authority across territory can thereby influence state strategies to address threats to particular territories. A less controversial claim is that states treat different identity sub-groups in an uneven manner. As a result, different identity sub-groups can possess different levels of economic and political power," able to hold states accountable differently,' 2 even be formally subjected to different laws (even in democracies), possess different organizational capacity,1 3 attract different levels of suspicion,14 and get integrated or excluded from a polity in varying ways.' Though the state is theoretically the product of a social contract, in societies with multiple identity or ethnic groups, 16 states-and their core institutions-rarely represent the people evenly in terms of their composition and interests. While third-parties are uniformly seen as outsiders (and often short on information and legitimate authority), a domestic incumbent by virtue of its localness is thought to safely bypass these problems. In fact, a domestic incumbent may be unevenly connected with certain sections of its population possessing better information, authority, or sensitivity towards some groups over others. Nor are all minorities equal as some may be valued and tethered to the state and elites while other groups remain marginalized and disenfranchised. This may owe to both historical contingencies like the fragmented polities inherited from colonial regimes as described above, as well as active cultivation by states. Either way, important institutions like security forces might become dominated by particular groups " Donald L. Horowitz, Ethnic Groups in Conflict, Berkeley: University of California Press, 1985. 12 Lily L. Tsai, Accountability Without Democracy: Solidary Groups and Public Goods Provision in Rural China, New York: Cambridge University Press, 2007. 13 Petersen, 2001; Staniland, 2014 14 For instance, the Japanese-Americans were legally treated with far greater suspicion that Italian-Americans during WWII. See Mary Elizabth Basile Chopas, "Law, Security, and Ethnic Profiling: Italians in the United States during World War II," PhD Dissertation, University of North Carolina, Chapel Hill, 2013. 15 Mylonas, 2013. 16 Following Horowitz (1985) and Liberman and Singh (2012), I use "ethnic group" for all groups based on ascriptive identity including race, language, religion, tribe, or caste. See also Ashutosh Varshney, "Nationalism, Ethnic Conflict, and Rationality," Perspectives on Politics, 1 (1), March 2003, pp. 85-99. 78 with identity allegiances and preferences that may play out in the actual conduct of conflict. As one scholar writes, "Placing the identity of perpetrators and victims at the center of our studies would open several new avenues of research."' 7 There is good reason then to think that amongst this uneven milieu of subnational identity groups, the state can selectively apply different types of treatment to different sub-groups that violently challenge their authority. III. Independent Variable 1: Territory and Effort The first dimension of the core-periphery theory examines the nature of threatened territory-whether it is core or peripheral to the state-and the effect of this on the state strategies of effort in counterinsurgency. The starting point for this argument is that states that are faced with rebellion are generally weak to begin with. These incumbents are often politically divided, beset by contentious politics, and face competing priorities from development challenges to major external threats, all of which may limit the resources the state can allocate to a costly, prolonged, and tedious counterinsurgency campaign that can stir a hornets' nest or devolve into a quagmire. Control of territory can provide material utility and strategic resources central for a state's physical survival,' 8 but these resources are not evenly distributed throughout a country or state's territory, with some regions offering few or none of these resources. Furthermore, if control of territory only pays off under certain conditions in industrial societies, then it will offer comparatively few returns in non-industrial, peripheral zones with far less productivity.19 17 Lyall, 2010a, p. 18. Toft, 2003, pp. 19-20. 19 Peter Liberman, Does ConquestPay? The Exploitation of OccupiedIndustrialSocieties, Princeton NJ: Princeton University Press, 1996. 18 79 There are a host of explanations as to why a state initially chooses to fight and/or selects into conflict (e.g. reputation or concentrated threats), but these can be distinct from what motivates a state to escalate to high effort strategies that pour significant resources in the form of money, manpower, and materiel, and make costly political and administrative actions over a sustained period of time. States will choose a decisive strategy to regain control when they believe the payoff of victory is worth the costs of the fight. Ultimately, my argument is that this boils down to a cost function that weighs the benefits of a high effort campaign to achieve control (or the costs of feeble effort to achieve containment) relative to the direct costs to conduct and sustain such a counterinsurgency strategy. Core territory fosters stronger incentives than peripheral territory to motivate high effort campaigns through material, strategic, ideational and psychological incentive mechanisms. Moreover, because of the degree of pre-war integration and development, core territory also imposes fewer operational costs than peripheral territory through distance, physical terrain, and human terrain operational cost mechanisms. It is perhaps because of these incentives and permissive conditions that conflict on core territory is on average of shorter duration. 20 These insights help us generate the following hypotheses: Hla: If a given incumbentfights a rebel group challenge in lower value peripheral territories,then the incumbent will be more likely to choose lower effort strategies of containmentlike enfeeblement and mitigation. Hib: If a given incumbentfights a rebel group challenge in higher value core territories, then the incumbent will be more likely to choose higher effort strategies of control like attritionandpopulation control. H Ic: If a given incumbentfights a rebel group challenge in lower value peripheral territories, but the rebel group is also able to project substantialforce on to higher value core territories,then the incumbent will be more likely to choose higher effort strategies of controllike attrition andpopulationcontrol. Note this finding is not statistically significant in their original article but it is in a correction the authors published on January 13, 2012, using the correct specification of the hazard function. See Corrigendum for Cunningham, Gleditsch, Salehyan. 2009 (http://privatewww.essex.ac.uk/-ksg/papers/CorrigendumCunninghametal2009.pdf). 20 80 Measurement Before detailing these mechanisms in depth, it is important to explain how I measure core and peripheral territory. Previous studies tend to rely on a dichotomous treatment of distance or borders as the measure of the peripheral, but the identification of peripheral territory is more complex than a computation of distance from the capital or dichotomous identification of borderlands. 2 ' Basra is much further from Iraq's capital of Baghdad but far more important to the country than the backwaters of al-Anbar province that literally border the capital. Likewise, Mexico's borderlands with the U.S. may be far from the capital but they are incredibly valuable for the economic value generated by manufacturing and trade of licit (and illicit) products. Thus an integrated index or composite score can be created based on accessible quantitative data to measure peripheral from core regions. The values of different regions would be calculated based on: 1) levels of production (agriculture, industry, resource extraction like energy and mining, and tax revenue); 2) levels of physical infrastructure (access to electricity and water, road density); 3) levels of social infrastructure (human development; health and education indicators); and 4) degrees of interconnectedness to other parts of the country including the core (communications technology density; frequency of news articles about the region). These indicators can be measured for all administrative units or regions at the very least for ordinal comparisons. On a relative scale, core territories would score high on these measures and peripheral territories would score low. Two other indicators of territory type-strategic value and ideational exclusion-can only be identified qualitatively. Strategic value can be accounted if there is a consensus on the strategic priority of a region for purposes of national defense. Ideational exclusion can be 2 Halvard Buhaug, Scott Gates, and Paivi Lujala, "Geography, Rebel Capability, and the Duration of Civil Conflict," Journalof Conflict Resolution, 53 (4), August 2009, pp. 544-569. 81 identified in a number of ways, including how frequently the region features in national discourse. Another way to identify the ideational periphery is if there is a dejure designation of unincorporated territory or indirect administrative rule where normal patterns of authority or rule of law do not apply (e.g. U.S. rule of Puerto Rico or Guam). Without formal designation, de facto indirect administration can be also be identified by close scholarly research (e.g. the type of indirect state authority wielded2 2 ). Mechanisms States will exert more effort in core territories than peripheral territories because the incentives for success are greater and operational costs affecting the probability of success are lower. Three types of incentive mechanisms are identified-material, strategic, and ideationaland three types of operational cost mechanisms-distance, physical terrain, and human terrain. Incentives. First, since not all territory is created equally, the relative value, utility, or importance of the territory under threat can provide incentives for greater effort (and/or raise the costs of feeble action). On average, core territory will spur greater effort through the four mechanisms of material, strategic, ideational, and psychological incentives, though material and strategic are likely the strongest. First, material incentives are central as states are generally interested in what they can extract from a territory.23 If the embattled or threatened region is valuable to the state in terms of economic production due to the presence of industry, capital, resources regularly being Adnan Naseemullah and Paul Staniland, "Varieties of Indirect Rule: Explaining Governance Beyond Weberian Sovereignty," Governance, forthcoming. 23 Margaret Levi, OfRule and Revenue, Berkeley, CA: University of California Press, 1988. 22 82 consumed (e.g. agriculture or water supply), 24 allowing this region to languish amidst a rebellion will prove extremely costly and motivate the state to expend considerable efforts in terms of money, manpower, and materiel to regain control of the region rather than merely contain the insurgency. By contrast, a region's low or non-incorporation into national patterns of economic production25 would offer few incentives for a high degree of effort. Material incentives can also be based on areas of high population density and human capital. As the US tried to manage resources in Afghanistan, it inevitably ceded ground to insurgents but predominantly in sparsely populated areas. 2 6 Second, the strategic value of a region may motivate the state to considerable action in interest of its survival. States are interested in survival and, particularly in acute threat environments, even a small a rebel threat to strategic territory can motivate a high effort response by the state because the rebel threat is compounded by the vulnerability of the territory that, if compromised, could have far greater implications than rebellion. Certain regions or geographic terrain are seen as particularly important for the strategic defense of a country for their natural barriers (e.g. mountain ranges), to maintain lines of communication, or as a buffer zone. For instance, Israel views the Golan heights as an essential fortification of its national integrity. A simple count of strategic assets may not be able to weight the relative strategic importance of regions. A measure of strategic value based on a composite score of any territory with borders, bases, ports, transport nodes, mountains, or attack routes 27 may render the concept all-encompassing and therefore meaningless. Strategic value must go beyond asset counts to get There is a difference in the costliness of regular productivity that a state relies on day-to-day and unexploited resource wealth that offers future returns. 25 Catherine Boone, "Decentralization as Political Strategy in West Africa," Comparative PoliticalStudies, 36 (4), May 2003: 365; Catherine Boone, PoliticalTopographiesof the African State: TerritorialAuthority and InstitutionalChoice, New York: Cambridge University Press, 2003. 26 Alissa J. Rubin, "Afghans' Distrust Threatens U.S. War Strategy," New York Times, May 12, 2010. 27 Walter, 2009, p. 73. 24 83 at something central to the state's ability to survive or at least defend itself. For instance, borders are generally important but Iran's border with Iraq is of far more strategic importance to Tehran than its border with Turkmenistan, since it fought a war on this border. Until 2013, Ukraine had not fought a war with Russia in the modem period but nevertheless this border was of greater strategic importance than its border with Poland whom it had fought. Third, territory or region can have a certain ideational, moral, or symbolic value for the identity of a nation-state and emotional attachment to it can spur it to dig in harder in the face of significant threats. Sometimes offered as the explanation for issue indivisibility, some regions might have significance to a polity, whether inherent or socially manufactured, that is sacred or central to their national or civic identity and can generate resources for mobilization, political actions, and conceivably, determined effort.2 8 For instance, while Sunni extremist violence had been ongoing for years after the U.S. invasion of Iraq and targeted a number of critical material assets, the bombing of one of the holiest sites in Shia Islam, the al-Askari Mosque in Samarra, which neither killed nor injured anyone, triggered outrage and a major escalation by the Shiadominated Iraqi state (and sectarian militias) outraged by the symbolism of the attack.2 9 Years later when Iraq was facing the rise of a new Sunni insurgency, it concentrated its efforts on protecting Shiite areas first and worrying about other regions as a secondary matter. 30 Ideational value may not be correlated with distance. Some distant territories might be more formally integrated, whether administratively or in day-to-day interactions with the rest of the country, while more proximate ones are not. Alaska and Hawaii are both less populated and Toft, 2003; Stacie E. Goddard, "Uncommon Ground: Indivisible Territory and the Politics of Legitimacy," InternationalOrganization,60 (1), Winter 2006, pp. 35-68; Ron E. Hassner, "'To Halve and to Hold': Conflicts over Sacred Space and the Problem of Indivisibility," Security Studies, 12 (4), 2003, pp. 1-33. 29 Robert F. Worth, "Blast Destroys Shrine in Iraq, Setting Off Sectarian Fury," New York Times, February 22, 2006. 30 Alissa J. Rubin and Michael R. Gordon, "Iraq's Military Seen as Unlikely to Turn the Tide," New York Times, June 23, 2014 2 84 more distant from the U.S. mainland and capital than Puerto Rico but they are still much more politically, economically, and culturally integrated into the United States. Though generally endogenous to the first three types of incentives, prior pre-war integration of the embattled region into the mainstream or day-to-day activities of a country's core might make a state more averse to losing control and settling for containment and therefore incent much more costly efforts to regain full control of territory that has fallen into contestation. By contrast, regions that were not well-integrated pre-war will not trigger the same concern, allowing the state to choose low effort strategy. Integration does not happen naturally and requires active state investments even prior to the conflict. It is important to point out that peripheral territories of the country are not valueless in absolute terms and may have extractable or strategic resources yet still may not be valued by the state to the same degree as its core regions that are more productive and better integrated. Correspondingly, a state be interested in a nominal militarized defense of these territories rather wholesale abdication of territory, but perhaps unwilling to commit to an intensive counterinsurgency campaign for something not worth the cost-a distinction lost on most scholarship focused on a state's decision whether to fight rather than how to fight. For instance, Indonesia marshaled almost the entirety of its domestic forces to achieve a decisive victory over Darul Islam in West Java because it "directly threatened the political survival of the Republic of Indonesia., 32 By contrast, though it did choose to fight the rebels in much more distant but resource rich West Papua, Indonesia deployed far less manpower and effort, with low quality and quantity of manpower acknowledged by local authorities,33 and the James C. Scott, Seeing Like a State: How Certain Schemes to Improve the Human ConditionHave Failed, New Haven, CT: Yale University Press, 1998. 32 Kilcullen, 2010, p. 85 33 Martin O'Hare, The IndonesianMilitary in IrianJaya, MA Thesis, Australian National University, 1991, p. 53. 31 85 military remained "relatively invisible" and "directed toward preserving the status quo,"3 4 treating the rebels as a "minor nuisance" and " leaving the guerillas... to rot in the jungle."3 5 Operational Costs. Even if territory offers modest value to a state, the embattled territory can shape the direct costs of a state's operation or campaign by enabling or obstructing the projection of military and political power. Combined with high operational costs, a territory of modest value may not warrant a high effort strategy. Generally peripheral territory imposes higher operational costs than core territory through the three mechanisms of distance, physical terrain, and human terrain. Projection of force to a distant region imposes one type of physical access cost. 36 The projection of force itself beyond its home base is both costly and manpower intensive as tremendous resources need to be allotted not only to the pointy end of the spear but also to logistics and supply. Therefore the loss-of-strength gradient will be steeper over greater distances3 7 and limit the exercise of political and military dominance. Thus, Gilpin writes, "the ability of a political center to radiate its influence is affected significantly by the cost of transportation."38 In addition to distance, the projection of force and state reach can be further encumbered by another physical access cost of the region, specifically the terrain and pre-war infrastructure. Insurgencies are, on average, correlated with difficult terrain, whether mountainous or jungle terrain but insurgents can operate elsewhere. Even the average rolling plain offers sufficient 34 Reynders, "Consular Trip to West Irian," Confidential Airgram, May 10, 1968, National Security Archives (http://www2.gwu.edu/-nsarchiv/NSAEBB/NSAEBB128/) 3 Quotation from Indonesia Defense Minister General Mohammad Jusuf in 1978. From Richard Chauvel and Ikrar Nusa Bhakti, The PapuaConflict: Jakarta'sPerceptionsandPolicies: Policy Studies 5, Washington DC: East West Center, 2004, p. 23 36 For colonial powers, Fearon, 2004 controls for whether the conflict is fought in the core/metropole or the ?eriphery/colony to get at this cost of distance, but he does not do this for regions/territories within states. 7 Kenneth Boulding, Conflict andDefense: A General Theory, New York: Harper and Brothers, 1962 pp. 245-47. 38 Robert Gilpin, War and Change in World Politics , New York: Cambridge University Press, 1981, p. 57. 86 cover and concealment for the projection of modem military power, and there are a substantial number of urban-based insurgencies (most recently in Iraq and Syria) to suggest that terrain is not the only factor. But physical infrastructure also plays a role in accessibility and projection of military force and state control. The variables of terrain and infrastructure are distinct but conceptually related to the physical accessibility of a region. Table 3.1: Summary of Territory Mechanisms Core Territory Incentives Material Strategic Ideational Operational Costs Distance Physical Terrain Human Terrain Observable Effort-enhancing Mechanism Valuable Production, Capital, Human Capital Defensive features or buffer territory to hedge against rival Territory critical to religious or national identity with symbolism; Historic integration with national mainstream; news mentions of a region; normal (rather than indirect) legal administration Materialist Interests Region in close proximity to capital urban hubs, or security force concentrations Minimally obstructive terrain (e.g. low mountains, forest cover) and high physical infrastructure (e.g. road density, electrification, and water access) Population accustomed to some central or state authority; social infrastructure (human development, population's Survival Emotional Attachment Logistics and Supply Physical Accessibility Political Authority Potential I education and health indicators) A third factor that raises the operational cost of counterinsurgency is the human terrain or society within which the state has to make war. Since certain strategies described in the previous chapter require higher levels of engagement with the population and even dependence on them for information, the pliability of a population that has been habituated to centralized authority will determine the effectiveness of this strategy. These costs can be higher in territories with 87 particularly strong societies,39 a history of ungoverned spaces, 40 illegibility of territory,41 an absence of "empirical statehood" 4 2 or limited "infrastructural power" that fails to penetrate, discipline and order society.4 3 Regions that have been previously penetrated, tamed, rendered legible, and "developed" by the state enables the incumbents' authority to be perceived as legitimate, which further "lowers the cost of social control (incentives, coercion, enforcement, monitoring)."" These human terrain features, measured by human development like education and health indicators, are not randomly distributed in a country but fostered by states (or their predecessors), potentially endogenous to the incentive structures described above. Qualifications to the Territory Hypothesis Change over Time. An incumbent's perception of what is the core and what is the periphery can certainly change over time but not in a short enough span of time to pose a reverse causality problem. A region's material value can gradually change over time with shifts in production, efforts at economic integration and development, and international economic shocks while its strategic value can change with the rise or fall of rival powers or inter-state crises. Global economic shifts have made Asia more economically valuable relative to Europe and the end of the Cold War along with the rise of China have shifted US strategic interests eastward as well. Thus it would be fair to say that while Europe may have been the core and Asia the periphery for US grand strategy, the economic and strategic value placed on these two regions 39 Migdal, 1988. 4* Anne L. Clunan and Harold A. Trinkunas, UngovernedSpaces: Alternatives to State Authority in an Era of Softened Sovereignty, Stanford, CA: Stanford University Press, 2010. 41 Scott, 1998. 42 Robert Jackson, "Juridical Statehood in Sub-Saharan Africa," JournalofInternationalAffairs, 46, 1992, pp. 1-16. 43 Mann, 1984. These concepts are also derived from Foucault's distinctions between disciplinary power and juridical power. See Michel Foucault, Discipline and Punish: The Birth of the Prison, Trans. Alan Sheridan, New York: Vintage, 1995. 44 Boone, "Decentralization...," 2003, p. 362. 88 are at least "rebalancing" if not reversing. Similarly, a state's economic and strategic interests within different parts of its country can change over time, but this type of change typically takes place across decades rather than rapidly within a counterinsurgency campaign. The state of California developed from an unincorporated frontier region to an economic juggernaut and the leading state contributor to US economic production but only over the course of decades if not a century. Projection. An important caveat to bear in mind is that the theory is based on what territories are under threat so if rebels generate the capacity to threaten numerous territories, that can also change the state's response. Rebels contesting peripheral areas alone, even highly concentrated groups in a region, are unlikely to elicit a high effort incumbent counterinsurgency response. However, if rebels based in peripheral territory develop the capability to project significant power, force, or violence outside its area of operations and "reach" core regions that the incumbent values (for instance, territorial encroachment, armed assaults, or bombings), then the incumbent can justfy the cost of a higher effort "control" strategy to defeat or disarm the rebel group (Hi c). Though projection may not impact direct operational costs, it does alter the incumbents' incentives because valuable territory is under threat. Conclusion. Given that even wealthy states struggle with resource allocations, weak states, which are more frequently beset by insurgency, regularly face competing priorities of development, security, and contentious political coalitions (even in autocracies) in their core regions. In the event of an insurgency in its frontier, such a state may want to retain it but have little interest in doing anything more than restoring the long-standing pre-existing order instead of costly efforts of comprehensive pacification and political transformation-whether through 89 substantial repression or population protection and public goods provision. They can afford to trade disorder of frontier space for a time and choose to simply contain the threat with less effort. IV. Independent Variable 2: Identity and Violence The second dimension of the core-periphery theory deals examines the "distance" of the rebels' social identity group in relation to the state and the effect of this relative distance on state strategies of counterinsurgency violence. While previous studies have talked about these as "identity" or "ethnic" wars (as opposed to "ideological" or "nonethnic") in dichotomous terms,4 5 in fact these terms are used as a proxy for social distance between the parties, which may be more accurately captured on a continuum as most quantitative indices attempt to do.46 Particularly in ethnic environments that are not bi-polar, even distinct minority ethnic groups may be viewed and categorized as part of the "in-group." Drawing on social psychology literature and social identity theory, distinct identity groups can be "closer" or "further" 47 from the state's "core group," 48 and their position on the social distance continuum can shape the state strategy of violence.4 9 (Throughout this work, "state" is used as a shorthand for those state 45 Roy Licklider, "The Consequences of Negotiated Settlements in Civil Wars, 1945-1993," American Political Science Review, 89 (3), September, 1995, pp. 681-90; Nicholas Sambanis, "Do Ethnic and Nonethnic Civil Wars Have the Same Causes? A Theoretical and Empirical Inquiry (Part 1)," Journalof Conflict Resolution, 45 (3), June, 2001, pp. 259-82; Alexander B. Downes, "'The Problem with Negotiated Settlements to Ethnic Civil Wars," Security Studies, 13 (4), 2004, pp. 230-279. 46 This is true of the ethnolinguistic fractionalization, ethnic polarization, or an ethnic concentration index. 47 The notion of relative distance from the state core is also expressed by Cynthia H. Enloe, Ethnic Soldiers: State Security in Divided Societies, Athens: University of George Press, 1980, pp. 23-25. 48 Mylonas, 2013, pp. 23-24. 49 Parsing types of conflicts beyond the identity/ideology dichotomy is important if increasing numbers of civil conflicts are over identity given "the eclipse of the left-right ideological axis" and the "marked ethnicization of violent challenger-incumbent contests." See Rogers Brubaker and David D. Laitin, "Ethnic and Nationalist Violence," Annual Review ofSociology, 24, 1998, pp. 424-25. 90 ) elites, 50 typically of the "core" ethnic/identity group that dominate the political and bureaucratic institutions of the state. 51 Simply put, the cost of violence for the state varies with the position of the group on the social continuum, incentivizing different strategies. Violence may be cheap against distant groups but it can be costly against proximate or core groups. Strategies involving high levels of violence against those proximate to or constituting the core group incur both direct costs on the state and its elites-morally, politically, or strategically-and indirectcosts (or opportunity costs) by foreclosing on tactical options enabled by trust and information. [Put another way, the state will accord "home-team discounts" to more proximate groups that rebel, and have certain home-team advantages in dealing with them.] Both these direct and indirect costs then serve to restrain the state and channel its efforts into less violent strategies. However, high violence against distant groups incurs fewer of these costs, unleashing the state against these distant groups. The history of counterinsurgency is littered with examples identity mediating violence, even amongst the professional militaries of great powers. In the cases of varying counterinsurgent violence described in chapter 2, a critical factor was the identity of the rebel group. More violence by Union forces was permissible against Missourans rather than Virginians because "the typical figure thrown up by the strife in Missouri turned out to be somebody remarkably, frighteningly different, the blood-soaked Quantrill." 52 Both the German brutality in the Eastern theater, which far outstripped the West, as well as the targets of violence and Enloe writes, "State elites here refers to those political actors who control the direction of the civil bureaucracy, the courts, the police, state corporations and military." See Enloe, 1980, p. 14. 51 This generally corresponds with the demographically dominant ethnic group though there are exceptions like the Alawite-dominated Syria, Sunni-dominated Bahrain, and Sunni-dominated Baathist Iraq (until 2003), Afrikaanerdominated apartheid South Africa, and white-settler dominated Rhodesia. 52 Joes, 2000, pp. 61-62. 5 91 collaboration on the Eastern theater had much to do with Nazi ideology and "racial factors." Soviet identity proximity to Slavic Ukraine and particularly the Greek Orthodox Volhynia region of Western Ukraine 54 relative to the Balts in Lithuania corresponds with comparatively less violence. 55 Even the highly professional British forces conducted their most brutal campaign in Kenya in defense of the "most openly racist regime in the British empire" and the violence perpetrated relied on identity perceptions that "damned their opponents as atavistic savages and animals." 5 6 The degree of identity or social distance of the rebel group-henceforth referred to as ethnic group 57-- can specifically influence state violence in two ways: first through group embeddedness based on its representation within critical state institutions, and second, through group worth (as perceived by state elites) resulting from prior group stigma and legitimacy. States will be less willing to employ strategies of high violence when facing less socially distant ethnic groups, measured in terms of their embeddedness in the state and perceived group worth. This core-periphery theory of identity translates into the following hypotheses on strategies of counterinsurgency violence: H2a: If a given state incumbentfights a rebellingidentity group with high group embeddedness in state institutions and high group worth, then the incumbent will be more likely to choose less violent strategies like mitigation andpopulation control. 5 Lieb, 2008, pp. 65, 75. 54 Petersen, 2001, p. 214. Statiev, 2010, p. 116. French, 2011, p. 72. 57 Like other political scientists, I treat ethnic groups as descent based ascriptive group identities including "religion, sect, language, dialect, tribe, clan, race, physical differences, nationality, region, and caste." Quote from Chandra and Wilkinson, 2008, p. 519. Others who explicitly defend this include Horowitz, 1985, pp. 41-54; Ashutosh Varshney, Ethnic Conflict and Civic Life: Hindus and Muslims in India, New Haven, CT: Yale University Press, 2002, p. 5; Evan S. Lieberman and Prerna Singh, "Conceptualizing and Measuring Ethnic Politics: An Institutional Complement to Demographic, Behavioral, and Cognitive Approaches," Studies in ComparativeInternational Development, 47 (3), 2012, p. 257. 5 56 92 H2b: If a given state incumbentfights a rebelling identity group with low group embeddedness in state institutions and low group worth, then the incumbent will be more likely to choose more violent strategies like enfeeblement and attrition. H2c: If a given state incumbentfights a rebelling identity group with high group embeddedness in state institutions but low group worth (or vice versa), then the incumbent will be more likely to choose a mixed or moderately violent strategies. In the following sections I will unpack the two forms of social distance-embeddedness and worth-how they are measured, and their four cost mechanisms that shape state strategies of violence against an insurgency. As a preview, embeddedness primarily operates through strategic mechanisms of vulnerability to exit and information, while worth primarily operates through affective mechanisms of empathy and trust. Group Embeddedness Ethnic structure or demography-simply measured by the degree or structure of heterogeneity or polarization in the polity (like an ELF score)-has not been linked to the onset of civil conflict because not all identity fault lines possess the same political salience. 58 However, there is some evidence to suggest "ethnic practice" does have an effect on conflict dynamics.5 9 In particular, social identity theory suggests the ethnic practice of institutionalizing identity can produce greater differentiation and violence: ... When institutions create or reify intergroup comparisons, this signals that a dividing line exists between "us" and "them," priming relational status concerns and shaping how subsequent facts are likely to be interpreted within a political context. Leaders, in turn, find it easier to successfully mobilize followers to commit acts of violence against ethnic others.60 David D. Laitin, Hegemony and Culture: PoliticsandReli gious Changeamong the Yoruba, Chicago: University of Chicago Press, 1986; Daniel N. Posner, "The Political Salience of Cultural Difference: Why Chewas and Tumbukas Are Allies in Zambia and Adversaries in Malawi," American PoliticalScience Review, 98 (4), November 2004, pp. 529-545. 59 Practice identifies which ethnic groups are empowered or which fault lines are salient - such as ethnic groups represented in colonial institutions see Chandra and Wilkinson, 2008 or in the executive branch Cederman et al, 2010. 60 Lieberman and Singh, "The Institutional Origins of Ethnic Violence," 2012, p. 2. 58 93 The corollary to this argument is that the groups that are highly included and embedded in state institutions may be more closely associated with the core or "in-group." Recent work has shown how the empowerment of certain groups within the executive branch, the Ethnic Power Relations dataset (EPR), might also shape conflict onset and dynamics.6 1 Power does not always conform to demographics-majorities can be rulers or "mass subjects" while minorities can be dominant elites, junior partners, or powerless. 62 In the EPR, representation in executive institutions mediates who is included within the state's "core group," and therefore prone to less conflict with the state than excluded groups.6 3 Building on these findings, one might expect that if a group empowerment in a state's executive branch can mediate the state's conflict strategies, so too might embeddedness in other critical state institutions.64 Group embeddedness in the state affords it greater access to and influence on state decisions and actions. 65 Representation in the bureaucracy and security forces are amongst the most critical components to a group's positional status in the ethno-political hierarchy,66 particularly since they are structurally more enduring than the executive. 67 Security forces are Cederman, Wimmer, and Min, 2010; Julian Wucherpfennig, Nils W. Mettemich, Lars-Erik Cederman, and Kristian Skrede Gleditsch, "Ethnicity, the State, and the Duration of Civil War," World Politics, 64 (1), January 2012, pp. 79-115. 62 R.A. Schermerhorn, Comparative Ethnic Relations:A FrameworkforTheory and Research, Chicago: University of 63 Chicago Press, 1970, p. 13. Horowitz (1985, pp. 147-49) identifies this concept in terms of "advanced" and "backwards" groups. Chandra and Wilkinson (2008, pp. 537-545) identify it as ethnic imbalance or concentration. Petersen (2011, pp. 54-55) describes this as "political status." I choose representation to narrow in on what specifically on how positions within the state shapes state choices while exclude other elements like economic status discussed by Horowitz and street names and the language of government discussed by Roger Petersen, Western Intervention in the Balkans: The Strategic Use of Emotion in Conflict, New York: Cambridge University Press, 2011. 6 Since the EPR is based exclusively on informed judgments by regional experts on the composition of the executive branch - President, cabinet, senior posts in the administration (Wucherpfennig et al, 2012, p. 97) -- but of course this cannot account for the non-elected branches of government like the bureaucracy and military which can be highly influential at the strategy and policymaking level as Horowitz (1985), Chandra and Wilkinson (2008), and Petersen (2011) emphasize. 65 Lily L. Tsai, 2007. 66 Horowitz, 1985; Schermerhorn, 1970. 67 For instance, even though the international community has sought to ensure ethnic balance in a number of formal political institutions and the executive branch in Iraq and Afghanistan, the composition of security forces in Iraq and Afghanistan is still hotly contested because that is fundamentally where power lies. 61 94 especially important because in the modem era, they have acquired a larger portion of state budgets and prestige, 68 and because they heavily influence security policy and decisionmaking. 69 Therefore, co-ethnics in positions of political or military power have the capacity to directly shape less violent strategies and/or compel other state elites to choose less violent strategies. The contention that minority group presence in institutions can modulate levels of violence in conflict is not a new one. In fact, some work shows how intergroup (interethnic) linkages to critical social (civic) institutions that are vulnerable to costly losses in the event of violence can provide sources of information, grievance expression, trust, and motives for restraint that contain or mitigate social (communal) violence.70 One could imagine that interethnic links to state institutions might serve the same purpose of restraining state violence or the type of violence they could employ. In fact, such a relationship has been demonstrated in a study linking higher levels of African-American participation and representation in the police and local government to lower escalation of race riots in cities. Measurement Group representation can have one of five values-highly overrepresented, overrepresented, represented, underrepresented, or highly underrepresented-measured against the group's demographic proportion of the country. While comprehensive time-series data on representation is not readily available for most states (particularly those most likely to be Enloe, 1980, p. 49. Posen, 1984; Michael C. Desch, Civilian Control of the Military: The ChangingSecurity Environment, Baltimore: Johns Hopkins University Press, 1999. 70 Ashutosh Varshney, "Ethnic Conflict and Civil Society: India and Beyond," World Politics, 53 (3), April 2001, pp. 362-98. 71 Stanley Lieberson and Arnold R. Silverman, "The Precipitants and Underlying Conditions of Race Riots," American SociologicalReview, 30 (6), Dec. 1965, pp. 887-898. 68 69 95 affected by insurgency), a well-informed qualitative judgment can be rendered for each group per decade in both the civil bureaucracy and security forces based on snapshot quantitative data compiled from numerous sources. Since these institutions are primarily composed of career professionals, representation figures are likely to be sticky and a single year snapshot can approximate representation for a decade. Mechanisms A group's representation in the state can mediate state strategies of violence through a number of ways. First, and most intuitively, there is a simple preference mechanism-powerful enough coethnic elites in the center may oppose high violence against their own group due to informal group norms of solidarity that hold them accountable,7 2 and can veto high violence strategies. Additionally, there are two strategic mechanisms: 1) strategic vulnerability, which restrains state violence by imposing direct strategic and political costs on the incumbent if it employs high levels of violence, and 2) information, which offers the state opportunities to fight rebellion with more discriminate, and therefore less violent tactics (constituting an indirect opportunity cost). Direct Cost: Strategic Vulnerability. The first mechanism is the state's sensitivity and vulnerability to the group's exit. While voicing dissent may be the preferred option, in the event that the state considers or begins to pursue a high violence counterinsurgency strategy, the coethnics within the state can threaten to exit (mass resignations, mutiny, defection), posing major strategic and political costs for the state and compelling adjustments or alternative approaches. 72 Tsai, 2007. 96 The mere threat of exit from highly placed loyalists (those co-ethnics represented within the state) can provoke a "jolt" and compel changes in policy or strategy (voice).73 For instance, an initial course of high violence (and corresponding collateral damage) against represented groups may engender political backlash from co-ethnic and other elites or the broader public such as voiced dissatisfaction, public rebukes, and threats to exit, thus raising political and strategic costs and compelling the state to make course corrections. Security bureaucracies relying on officers and soldiers to fight their co-ethnics can be particularly vulnerable and sensitive to the risk of desertion, declines in morale and cohesion, or even mutiny as well as the political friction that results from it. Such processes would generate observable evidence of internal dissent, strategic options that are overruled, voiced concerns about blowback and strategic costs, and even rapid adjustments and adaptations following any blowback, ultimately resulting in less indiscriminate violence or civilian casualties and greater inclusion of substitute non-kinetic tactics. The mechanism of strategic vulnerability is not dependent on group worth. Despite the prevalence of racism, stigma, and discrimination against Irish during the Victorian era, they nevertheless benefitted from stereotypes of martial prowess and became a critical element in the British Army and Navy by the late 18th century. This embeddedness within key security institutions may have subsequently altered state responses to rebellion as the English commander-and-chief expressed reluctance to dispatch forces to put down the Irish rebellion in 1797 because of his Irish-filled regiments. Albert 0. Hirschman, Exit, Voice, Loyalty: Responses to Decline in Firms, Organizations,and States, Cambridge, MA: Harvard University Pres, 1970, p. 117. Most of these concepts are drawn particularly from Chapter 7, "A Theory of Loyalty." 74 Enloe 1980, p. 46 7 97 Indirect Cost: Information. Group embeddedness shapes state violence through a second mechanism of information. The rebelling identity group's presence in the state affords policymakers and military planners unique access to better information on the group-a distinct "home team advantage." Since co-ethnicity provides a conveyer belt for information and shared communication through language, experience, modes of social interaction, and technologies of "findability," 75 state co-ethnics are more likely to cultivate better informants and intelligence gathering networks and provide insight on rebel organization, leadership, behavior, and aims. Information is primarily conceived of in tactical terms. Better information can offer insights for better intelligence, planning, and situational awareness, which in turn can minimize violence of indiscriminate strategies.76 Access to information and intelligence can also minimize violence by distinguishing hardcore from reconcilable insurgents77 and enabling selective violence.7 8 A recent study shows co-ethnic security forces are able to carry out more discriminate force due to their ability to distinguish insurgent from civilian.7 9 But information is central to a strategic understanding of insurgents' political aims. While the British counterinsurgents in the post-WWII era recognized the importance of countering political subversion in its colonies as much as killing insurgents, "if the British did not understand, or take seriously, the political aims of their opponents, their counter-insurgency strategy was likely to be misconceived." 80 Finally, co-ethnic ground-level commanders and operators would have better "situational awareness," particularly of human terrain and language, 75 Karl Deutsch, Nationalismand Social Communication, Cambridge: MIT Press, 1966; Russell Hardin, Onefor All: The Logic of Group Conflict, Princeton, NJ: Princeton University Press, 1995; James Habyarimana, Macartan Humphreys, Daniel N. Posner, and Jeremy M. Weinstein, "Why Does Ethnic Diversity Undermine Public Goods Provision?" American PoliticalScience Review, 101 (4), November 2007, pp. 709-725. 76 Simpson, 2010; Kalyvas, 2006. 77 Kalyvas, 2006, pp. 89-91. 78 Kalyvas, 2006, pp. 169-171. 79 Lyall, 2010a, p. 18. 80 French, 2011, pp. 72-73. 98 to effectively communicate and maneuver without generating as much friction, dislocation, or hostility amongst the civilian population. 8 ' A failure to harness this opportunity with co-ethnics would then constitute an opportunity cost. All of these information-gathering processes would manifest themselves in observable evidence of larger and more successful intelligence generation efforts (especially human intelligence) as well as discriminate, intelligence-driven targeting for both kinetic and non-kinetic operations. While over time, counterinsurgents are expected to evolve from indiscriminate to selective violence with greater local information,82 it should be noted that the pace and degree of this transition can still be mediated by the rebel identity group's representation within state (and secondarily its group worth), particularly since such a transition can be quite costly. In other words, while most states and their security forces are learning organizations, I expect to observe an "opposing team handicap" when states try to generate information on identity groups that are not embedded, revealing observable differences in behavior toward different groups in the early years of the conflict. Group (Moral) Worth In addition to identity group embeddedness within the state, a second factor that shape's the state's use of violence is state elites' perceptions of the group's character, what some have termed the perceived "worth" of the group. "Groups are felt to have different mixes of attributes.. .evoked in behavior and subject to evaluation," which Horowitz terms a group's James Worrall, "Bringing the Soil Back in: Control and Territoriality in Western and Non-Western COIN," The New Coutner-InsurgencyErain CriticalPerspective, Ed. David Martin Jones, Celeste Ward Gventer, and M.L.R. Smith, New York: Palgrave Macmillan, 2014, p. 140; Kilcullen, 2010, p. 43, 70. 82 Kalyvas, 2006, pp. 170-71. 8 99 "moral worth." 83 Group worth also shapes perceptions of its legitimacy and loyalty, all assessed by comparison and this in turn can mediate the intensity of conflict and the acknowledgement of challengers' claims.8 4 Group worth is essentially capturing something about affective social distance-how state elites perceive the rebel group and its values, regardless of its embeddedness within the state, and whether it fits with the state's core "in-group." In laboratory experiments, lowering the social distance has been found to yield greater "other regarding behavior" and efforts at fairness. 85 Thus in a civil conflict often over the distribution of power in some form, one could expect greater levels of state empathy, grievance recognition, and tactical substitutes for violence against groups that are less socially distant and of greater moral worth. State elites'-primarily composed of the dominant group within the countryperceptions of a group's moral worth may stem from specific ethnic practices such as a group's participation within the state or during its origins (such as anti-colonial struggles) or simply be a product of historical contingency in colonial regimes or national myths. Though I expect this may track closely with representation, theoretically, even if the rebel group commands little representation within the state, state elites' ex ante beliefs about the group's "moral worth" may also enable the state to treat it as an "in-group" and minimize the violence it is willing to employ. There are certainly cases of high worth groups that command respect even with low representation in the state and security forces. For instance, Israel's "ultra-Orthodox" sect, the "Haredim" (-10% of the Israeli population) have high moral worth but low representation within the state. The Haredi command a certain "moral authority" as they are seen by the broader 83 Horowitz, 84 1985, pp. 142-43. Horowitz, 1985, pp. 209-15. 85 Elizabeth Hoffman, Kevin McCabe and Vernon L. Smith, "Social Distance and Other-Regarding Behavior in Dictator Games," The American Economic Review, 86 (3), June 1996, pp. 653-660. 100 Jewish population to be more "authentic."8 6 Despite poor representation in the civil service, security forces, and workforce in general, and only modest political power, government and public respect for the minority group affords them tremendous public sympathy, exemptions, and government welfare. 87 Afghan ethnic Hazaras offer another example as a group recently well-embedded in the state despite remaining socially marginalized. Hazaras have been historically discriminated and subordinated for their different ethnic features and minority religious sect as Shias. After the 2001 invasion of Afghanistan and ouster of the Taliban regime, their levels of representation in the state and security forces increased proportionate to their population," but the historical prejudice, discrimination, and stigma against them remain strong. 89 By contrast, Pashtuns are judged to still have high "value" 90 or moral worth even if they are now underrepresented in the Tajik-dominated state security forces, police, and intelligence agencies. 91 Moreover, there are numerous cases of indigenous "sons of the soil" groups that were deemed the worthy inheritors of the post-colonial state, but took several decades to "catch-up" in populating state institutions. Kikuyus, the largest Kenyan demographic group, quickly emerged 86 Nora L. Rubel, Doubting the Devout: The Ultra-Orthodoxin the Jewish American Imagination,New York: Columbia University Press, 2010, p. 148; Shahar Ilan, "The Myth of Haredi Moral Authority," Haaretz, July 12, 2012 (http://www.haaretz.com/the-myth-of-haredi-moral-authority-1.450670) 87 Moti Bassok and Haim Bior, "Israel's Cabinet Creates Two Panels on Haredi, Arab Integration," Haaretz, May 26, 2014. (http://www.haaretz.com/news/national/.premium-1.592752) 88 Scott Baldauf, "Afghan Army Gets Ahead by Getting Along," ChristianScience Monitor, February 11, 2001 (http://www.csmonitor.com/2003/021 1/p07s01-wosc.html); Hamid Wahed Alikuzai, A Concise History of Afghanistan in 25 Volumes, Volume 14, S.I: Trafford, 2013, p. 773. 8 Nic Robertson and Matiullah Mati, "Minority Group Emerges as a Force in New Afghan Parliament," CNN, December 2, 2010 (http://www.cnn.com/2010/WORLD/asiapcf/12/02/afghanistan.elections/); Mark Sappenfield, "Afghanistan's Success Story: The Liberated Hazara Minority," ChristianScience Monitor, August 6, 2007 (http://www.csmonitor.com/2007/0806/p06s02-wosc.html). 90 "Ethnic Discrimination Infests Afghan Army, Soldiers Say," Agence France Press, January 3, 2011. 91 Selig Harrison, "Discarding an Afghan Opportunity," Washington Post, January 30, 2007 (http://www.washingtonpost.com/wp-dyn/content/article/2007/01/29/AR2007012901448.html); Selig Harrison, "Ethnic Hostility is a Big, Maybe the Biggest, Part of the Afghan War," The Globe and Mail, Nov 2, 2009 (http://www.theglobeandmaii.com/globe-debate/ethnic-hostility-is-a-big-maybe-the-biggest-part-of-the-afghan- war/article1348333/); Heidi Vogt, "Afghan Army Struggles with Ethnic Divisions," Associated Press, July 27, 2010 (http://www.nbcnews.com/id/38432732/ns/worldnews-southandcentralasia/t/afghan-army-struggles-ethnicdivisions/#.U4zWLseLGUc) 101 as the political leaders of the post-colonial landscape but remained significantly underrepresented in the security forces and officer corps even a decade or two after independence.92 Measurement Group worth is not easy to identify but in contentious or hostile multi-ethnic environments, group worth is often distinguished on the basis of stigma, where a group is seen as "inherently inferior or deficient." Stigma can also generate evidence of contempt and prejudicial ideas as negative qualities are ascribed to the group while positive qualities are not. 93 Similar to in-group identification, prejudice can also filter information during conflict dynamics, assigning certain motives to particular actions and perceptions of loyalty, making greater violence against the group more permissible. Stigma is discernible through a range of social and cultural indicators such as intermarriage rates, residential segregation, and prevailing negative stereotypes or derogatory slurs.94 It can also be expressed in surveys and experiments attempting to gauge social distance. After reviewing and aggregating evidence on intergroup relations and potential stigma, a qualitative judgment can be produced assigning the group high, moderate, or low group worth. Mechanisms A group's worth in the eyes of the state can mediate state strategies of violence through two affective mechanisms: 1) empathy, which restrains state violence by imposing a moral cost 92 Enloe, 1980, p. 113; Timothy H. Parsons, The 1964 Army Mutinies and the Making of Modern East Africa, Westport CT: Praeger, 2003 p. 175; Kenneth Kwama, "Why the Colonialists Banned Kikuyu, Meru from Joining the Army," The Standard(Kenya), Ocotber 8, 2013. (http://www.standardmedia.co.ke/?articleID=2000095098#) 93 Petersen, 2011, p. 57. 94 Petersen, 2011, pp. 56-57. 102 on high violence, and 2) trust, which offers the state opportunities to contend with rebellion through less violent non-kinetic substitute tactics. Direct Cost: Empathy. The first affective mechanism is empathy with the co-ethnic population suffering the brunt of violence, which might constrain the strategy employed by decision makers and leaders to minimal force or even active protection. Empathy could certainly be voiced by co-ethnics within the state to lobby state decision makers and shape policy, feeling bound by certain moral obligations to the solidary groups to whom they belong, 95 but empathy could also be shared and expressed by other non-co-ethnics as some social psychology literature suggests. These state elites might -after extensive contact, equal-status interaction, or inclusion with the co-ethnics assimilated in these critical institutions-come to "recategorize" rebel identity group as part of the state's core, revise their in-group identity to be more inclusive, or see each other as "members of the same group at the superordinate level" - i.e. "all playing on the same team."9 6 Regardless of their representation levels, this recategorization of high worth groups as part of the same team can filter perceptions of attribution and responsibility and endow that group with a "home team discount." 97 Such a discount filters how "victim"-in this case the state-interprets the group's past behavior as well as anticipates the group's future behavior. 98 State elites can see aggressive or negative actions (like rebellion or terrorism) as situational (compelled by grievances) rather than inherent (that is, misguided rather than disloyal)95 Tsai, 2007, pp. 18-19, 93-95. 96 Quote from Hewstone 591-92; and evidence of contact and recategorization from an out-group to an in-group reducing bias and increasing empathy is described in Miles Hewstone, Mark Rubin, and Hazel Willis, "Intergroup Bias," Annual Review ofPsychology, 53, 2002, pp. 575-604. Evidence of causal effects for the contact hypotheses appear to be the most compelling in recent studies including lab and field experiments. See Elizabeth Paluck and Donald Green, "Prejudice Reduction: What Works? A Critical Look at Evidence from the Field and the Laboratory," Annual Review ofPsychology, 60, 2009, pp. 339-67. 97 Jason Lyall, Graeme Blair, and Kosuke Imai, "Explaining Support for Combatants During Wartime: A Survey Experiment in Afghanistan," American PoliticalScience Review, 107 (4), November 2013, p. 681. 98 Lyall et al, 2013, p. 681; Lieberman and Singh, 2012, p. 5. 103 allowances not afforded to out-groups who are punished for transgressions. With such a discount, high violence would seem unwarranted and immoral. Empathy then may lead incumbents to view certain rebel groups and/or their bases as semi-loyal, albeit violent, rather than totally disloyal opposition groups9 9 warranting more careful and selective violence. For instance, archival research has revealed that even as a third-party, British counterinsurgency forces in Northern Ireland unevenly targeted the activities of the ethnically distant Catholic militias while discounting or ignoring the more proximate Protestant militias, particularly in the early years of "the Troubles." 100 By contrast, groups of low moral worth or greater social distance and subject to resentment or disdain can bias the incumbent's impressions of the insurgency and its identity base (like their aims, cohesion, strength, legitimacy, and popularity). This bias may allow wholesale categorization of the group as disloyal and culpable, delegitimizing victims and placing them outside the boundaries of morality, fairness and justice.' 0 1 Recent qualitative work has shown how such social distance facilitates dehumanization and allows the prospect for massive violence, even genocide, against a particular group. 0 2 Historian David French notes of the British during the Kenyan emergency (1952 - 1960), "Labeling their enemies as wild animals, rather than merely 'wogs' or 'gollies', helps to explain why it was possible for the [British] security forces to employ a far higher degree of coercion in Kenya than elsewhere."1 0 3 99 Juan J. Linz and Alfred Stepan, The Breakdown ofDemocraticRegimes, Baltimore, MD: Johns Hopkins University Press, 1978, pp. 28-33. 100 Owen Bowcott, "MoD took Softer Line on Loyalist Paramilitaries, Secret Files Reveal," The Guardian,October 10, 2010. 101 Such Intergroup bias in conflict is briefly reviewed by Hewstone et al, 2002, p. 594; Ervin Staub, "Ethnopolitical and Other Group Violence: Origins and Prevention," Ethnopolitical Warfare: Causes, Consequences, and Possible Solutions, Ed. Daniel Chirot and Martin E. P. Seligman, Washington, DC: American Psychological Association, 2001, p. 292 Kristen Renwick Monroe, Ethics in an Age of Terror and Genocide: Identity and Moral Choice, Princeton: Princeton University Press, 2011. 102 103 French, 2011, p. 72. 104 Both tacit and deliberate civilian victimization can become permissible to ensure sufficient defeat or deterrence of an insurgent combat power, escalating and increasing levels of violence. Concern for both the rebels' civilian base and efforts at restraint would manifest themselves through observable efforts at minimizing collateral damage, actively protecting civilian population centers, and emphasizing surrender and rehabilitation above arrests and kills to deplete insurgents. Indirect Cost: Trust. The second affective mechanism again operates as an opportunity rather than a direct cost. When social distance between the state and the group is lower, trust is likely to be higher, 104 enabling state strategies involving more non-kinetic substitutes. The counterinsurgents' trust problem needs to be described as it poses a fundamental challenge for both sides in civil wars. 105 Though state strategy is often conceived of as a unidirectional move, in fact some approaches to counterinsurgency may need to be thought of as "two-way politics,,106 with simultaneous violence and bargaining in an iterative game that requires reciprocal moves by the opposition in order for the initial state move to have any meaning or utility. While pure kinetic activities may not require participation of the opposition, kinetic substitutes addressing rebel demands-like concessions to redress grievances, public goods provision, devolution of authority, and side-deals or side-payments to induce defectionare forms of bargaining requiring some degree of rebel or rebel-base input, participation, and/or acceptance. 107 Hewstone, 2002, pp. 578, 592. Downes, 2004; Barbara F. Walter, "The Critical Barrier to Civil War Settlement," InternationalOrganization, 51 (3), Summer 1997, pp. 335-364. 106 Jackson, 2008, p. 11. 107 Although state-orchestrated ethnic defection is frequently attempted in asymmetric conflict, since the state is stronger, the extent and viability of this tactic can vary based on the rebel identity group's previous status (as groups will be more assured and confident in their defection if they are already esteemed by and represented within the state). See Stathis N. Kalyvas, "Ethnic Defection in Civil War," ComparativePoliticalStudies, 2008, 41 (8), pp. 1043-68. Furthermore, as Sambanis et. al. (2012) point out, "It is clear, however, that short-term, strategic alliances 104 105 105 In order to initiate less violent substitutes, the state must first trust that addressing some rebel demands through concessions and offers will satisfy them and mitigate violence rather than stoke their greedy appetite.1 08 In other words, the state must have some trust that the rebel base is committed to the state once these disputes (over power, autonomy, political control, resources, respect, etc.) are resolved. For instance, despite repeated calls by experts, the U.S. distrust of Taliban intentions and goals inhibited its willingness to seriously bargain with Taliban commanders amidst its Afghan counterinsurgency campaign. 109 When the state has prior high esteem for the group, judging its moral worth to be high (and reducing social distance), the "home-team discount" is more likely to impute other qualities like trustworthiness, at least in relative terms. Owing to this recognition, the state may have greater confidence in non-kinetic substitute tactics that wean rebels away from the insurgency and be more willing to initiate forays to address the demand side of rebellion like public goods, side deals, and political bargaining. Because this is an interactive game, the second move requires that the rebel group must trust the state for these alternative tactical forays to be effective and expand into a formidable role in strategy. Any sort of wartime bargaining or substitutes expose both sides-but in particular the rebel side-to risks of political outbidding (and outflanking) or targeting by extremist factions. These risks create credible commitment problems at the stage of conflict across group boundaries are no indicator of parochialism's absence, although sometimes interpreted as such. Pragmatic alliances forged during conflict rarely reflect deep convictions." See. Nicholas Sambanis, Jonah Schulhofer-Wohl, and Moses Shayo, "Parochialism as a Central Challenge in Counterinsurgency," Science, 336 (6083), May 18, 2012, p. 808. 108 This is a loose application of Glaser's typology of greedy versus security-satisfying states to rebel groups' motives based on "unit-level information" of the group. See Charles L. Glaser, "The Security Dilemma Revisited," World Politics, 50 (1), October,1997, pp. 171-201. However, I would contend this information on the group's motives can certainly mediated by social distance. 1 09 James Shinn and James Dobbins, Afghan Peace Talks: A Primer, Santa Monica: RAND, 2011, p.xiv 106 termination but they also arise during "warring group interactions" in the midst of conflicts."' In these situations, reciprocal trust is critical for a healthy exchange of information for public or club goods between the population and the state security forces. It is also essential for low-level commanders to flip and integrate into state forces, surrender, or enter into rehabilitation. The state can initiate attempts to engage the population and/or address this "demand-side" of rebellion, but if those attempts at "two-way politics" and "violent negotiation" are mistrusted and left unmet, they will quickly collapse before getting off the ground and serve no role in the state's overall strategy. 1 ' For instance, even years after the U.S. attempted to shift to population control and negotiations in Afghanistan, the Taliban remained deeply distrustful of US sincerity as well as long-term intentions, while the broader Afghan public remained distrustful of the U.Sbacked central government, limiting real traction on these tactic.' 2 Group embeddedness also engenders trust from the perspective of the rebels and their population base. Like the home-team discount that mitigates the direct use of violence, these "home-team advantages" that offer alternative routes to ending an insurgency indirectly mitigate violence. Co-ethnic representation within the state can serve as a bridge for trust and loyalty," as well as reduce rebels' cost-tolerance," 4 3 to make the state appear as a more trustworthy partner. Co-ethnics provide the assurance that rebel moves towards de-escalation will be met by the state and commitments will be upheld, based on several formal and informal norms, in-group Christia,2012, p. 6. "1 Negotiated settlements are more likely to collapse with a higher degree of social distance between the parties. See Licklider, 1995. 112 Shinn and Dobbins, 2011, p. 25; Alisa J. Rubin, "Afghans' Distrust Threatens U.S. War Strategy," New York Times, May 12, 2010; Carlotta Gall, "Peace Talks with Taliban Top Issue in Afghan Vote," New York Times, August 17, 2009. "3 Habyarimana, Humphreys, Posner, and Weinstein, 2007. 114 Wucherpfennig et. al., 2012, p. 97. 110 107 policing, and a history of repeated interaction. 15 Rebel and base co-ethnics with a higher level of trust will be relatively less fearful of being letdown or double-crossed by the state, affording these alternative tactics more space and time to become sustainable substitutes to counter insurgency rather than "one and done" failed attempts. Real grievance acknowledgement and serious efforts to end the insurgency through nonkinetic substitutes can be found in observable evidence of sustained public goods provided, escalation management, grievance redressal and concessions, rebel side-deals, and continuous efforts at political deals and accommodations, even simultaneous to intense violence. Qualifications Overlapping Mechanisms. The logic of group representation and group worth along with their subsequent mechanisms are summarized in Table 3.2. The four specified mechanisms reveal ideal-type pathways for identity to shape or constrain violence, but in reality, the two are quite messy and can overlap. While the affective mechanisms identified above are primarily tethered to group worth and the strategic mechanisms to group embeddedness, in fact there is overlap. While empathy and trust are primarily a function of group worth, they can also be a function of group representation as both empathy and trust can increase with intergroup contact, inclusion, and joint problem solving.116 Simultaneously, group worth can also influence strategic mechanisms like vulnerability to exit, especially if group worth is recognized by a sympathetic national audience, potentially imposing significant costs to the use of high violence against them. Only information seems to be exclusively the domain of group embeddedness. 115 Habyarimana et. al., 2007; Tsai, 2007; James D. Fearon and David D. Laitin, "Explaining Ethnic Cooperation," American PoliticalScience Review, 90 (4), 1996 pp. 715-735. 116 Staub, 2001, p. 297. 108 Change Over Time. While a large constructivist literature makes the case that identity groups can shift allegiances in times of highly intense violence,117 group worth and embeddedness can be measured prior to conflict onset, and are not easily malleable in the short to medium term, therefore mitigating concerns that social distance is purely endogenous to the conflict and strategy itself. The type of meaningful representation in state institutions is unlikely to substantially change at the national level in the short or medium term absent a major regime change and/or status reversal of a group, but it can change gradually through deliberate (affirmative action) or exogenous forces (structural economic changes). In theory, group representation can shift rapidly in the executive branch, as evidenced by swings in ethnic power over short periods captured in the EPR. But because this theory primarily rests on ethnic representation in enduring, "careerist" state institutions like the civil bureaucracy and security forces, this variable is not subject to easy or rapid shifts. While the executive branch may play a more formidable role in challenges of distributional politics, such bureaucratic institutions may be more relevant managing the day-to-day responsibilities of crises like civil wars. Additionally, group worth is also particularly durable because stigma, like the residue of history, is not malleable in the short or medium term, but it too can erode over time." 8 Non-independence of Worth and Embeddedness. It is apparent from the preceding descriptions that group embeddedness and group worth are not mutually exclusive and likely constitutive of one another as some proponents of social identity theory contend.1 9 Group worth and perceived loyalty likely factor in to the selection of groups to fill nontrivial roles in critical state institutions. Enloe points out that state elites have in their minds "ethnic security maps" based on the dependability of various groups, which they then use to structure military and 17 Kalyvas, 2008. Petersen, 2011, pp. 59-60. 119 Singh and Liberman, 2012. 1 1 109 ' security forces.12 0 And if loyalty stems from participation in and the benefits of critical state institutions, then representation will shape perceptions of group worth and future expectations of loyalty. Still, it is conceivable to think of representation and group worth as partially exogenous-and thus further mitigating endogeneity concerns-since they are often inherited from historical contingencies like colonial institutions, and are highly path dependent. Furthermore, many colonial powers invented myths based on a both group character as well as political expediency, which may have elevated some otherwise backward groups.1 2 I expect that group representation and group worth tightly interact and may track closely with each other (as I observe in much of my empirical work). Nevertheless, I have attempted to identify theoretically distinct causal pathways for greater precision as well as for the theory to better travel to other regions. In cases where worth and representation diverge, I would expect that each would still impose some constraints on state violence, with embeddedness providing the slightly more powerful restraint than group worth since it can influence violence through all four mechanisms while worth may only operate through three. Divergences in Worth and Embeddedness. While intuitively embeddedness and worth should co-vary (and often they do), there are certainly plenty of cases of divergence, especially due to political upheavals of colonial policies and post-colonial transitions. The British, French, and Ottomans all managed colonial empires by recruiting weak, low-status minority groups into their security institutions to cultivate loyalty and neutralize mass insurrections, and eventually some of these groups eventually advanced because of it.' 22 While these groups may have been 1 20 Enloe, 1980, pp. 15-16. Horowitz, 1985, pp. 160-66; Enloe, 1980, p. 36. 1 For example low-caste Punjabis were heavily recruited into the British Indian Army as a buffer after the Sepoy rebellion of 1957 (Yong) while rural Alawites were recruited into the Ottoman military because Sunni notables saw military service as beneath them (Khoury). See Tan Tai Yong, The GarrisonState: The Military, Government and Society in ColonialPunjab, 1849-1947, New York: Vanguard Books, 2005; Philip Khoury, "Interview: Philip Khoury, Associate Provost," Precis (MIT Center for International Studies), Fall 2012. 121 110 perceived by the dominant socio-economic and political classes to be of low worth, the colonial power could count on these weak groups' sufficient fear of the dominant group to ensure loyalty, and therefore recruited them into the internal security forces.1 23 Eventually this process of being a part of the security forces could elevate the group out of the "backward" status designation but this could occur gradually after many decades during which time group worth and representation did not co-vary. Other colonial policies allowed for several paths to group legitimacy and worth even without advanced aptitude for government service. The most powerful of these claimsthat of indigenousness-empowered a group with a "special position" that could remain long after colonial departure.1 2 4 One example of this misalignment is Kenya's Kikuyu who were more advanced and educated and fought for independence,1 2 5 became the leading political class in the post-colonial environment but remained underrepresented in its security institutions for decades. There are also strange contrasts in the opposite direction where a group has low worth but high representation, such as the case of the Irish in the British military. Another is that of Baathist Iraq: while Shias and Kurds were both of low group worth relative to the Sunnis in the Iraqi Baathist regime and revolted in 1991, Shias at least had modest if not proportionate group representation in critical institutions including the Baath Party, Command Councils, and the enlisted military.1 2 6 Horowitz, 1985, p. 446-449. Quote from Horowitz, 1985, pp. 201-209. 125 Horowitz, 1985, pp. 153, 207, 452. 126 "Sunni-Shia Relations in Iraq," Iraq: A Country Study, Ed. Helen Chapin Metz, Washington DC: GPO for the, 123 14 Library of Congress, 1988. (http://countrystudies.us/iraq/38.htm); Faleh A. Jabar, "Why the Uprisings Failed," Middle East Report, 22 (176), May/June, 1992 (http://www.merip.org/mer/merI76/why-uprisings-failed). 111 Table 3.2: Summary of Identity Mechanisms Independent Variables Mechanism Utility Calculation State Internal Behavior State Observable Behavior Empathy Direct Cost: Violence is Grievance Acknowledgement Selective Targeting Concern for Collateral Damage Minimal Collateral Damage and Civilian Casualties Restraint; Narrow Definition of Insurgent (Fear of Group Moral Sanction) Grievance Acknowledgement Rehabilitation, Surrenders, and Arrests favored above Kills Active Protection of Civilians Grievance Redressal/Concessions Formulation of Substitutes Public Goods Efforts to Address "Demand Side" of Rebellion Insurgent Demobilization and Defection Morally Costly (Home-team discount) Group Moral Worth (Observables: Slurs, stereotypes, public opinion, treatment prior to conflict, intermarriage and residential mixing with dominant group) Trust Indirect/Opportunity Cost: Non-kinetic substitutes available cheaply (Home-team discount) {Player 2 Trust - Rebel Trust) (Home-team advantage) Rebels Accept/Trust State Offers as Genuine Political Deals/Accommodations Expansion of Non-kinetic activities scale and proportion of COIN portfolio Vulnerability Direct Cost: Violence is Strategically or Fear of Strategic/Tactical Blowback Internal Debate and Dissent Fear of Political Costs and Friction Reflection on actual blowback and strategic costs Politically Costly Group Embeddedness (Observables: Representation levels in Civil Service and Security Forces) Information Indirect/Opportunity Cost: Non-kinetic substitutes available cheaply (Home-team advantage) Insights on Strategic Behavior of Rebel Group; Leadership; Organizational Structure Intelligence on Location and Identification of Insurgents Adjustments and Adaptation (Human) Intelligence collection Discriminate, Intelligence-Driven Targeting (for kinetic and nonkinetic/side deals) Lower Friction with and Dislocation of Civilians Cl V. Predictions and Alternative Explanations Predictions The two intersecting sets of hypotheses on territory (H la-Hi c) and identity (H2a-H2c) produce the following predictions for selection of counterinsurgency strategy; * When an incumbent faces a violent challenge from a marginalized identity group in its less-valued or frontier regions, it will respond with enfeeblement. * When the incumbent faces that same type of challenger that threatens valued or core territory, it will respond with massive force and an attrition strategy, sometimes bordering on collective punishment. e By contrast, an incumbent facing a violent challenge over low valued territory from a co-identity group with linkages to the state, the incumbent will adopt a strategy of mitigation. This would include a mix of some force along with wedge or divide-and-rule tactics through money and the backing and utilization of warlords, tribal leaders, and militias to contain the challenge. e And when the incumbent faces a violent challenge from a co-identity group in valuable territory or core regions, it will need to employ a strategy ofpopulation control to exert control while applying more discriminate force. 12 7 The question may be asked why would these privileged groups with ties to the state and residing in core territory ever rebel in the first place? Though not the modal type of insurgency, it certainly happens, since the drivers of civil war all multi-causal. See Horowitz, 1985, p. 234. Moreover, advanced groups in advanced regions may have fewer grievances but also fewer obstacles to solving the collective action problems of rebellion. The cases of marginalized groups in valuable or core areas (not simply with natural resources but real industrial development) is also possible. State investment in a backwards region may not benefit the local people if they are not party to the downstream effects of the industrial development such as employment. Simultaneously, measures to homogenize and integrate territory (without adequately distributing the benefits to the locale) may smack of internal colonialism that can generate resentment and separatist impulses. See Horowitz, 1985, p. 241; Hechter, 1975. 127 113 It is important to explicitly state that in all these predictions, there is an assumption that the actor selecting the strategy-that is, weighing the incentives and costs of the campaigns effort as well as the direct and indirect costs of violence-is the state incumbent primarily composed of and driven by a set of state leaders and elites. Again, this is a well-worn and defensible assumption in studies of state (or group) strategy and behavior based predominantly (though not always exclusively upon the) leadership, an assumption also shared by all the rival explanations described below. 128 Figure 3.1: Predicted Values of Center-Periphery Theory Territory Value and Access Periphery (Low) Rebel Positional Status (Identity Group Low/Peripheral Core (High) Enfeeblement Attrition (low effort, high (high effort, high violence) violence) Mitigation (low effort, low violence) Population Control (high effort, low violence) Wirth and Embeddedness) High/Core Alternative Explanations There are broadly two sets of competing explanations to my theory of counterinsurgency strategy. The first set of strategic interaction explanations directly tries to account for subnational variation, while the second set of macro-level explanations focuses on cross-national variation but offers little purchase on sub-national variation. As a result, the first set is treated as Anyone who studies strategy has to make this assumption or reduction. This is done with states (Posen, 1984) militaries (Mearsheimer, 1983), and sub-state actors (Petersen, 2011; Christia, 2012). Enloe writes, "State elites here refers to those political actors who control the direction of the civil bureaucracy, the courts, the police, state corporations and military." (Enloe, 1980, p. 14). 128 114 competing hypotheses directly taken on and the second set as explanations controlled for with the research design. Competing Hypotheses. A number of competing hypotheses purport to explain some aspects of sub-national variation in state strategy, though they are not perfect competitors to my argument as they tend to be focused on slightly different units of analysis (only separatists conflicts) or slightly different dependent variables (whether to fight, levels of accommodation, or intensity of violence, but not any evaluation of effort independent of violence). Nevertheless, I try to extrapolate predictions of these three potentially competing theories and, when possible, evaluate them against my own. Theories of internal deterrence and reputation contend that incumbents with multiple potential self-determination groups are more likely to fight, and also contend incumbents will fight harder against early, concentrated identity-group challengers in order to avoid setting a precedent of concessions and establish a reputation for toughness. 129 Theories of external deterrence expect incumbents to fight harder against insurgents that have links to and receive support from external rivals and might exacerbate the incumbent's external threat environment.'30 However, both of these deterrence theories are primarily limited to explaining whether and when states fight rather than specifically how they fight rebellion. Nevertheless, one can extrapolate how these theories might explain varying levels of violence and effort after conflict initiation, specifically for incumbents that have fought multiple campaigns. 129Toft 2003, pp. 26-29; Walter, 2009; This is measured in terms of the number of potential rebel groups but the problem with this is that the number of salient as well as total identity groups are not fixed but also endogenous to conflict and politics making it methodologically difficult to evaluate the number of initial groups 13 Mylonas, 2013; Ahsan Ishaq Butt, "Goodbye or See You Later: Why States Fight Some Secessionist But Not Others," PhD Dissertation, University of Chicago, 2012. 115 Theories of insurgent behavior-whether based on insurgents' (separatist) aims, strength and size, 32 or victims (civilian or government)1 33 -contend higher insurgent threats motivate tougher incumbent strategies to counter rebellion. Again, these have been solely used to explain violence but predictions on effort can again be extrapolated. It is important to note that prevailing theories of insurgent threat primarily focus on latent capabilities and aims of the insurgents' themselves, and then link them to abstract features of the conflict like duration, but discount what they actually threaten. Furthermore, research on British campaigns reveals incumbent perceptions of insurgent aims often did not match reality, but nevertheless, the British aced on these perceptions.134 Therefore, my theory offers a competing focus on the state incumbent's perceptions based on what rebels actually do, and specifically HIc predicts greater effort when insurgents actually project force on core territory, regardless of their claims or intent. My contention is not that these alternative explanations of strategic interaction are completely wrong, but that they are insufficient or under-specified to explain the observed subnational variation and still are largely bounded by my theory of territory and identity. In my empirical research, when possible, I attempt to leverage my research design (involving tightly paired comparisons) to hold these factors constant and evaluate my theory. When it is not possible to control for them, I evaluate the deterrence and insurgent threat theories against my own. While I do not expect the core-periphery theory to explain all sub-national variation, I expect it to perform better than these alternatives. Stanton, 2009. Valentino et. al., 2004; Cunningham et. al., 2009. 133 Andrew H. Kydd and Barbara F. Walter, "The Strategies of Terrorism," InternationalSecurity, 31 (1), Summer 2006, pp. 49-80. 134 French, 2011, p. 58. 131 132 116 Control Variables. There is substantial evidence on the influence of macrostructure' 35 and variables like state capacity or strength,1 3 6 political institutions or regime type,1 3 7 and civilmilitary relations and organizational/strategic culture' 3 8 predominantly on counterinsurgency outcomes, and to a lesser degree on strategies of violence and accommodation. However, few empirical studies actively try to explain an incumbent's level of effort and typically conflate it with levels of violence.1 3 9 There is little doubt these theories have some influence in shaping strategy, but these macro-structural theories cannot account for visible sub-national variation or instances of the same incumbent's varying counterinsurgency strategies in different campaigns. Therefore, I develop a research design that controls for these factors and holds them largely constant in order to test my theory and explore sub-national variation. Following this, I eventually relax some of these controls in empirical sections using cross-temporal analysis to weigh the relative explanatory power of these theories against my core-periphery theory. 4 0 VI. Research Design Since prior large-N and single case study work thus far has been incapable or not ideally suited to explain sub-national variation in incumbent counterinsurgency strategies, I employ a research design involving medium-N study along with two sets of controlled, focused comparisons to test the theory of core-periphery relations and evaluate it against other plausible Byman, 2006 offers a theory that combines all these structural variables, but principally focuses on state strength and political institutions. 136 Watts et al, 2014; Fearon and Laitin, 2003. 137 Walter, 2009; Stanton, 2009; Valentino et al, 2004; Bethany Lacina, "Explaining the Severity of Civil Wars," Journalof Conflict Resolution, 50 (2), April 2006, pp. 276-289. 138 Krepenivich, 1986; Nagl, 2002; Jackson, 2008. 139 Toft, 2003; Butt, 2012; Valentino et. al., 2004; Walter, 2009 0 By this I mean that in cross-sectional analysis, these variables will be held constant but cross-temporal examination of a single country can identify some, though less stark variation over time. In cross-national terms, the U.S. is a democracy with strong civilian control of the military but in cross-temporal terms, the U.S. is more democratic after the civil rights act and its civilian control of the military declines during periods of low external threat (e.g. after the Cold War). See Desch, 1999. 135 117 explanations. The design first requires ample room for qualitative methods. Assessing the core-periphery theory requires far more in-depth information for conceptual validity than a large-N data collection project can offer. Strategy is both highly complex and difficult to pin down along with the factors that shape perceptions of territory and group identity. Since neither the theory nor the dependent variable to be explained can be easily captured with off-the-shelf variables or proxies, qualitative case studies are necessary to "identify and measure the indicators that best represent the theoretical concepts."'141 Furthermore, tracing the mechanisms and processes through which territory and identity incent and constrain incumbent strategic choices requires much more than thin or highly aggregated empirical evidence. At the same time, single case studies alone with invariant designs have proved insufficient to distinguish the relative effects of different structural features and incumbent perceptions.14 2 A medium-N research design retains important properties of case studies-such as close attention to concept validity, causal linkages, and the measurement of complex variables-but also enables sufficient variation on the independent and dependent variables as well as some key control variables. 43 In other words, the design allows me to test the hypothesized effects of my theory on counterinsurgencies conducted under conditions of both strong and weak state incumbents, militaries, civilian control, democracy, and external threat environments. Such a design could also provide opportunities for tightly paired comparisons and in-depth qualitative Alexander George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences, Cambridge: MIT Press, 2005, p. 19. This is a point Lyall makes about much of the COIN literatures assessment of effectiveness in order to justify a large-N sudy. However the same argument applies to strategic choice as well. See Jason Lyall, "Do Democracies Make Inferior Counterinsurgents? Reassessing Democracy's Impact on War Outcomes and Duration," International Organization, 64 (1), Winter 2010, pp. 167-92. (Henceforth referred to as Lyall, 2010b). 143 For a defense of the medium-N approach, see Michael L. Ross, "How Do Natural Resources Influence Civil War?" InternationalOrganization, 58, Winter 2004, pp. 35-67. An earlier draft of the paper offers a more thorough defense of the method, "How Do Natural Resources Influence Civil War? A Medium-N Analysis," UCLA Dept of Political Science, January 2, 2003. 141 118 case studies for a richer examination of the research question. An ideal test of my hypotheses would draw on a series of observations of counterinsurgency campaigns by the same incumbent to control for a variety of observed and unobserved confounders. Examining multiple incumbents, each having fought a few distinct campaigns, could generate a sufficient number of observations to explore the range of variation and identify real patterns and external validity. The approach could also help to foster new hypotheses and variables to construct future large-N tests. At its core, the design relies on complementary methods of process tracing and tightly paired, structured focused comparisons. The comparison is "focused because it deals selectively with only certain aspects of the historical case... and structured because it employs general questions to guide the data collection and analysis in that historical case."144 Furthermore, tightly-paired comparisons, basically a most-similar design using Mill's method of difference, are closely matched on a number of key structural variables to hold these explanations constant and isolate the effects of my theory.1 45 Process tracing allows me to examine the causal chain of events and complex mechanisms hypothesized by my theory in order to assess their causal role relative to other factors. Moreover, the study relies on detecting broader patterns across multiple cases within the country and amongst a number of incumbents within the region. All these methods require longitudinal analysis of decisions and actions, content analysis of official and unofficial statements and reports, qualitative accounts of campaign strategies, and quantitative measures (with interval and ratio data) of the independent variables (identity and territory) and dependent 144 Alexander L. George, "Case Studies and Theory Development: The Method of Structured, Focused Comparison," in Diplomacy: New Approachedin History, Theory, and Politics, ed. Paul G. Lauren, New York: Free Press, 1979, pp. 43-68. Quotation from pp. 61-62. 145 Sidney Tarrow, "Bridging the Quantitative-Qualitative Divide," Rethinking Social Inquiry: Diverse Tools, Shared Standards, Ed. Henry E. Brady and David Collier, Lanham, MD: Rowman and Littlefield, 2004: 178. 119 variable (effort and violence), to identify whether and how core-periphery relations incent and constrain state's choice of strategy. Case Selection Though the theory identifies factors that shape the counterinsurgency strategies of all incumbents, because each country's security forces fight uniquely (based on the interplay of state capacity, regime type, civil-military relations, and organizational/strategic culture), the effects of my theory's hypotheses would get washed out without examining substantial "within-case variance." 146 Since Lyall and Wilson question, "whether strategy is a function of an actor's attributes rather than an independent variable in its own right," 147 I sought incumbents with substantial within-country or sub-national variation to control for these attributes and maximize opportunities to study the same incumbent in different counterinsurgency campaigns. This selection criteria also enabled a better test of the alternative explanations of internal and external deterrence and insurgent behavior that could only be evident across multiple cases. Starting from the Lyall and Wilson/Friedman dataset on modern counterinsurgencies (post-WWI)1 48, case selection began by selecting from a pool of 35 incumbents that had fought more than one modern counterinsurgency (totaling 130 conflicts). The pool was further narrowed by additional criteria to select a set of cases that were relevant to contemporary conflicts (post1945), filled a gap in the literature (domestic incumbents), and offered enough variation on the independent and dependent variables but also sufficiently tight controls. Because my theory's 146 Stephen Van Evera, Guide to Methods for Students ofPoliticalScience, Ithaca: Cornell University Press, 1997, p. 77. 147 Lyall and Wilson, 2009, p. 81. 148 Lyall and Wilson (2009) identify 286 campaigns. Friedman, (2011, p. 568) examines the 173 campaigns after WWI, which both he and Lyall and Wilson describe as the "modem period" of counterinsurgency. It is from this pool of 79 incumbents and 173 conflicts that I began. 120 key variables are centered on territory and identity, this last criteria, required incumbents that had fought both territory-seeking and center-seeking rebels 149 as well as identity and non-identity wars. Table 3.3: Case Selection Criteria # Selection Criteria 1 Total Modem (Post WWI) # Incumbents 79 # Conflicts 173 Incumbents (States) See Lyall 2010 or Friedman 2012 for Full List Afghanistan, Algeria, Burundi, Cambodia, Chad, China, Colombia, DRC, France, Germany, India, Indonesia, Iran, Iraq, Israel, Italy, Japan, Liberia, Nicaragua, Pakistan, Philippines, Portugal, Russia (Soviet Union), Rwanda, Serbia (Yugoslavia), South Africa, South Vietnam, Spain, Sri Lanka, Sudan, Turkey, UKG, USA, Yemen, Zimbabwe All the above except Italy, Germany, Japan, Spain, and Turkey All the above except France, Portugal, UKG, and USA China, DRC, India, Indonesia, Iran, Pakistan, Philippines, South Africa, Sri Lanka, Sudan, Yemen 2 Counterinsurgencies Criteria #1 + with Multiple Distinct Insurgencies 35 130 3 Criteria #2 + only Post-1945 31 91 4 Criteria #3 + Fought by Domestic Incumbents Criteria #4 + Includes conflicts of both govt and territory incompatibility (PRIO distinction) Criteria #4 + Includes both identity and non-identity wars (Friedman distinction) Criteria #5 and #6 combined -- includes conflicts of both govt and territory incompatibility and both 26 71 11 35 7 23 China, India, Indonesia, Pakistan, Philippines, Russia/USSR, Sri Lanka 6 20 China, India, Indonesia, Pakistan, Philippines, Sri Lanka 5 6 7 identity and non-identity I Having narrowed this down to six potential incumbents, I selected two countries in South Asia-India and Pakistan-for the variation they offered (see table 3.4), but also because they both had periods where at least two simultaneous campaigns were being fought allowing for highly controlled comparisons and holding constant key structural variables to better test my theory. Selecting from a single a region also enables one to study regional dynamics and control 149 The UCDP/PRIO Armed Conflict Dataset distinguishes these to incompatibilities of territory and of government. 121 for broad economic and colonial institutions to test new hypotheses. 150 The continuous interaction between the two countries has generated a richer set of empirical research materials and therefore offers synergistic value. The South Asian cases present a hard test for the theory's core predictions since both states inherited institutions, including its armed forces, which were created and developed by the British. Thus if any post-colonial domestic incumbents should adhere to the principles of the "counterinsurgency ideal," 15' it should be these states. Table 3.4: South Asia Incumbent Macro-level Variation State Strength Strong/Professional Military Regime Type Civilian Control (CivilMilitary Relations) External Threat India Medium High Democracy High Pakistan Low High Autocracy Low Medium High Beyond the methodological appeal, studying the region of South Asia allows an opportunity to make significant empirical contributions. South Asia exhibits many of the motivating puzzles of this project such as enduring conflicts, draws, and state brutality, including a number of contemporary and ongoing conflicts that have real world significance for policymakers, yet they are under-examined in much of the civil conflict and counterinsurgency literature, a fact revealed by the consistent miscoding and underestimation of civil conflicts in South Asia. 5 2 Finally, aside from being a highly consequential for international security, South "0 While the Philippines, Indonesia, and met some case selection criteria, their varied colonial pasts would not afford me the same opportunity as the South Asian cases to control for colonial legacies which have a clear impact on the variables being studied. See Boone, 2003; William F. S. Miles, HausalandDivided: Colonialism and Independence in Nigeriaand Niger, Ithaca, NY: Cornell University Press, 1994; Alexander Lee and Kenneth A. Schultz, "Comparing British and French Colonial Legacies: A Discontinuity Analysis of Cameroon," Quarterly JournalofPoliticalScience, 7 (4), 2012: 365-410. 151 Byman, 2006, p. 81 12 See the appendices on case selection and what cases meet the criteria but are excluded. For instance, Fearon and Laitin (2003) simply group all of India's Northeast insurgencies as one single observation and ignore all the Naxal 122 Asia's relative openness, inherited institutions of British civil bureaucracy (including meticulous record keeping), and prevailing use of English in government documents and amongst officials makes it possible for scholars to conduct deep qualitative analysis to unearth "better data and better controls, thus contributing to better theorizing."15 3 Data Sources Because these South Asian counterinsurgency campaigns remain systematically understudied by scholars or are very recent, there is a dearth of scholarly history and detailed understandings of the contours and shifts in state strategy are yet unknown.154 Extensive data collection and detailed historical information were required for this project, warranting intensive primary field research, especially to offset the "urban bias" in civil war research that tends to privilege information and historical accounts from a narrow set of sources. Case studies and data collection absent "careful historical or ethnographic research" or without casting "as wide a net as possible with regard to sources" will be vulnerable to a systematic source bias. 155 The arguments and evidence presented in this project are underpinned by more than five months of field and archival research in Pakistan, India, the Kashmir Valley, and the British archives contributing to over 140 interviews conducted, roughly half with current or retired officials, military officers, civil servants, and politicians and the rest with knowledgable scholars, journalists, and civil-society leaders. While data and information were collected from a number of secondary sources, including various historical accounts, this research is supported by numerous primary sources, mostly collected during field research. These include: government insurgencies. Lyall and Wilson (2009) confuse the Assam and Mizo insurgencies as well as JVP and Tamil insurgencies and incumbents in Sri Lanka. Both miss a number of early civil wars from Telangana and Balochistan. 153 Stathis N. Kalyvas, "The Urban Bias in Research on Civil Wars," Security Studies, 13 (3), Spring 2004, p. 31. 154 There are some recent exceptions, which will be addressed in the specific chapters. 155 Kalyvas, 2004, p. 30. 123 documents and reports; military service journals and doctrinal works; memoirs, biographies, published interviews, and statements of former officials; local journalistic accounts and thinktank analysis; and quantitative micro-level data from a number of different local sources and datasets on conflict in South Asia. VII. Conclusion In the previous chapter, the central argument was to establish that not all military strategies to counter rebellion are the same and identify the different dimensions and potential types of counterinsurgency strategy. In this chapter, I have advanced a theory to explain a state incumbent's selection of counterinsurgency strategy along those two dimensions of effort and violence based on the patterns of uneven pre-war relationships of the state to its various territories and identity sub-groups. The theory of core-periphery relations-that is a state's relative value of territory and the embattled/rebel group's positional status within the state-will determine the type of strategy a state adopts when faced with rebellion. I expect that state incumbents will invest more effort when core territory is contested or threatened, based on a set of mechanisms that provide strong incentives and enable accessibility. On the second dimension I expect that state incumbents will employ less violence when core groups constitute the rebel identity base through a set of mechanisms triggered by the group's worth and embeddedness in state institutions. I also expect parallel predictions of less effort on peripheral territory and more violence against peripheral groups. To test my theory, I have chosen to employ a medium-N research design involving comparative methods and process tracing of multiple (29) campaigns by two South Asian 124 incumbents that would allow sufficient variation to test my theory against leading alternative explanations of strategic interaction and largely control for macro-level explanations of strategy Furthermore, I have developed a number of qualitative and quantitative measures of my independent variables for the medium-N analysis that enable better replication of the study, testing of the theory in other regions, and greater inter-coder reliability. At the same time, I have also identified a number of distinct mechanisms and processes to look for in the structured focused comparisons and case studies. Based on extensive new materials and data underpinned by extensive fieldwork, the subsequent chapters will proceed to test my theory on 14 campaigns by the Indian state and 15 campaigns by the Pakistan state. 125 Table 3.5: Full List of Counterinsurgencies to be Studied Date Start 1948 Date End 1948 Other Dataset Inclusion & Criteria Meets Fearon Lyall /Laitin Lyall (10OObd) PRIO X X ID # IND-01 Incumbent India Insurgent Razkars led by Qasim Rizvi War Hyderabad IND-02 India CPI Telangana 1948 1951 IND-03 India Nagas/Federals 1955 IND-03b India NSCN - IM, NSCN-K Naga Rebellion Naga II IND-04 India ongoi ng ongoi ng Mizo, MNF Mizo Revolt 1966 1986 X IND-05 India Naxalite (CPI-ML) India v. 1969 1971 X IND-06 India TNV, ATTF, NLFT, Tripura 1978 2006 Naxalite I 1992 India IND-08 India PLA, UNLF, KCP, KNF, PREPAK Sikh Militant Groups -Panthic Committees I, II & III - KCF, KLF, BTKF, BKI, etc IND-09 India IND-10 India IND-11 India IND-12 India IND-13 India CPI-M PAK-01 Pakistan Khan of Kalat, Prince Manipur Insurgency India-Sikh X X SATP data X X Fearon identifies the X X X X X MHA data; SATP data 1982 ongoi ng ~ 1983 1993 X X MHA data; SATP data X Insurgency Naxal II Andhra 1987 2004 ? JKLF, HM, HuA, LeT, JeM. And other Kashmiri groups Kashmir 1989 ongoi ng NDFB, ABSU Bodoland Insurgency 1989 ULFA Assam 1990 Insurgency Pakistan Nauroz Khan & Baluch rebels PAK-03 Pakistan BPLF Note/Reason for Inclusion ? X Naxal III Baloch I X X SATP data (Sahni article) X X SATP data ongoi ng? X X SATP data ongoi X X MHA data; X X X ng 2004 ongoi 1948 ng 1948 MHA data; X SATP data Estimates based on Harrison, Bansal, Khan, Baloch II Kundi 1958 1959 X Estimates based on Baloch III 1963 1969 (NI SATP data rebels PAK-02 battle death criteria Sunderal Report, Noorani Book; Sundarayya X ? PWG Karim & Baluch X Proof of Northeast as a catch-all for all NE insurgencies Insurgency IND-07 X X Meets PRIO (25bd) X X Described in depth as a COIN campaign by a Pak LtGen (Akbar Khan, Dawn, 1960) Described in Pakistan Doctrinal Work as COIN Harrison, Bansal, Bennett Jones, Kundi Mass Violence (Shah, GB, 2002; Rehman, GB, 2002; Ghani, GB, 2002) Described in Pakistan report, Lionel Baixas, 2008, p. 10 of 33 Doctrinal Work as COIN (Shah, GB, 2002; Rehman, GB, 2002; Ghani, GB, 2002) -',, Waziristan 1971 1974 1975 1971 1977 1976 Sindh MRD 1983 1983 PAK-04 PAK-05 PAK-06 Pakistan Pakistan Pakistan Shanti Bahini Baluchi Wazir Mullah East Pakistan PAK-07 Pakistan MRD Baloch IV I x x x x I x x x Akbar Ahmed x Tahir Amin; Described by Pakistan Civil Service Officer as rebellion x Aijaz Ahmed Described in Pakistan Doctrinal Work as COIN (Haq, GB, 2002; Ullah, GB, 2002; Tabbassam, GB, 2002; Shah, GB, 2002; Abbasi, GB, 2002; Rehman, GB, 2002; Ahmed, GB, 2002; Javed, PAJ, 2008) PAK-08 Pakistan MQM PAK-09 Pakistan TNSM MQM Uprising Swat I 1990 1994 1996 1994 PAK-10 Pakistan Taliban/Al Qaeda FATA I 2002 2007 x PAK-11 Pakistan BLA, BRA Baloch V 2004 ongoi x x x x x x x x x x x x x Jaffrelot, 2002 x SATP data; PIPS data SATP data; PIPS data SATP data; PIPS data SATP data; PIPS data SATP data; PIPS data ng PAK-12 Pakistan TNSM Swat II 2007 ongoi ng PAK-13 Pakistan TTP-Mehsud FATA II-SWA 2008 ongoi ng PAK-14 PAK15 Pakistan Pakistan TP-allies TTP-Other FATA II Other FATA 2008 FATA II - NWA 2008 ongoi x x x x ng ongoi ng , SATP data; PIPS data Described in Pakistan Doctrinal Work as COIN (Ullah, GB, 2002) Chapter 4: A Medium-N Analysis of Pakistan' Counterinsurgency Campaigns I. Introduction The Pakistan government has faced a lengthy history of insurgency, which began immediately after its birth in 1947 and has continued into the present. While Pakistan's external challenges have generally attracted more attention and ink than their largely neglected internal challenges, since the 9/11 and the 2001 war in Afghanistan, scholars began to focus on the strengths and limits of Pakistan's counterinsurgency efforts. Analysts outside and even inside the country had a very difficult time trying to comprehend what Pakistan was doing with the wide range of strategies, tactics, and operations employed in different parts of the country. Faced with insurgency in its border region analysts were baffled by the dizzying range of Pakistani strategies that appeared incoherent, irresolute, and inconsistent bordering on schizophrenic. Secretary of State Hillary Clinton accused Pakistan of "abdicating to the Taliban,"I while others suggested the state was only trying to "contain rather than eliminate militancy," 2 with operations that were "sporadic" 3 and "incomplete, inconclusive, and insincere." 4 At the same time, the Pakistan was praised for their "presence-oriented approach" 5 that General Petraeus described as "really quite impressive" and a model of "clear, hold, build, and transition," 6 focused on protecting the population and ensuring civilians "were equally 1Glenn Kessler, "Clinton Says Pakistan is Giving in to Extremists," Washington Post, April 23, 2009. Ijaz 2 14. Khan, "Challenges Facing Development in Pakistan's FATA" NBR Analysis, Vol. 19, No. 3, August 2008, p. 3 Ejaz Haider, "Counterinsurgency: The Myth of Sisyphus?" Pakistan'sCounterterrorism Challenge, Ed. Moeed Yusuf, Washington, DC: Georgetown University Press, 2014, p. 69. 4 Marvin G. Weinbaum, "Hard Choices in Countering Insurgency and Terrorism Along Pakistan's North-West Frontier, "JournalofInternationalAffairs, Vo. 63, No. 1, Fall/Winter 2009, p. 76. 5 Haider Ali Hussein Mullick, "Lions and Jackals: Pakistan's Emerging Counterinsurgency Strategy," Foreign Affairs (online), July 15, 2009. 6 General David H. Petraeus, "CENTCOM in 2010: Views from General David H. Petraeus," Event Transcript, Institute for the Study of War, Washington, DC, January 22, 2010. 128 involved in operations." 7 All the while Pakistan was alleged to be conducting a "dirty war" 8 involving "unfettered brutality" 9 and "kill and dump" policies,' 0 though Pakistani leaders described those efforts as "not serious," tolerating more chaos and instability, with security forces regularly "tak[ing] breathers." 1 In sum, Pakistani counterinsurgency appeared selectiveat times, reluctant and restrained, and at other times, intensive and brutal. Interested observers and policymakers struggled to make sense of Pakistan's wildly varying approaches that did not conform to expectations. Different strategies appeared to be deployed by the same state despite the fact that it was, all at once: (1) recently bolstered by increased capacity from significant U.S. assistance; 12 (2) unhampered by democratic interference due to a military-dominated government;1 3 (3) guided by a strategic culture characterized in part by the primacy of defense and stability on its eastern borders;14 and (4) fighting ethnic-based rebellions of comparatively similar size.1 5 This chapter draws on the core-periphery theory advanced in chapter 3 to explain some of these puzzling inconsistencies of campaigns by disaggregating Pakistan's varied history of ongoing and past counterinsurgency campaigns and strategies. 7 Tara McKelvey, "A Return to Hell in Swat," Foreign Policy, March 2, 2011. 8 Declan Walsh, "Pakistan's Secret Dirty War," The Guardian,March 29, 2011. 9 Robert G. Wirsing, Baloch Nationalism and the Geopoliticsof Energy Resources: the Changing Context of Separatism in Pakistan, Carlisle, PA: Strategic Studies Institute, April 2008, p. 30. 10 Saeed Shah, "In Remote Baluchistan, Pakistan Fights Shadowy War," McClatchy News, March 29, 2012. " Senator Raisani quoted in "We Only Receive Back the Bodies," The Economist, April, 7, 2012; Author Interviews, Islamabad, Pakistan, October 14, 2011 (#35). 1 Susan B. Epstein and K. Alan Kronstadt, "Pakistan: U.S. Foreign Assistance," Congressional Research Service, R41856, October 4, 2012, p. 40. 13 Aqil Shah, 2014; Ayesha Jalal, The State of MartialRule: The Origins of Pakistan'sPoliticalEconomy of Defense, Cambridge: Cambridge University Press, 1990; Hussain Haqqani, Pakistan: Between Mosque and Military, Washington, DC: Carnegie Endowment for International Peace, 2005; Ayesha Siddiqa, MilitaryInc.: Inside Pakistan'sMilitary Economy, London: Pluto Press, 2007. 14 Peter R. Lavoy, Pakistan'sStrategic Culture, Washington, DC: SAIC, October 31, 2006. 15 The non-state actor dataset from Cunningham et. al., 2009 shows the TNSM at 5,000, the TTP at 6,000, and the BLA at 5,000. 129 1.1. Literature Though systematic accounts of Pakistani counterinsurgency are few and far between, the literature tends to fall into one of two camps. The first suggests all strategy is largely uniform, guided by the state dominated by the military with the objective to crushing rebellions with brute force due to the strategic culture of the Pakistani state and the conventional orientation of the military. 16 A second camp offers a bit more nuance, arguing that the state has fought brutally and violently against separatism but has indulged violent political challenges to state authority emanating from groups more ideologically aligned with the state, including religious-based militant organizations.17 A third, newer explanation for relatively recent campaigns argues the Pakistani state has permitted insurgencies to endure in order to milk the international community of resources.' 8 All these existing explanations struggle to explain some part of the variation observe in Pakistan's full range of counterinsurgency campaigns. In many cases, the military has been more limited in its efforts and more measured in the use of force, as in the cases of Karachi and Malakand. Meanwhile, civilians have directed rather harsh campaigns such as Balochistan IV in 1973, not to mention actively supported other high-violence campaigns in the Baloch I, East Pakistan, and the South Waziristan (2009) conflicts. The ideology explanation has a hard time accounting for the low effort used to combat the numerous separatist insurgencies in Balochistan, the low violence against the MQM-Karachi insurgency, and the low effort and violence against Pashtun-based rebellions suspected by the state to have separatist inclinations such as the South 16 Haris Gazdar, "'Counter-Insurgencies' in Pakistan," Economic andPolitical Weekly, May 20, 2006, pp. 1952-53; C. Christine Fair and Seth G. Jones, "Pakistan's War Within," Survival, Vol. 51, No. 6, December 2009-January 2010. 17 This point is subtly made by Lieven, 2011. 1 Ayesha Siddiqa, "Pakistan's Counterterrorism Strategy: Separating Friends from Enemies," Washington Quarterly, 34 (1), Winter 2011, pp. 149-162. See especially pp. 154-157. 130 Waziristan 1975 campaign and the FATA I campaigns. Finally, the notion that Pakistan has milked insurgencies to portray a crisis or frontline state and acquire international support cannot explain some of the recent high-effort attempts to fully resolve conflicts and control territory as in Malakand and South Waziristan, nor can it explain the varying behavior of high and low-effort campaigns prior to 2001 before the international community was at all interested. While these current explanations are not wrong, they are still quite incomplete due to present-day bias and negligent examination of the full range of counterinsurgency variation throughout Pakistan's modern history. Figure 4.1: Timeline of Pakistan's Counterinsurgency Campaigns Ma"MjaU, Baloch V FATAI&II Malakand I Karachi Sindh S. Waziristan &"h IV E.engal Raloch III Upch II Bloch I 1945 1955 1965 1975 1985 1995 2005 2015 1.2. Recap of the Theory and Empirical Strategy Instead of fighting hard and violently against all insurgencies, the core-periphery theory expects that in the event of a rebellion, state incumbents will employ minimalist, low effort strategies like enfeeblement or mitigation to contain insurgencies threatening peripheral territory 131 with few incentives and high costs, unless rebels manage to project violence on core territory. Furthermore, it expects affective and strategic motives will restrain state incumbents to low violence strategies like mitigation or population control if the rebel identity group is perceived to have high moral worth or is highly embedded in state institutions. To this end, the country of Pakistan provides a valuable opportunity to test this theory because of its tremendous sub-national variation while controlling for confounding factors. The state of Pakistan has presided over an extremely diverse and contentious ethnic and political landscape carved out of British India in 1947. It has fought fifteen counterinsurgency campaigns against both separatist and ideological insurgencies' 9 against nearly every ethnic group in the Figure 4.2: Map of Pakistan's Provinces and Territories 19 These campaigns are identified with UCDP/PRLO criteria but include a number of cases missed by the Armed Conflict Dataset that are defined by Pakistani academic and government sources as counterinsurgency campaigns. 132 Table 4.1: List of 15 Pakistan Counterinsurgency C ampaigns Inclusion in other Datasets War Insurgent ID # PAK-01 Khan of Kalat, Prince Baloch I Karim & Baluch rebels Baloch II PAK-02 Nauroz Khan & Baluch rebels PAK-03 BPLF Baloch III PAK-04 Shanti Bahini East Pakistan PAK-05 Baloch Tribal Rebels Baloch IV Waziristan PAK-06 Mullah Noor Muhammad Sindh-MRD PAK-07 MRD PAK-08 MQM PAK-09 TNSM PAK-10 PAK- 1 PAK-12 PAK-13 PAK-14 Taliban/ al- Qaeda BLA, BRA TNSM TTP-Mehsud TTP-allies PAK-15 TTP-Nazir, Gul Bahadur, others Date Start 1948 1958 1963 1971 1973 1975 Meets Meets PRIO Proof of battle Lyall Fearon Date End /Laitin Lyall (1,000bd) PRIO (25bd) death criteria 4 Estimates based on 1948 Harrison, Bansal, Khan, Kundi 4 Estimates based on 1959 Harrison, Bansal, Bennett Jones, Kundi Mass Violence 9I 1969 report, Lionel Baixas, 2008, p. 10 of 33 1971 1 1977 I 9 Akbar Ahmed 1976 1 1983 1983 Karachi - MQM Malakand I 1992 1994 1996 1994 FATA I Baloch V Malakand II FATA II-SWA FATA II Other FATA FATA II - 2002 2004 2007 2008 2008 2007 ongoing ongoing ongoing ongoing 1 1 2008 ongoing 4 NWA *GB= Pakistan Army Green Book 1 9 1 Note/Reason for Inclusion* Described in depth as a COIN campaign by a Pak Lt. Gen. Akbar Khan, Dawn, 1960) Described in Pakistan Doctrinal Work as COIN (Shah, GB, 2002; Rehman, GB, 2002; Ghani, GB, 2002) Described in Pakistan Doctrinal Work as COIN (Shah, GB, 2002; Rehman, GB, 2002; Ghani, GB, 2002) Described by Pakistan Civil Service Officer as rebellion Tahir Amin; Aijaz Described in Pakistan Doctrinal Ahmed Work as COIN (Haq, GB, 2002; Ullah, GB, 2002; Tabbassam, GB, 2002; Shah, GB, 2002; Abbasi, GB, 2002; Rehman, GB, 2002; Ahmed, GB, 2002) 1 Jaffrelot, 2002 1 1 1 I '1 SATP SATP SATP SATP SATP & PIPS & PIPS & PIPS & PIPS & PIPS Described in Pakistan Doctrinal Work (Ullah, GB, 2002) data data data data data SATP & PIPS data country, and at certain points in time it has prosecuted multiple counterinsurgency campaigns simultaneously allowing for unique tests and structured, tightly controlled comparisons. For nearly all of its time since independence, Pakistan has been a weak and nondemocratic state (by most standardized measures) with a very large and capable professional military that has domineered over civilians and fought four major wars with India, its external threat and rival since its birth. These features allow me to hold constant country-level variables like regime type, state strength, relative strength to insurgents, and threat environment, and examine how the coreperiphery relations theory fares against other ones that claim to explain sub-national variation. Eventually even these controls can be relaxed to compare periods when Pakistan was relatively stronger or more democratic. The dynamics of Pakistan's different civil conflicts also vary, potentially offering greater generalizability. Some counterinsurgency campaigns have been quite short, one-off interactions (East Pakistan, Sindh, Mullah Noor Mohammed), while other conflicts with the same insurgent organizations have recurred due to the insolvency of the previous campaign (Balochistan I-V, Malakand I-II). In some campaigns, the Army has played the lead counterinsurgent role (East Pakistan, Malakand II), while in others it has deferred to the locally based policing forces (Karachi) or paramilitary (FATA I, Balochistan V). Money appears to be no object in some cases as resources are poured in (East Pakistan, Sindh), while other campaigns appear starved of them (Balochistan, FATA). And in some simultaneous counterinsurgencies, the state has offered significant political accommodations (Malakand II) while in others, comparatively little (Balochistan V). Pakistan's structural conditions have been largely constant since independence. It has been a weak and mostly non-democratic state (by most standardized measures), fielded a very 134 large and capable professional military that has mostly dominated civilians, politics, and national security decision making, and fought four major wars with India, its external threat and rival since its birth. These features allow this medium-N analysis to hold constant country-level variables like regime type, state strength, strength relative to insurgents, and threat environment, and examine how the core-periphery relations theory fares in explaining sub-national variation. Later, even these controls can be relaxed to compare periods of relative difference in Pakistan's structural conditions. 1.3. Roadmap The argument made in this chapter seeks first to convincingly establish a relationship between patterns of counterinsurgency strategy and the hypothesized independent variables of territory and identity. The second section draws on qualitative and quantitative data to identify the specific values of the theory's independent variables-the importance of various territories and identity sub-groups-across Pakistan's regions and the decades since independence in order to generate predictions. The third section identifies the strategy employed in each of the 15 counterinsurgency campaigns using quantitative and qualitative measures of the dependent variable as well as descriptive accounts of the effort, violence, tactics, and political authority. Finally, the fourth section evaluates the performance of the core-periphery theory in explaining variation in strategy across different campaigns and weighs it against alternative explanations, and the fifth section offers some conclusions on what this case tells us about counterinsurgency. The following chapter will then pick up on the arguments made here and proceed to a thicker qualitative examination of certain data-rich Pakistan cases from East Bengal, Malakand, FATA I, 135 and Balochistan V for evidence of hypothesized mechanisms linking territory to effort and identity to violence. II. Independent Variables: The Value of Pakistan's Territory and Identity Groups Pakistan's core has generally been defined as the people and territory of Punjab given their centrality to Pakistan's identity, economy, and politics. Other territories like Bengal and Sindh also offered great economic and strategic value, while the Western half of the countryBalochistan, the Federally Administered Tribal Areas (FATA), and the Federally Administered Northern Areas (later renamed Gilgit-Baltistan) possessed little value as peripheral backwaters. The North West Frontier Province (later renamed Khyber Pakhtunkhwa or KPK) and Azad Jammu & Kashmir (AJK) fell somewhere in between the two poles and would command value at different points in time. In terms of identity group value, the Punjabis certainly dominated the country but worked in partnership with the Mohajirs, and later the Pashtuns. Sindhis and Baloch, along with other religious minorities, were regarded of low worth and were poorly embedded within enduring core institutions like the civil bureaucracy and military, while Kashmiris and Hazaras had some degree of embeddedness but were for most of Pakistan's history, were too small to be politically salient groups. 2.1. The Incentives and Operational Costs of Pakistan's Regions and Territory The Government of Pakistan has presided over a crudely-stitched patchwork of territories and peoples, much like many post-colonial states, and the GOP has tended to look more favorably on some territories and sub-groups than others. In terms of territory, Punjab has always been the core, with other regions possessing lesser degrees of importance and peripherality. Prior 136 to the separation of East Pakistan in 1971, the state highly valued the territory of East Bengal, (though not the people), contributing about half of Pakistan's GDP during its first 25 years that disproportionately benefitted the Punjabi core. Sindh had been something of a core region prior to this but after it gained far more prominence as part of Pakistan's territorial core both its industrial and resource capacity. Meanwhile, the regions of Balochistan, FATA, the Northern Areas (now Gilgit-Baltistan), and AJK were largely relegated to the periphery, where KPK also resided until the turn of the century. In order to identify whether a particular territory was core or periphery, I drew on qualitative and quantitative data to systematically measure the incentives and costs of projecting force in a particular region. To estimate the relative material importance of each region, historical data was collected to measure material incentives based on production (GDP) and wealth (GDP/capita), as well as cash-crop and industrial value-added per capita. Strategic value was primarily estimated based on expert qualitative judgments of the region, and ideational value was based on content analysis/frequency of mentions in Pakistan news headlines. To measure the relative operational costs of projecting force on to each region, I collected data on the distance from the capital, physical terrain or infrastructure (road and electrification density), and human terrain or social infrastructure (human development index scores). Since states will actively invest in the physical and social infrastructure of regions that are materially or strategically valuable, thereby reducing future operational costs, these provide further indicators of the territory's importance. Based on these measures, Punjab, East Bengal, Sindh (both rural Sindh and the urban metropolis of Karachi), and are clearly core territories, while Balochistan, FATA, and Gilgit undoubtedly constitute Pakistan's periphery. KPK was more of a peripheral territory until it 137 experienced exogenous changes that became apparent by the early 2000s. 20 Meanwhile AJK is something of a mixed case, far less integrated into the normal Pakistani institutional architecture (like FATA, Balochistan, and Gilgit) with low value, modest costs, but potentially high ideational value should a conflict emerge. Some of the quantitative measures do not adequately capture the importance of territory and subsequent qualitative descriptions of these regions based on expert and practitioner accounts help substantiate and confirm the initial measures. A territory dubbed "peripheral" does not mean it is value-less, only that the costs of a high-effort counterinsurgency campaign would not be worth the returns accrued. For instance, Balochistan's mineral wealth, size, and ocean access are often cited as reasons why the state is of major strategic value, but the state does not need to assert and exercise the type of Weberian control to dominate and transform a region in order to leverage these types of resources. As a result, the state has tended to wield a type of political authority closer to administrative occupation rather than usurpation exercised in Sindh. 2.1.1. Punjab The territorial heart of Pakistan has always been the province of Punjab, home to the majority of its population, the largest share of GDP (after the separation of East Bengal in 1971), arterial road networks, and its concentration of urban and industrial assets. It is upon this territory that Pakistan's grand strategy and national security assessments were built. Cohen explains: Officers attending the Staff College (Quetta) are taught in their geopolitics course that every country has a core area, which contains the strategic centers of population, political authority, and the basic sinews of economic life, "the military loss of which would normally result in the One distinguished analyst pointed out that by the outbreak of the Swat conflict in 2007, the NWFP economy had become "totally integrated with Pakistan." Author Interview, Pakistani analyst, Lahore, Pakistan, October 1, 2011, (#18). 20 138 collapse of national resistance." Of course, if Pakistan has a "core area" it is Punjab, and the other three provinces constitute "invasion routes." Consequently, as its economic, strategic and ideational center, and because the greater Punjab region was a vital region of contestation for India and Pakistan,22 Pakistan's core military assets as well as six of the Pakistan Army's nine corps commands and five of eleven air bases were located in Punjab. Though the state has never fought an insurgency in Punjab, the prospect would pose little operational trouble for the government due to highly manageable physical terrain and strong state reach a state, and a long history of legitimate authority.24 2.1.2. East Pakistan (East Bengal) Prior to the 1971 war, Pakistan principally valued East Bengal (later renamed East Pakistan) as a core territory for what it materially contributed to the country. Like Punjab, Bengal was one of the crown jewels of British India, which was split in the 1947 partition of India and Pakistan. East Bengal contained more than half the Pakistani population, contributed over 50% of GDP through 1955,26 between 43 - 46% of Pakistan's GDP in the 1960s, 27 and 23% of central revenue receipts. The province housed two major ports critical to Pakistan's foreign trade,2 8 where it ran a trade surplus exporting jute and earned 60% of Pakistan's foreign currency Stephen P. Cohen, The PakistanArmy, Berkeley, University of California Press, 1994, p. 45. Gul Hassan Khan, Memoirs ofLt. Gen. Gul Hassan Khan: The last Commander-in Chiefofthe PakistanArmy, Karachi: Oxford University Pres, 1993, p. 174. 2 PAF Official Website (http://www.paf.gov.pk/active_bases.html). 2 Syyed Vali Reza Nasr, Islamic Leviathan: Islam and the Making ofState Power, New York: Oxford University Press, 2001. Gazdar, 2005, p. 1953 concurs with this judgment. 25 Perry Anderson, "Why Partition?" London Review ofBooks, Vol. 34, No. 14, July 19, 2012. 26 Feroz Ahmed, Ethnicity and Politics in Pakistan, Karachi: Oxford University Press, 1998, p. 16. 27 Based on data from 1959/60 to 1969/70 from S.M. Naseem "A Consistent Series of National Accounts for East and West Pakistan: 1949-50 to 1969-70," PakistanDevelopment Review, 1, 1975: 33-52. 28 Hasan Zaheer, The Separationof East Pakistan: The Rise and Realization of Bengali Muslim Nationalism, Karachi: Oxford University Press, 1994, p. 186, 223. 21 2 139 to offset West Pakistan's deficit, 29 and held substantial industrial assets, owned and managed by West Pakistanis.30 It is estimated that in the course of 20 years of this colonial relationship, East Pakistan transferred to West Pakistan resources estimated at $31 billion by 1970 or $187 billion in 2013 dollars.3 ' The resources and port access Bengal afforded Pakistan also offered strategic value. Multiple ports gave Pakistan access to two different maritime theaters while two land theaters allowed the Pakistani state to contest or defend against India on multiple fronts. Additionally, East Bengal contributed ideational value by validating Pakistan's founding principle of the two-nation theory despite significant ethnic and cultural differences. This narrative espoused the primary identity of Muslims on the subcontinent was their religion rather than their language or ethnicity or caste, thereby justifying the partition of British India and the initial creation of Pakistan.32 Controlling Bengal was not cheap. The distance and physical separation of the East Bengal territory from West Pakistan by 1600 km of Indian territory imposed the highest costs of operational control. However the physical terrain was not as challenging with much higher road density connecting a number of its urban hubs, no mountainous territory and only moderate forest cover. Human development and standards of living were lower in Bengal than the average of West Pakistan (though literacy was higher), but Bengal's human terrain was still more manageable as Bengal had been the epicenter of British India, and long subject to direct state authority in contrast to the unruly tribes of Pakistan's Western and Northwest regions. Akbar Ahmed, Jinnah, Pakistanand IslamicIdentity: The Searchfor Saladin, London: Routledge, 1997, p. 261. Zaheer, 1994, pp. 188, 216. 31 Feroz Ahmed, 1998, p. 20. Figure then converted to present value using inflation calculator. 32 Haqqani, 2005, p. 4; Samina Mallah, "Two-Nation theory Exists," Pakistan Times, August 6, 2007; Robin W. Winks, Alaine M. Low, The Oxford History ofthe BritishEmpire: Historiography,Oxford University Press, 2001. 3 Human development based on UNDP figures from 1980. East Pakistan literacy rates outstripped West Pakistan 21% to 15% (1951) and 17% to 12% (1961). Calculated from 25 Years ofPakistan Statistics, 1947-72, Karachi: Government of Pakistan, 1972, p. 12. 29 30 140 2.1.3. Sindh Following Punjab and East Bengal, the state considered Sindh the next most valuable territory, which only grew after the secession of East Bengal in 1971. The fertile and resourcerich Sindh province also hosted Pakistan's largest and most important industrial port and commercial center, Karachi, which served as Pakistan's capital for the first two decades. In the 1990s, Karachi contributed at least 20% of national GDP, 40% of national tax revenue, and 80% of Sindh tax revenue-and its relative contribution has continued to rise.34 Compared to the other provinces, "Sindh is much more organically connected with the economy in Punjab..." and "the economies of the two provinces.. .are wholly integrated." 35 Besides Karachi and its other urban hub of Hyderabad, Sindh's natural assets and fertile land along the Indus river plain has historically made a significant contribution to the economy, after irrigation projects in the 1930s dramatically increased agricultural productivity. 3 6 Despite all the attention accorded to the Balochistan's natural resource wealth (due to the absence of anything else), Sindh possessed 90% of Pakistan's coal reserves, including the world's third largest coal deposit in the world,37 and 70% of the country's natural gas reserves.3 8 In strategic terms,Sindh is also critical, affording Punjab and most of Pakistan its only major port access and naval base until the recent development of naval stations in Balochistan beginning in the 1990s. In addition, interior Sindh hosts two national highways that link Pakistan's North and South, and its border with India alongside Punjab, has featured prominently in the country's defense, particularly during the Rann of Kutch dispute in 1965. In 2005, then Chairman Joint 34 Lt. Gen. Jahan Dad Khan, PakistanLeadership Challenges, Karachi: Oxford University Press, 1999, p. 290. By 2010 it was estimated to provide a quarter of Pakistan's GDP and state revenue and contain half of Pakistan's banking assets and a third of its industry. See Lieven, 2011, p. 309. 35 Aijaz Ahmad, "The Rebellion of 1983: A Balance Sheet," South Asia Bulletin, 4 (1), Spring 1984, p. 43. 36 Peter Blood, "Sindhis," Pakistan:A Country Study, Washington: GPO for the Library of Congress, 1994. 37 Pakistan'sThar Coal Power Generation Potential,Private Power and Infrastructure Board, July 2008; Sumit Moitra, "Pakistan Seeks Coal India Help to Mine Thar," DNA India, April 29, 2013. 38 "Gas Reserves Discovered in Sindh: Report," Express Tribune, June 19, 2013. 141 Chiefs of Staff Committee (CJCSC) Gen. Ehsan ul Haq reaffirmed its value stating Sindh would continue to "play a great role in the defence, stability and progress of Pakistan." The CJCSC also went on to remind people of Sindh's ideational value having served as "the gateway for the prevalence of Islam in South Asia." 3 9 The high incentives for firmly controlling Sindh were magnified by the low operational costs the territory posed. In urban Sindh (Karachi and Hyderabad), high road density, limited physical obstacles, and a highly literate and strong a middle class population accustomed to state authority rendered these costs negligible. Moving to the rural areas of Sindh, the physical terrain became slightly more challenging primarily due to lower road density, but limited mountainous or forested terrain and relatively high levels of electrification and access to water would help sustain supply lines. And despite their frequent grievances, the rural Sindh population, much like in Bengal, generally accepted strong central authority, due to the long history a few feudal landlords dominating the vast number of poor tenant farmers. 2.1.4. Khyber Pakhtunkhwa (KPK)4 1 With a population and production capacity far less than Punjab and Sindh, and significant challenges to physical and human terrain, KPK is something of an intermediate value, but its importance has steadily grown to join the core of Pakistan as it entered the 21st century. Though small relative to the other high value provinces, KPK steadily grew each decade. Between 1970 - 2000, KPK had increased its provincial share of GDP while other provinces had lagged, and state leaders invested, particularly in the Malakand region, to upgrade physical and social "Pakistan occupies strategic importance: Gen Ehsan," January 14, 2005 (posted at PakistaniDefence.com, Last accessed September 2, 2014). 40 Literacy amongst urban Mohajirs was estimated to be as high as 80% in the late 1980s. See Lt. Gen. Jahan Dad Khan, 1999, p. 289. 41 Formerly known as the North-West Frontier Province (NWFP). 39 142 infrastructure and expand the writ of the state.42 Economic growth of KPK stemmed from the economic empowerment and industriousness of its primary population, the Pashtuns, the spread of resources during the Afghan jihad, and the rise in migrant Pashtun labor remittances fueled by the Gulf oil and construction booms.4 3 Former finance minister and current national security advisor, Sartaj Aziz, specifically noted KPK's economic bump around the turn of the century." Strategically, KPK was not central for the defense of the state from the East, but KPK did contain many arterial routes and access points that led directly to Pakistan's capital and nerve center of Islamabad and Rawalpindi. Given the history of conflict on Pakistan's western border and the numerous Afghan and tribal incursions, KPK would offer moderate strategic value in the 1980s and this would increase dramatically after the 2001 U.S. invasion of Afghanistan redirected some Pakistani strategic attention to the instability on its Western border. KPK's modest ideational value also increased over time. The Malakand division in particular had great symbolic value as its famous Swat Valley became a tourist attraction and was once referred to by Queen Elizabeth II as "the Switzerland of the East." 45 KPK also gradually became incorporated into more "normal" patterns of legal authority. Under the British, the KPK was considered a settled area relative to FATA and much of Balochistan, but as evidenced by the word frontier in its name (NWFP), it was still a land apart from Punjab. Until 1969, almost a third of the province, the Malakand division, held semi-autonomous status 42 Daud Khan Khattak, "The Battle for Pakistan: Militancy and Conflict in the Swat Valley," Counterterrorism Strategy Initiative Policy Paper, New America Foundation, April 2010, p 1; Akmal Hussain, PakistanNational Human Development Report 2003: Poverty, Growth and Governance, Karachi: UNDP, 2003, p. 11-14; Kaiser Bengali and Mahpara Sadaqat, Regional Accounts ofPakistan:Methodology and Estimates 1973-2000, Karachi: Social Policy and Development Centre, 2006. 43 Robert Nichols, A History ofPashtunMigration, 1775-2006, Karachi: Oxford University Press, 2008, pp. 140176. 44 Sartaj Aziz, "The Economic Cost of Extremism," Pakistan'sQuagmire:Security, Strategy, and the Future of the Islamic-nuclearNation, Ed. Usama Butt and N. Elahi, New York: Continuum International Publishing, 2010, p. 8587. 45 "Queen Elizabeth II in Swat," The Friday Times, April 4, 2014. 143 similar, and even after formal incorporation, Pakistani authorities willingly accepted a different standard of governance and violence in KPK than would be permitted in Punjab. By the turn of the 21st century, the notion that KPK was a "settled" area began to have real staying power and state leaders began treating it qualitatively differently than the tribal regions it bordered.4 6 The costs of operating in KPK were also lower than FATA and Balochistan, but higher than Sindh and Punjab. The physical terrain of the KPK was both mountainous and forested but this was slightly offset by improving road density and electrification that had increased in the 1990s onward.47 Close proximity to the capital as well as the garrisons of Punjab also reduced the burdens of any logistical tail. For much of its history, KPK human terrain was on average closer to Balochistan and FATA with lower human development, stronger tribal bonds, and a strong social culture of Pashtunwali, which is much more resistant to state and legal authority than in Sindh and Punjab.4 8 One author described the region of Malakand in the early 1990s as "one of the most backward in the country." 49 However, some pockets of KPK gradually became more "settled" and normalized during the 1990s - 2000s following economic growth spurts as evidenced by the rapid rise in human development that far outstripped the growth of any other province or the national average. 50 By 2006, KPK urban human development had moved closer to urban Punjab and Sindh and rural human development lagged only behind rural Punjab. This difference was first expressed to me by two different retired generals. Author Interview, phone, December 17, 2009 (#2); Private Meeting, April, 2011; Author Interview, Islamabad, Pakistan, October 17, 2011 (#39). Many others have since repeated this refrain. Along with this shift, there appears to have been some legal shifts in the late 1990s that may have been related to this ideational understanding. See International Crisis Group, Pakistan: CounteringMilitancy in PA TA, Asia Report No. 242, Brussels/Islamabad: ICG, January 15, 2013, pp. 3-7. 47 Robert LaPorte, Jr., "Direct Development: An Analysis of the Approach and Methodology of the Narcotics Affairs Section/Peshawar of the U.S. Embassy Islamabad," Unpublished Paper, August 16, 2008. 48 Lieven, 2011: 381, 385; Thomas H. Johnson and M. Chris Mason, "No Sign Until the Burst of Fire: Understanding the Pakistan-Afghanistan Frontier, InternationalSecurity, 32 (4), Spring 2008, pp. 41-77. 49 Kalim Bahadur, Democracy in Pakistan: Crises and Conflicts, New Delhi: Har-Anand Publications, 1998, p. 310 50 Calculated the average and growth rates from 1998 and 2005 HDI scores for each district. Data from Haroon Jamal and Amir Jahan Khan, "Trends in Regional Human Development Indices," Social Policy and Development Centre, Paper No. 73, July 2007, pp. 12-15. 51 Lieven, 2011: 519 46 144 2.1.5. Balochistan Accounts of Pakistan often describe Balochistan as a "wasteland of deserts and mountains,"5 2 populated by "villages of yellow-grey mud, which from the outside could be the first human towns of the Middle East 12,000 years ago." 5 3 The British had largely left this "poor and desolate wilderness" untouched as a buffer zone for its empire. 54 Despite significant resource deposits, these alone are not sufficient to substantially alter the cost-incentive calculation with which the state views Balochistan." What little development that exists is concentrated in the mostly Pashtun-inhabited Northeastern region while Western Balochistan remains a true frontier, "a haven for spies, smugglers, and refugees." 56 One report goes further: The land of Balochistan is exceedingly inhospitable; geologists have even compared the landscape with Mars. A Pakhtu expression, reflecting on ethnic relations as well as on geography, describes Balochistan as "the dump where Allah shot the rubbish of creation."5 7 Economic data only reaffirms the low ideational valued placed on the region. In a three different studies by Pakistani economists developing comparative index of Pakistan's 46 districts in terms of infrastructure, production, and social development, nine of Baluchistan's ten districts ranked dead last with only its capital of Quetta ranking higher, a regional imbalance that remains today. 58 A comprehensive study of Pakistani economic growth across its four provinces between 1972 - 2000 found GDP per capita grew 2.4% in Punjab, 2.2% in KP, 1.7% in Sindh, but lagged significantly in Balochistan at a mere 0.2%, "regressing further into under-development." 59 One analyst remarked Balochistan was "... isolated precisely because of the peripheral character of its Alok Bansal, Balochistan in Turmoil: Pakistan at Crossroads,New Delhi: Manas Publishing, 2010, p. 187. 53 Lieven, 2011, p. 340. 54 Lt Col Syed Iqbal Ahmad, Balochistan and its Strategic Importance, Karachi: Royal Book Company, 1992, p. 8, 77. 55 Syed Iqbal Ahmad (1992, p. 7) points out,scarcity of water imposed a constraint on all development including mining. 56 Syed Iqbal Ahmad, 1992, p. 47. 57 Peter Blood, "Baloch," Pakistan:A Country Study, Washington: GPO for the Library of Congress, 1994. 58 Akmal Hussain, 2003, p. 12; Shahid Kardar, "Polarization in the Regions and Prospects for Integration," Regional Imbalances and the National Question in Pakistan, Ed. Akbar S. Zaidi, Lahore: Vanguard Books, 1992, p. 322-23. 59 Wirsing, 2008, p. 24; Bengali and Sadaqat, 2006, pp. 74-75. 52 145 economy and its territorial isolation from the system of economic exchanges in the rest of Pakistan." 60 Strategically, Pakistani military officers have argued Balochistan "does not offer any worthwhile objective of military significance" other than serving as a buffer. 61 The region is also not easily vulnerable to external aggressors, who would face a daunting challenge trying to control it and require strong footholds in Afghanistan, Iran, and the Persian Gulf, Balochistan. It is then a place where "no major military operations could normally be conducted or conceived," whether by external forces or domestic counterinsurgents.6 2 Meanwhile, the costs of operating in Balochistan were exceedingly high and required a "big quantum of military forces" given this treacherous physical terrain. 63 The scarcity oof infrastructure, particularly water, and the difficulty of maintaining lines of communication posed "a major constraint on the deployment of forces and in selection of approaches." 64 Furthermore, one officer explains that Baloch social relations, characterized as "feudal militarism,"6 5 rendered the human terrain just as challenging. The "backward" population with the lowest human development, 66 full of "wild tribesmen" that the British described as "robbers and marauders" had a Spartan-like penchant for "resist[ing] aggression" and "wag[ing] war for decades."6 7 Therefore, 95% of Balochistan province (the rural areas where most of the population lives) has 60 Aijaz Ahmad, "The Rebellion of 1983: A Balance Sheet," 1984, pp. 27-44 (quote on p. 44). Syed Iqbal Ahmad, 1992, p. 11. 62 Syed Iqbal Ahmad, 1992, pp. 11, 17. 63 Syed Iqbal Ahmad, 1992, p. 262. 6 Syed Iqbal Ahmad, 1992, p. 6. 65 Blood, "Baloch," 1994. 66 Of the four provinces HDI, Balochistan is the lowest. Punjab (.557) followed by Sindh (.540), KPK(.5 10), and Balochistan (.499). See Akmal Hussain, 2003, p. 11; Ghulam Mohammad Arif, "Pakistan Poverty Assessment Update," Asian Development Bank, Project # 37717, Feb. 2006, p. 8. Furthermore Balochistan is much worse than the numbers let on. The percentage in high deprivation districts is 88% compared to KPK (51%), Sindh (31%), and Punjab (25%). 67 Syed Iqbal Ahmad, 1992, pp. 16, 18, 43, 77. 61 146 been administered as "B areas," where the provincial police have no jurisdiction and law and order is administered by tribal levies controlled by Baloch Sardars (traditional tribal leaders).68 2.1.6. Federally Administered Tribal Areas (FATA) FATA, like much of Balochistan, is another frontier where exercising state control and authority came at a much higher price than it was worth. After joining the state of Pakistan in 1947, FATA remained un-integrated with the rest of the country, offering "no revenue source"69 with little industrial or agricultural productivity and depending on government allowances to tribal elders, NGO aid, illicit trading and taxing of products, and migrant worker remittances, none of which was extractable. Some investments in development were promised to the region beginning in the 1970s but they were never seriously implemented.71 Little material value and limited ideational connection to the Pakistani state also matched the region's minimal strategic value other than serving as buffer between settled areas of Pakistan and the unruly aggressions of Afghanistan. The costs of controlling FATA far outstripped its limited value. Much of the territory remained highly inaccessible, unruly, and to borrow Scott's term, "illegible." This was partly related the "harsh and rugged terrain"72 of the Hindu Kush Mountains and the Safed Koh range running through the region, along with the hundreds of mountain passes and trails for infiltration, 68 Author Interview, former Punjab Police Inspector General, Lahore, Pakistan, October 8, 2011 (#26). 69 Hamid Hussain, "Back in the Game - Tribes and Society in Pakistan," Defence Journal(Karachi), April 2007 70 Amina Khan, "FATA: Voice of the Unheard - Path-Dependency and Why History Matters," Strategic Studies (Islamabad), Vol 31, No 1 and 2, Spring 2011, pp. 40-74; Ainslee T. Embree, "Pakistan's Imperial Legacy," Ed. Ainslee T. Embree, Pakistan'sWestern Borderlands:The Transformation of a PoliticalOrder, Karachi: Royal Book Co., 1979, pp. 24-41; Naseemullah, "Foundations of the Frontier," 2011. 71 jaz Khan, 2008, p. 16; LaPorte, 2008. 72 Amina Khan, 2011, p. 45. 147 making the border impossible to control.7 3 The physical terrain gave rise to a human terrain that was also very difficult to subdue. FATA was known for being exceedingly difficult to pacify, harboring "the perfect insurgents." 74 The culture of Pashtunwali,combined with trying physical terrain and low human development,7 5 frustrated nearly all historical any effort to assert central authority or control over the region. The British considered these tribal groupings "too heavily armed, too independent-minded, and too inaccessible in their steep and entangled mountains to be placed under regular administration." 7 7 Instead of the British employed a cheaper method of authority with an implicit non-interference policy that ensured "de facto independent status," governed by a loose system of political management and draconian collective punitive measures if any group stepped too far out of line. 78 Since "permanent stationing of large number of troops to keep turbulent tribes in check was not considered a cost effective," 79 the GOP maintained "minimum state penetration," 80 and governed FATA indirectly, outside the normal patterns of Pakistani authority. Unruly violence was tolerable below a certain level. For instance, in March 1998, tribesmen in Kurram agency battled each other for several days with mortars and surface-to-surface rockets until state forces finally intervened to de-escalate the conflict.81 FATA's system of isolation acquired a pathdependence as numerous power brokers in the region profited from the state of semi-autonomy, Sameer Lalwani, "Pakistani Military Capabilities for a Counterinsurgency Campaign: A Net Assessment," CounterterrorismStrategy InitiativePublication,New America Foundation, September 2009, pp. 22-23; Pervaiz Iqbal Cheema, "Challenges Facing a Counter-Militant Campaign in Pakistan's FATA," NBR Analysis, Vol. 19, No. 3, August 2008, p. 24. 74 Johnson and Mason, 2008, p. 50. 7 Though available data from FATA does not stretch as far back as the four major provinces, the current data suggests it is located somewhere on par with Balochistan. 76 Amina Khan, 2011; Johnson and Mason, 2008. 7 Lieven, 2011, p. 382. 78 Amina Khan, 2011, p. 40. 79 Hamid Hussain, April 2007. 80 Ijaz Khan, 2008, p. 16. 8 Anthony Davis, "Pakistan: State of Unrest," Jane'sIntelligence Review, 11 (1), January 1, 1999, p. 33. 73 148 developed a vested interest in maintaining their privileges, and integration plans further fell by the wayside due to prohibitive economic and political costs. 82 2.1.7. Other Areas Other territories, like Gilgit and AJK, have not been subject to major internal political violence but would fall into the same peripheral category as FATA and Balochistan given their physical separation, sparse population, neglected development, and frontier-like administration apart from the Pakistani mainland. Underdevelopment and poverty levels in Gilgit closely rival FATA (-40%) far below the national average.83 Much like FATA and Balochistan's "B areas," both non-provincial territories have exhibited a political authority closer to suzerainty than modern state governance. Gilgit-Baltistan was also governed directly by the federal government outside the constitution until 2009. AJK is similarly outside the normal system of governance as a self-governing state under Pakistani control, but it offers modest value similar to KP. 84 AJK's material value is rather limited but since 1947 it has held great strategic and ideational value for the GOP, much like the state of J&K for the GOL. ASK is predominantly rural though not particularly fertile, mountainous, and lacks an industrial base. While limited in production, ASK does not suffer the same degree of socio-economic destitution as the other peripheral regions. It possesses poverty levels on par with KPK (-27%), relatively higher literacy levels and primary Ijaz Khan, 2008, p. 20; Amina Khan, 2011, p. 56. FoodInsecurity in RuralPakistan, 2003, Ed. Amanur Rehman Khan and Shafqat Munir, Islamabad: World Food Program (WFP)/Sustainable Development Policy Institute (SDPI), 2004, Table 6.5, p. 138; "Pakistan Development Forum 2010" Reform Agenda-Road to Sustainable Economic Development, Govt of Gilgit-Baltistan, (citing Economic Survey 2009-10, NIPS, Assessment of Socio Economic Trends in GB 2005 (AKRSP), Annual Plan 200910 (Planning Commission) & GB Govt.). 84 Watts et. al., 2014, pp. 114-115. 82 83 149 school enrollment rates, and income levels bolstered by overseas remittances as well as almost 30% of the workforce in government service.85 Figure 4.3: Pakistan's Concentration of Economic Production (1973) (Source: "Pakistan - Industry, Mining, Power Centers," Map No. 501383 1973, Perry-Castaneda Library Map Collection, University of Texas) Food Insecurity in Rural Pakistan, 2004, p. 138; Teresita C. Schaffer, Kashmir: The Economics ofPeace Building, Washington DC: Center for Strategic and International Studies, 2005, pp. 29-32. 85 150 Table 4.2: Cost/Incentives and Predictions of Territory Incentives* Conflict 1. Baloch I Province/ Region Balochistan Costs* Cost-Benefit/ Expected Return Date(s) Material Strategic Ideational Distance Physical Human Terrain 1948 Low Low Low Moderate High High -5 (Kalat) 2. Baloch II Balochistan (Jhalwan) 19581959 Low Low Low Moderate High High -5 3. Baloch III Balochistan (Jhalwan Kohlu) East Bengal/ East Pakistan 19631969 Low Low Low Moderate -High High High -5.5 1971 High High High High LowModerate High 1.5 5. Baloch IV Balochistan Moderate Low Low High High High -5 6. Waziristan FATA (South 19731977 1975- Low Moderate Low Moderate High High -4 Waziristan) 1976 Sindh (Dadu, 1983 Moderate High High High Low- Moderate 2 4. East Pakistan 7. Sindh-MRD Nawabshah) 8. Karachi MQM 9. Malakand I Sindh (Karachi) KPK (Swat/ -High 19921996 1994 Malakand) 10. FATA I Moderate High High High High Low Low 4 Low- Low Moderate Low Moderate- High -2 High High -3.5 ModerateHigh Moderate High -5 Low- 2.5 Moderate High FATA 2002- (Waziristan) 2007 11. Baloch V Balochistan 12. Malakand II KPK (Swat/ 20042013 2007- Malakand) 2013 FATA (South Waziristan Agency) FATA (all other agencies) FATA (North Waziristan 20082013 Low Moderate Low Moderate High High -4 20082013 Low Moderate Low Low High High -3 20082013 Low Moderate Low Moderate High High -4 13. FATA II SWA/Mehsud 14. FATA II Other FATA 15.FATA II NWA Low Moderate Low LowModerate Low Low Low Moderate Moderate High Moderate -High Low Moderate Agency) *Value assigned to each incentive/cost based on following scale: High=2, Moderate=1, Low=0; Cost-Benefit therefore ranges from -6 to 6 2.2. The Moral Worth and Embeddedness of Pakistan's Identity Groups The dominant cleavage in Pakistan's identity politics has been ethnicity. In a country founded on the basis of cultural Islam, which became increasingly religious over time, the small populations of religious minorities left behind in partition (Hindus, Sikhs, Christians, and Buddhists) were easily marginalized. The history of Pakistani internal conflict is then one of ethnic dominance relations and "ethnic subjugation."8 6 Horowitz describes the Punjabis as the "advanced" ethnic community at the center, the Bengalis, Sindhis, and Baloch as the "backwards" groups on the periphery, and others residing somewhere in between.8 7 When Pakistan was formed in 1947, Punjabis emerged as the dominant elite alongside Mohajirs, Urduspeaking immigrants mostly from Northern India who settled in Pakistan, 8 while the majority Bengalis commanded very little institutional power, status, or respect. Punjabis eventually bcecme the dominant majority instead of the dominant minority after the exit of the Bengalis in the secession of East Pakistan (now Bangladesh). Over this time, Mohajirs receded to a junior partner role as Pashtuns rose in their embeddedness within enduring state institutions (e.g. the military and bureaucracy) while the Baloch and Sindhis continued to remain at the bottom despite periods of heightened political power and representation. Punjabi dominance stemmed from their overrepresentation in Pakistan's most powerful and professional institution, its military.8 9 While not purely exogenous, Punjabi dominance of the military is mostly attributable to accidents of history. Colonial policies stemming from the 1857 Sepoy rebellion that threatened the British Empire gave them a privileged status in the British Charles H. Kennedy, "Managing Ethnic Conflict: The Case of Pakistan," RegionalPolitics & Policy, 3 (1), Spring 1993, p. 123. 87 Horowitz, 1985; Kennedy, 1993, pp. 127-128. 88 The Mohajir appellation continues to be applied to their descendents, even those born in Pakistan. 89 Lieven, 2011; Jalal, 1990. 86 152 . Raj, where they at times accounted for nearly 50 - 60% of the British India Army.90 Punjabis, and to a certain extent Pashtuns were overrepresented in the military because they were deemed "martial" races while others larger groups like Bengalis were not, resulting in disproportionate recruitment levels. 9 1 Part of this stemmed from stereotypes of Bengalis as "clever, but untrustworthy" while Punjabis were thought "none too bright, but loyal." 92 In fact, the British created this myth in part to hedge against another mass uprising like 1857. Punjabis, Pashtuns, and other "martial" races were deemed loyal to the British Raj while Bengalis were not and so the British altered recruitment patterns using the trope of martial races to better secure the Raj from another mass mobilized rebellion. 93 Nevertheless, by partition, the inherited Pakistan military was disproportionately Punjabi and somewhat disproportionately Pashtun. Bengalis, Sindhis, and Baloch were woefully underrepresented while Mohajirs managed to have something like proportional representation early on. The well-represented groups are also reflected in who has occupied the position of Chief of Army Staff (effectively the highest position in the military and, for the most part, in the country) where there have been 5 Punjabis, 4 Pashtuns, 2 Mohajirs, 1 Kashmiri, and 1 Hazara. Other key positions in the military (Army Vice-Chief, ISI Chief, Chairman of Joint Chiefs of Staff, and Corps Commanders) also reflected this distribution with high numbers of Punjabis and modest numbers of Mohajirs and Pashtuns. The Pakistan civil service was also highly ethnically imbalanced, again as an accident of history. When the Indian Civil Service (ICS), which was not structured in such a disproportionate manner as that of the British Indian Army, split between India and Pakistan, 90 Yong, 2005, p. 17-18. Kennedy, 1993, p. 12 5 92 Charles H. Kennedy, "Pakistan: Ethnic Diversity and Colonial Legacy," The TerritorialManagement ofEthnic Conflict, Ed. John Coakley, 2003, p. 146. 93 Anderson, 2012. 91 153 officers were given their choice of country. Though Bengalis were overrepresented in the ICS, only one chose to come to Pakistan, leaving the rest of Pakistan civil service populated mainly by Punjabis and Mohajirs.94 Quotas were introduced under Prime Minister Zulfikar Ali Bhutto in an attempt to rectify the imbalance, particularly for Sindhis, but fell well short of achieving this goal. Historical British policies and byproducts of partition ultimately had a major path dependent effect on the embeddedness of groups in Pakistan's core state institutions, though discrimination towards particular ethnic groups-most also inherited from British colonial practices-certainly contributed to this exclusion. Conscious collusion seems unlikely given that the Punjabi bureaucratic elite and the landlord/politician-elite also had a deeply conflictual relationship that made collusion unlikely and difficult to sustain.95 Ethnic divides remained strong due strong kinship ties, even amidst growing urbanization, and very low inter-ethnic marriage rates, 96 but the dominant Punjabis still maintained some esteem or at least grudging respect for their junior partners while harboring intense disdain for groups like Sindhis and Baloch. Consequently, the state elites' estimations of group worth generally tracked closely with the degree of group embeddedness within state institutions. Of the smaller groups that were not involved in rebellion, Kashmiris gradually acquired both moderate to high group worth and some degree of embeddedness, Hazaras achieved some embeddedness but still low group worth due to their sect (at least after the rise of more conservative Sunni Islam in the 1970s onward), and religious minorities (Hindus, Christians, Buddhists, and Ahmadis, and others) possessed neither group worth nor embeddedness. 94Kennedy,2003,P.160. 95 Jalal, 1990, p. 223. 96 Lieven 2011: 18; Liquat Ali Khan, "Understanding Pakistani Geopolitics," Huffington Post, Nov 11, 2013. 154 Table 4.3: Population of Pakistan's Ethnic Groups (Note: because these are estimates they ma not total to 100 %) 1947 Punjabi 28 Siraiki Sindhi 6 7 Pashtun Mohajir 5 Baloch 2 Kashmiri Bengali 52 Other Total 100 (Sources: Multiple) 1951 29.9 1961 29.6 5.7 3.6 3.1 1.4 2.6 55.4 5.6 3.8 3.4 1.5 57.3 101.7 101.2 1981 48.1 9.8 11.8 13.1 7.6 4.2 2.4 0 2.9 99.9 1996 56 17 16 6 3 0 2 100 1998 45.4 10.9 14.6 13 7.8 3.5 2.8 0 2 100 2008 44.1 10.4 15.4 15.4 7.6 3.6 2.6 0 2.1 101.2 Table 4.4: Population of Pakistan's Provinces/Regions Province/Region Pakistan Punjab Sindh KPK Balochistan Fed. Captl. (Islamabad) FATA AJK (Sources: Multiple) % 100 61.1 17.8 13.6 3.3 0.3 3.9 2.6 1951 Million 33.7 20.6 6 4.6 1.1 0.1 1.3 .89 % 100 59.4 19.5 13.2 3.0 0.2 4.3 1961 Million 46.1 27.4 9 6.1 1.4 0.1 2 % 100 57.6 21.6 12.9 3.7 0.3 3.8 2.5 1972 Million 65.3 37.6 14.1 8.4 2.4 0.2 2.5 1.6 % 100 56.1 22.5 13.2 5.2 0.4 2.6 2.4 1981 Million 84.3 47.3 19 11.1 4.4 0.3 2.2 2 % 100 55.6 23 13.5 5.0 0.6 2.4 2.8 1998 Million 130.6 72.6 30 17.6 6.5 0.8 3.1 3.6 Values for each rebel group's worth and embeddedness can be estimated for each decade using a range of empirical support from scholarly research, government reports, speech evidence, and data on group representation within state institutions. Group worth is largely based on qualitative evidence as well as some data on prejudicial attitudes and discrimination, while group embeddedness is much more grounded in systematic quantitative data of group representation in critical state institutions including the military and civil bureaucracy (which is surprisingly more transparent and accessible than in India).9 7 Based on these qualitative and quantitative measures 97 Representation figures of senior level generals, the officer corps, the enlisted personnel, and for the civil service and drawn from over 30 data tables of ethnic representation (reported either based on residential location and/or language/mother-tongue) from over 30 sources. 155 , of the position of various groups (summarized in Table 4.5), the core-periphery theory would then expect strategies of high violence against Bengalis, Baloch, and Sindhis and it would predict strategies of low violence against Mohajirs and Pashtuns. Table 4.5: Identity Group Position by Decade (High, Moderate, 160s Low) 195s 1940 1940s Worth Punjabis Pashtuns High High Mohajirs Sindhis Baloch Kashmiri Bengalis High Embed Civil I Mod Low Milt High High High Low Mod Low Low Low Low Low Low Low Low Low High Low Worth High High High Low Low High Low 1980s Worth High Low High Low Low Low Low Low Low Low 970 1960s 1970s Worth Embed Worth Embed Civil Milt I Civil Milt Mod High High High High High Low High Low- High High High Mod High High High High High High Low Low Low Low Mod Low Low Low Low Low Low Low Low Low Mod Low High High Low Low Low Low Low Low 1990s Embed Civil Milt Punjabis Pashtuns Mohajirs High High High High High High Sindhis Baloch Kashmiri Low Low High Low Low Mod High High Mod High Low Low Mod - - -- Bengalis 1950s Embed Civil milt High High Low High Worth 2000s Embed Civil Milt High High Mod High High High Low Low High Low Low Mod High High ModHigh Low Low High Worth 2010s Embed Civil Milt High High Mod High High High Low Low High Low Low Mod High High Mod High Low Low High Worth Embed Civil Milt High High Mod High High High High High Mod Low Low High Low Low Mod Low Low High .... 2.2.1. Peripheral Groups Baloch The Baloch people were considered of low worth and had almost zero embeddedness within state institutions, being "the least developed and the least privileged of all Pakistan's ethnic groups."98 Group Embeddedness. Baloch exclusion and weak positional status is reflected in underrepresentation or no representation in nearly all of the state bureaucratic institutions. Part of 98 Lieven, 2011, p. 344. 156 this was endogenous to prior strategies of territorial management. From the very beginning of Pakistan and dating back to British policies under the Raj, the Baloch people were never integrated into state institutions, or for that matter the regular economy, instead retaining antiquated landlord-peasant relations. 99 The British found that "keeping Balochistan poor and illiterate served British imperial interests" and treated the Baloch much more harshly than Pashtuns.1 00 By accepting and preserving the feudal Sardari system peculiar to Baloch culture, the Pakistani state inadvertently reified Baloch exclusion from state bureaucracies. 10 1 Despite episodic claims by the state to recruit Baloch into the security forces based on regional quotas, other ethnic groups residing in the region tend to fill these slots. The Baloch "[continue] to be grossly underrepresented in all the organs of the Pakistani state and as a result people find it difficult to identify themselves with the government." 102 Group Worth. In the opinion of state leaders, the Baloch were of low worth and viewed as "stumbling, backward provincials." 10 3 President Musharraf was known for his disdain of Sardar tribal leaders but this contempt dated back to Ayub Khan who wrote in his diary of Akbar Bugti, "No bigger a scoundrel could be found anywhere." 04 Outside analysts noted a degree of ethnic arrogance by the dominant Punjabi community that possessed "a desire to show the 'primitive' Baloch tribesmen who is master" and looked down on them "with condescension and contempt."1 05 At best, attitudes towards the Baloch were "dismissive" 106 treating them like 99 Nina Swidler, "Brahui Political Organization and the National State," Pakistan'sWestern Borderlands: The Transformationof a PoliticalOrder, Ed. Ainslie T. Embree, New Delhi: Vikas Publishing, 1977 p. 107. 100 Syed Iqbal Ahmad, 1992, pp. xix; 53. 101 Syed Iqbal Ahmad, 1992, p. 154. 102 Bansal, 2010, p. 186; Feroz Ahmed, 1998, p. 253. 103 Kennedy, 2003, p. 143. 104 Craig Baxter, DairiesofFieldMarshallMohammadAyub Khan: 1966-72, New York: Oxford University Press, 2008, p. 343. 105 Selig S. Harrison, "Ethnic Conflict in Pakistan: The Baluch, Pashtuns, and Sindhis," Conflict and Peacemaking in Multiethnic Societies, Ed. Joseph V. Montville, Lexington Ma: Lexington Books: 1990, pp. 308, 312. 106 Stephen Cohen, The Idea ofPakistan, Washington DC: Brookings Institution Press, 2004, p. 224. 157 . "dalits" (lowest-caste Hindus) or the "Red Indians" of Pakistan, 107 while at worst the state establishment viewed the Baloch people as "a permanent threat and disloyal citizens." 108 One of the few ethnic Baloch to rise in the military ranks went even further alleging that in the state's view "Every Baloch is viewed suspiciously. They [bureaucracy] believe every Baloch is antinational and a nationalist."' 09 As further evidence of perceived Baloch disloyalty, the state actively supported Pashtun settlers in the northeastern part of Balochistan, whose loyalty they believed was assured." 0 Bengalis Similar to the Baloch, the Bengalis of East Pakistan were regarded as a low worth group, under the thumb of the Hindu Bengalis that made up 20% of East Pakistan's population,"' and were minimally embedded in state bureaucratic institutions but their exclusion was far more pronounced since they constituted the majority of the population. Though their representation in political institutions was proportionate to their group size, their absence from the security institutions and the bureaucracy proved quite consequential. Group Worth. The elites of West Pakistan could never conceal their beliefs in the social distance and "racial inferiority" of Bengalis."2 Bengalis were first considered a group emotionally apart from Pakistan, possessing "a very different ethnicity, history, tradition, language, and culture."' 3 In fact in the 1930s, Muslim leaders' initial imagination of the nation Saleem Shahid, "Action Group on Balochistan Urged," Dawn, September 10, 2012; "Baloch in Karachi Ask Not for Entitlement, Only Equal Rights," Express Tribune, September 24, 2012. 108 Sanaullah Baloch, "Baloch Demands Still Unmet," Dawn, February 19, 2009. 109 Lt. Gen Abdul Qadir Baloch, quoted on GeoTV, January 21, 2007. From Malik Siraj Akbar, "The Disloyal Baloch," January 22, 2007. 10 Lieven, 2011, p. 34 1 " Baxter, Dairiesof FieldMarshallMohammad Ayub Khan 2008, p. 474. 112Akbar Ahmed, Jinnah, Pakistanand Islamic Identity: The Searchfor Saladin, London: Routledge, 1997, p. 260. 113Ian Talbot, Pakistan:A Modern History, New York: St. Martin's Press, 1998, p. 186. 107 158 of Pakistan-which had got its name after the different Muslim peoples of the subcontinent-left out Bengalis. 1 4 The state elite treated Bengalis with more contempt than any of Pakistan's ethnic groups. Prior to the conflict, Bengalis were the victims by "unspeakable outpourings of racial hatreds and stereotyping."' 15 West Pakistanis had inherited British stereotypes of Bengalis as "feminine races" that were "feeble and spineless,"' 16 and "in thrall to 'Hindu' culture." 17 The private diaries of Field Marshall Ayub Khan, Pakistan's military dictator who ruled 1958 - 1969, revealed this deep-seated sentiment in his descriptions of Bengalis as non-rational "whose minds are totally shut to reason," "crooked," having "no culture or language of their own," "vulnerable to Hindu culture," choosing to "isolate themselves" and "cut themselves off from Muslim culture and thought." 18 Of West Pakistanis' perceptions of Bengalis, Ayub wrote that their "feelings of aversion are growing fast" and that they considered Bengalis as "unreasonable and insufferable companions."' 19 On the Bengalis chief political leader, Ayub wrote, "Mujibur Rahman is a pure and simple agitator. He has no brains, no administrative acumen, is unreliable, impulsive and emotional. He is a wrecker not a builder. Besides, his loyalty to Pakistan is extremely doubtful." 120 According to one Pakistani official, West Pakistani liberal elites, senior civilian bureaucrats, and political leaders all shared in this contempt of Bengalis. 12 1 "4 Pakistan was formed as an acronym P-A-K-S-TAN for Punjab, Afghan (referring to Pashtuns), Kashmir, Sindh, and BalochisTAN but did not include Bengalis. See Ahmed, Jinnah, Pakistanand Islamic Identity, 1997: p. 261. "1 Talbot, 1998, p. 212. 116 Barbara D. Metcalf and Thomas R. Metcalf, A ConciseHistory ofIndia, New York: Cambridge University Press, 2002, p. 111. 117 Saadia Toor, State ofIslam: Culture and Cold War Politics in Pakistan, London: Pluto Press, 2011, p. 28. 118 Baxter, DairiesofField MarshallMohammadAyub Khan 2008, pp. v, 132, 210, 259. 119 Baxter, Dairies ofFieldMarshallMohammadAyub Khan 2008, p. 461. 1* Baxter, Dairies of FieldMarshallMohammadAyub Khan 2008, p. 361. 121 Akbar Ahmed, 1997, pp. 260-61; Hamid Hussain, "Demons of December - Road from East Pakistan to Bangladesh," Defence Journal(Karachi), December 2002. 159 Group Embeddedness. Despite constituting a slight majority of the population, Bengalis were also staggeringly underrepresented in Pakistan's core institutions from the inception of the Pakistani state, constituting only 1% of military officers at independence, 4% by the mid 1950s, and while their numbers increased in the lesser air force (16%) and naval (10%) branches, Bengalis only made up 5% of Army officers by the mid 1960s, mostly concentrated in East Pakistan regiments. 122 In the mid 1950s, Bengalis held only 7% of the more than 700 top civil service posts and "by 1970,123 out of the twenty secretaries of the government of Pakistan, only three acting secretaries were East Pakistanis... [and] only one of the senior thirty-five generals was East Pakistani." 24 Sindhis Group Worth. Having played some role in Pakistan's independence and giving rise to a number of powerful political figures, Sindhis were not completely reviled the way the Baloch and Bengalis were but they were looked down upon with attitudes of "contemptuous." 12 Both Mohajirs and Punjabis "stereotyped Sindhis as backward, lazy, and provincial" and unwilling to leave their villages whether for migrant labor or martial service.1 2 6 Some ruling castes "made Sindhis the target of their modernist disdain" and "looked down upon Sindhi culture as 'feudal, primitive and backward," 1 27 composed predominantly of "peasants dominated by perverse feudal Kennedy, 1993, p. 128; Hasan Askari Rizvi, The Military andPolitics in Pakistan, 1947-86, Lahore: Progressive Publishers, 1987. 1 Talbot, 1998, p. 162. 14 Akbar Ahmed, 1997, p. 261. 125 Cohen, 2004, p. 224. 126 Theodore P. Wright Jr., "Center-Periphery Relations and Ethnic Conflict in Pakistan: Sindhis, Mohajirs, and Punjabis," ComparativePolitics, 23 (3), April 1991, p. 304. 1' Adeel Khan, Politics ofIdentity: Ethnic Nationalism andthe State in Pakistan, New Delhi: Sage Publications, 2005, pp. 139, 148. 2 160 chieftains." 128 It was this feudalism combined with perceived corruption of Sindhis and their lack of civic sense and morality that made other elite groups pity Sindhi peasants as "slaves," and revile their weak-willed feudal leaders.129 Ethnic prejudice persists and journalists have detected it has been directed even at powerful elites Sindhis, recently manifested in the vitriol towards the Pakistan People's Party (PPP) leader, Asif Ali Zardari.13 0 Besides prejudice and state employment, Sindhis were losers in the politics of distribution. The fact that Sindh produced an agricultural surplus, yet the rural districts were some of the most impoverished in Pakistan, indicating the state held little regard for much of the population even as it valued the land."1 One Pakistani ethnic politics scholar contends Sindhis suffered more severe national oppression than the Bengalis, which has reduced them to the status of Palestinians or American Indians.1 32 State leaders often suspected Sindhis of disloyalty given their interactions with Indian Sindhis (often labeling authors who published in or traveled to India as 'Indian agents'),1 33 commonalities with Hindus due to shared language, critical to Sindhi identity, as well as the notable absence of Hindu-Muslim violence during partition in 1947 that was formative for other areas like Punjab.1 3 4 Group Embeddedness. Sindhis were generally not embedded within state institutions for most of Pakistan's history. "Sindhi peripherality and subordination" went "unheeded by foreign Kennedy, 2003, p. 143. Lieven, 2011: 338; Taimur Arbab, "Wearing a Sindhi Topi and Ajrak Won't Rebuild a Backward Province," Express Tribune Blog, December 10, 2013. 130 Declan Walsh, "Pakistani Premier Faces Verdict in Clash with Court," New York Times, April 25, 2012. 131 Adeel Khan, "Pakistan's Sindhi Ethnic Nationalism: Migration, Marginalization, and the Threat of "Indianization" Asian Survey, 42 (2), March/April 2002, p. 218. 132 Feroz Ahmed, 1998, p. 69. 133 Feroz Ahmed, 1998, pp. 76-77. 134 Iftikhar H. Malik, State and Civil Society in Pakistan:Politics ofAuthority, Ideology, and Ethnicity, New York: St. Martin's Press, 1997, p. 216. 128 29 1 161 scholarship,"1 35 partly because they were not as glaringly absent as other groups like the Baloch or Bengalis, and also due to periods of political power some Sindhis wielded. Sindhis were tremendously underrepresented in core bureaucratic institutions save for short periods when political leaders intervened by changing the civil service and creating parallel security services. Based on his time serving in Pakistan, one retired British diplomat estimated that, "of the four nationalities in the country, the Baloch and the Sindhis were essentially excluded from the machinery of the state."1 Another scholar of Pakistani ethnic politics similarly concludes, "[Sindhis] have been virtually unrepresented in the higher echelon of the military and greatly underrepresented among top bureaucratic elite."13 7 One estimate just prior to the Sindh uprising in 1983 placed Sindhis at 2% of the armed forces and only 5% of the federal civil service. Despite being 60% of the population in the province, ethnic Sindhis only made up 40% of provincial government service, controlled 25% of the 2,000 industrial units within the province, and held few of the highest-level administrative positions. When Sindhis were employed within the state, they tended to occupy lower positions while outsiders dominated the senior ranks. In senior administrative and police postings in the province, Sindhis made up only 27 - 28% in the first two decades of independence. Sindhis were also poorly represented in private industry within Sindh. Though Sindhi representation in the bureaucracy increased in the 1970s under the democratically elected PPP government headed by Zulfikar Ali Bhutto (himself a Sindhi), the return of the military government in 1977 purged thousands of Sindhis from government jobs and public enterprises. 138 13 5 Wright Jr, 1991, p. 301. Sir Morrice James, Pakistan Chronicle,New York: St. Martin's Press, 1993, p. 199. Tahir Amin, Ethno-NationalMovements ofPakistan, Islamabad: Institute of Policy Studies, 1988, p. 197. 1 1 Khalid Bin Sayeed, "Pakistan in 1983: Internal Stresses More Serious than External Problems," Asian Survey, 24 (2), February 1984, p. 221.; Feroz Ahmed, 1998, pp. 71, 75. 136 1 37 162 2.2.2. Core Groups In addition to the dominant group of Punjabis, the Mohajir and Pashtun ethnic minoritiy groups also constituted part of Pakistan's core. Mohajirs Mohajirs commanded both high worth and embeddedness as the dominant group and founders of the country who eventually receded to a junior partner of the Punjabi majority. Muslims of North India had led the movement for the creation of Pakistan and when they immigrated to the country as Mohajirs, they were able to make the national language their native Urdu even though very few native to the territory of Pakistan spoke it. Because it was not easy for poor Muslims to relocate from India, Mohajirs who arrived in Pakistan were disproportionately literate, urban, skilled, and upwardly mobile middle class, and Pakistan's new leaders offered their co-ethnic Mohajir migrants numerous spoils of partition such as government posts and property abandoned by Hindus.13 9 Group Worth. Rather than being looked down upon as outsiders, Mohajirs were closely associated with the founding ideology of Pakistan-the "two-nation theory." They were "model Muslims" and "the most pro-Pakistan" of all Muslim groups.1 40 The myths surrounding Mohajirs that pervaded the newspapers were that "Mohajirs have made great sacrifices for Pakistan" and "Mohajirs have created Pakistan."141 Consequently they carried a chauvinism and sense of superiority as a dominant group even as they were reduced to the junior partner of the Punjabis after a decade. Mohajirs "continued to share the rule politically and economically by reason of their high literacy (70 percent), relative modernity, and historic role in the Pakistan Lt Gen Jahan Dad Khan, 1999, p. 295. 140 Feroz Ahmed, 1998, p. 104-105; Cohen, 2004, p. 203. 141 Feroz Ahmed, 1998, p. 154. 139 163 movement." 14 2 Even as Mohajirs began to agitate in politics and challenge the state due to their declining advantage, their rhetoric about being a "free" middle class rather than "feudal slave[s]" resonated with and commanded the respect of other social and ethnic groups. 14 3 Group Embeddedness. Almost overnight, Mohajirs became the majority in the nation's largest city and capital for the first two decades, Karachi. Mohajir immigrants composed a disproportionate share of the civil bureaucracy and urban intelligentsia in the early decades of Pakistan's founding when they wielded enormous influence over the politics of distribution and ideology. For instance Mohajirs held prominent positions in the cabinet, bureaucracy, and the leading party, the Muslim League that produced early leaders like Jinnah and Liquat Ali Khan. 44 As the dominant group within the state until 1958 and co-equal partners with the Punjabi elite, at least through 1971,145 "Mohajirs not only dominated politics and bureaucracy but also business" 146 and seven of the twelve biggest business houses were controlled by Mohajirs.1 47 Even after the relative decline of Mohajirs in the 1970 - 80s, one that was highly exaggerated by the Mohajir Quami Movement MQM, 148 they remained a powerful junior partner and "still disproportionately the most privileged of all the ethnic groups in urban Sindh."1 49 In terms of political power, the MQM virtually controls urban Sindh and because of this, it has been brought in as a coalition partner to every elected government since 1988, whether led by the PPP, Nawaz Sharif's PML-N, or Musharraf's military government.1 50 Wright, Jr, 1991, p. 300. Feroz Ahmed, 1998, p. 145. 44 1 Feroz Ahmed, 1998, p. 96-101. 141 Wright, 1991, p. 300. 14 6 Mohammad Waseem, "Ethnic Conflict in Pakistan: The Case of the Mohajir Qaumi Movement," Regional Security, Ethnicity and Governance: The Challengesfor South Asia, Ed. Justus Richter and Christian Wagner, New Delhi: Manohar, 1998, p. 78. 147 Cohen, 2004, p. 215. 148 Waseem, 1998, pp. 84-85. 149 Malik, 1997, p. 255. 150 "Timeline: A History of the MQM," Dawn, May 25, 2013. 142 143 164 Their high degree of embeddedness within the state was reflected in punching far above their population weight in senior civil service positions and the top military elite, producing two Army Chiefs (Gen. Mirza Aslam Beg and Gen. Pervez Musharraf) within a ten-year span in the 1990s.1 5 1 The man who led Pakistan's nuclear program, A.Q. Khan, and the man who transformed the ISI into a dominant domestic and international force, Gen. Akhtar Abdur Rehman, were also Mohajirs. Mohajir anger backlash stemmed from status decline relative to what they had been, and a desire to preserve their disproportionate position, no matter how unreasonable it may have been. Even though they constituted only about 20% of the population in Sindh, they insisted on 50% of the top provincial administrative posts.152 Today Mohajirs still hold high positions of power but anticipate their future in Pakistan is dimmer than their past."5 3 Pashtuns Pashtuns were of two minds, resisting the creation of Pakistan as a nation and periodically flirting with self-determination while Pashtuns had a firm foothold within state institutions, principally the military institutions owing to British martial race recruitment policies. However this ambivalence would gradually wane, particularly during the 1970s onward as their embeddedness and group worth grew, solidifying their position as another junior partner. Group Worth. Pashtuns were a minority that was sometimes looked down upon as being "lawless, rough and anti-intellectual" but these stereotypes also helped advance Pashtuns within society, government, and most importantly the security services.15 4 Pashtuns were likened to the "eighteenth-century Scots without alcohol," stereotyped as proud, brave, shrewd, independent 151 Waseem, 1998, p. 77-78; Tahir Amin, 1988, p. 174. 152 Adeel Khan, 2002, p. 223; Another source claims Mohajirs comprised 50% of Sindh population in the 1980s. See Jahan Dad Khan, 1999, p. 276. 153 Cohen, 2004, p. 217. 154 Kennedy, 2003, p. 143. 165 political realists with forceful character, who at the same time "draw tremendous advantages from membership of [the] state." 155 Stereotypes depict Pashtuns as middlemen minorities, "ambitious and hardworking" but also at times "shrewd and unethical."15 6 Scholars acknowledge that minority groups like the Baloch were much more alienated from the dominant PunjabiMohajir establishment while the Pashtun minority encountered "a more ambivalent Punjabi attitude" 157 and "[did] not feel a comparable sense of complete exclusion."1 58 The state often read the Pashtuns as loyal, even amidst a strong Pashtunistannationalist movement, and Pashtuns intense religious conviction made them appear as "exemplary Muslims."1 59 Pashtuns were the vanguard of Pakistan's effort to capture Kashmir in 1947, and during tensions with Afghanistan in the 1960s, Pashtun tribal councils in Balochistan pledged their support for Pakistan and condemned Afghanistan's actions.1 60 The claim that people from Punjab, particularly its northwest, "have strong cultural affinities with the Pashtuns,"161 has also been substantiated by recent survey data. There are few scholarly sociological studies on ethnic attitudes in Pakistan but one recent study sought to estimate prejudice of Pakistan's dominant Punjabis towards other major ethnic groups. The study found that the Punjabi majority viewed Pashtuns as the least socially distant, most proximate, Lieven, 2011, p. 371. Also see Feroz Ahmed, 1998, p. 227. Paul Titus, "Honor the Baloch, Buy the Pushtun: Stereotypes, Social Organization and History in Western Pakistan," Modern Asian Studies, 32 (3), July 1998, p. 674. 157 Selig S. Harrison, "Ethnicity and the Political Stalemate in Pakistan," Regional Imbalances and the.National Question in Pakistan, Ed. Akbar S. Zaidi, Lahore: Vanguard Books, 1992, p. 248. 158 Harrison, "Ethnic Conflict," 1990, p. 312. 159Titus, 1998, p. 674. 160 Paul Titus and Nina Swidler, "Knights, Not Pawns: Ethno-Nationalism and Regional Dynaics in Post-Colonial Balochistan," InternationalJournalof Middle East Studies, 32, 2000, p. 57. 161Anatol Lieven, "Pakistan's Paradoxical Survival," Pakistan'sCounterterrorismChallenge, Ed. Moeed Yusuf, Washington, DC: Georgetown University Press, 2014, p. 198. 155 156 166 and having "desirable physical and social attributes," while Sindhis and Baloch were regarded as equally socially distant.1 6 2 Group Embeddedness. Pashtuns have played a prominent role in the country's key institutions including the military and the civil bureaucracy. They started out early as heavily overrepresented in the military composing 35% of the Army and 45% of the most senior generals in the first two decades.1 63 As early as the 1960s, Field Marshall Ayub Khan, a Pashtun himself, began to empower his fellow brethren through supporting Pashtun migration to Karachi. 64 By the 1970s, a concerted effort at "the co-optation of the Pashtun elite through the Army into national power," was noticeable1 65 and Pashtuns eventually "acquired a strong junior position."1 66 Middle class Pashtuns' urban ranks began to integrate into Punjab, and their rural ranks "were tied to the Pakistani state through a long history of military service and even higher percentage of soldier than the Punjab."1 67 Because of their trade links across provincial and national borders, Pashtuns thrived in commercial life as well as in smuggling networks and the transportation sector. 16 Pakistan's participation in the Afghan insurgency helped to elevate the power and status of Pashtuns, and by the late 1980s, the Pashtun elite was "well-represented among the state elite" and "continued to find access to the higher echelon of the military and 162 Rehan Mullick and Joseph Hraba, "Ethnic Attitudes in Pakistan," InternationalJournalofIntercultural Relations, 25 (2), March 2001, pp. 165-179. In the study Mohajirs were viewed as the most socially distant and the authors suggested this was a function of their direct competition for top positions in the bureaucracy, but as the survey was conducted in 1996, perceptions of Mohajirs were undoubtedly influenced by ongoing insurgency led by the MQM. 163 Alok Bansal, "Factors Leading to Insurgency in Balochistan," Small Wars & Insurgencies, 19 (2), June 2008, p. 185; Amin, 1988, p. 81. 164 Author Interview with Pakistani analyst, Lahore, Pakistan, October 6, 2011 (#23). 165 Talbot, 1998, p. 252. 166 Feroz Ahmed, 1998, p. 69. 167 Cohen, 2004, p. 218. 168 Paul Titus, 1998, pp. 676-77. 167 bureaucracy."1 69 bureaucracy." 170 The "Punjabi-Mohajir-establishment" was now the "Pathan - Punjabi - Mohajir In the decade during major Pashtun-led rebellions in Pakistan's northwest, Pashtuns were "represented disproportionately in the country's power structure,"171 central to the decision making process, and their "influence runs from FATA to Peshawar all the way up to 'Pindi."' 7 2 Five out of the 13 Pakistan Army Chiefs of Staff have been Pashtuns, and they are believed to presently compose between 15 - 25% of the Army, as high as 22% of the officer corps, and a dominant share of the powerful Inter-Service Intelligence (ISI).m73 2.2.6. Pashtuns vs. Pakhtuns Some have alleged there is also a difference in the way in which settled Pashtuns and tribal Pashtuns are viewed by and embedded within the state with slightly greater social distance with tribal Pashtuns and greater proximity to settled Pashtuns. This distinction had a tribal basis to it as settled Pashtuns were from major tribes in KPK (e.g. Khattaks and Bannuchis) and were more accustomed to state authority while the "independent tribes," most notably the Wazirs and Tahir Amin, 1988, p. 239; Feroz Ahmed, 1998, p. 120. Harrison, 1990, p. 312; Syed Iqbal Ahmad, 1992, p. 301. Pathan is another word for Pashtun that the British popularized. 171 Robert Boggs, "Pakistan's Pashtun Challenge: Moving from Confrontation to Integration," StrategicAnalysis, 36 (2), 2012, p. 209. 172 This quotation refers to Rawalpindi, the home of the Army's General Headquarters. Author interview with Pakistani journalist, Islamabad, Pakistan, October 20, 2011 (#46); Author interview, Pakistani civilian analyst, Lahore, Pakistan, October 1, 2011, (#18). 173 Shuja Nawaz, CrossedSwords: Pakistan, Its Army, and the Wars Within, Karachi: Oxford University Press, 2008, p. 570; Hassan Abbas, "Musharraf Contends with the Pashtun Element in the Pakistani Army," Terrorism Monitor, 3 (42), Nov 1, 2006; In a presentation, Abbas says Pashtuns are 20-25% of the Army. See Hassan Abbas, "Pakistan's Volatile North-West Frontier: Pashtun Tribes, Taliban, and al-Qaeda," Presentation to the Jamestown Foundation, Washington DC, December 14, 2006. Dorronsoro (2002, p. 170) says Pashtuns are dominant in the ISI. See Gilles Dorronsoro, "Pakistan and the Taliban: State Policy, Religious Networks and Political Connections," Pakistan:Nationalism Without a Nation, Ed. Christophe Jaffrelot, New York: Zed Books, 2002, pp. 161-178.. Others estimate closer to a quarter. See Mark Silinsky, "The Problem with Pakistan," inFocus Quarterly, Winter 2010. 169 170 168 Mehsuds, that straddled the Durand Line between FATA and Afghanistan. 174 One scholar described it as a subtle difference between Pakhtun and Pashtun.1 75 Pashtuns were better integrated, not only in the armed forces, but also in the Pakistan civil bureaucracy, economy, and politics. Pakhtuns were less disciplined, more unruly, and in some ways were wards of the state, receiving free productive land, electricity and some services without paying taxes. 176 Even when Pakhtuns increasingly participated in the state, they still maintained a desire for independence and aloofness, a tension that would remain and keep them more distant than Pashtuns.1 77 Still Pakhtuns demonstrated their group worth with their fierce loyalty early on in the 1947 "jihad" in Kashmir. 78 And their leader, Maj. Gen Akbar Khan praised their valor and willingness to fight for country, stating, "Such are these people who today constitute one factor in our strength." 17 9 Pakhtun tribesmen were afforded opportunities for embedding in government institutions, both in the military and in the civil service with a number ascending to high-level posts. Pakhtuns were regularly recruited into the armed forces as well, particularly from the tribal areas. Both Mehsuds and Wazirs were well-represented in the Army as well the Frontier Corps, 180 and it is estimated that of the 90,000 Pakistani soldiers captured in East Pakistan in 1971 who would later be repatriated, 20% were tribals from FATA, making them roughly 5% of the Army at the time.18 ' Nevertheless, the distinction and relative position of Pashtuns and Pakhtuns would prove important. Selig Harrison, 1990: 315; Also based on Author Interviews. Author Interview, Pakistani analyst, Washington, DC, June 17, 2013 (#130). 176 Hamid Hussain, 2007. 177 Hamid Hussain, 2007. 178 Lieven, 2011, p. 383. 179 Akbar Khan, Raiders in Kashmir, Lahore: Jang Publishers, 1992, p. 196. 180 Imtiaz Gul, The Most DangerousPlace: Pakistan'sLawless Frontier, New York: Viking, 2011, pp. 36-38; Akbar S. Ahmed, Resistance and Controlin Pakistan,London: Routledge, 2004, p. 34. 181 Akbar Ahmed, 2004, p. 55; The Military Balance (London: IISS, 1971) in 1971 reports 365,000 Army troops. 174 175 169 2.2.3. Distinctions from the EPR Index Similar to the recent ethnic power relations index (EPR),182 I try to measure the relative embeddedness and strength of ethnic groups, based on multiple sources and types of data, and assign a value for each relevant group per decade. Though there is substantial correspondence, when these values sometimes differ from the EPR, my qualitative evidence provides extensive support for the assigned values. 183 Part of the problem is that indices like the EPR treat periods of formal civilian rule as substantively meaningful when a closer understanding of Pakistani politics reveals this is not the case. It also appears Large-N datasets like the EPR or even the Minorities at Risk dataset tend to overestimate formal positions of power and look exclusively at the executive branch without paying close attention to critical bureaucracies that carry out the day-to-day activities of the state. This is why they are able to code a group as powerless one year and a junior partner the next. Furthermore they have a hard time grappling with the unique status of this "elite" minority even when executive control is in different sets of hands. While I concur with the EPR that Punjabis were senior partners, and Bengalis were discriminated against, the index underestimates the power of Pashtuns and Mohajirs throughout stretches of Pakistan's history, underestimates the discrimination against the Baloch that remained fairly consistent throughout all of Pakistan's history, and overestimates the strength of Sindhis during periods of democratic rule. The evidence presented above and summarized in table 4.5 (based on historical data), shows Pashtuns had enduring power in core institutions, and continuous recruitment into the Cederman et. al., 2010. The EPR score for a country is based on a single expert's judgment, specifically weights formal rather than substantive power, and is admittedly limited to the executive branch rather than other institutions like the civil and security bureaucracies. 182 183 170 military and civil bureaucracy even when the military. Thus Pashtuns wielded power even when the military was not formally controlling government both because Pashtuns were diversified in other institutions (civil bureaucracy, the economy, and politics with a growing population share) and the military still wielded immense power behind the scenes of civilian governments.184 Likewise, Mohajirs did not lose their position just because of democratic elections in 1972 as the EPR implies. Though their influence and power was certainly watered, they still dominated the bureaucracy with over 30% of class I federal civilian bureaucrats and over 40% of senior executives in public corporations, and even though this would come down in Bhutto's reforms, they would still remain more than double their population share of 7.5%.185 The EPR codes the Baloch as powerless rather than discriminated, but throughout all these periods, qualitative accounts suggest not just that the Baloch were disregarded but in fact systematically discriminated against, even in their own province. Furthermore, Sindhis position certainly improved under Bhutto but their share of civil service positions was still roughly only half to two-thirds of their population share and they remained horribly underrepresented in a military that would dominate politics for the next three decades, even during the resumption of civilian rule. 186 2.3. Predictions Based on my core-periphery theory of how territory and identity shape state counterinsurgency strategy and combining the measures of the independent variables in the Jalal, 1990; Lieven, 2011; Siddiqa, 2007; Aqil Shah, 2014. Tahir Amin,1988, p. 82; Kennedy, 2003, p.163, Table 7.4; Selig Harrison, 1992, p. 228, table 1. 186 Sindhis were 12.7% of the population in 1981 but their overall representation in the federal bureaucracy was only 5.4% in the early 1980s. See Charles H. Kennedy, "Policies of Redistributional Preference in Pakistan," Ethnic PreferenceandPublic Policy in Developing States, Ed. Neil Nevitte and Charles H. Kennedy, Boulder, CO: Lynne Rienner, 1986, p. 79. 184 185 171 - Pakistan cases, I am able to generate a set of predictions. I expect the Pakistani state incumbent to employ strategies of higher effort in the regions like East Bengal, Sindh, Karachi, and KPK after it became more developed by the 2000s, and lower effort in Balochistan and FATA. Furthermore, I expect the Pakistani state to rely on strategies of higher violence strategies and brute force in conflicts against the Baloch, Bengalis, and Sindhis and less violent strategies with more non-kinetic substitutes and accommodations in conflicts facing Mohajirs and Pashtuns (additionally with an expected difference between settled Pashtuns and tribal Pakhtuns). 8 7 The prediction then is for attrition (high effort, high violence) in the cases Bengal and Sindh, population security (high effort, low violence) in the case of Malakand II and Karachi, enfeeblement (low effort, high violence) in the cases of Balochistan, and mitigation (low effort, low violence) in the cases of FATA I, Malakand I and FATA II. Figure 4.4: Distribution of Predicted Strategies for Pakistan Peripheral (Low) Territory Value Periphery Core (Low) (High) Enfeeblement Attrition Baloch I, II, III, East Pakistan IV, V FATA II-Other Sindh-MRD FATA II SWA/Mehsud Group Positional (WSothtd Embeddedness) Core (High) Mitigation Population Security Waziristan Karachi - MQM Malakand I Malakand II FATA I FATA II - NWA 8 Should the scenario ever arise, I would expect the state to also use low violence against Punjabis and Kashmiris and high violence in the event that any religious minorities rebelled. 172 III. Pakistan's Counterinsurgency Campaigns (1947-2012) To evaluate the predictions of the core-periphery, this next section depicts the values of the dependent variable in order to estimate how many Pakistan counterinsurgency case the theory can explain. This next section summarizes the backgrounds of each of these conflicts by region, describes each state counterinsurgency campaign, and assesses the strategy employed to identify a value for the dependent variable. I assess strategy based on careful estimations of the strategy's level of effort and violence based on critical decision points, tactical choices, and other qualitative information on the factors described in the theory chapter 2 (summarized in Table A2.2). The material used for this analysis is drawn from historical literature and journalistic accounts, memoirs, doctrinal works and military publications, interviews, and even some quantitative data on recent campaigns The Pakistani state has faced insurgencies in five distinct regions-Balochistan, Bengal, FATA, Sindh, and KP, and conducted fifteen campaigns to fight rebellion, one of which it lost outright.188 Organizing the counterinsurgencies by province/territory, I provide a narrative background and account of each campaign seeking to describe the levels of effort, violence, dominant tactics, and type of political authority wielded. I then use these accounts to specifically measure effort and violence for each strategy in qualitative terms and assign each a value for comparison. Finally, I seek to confirm each coding of strategy using observable quantifiable proxies for effort and violence collected over the course of this research. Though some try to describe a few of my observations as terrorism as opposed to insurgency, my discussions with Pakistani military officers and security analysts suggest the distinction may be arbitrary and unhelpful. 173 3.1. Counterinsurgency in Balochistan The ethnic Baloch people, concentrated in Balochistan's central and western regions, have fought four separatist rebellions against the Pakistani state and are currently engaged in another ongoing one. Throughout all these nationalist insurgencies often led by tribal leaders, Pakistani state responses have relied on the brutal and indiscriminate force of artillery, air power, and armored brigades to crush the rebels militarily and coerce their civilian support base into submission. In these campaigns, the state has offered little meaningful tactical or strategic substitutes to minimize violence via political accommodations or improving public welfare. Fi ure 4.5: Baloch I-IV Insurency Zones (Source: Harrison, 1981) While consistently highly repressive, the campaigns have been rather minimalist and reluctant, involving limited and/or intermittent troop commitments, materiel, or political resources, and repeating the same kinetic-centric tactics without developing new techniques for more sustained, durable control. General Zia ul Haq once expressed an opinion that helps explain this casual, low-effort approach, suggesting the GOP never worried about the "Baloch problem" 174 in its own right but principally feared external aggression or predation.1 89 In four of the five campaigns against Baloch insurgents, the Pakistani state pursued enfeeblement strategies (including an ongoing campaign) only attempting to disrupt and deter but never defeat the insurgency through supply- or demand-side efforts. The GOP may have exerted more effort towards an attrition strategy in one of the Baloch campaigns (IV), but even this is not definitive. 3.1.1. Baloch I (1948)90: Enfeeblement Background. Ever since Pakistani independence in 1947, the underdeveloped and indirectly governed territories under British rule that formed the province of Balochistan (specifically the Eastern regions encompassing parts of Kalat, Khuzdar, and Jhalwan) have witnessed a number of uprisings against the state including one ongoing since 2004. Resentful that the kingdom of Kalat had been absorbed into the country of Pakistan, the Khan of Kalat's brother Prince Karim led a secessionist attempt in 1948 involving a 1,500-man militia backed by the Zarakzai clan that took to Afghanistan and launched raids to spark a larger nationalist movement. Karim had taken his force to the Afghan border, established a new state cabinet, 191 and sought the support of Afghan forces and some in the Pakistan state believed his forces numbered in the thousands. Effort. Disregarding calls for a larger force based on much higher estimates of the rebel size and local tribal support, Lt. Gen. Akbar Khan tasked with managing the rebellion dispatched 189 Harrison, 1992, p. 255. 190 Accounts from Bansal, 2010: 36-38; Titus and Swidler, 2000; Selig S. Harrison, In Afghanistan's Shadow: Baluch nationalism and Soviet Temptations, New York: Carnegie Endowment for International Peace, 1981, pp. 2527; Mansoor Akbar Kundi, "Insurgency Factors in Balochistan," CentralAsia Journal,No. 64, Spring 2010; Maj. Gen. Akbar Khan, "Early Days of Freedom: Reminiscences of a Solider," Dawn (Independence Day Supplement, Karachi), August 14, 1960, p 1-2, 8; Maloy Krishna Dhar, "Balochistan - Cruces of History," M.K. Dhar blog, April 25, 2008; Martin Axmann, Back to the Future: The Khanate ofKalat and the Genesis of Baloch Nationalism 1915- 1955, Karachi: Oxford University Press, 2009, pp. 234-35. 191 Akbar Khan, 1960, p. 8. 175 a depleted 7th Baloch Battalion to counter Karim's rebellion. Even while other security advisors recommended a much larger force, Khan refused, believing the battalion would prove sufficient. Resources may have been constrained by the ongoing war with India but by April 1948 when government forces were dispatched after Karim, the intensity of that war had mostly died down. Within three months the unit succeeded in killing some insurgents and capturing Karim and 142 of his men, after which GOP forces wound down the operation, declaring victory, and allowing the insurgent remnants to retreat and regroup to the Baloch-Afghan border region where they continued some periodic low-level violence and ambushes. Violence. The state's actions were considered harsher than necessary and clearly antagonized the Baloch people. Additionally, the state made little attempt to bargain with Karim, work through the Khan of Kalat, or solicit the cooperation of local tribes with inducements. Instead, the state strategy relied solely on use of superior firepower and decapitation of the insurgent leadership. 3.1.2. Baloch H (1958 - 60)192: Enfeeblement Background. The absence of pursuit or consolidation of gains with a sustained presence would allow Baloch insurgents to rise again in 1958. In response to perceived state interference in Baloch provincial affairs and the arrest of the Khan of Kalat, 80-year old Nauroz Khan led his Zehri tribe in a rebellion. An estimated 500 - 1,000 Baloch guerillas took to the mountains, and an Army brigade including tanks and artillery moved in to trap them. The insurgents began 192 Bansal, 2010, pp. 38-40; Kundi, 2010; Harrison, 1981, pp. 27-29; Syed Iqbal Ahmad, 1992, p. 169; Dhar, 2008; Axmann, 2009, pp. 285-88; Titus and Swidler, 2000, p. 52; A.B. Awan, Baluchistan:HistoricalandPolitical Processes, London: New Century Publishers, 1985, pp. 227-29; Owen Bennett Jones, Pakistan:Eye of the Storm, New Haven: Yale University Press, 2002, p. 67; Sherbaz Khan Mazari, A Journey to Disillusionment, Karachi: Oxford University Press, 1999: 83-85. 176 ambushing Army and frontier corps troops and the insurgency grew and spread to other tribes across the Jhalwan subdivision. Effort. Again the state refused to dispatch more than a brigade to deal with the insurgents that had grown in size. The Army made few changes to previous campaign plan, nor developed new specialized capabilities to deal with what was evidently a recurring problem. Though it remained at it for a little more than a year, after the capture of Nauroz Khan, the army again abruptly ended operations. This time the government did establish a new garrison to maintain some degree of a state footprint in the region but it did little to sustain the campaign or prevent insurgents from regrouping and rearming before the next conflict. Violence. State forces grouped into guerilla-hunting parties aggressively swept sections of the Jhalwan district and sought out conventional encounters with large lashkar (militia) parties in order to "inflict heavy casualties on the guerillas."1 9 3 State forces later turned to highly punitive measures like aerial bombing and strafing of villages to coerce the Baloch population into submission. After a year and a half into the operation, a group of 163 insurgents led by Nauroz Khan surrendered (supposedly to Lt. Col. Tikka Khan) with the assurances of their conditions being met. Instead all were executed except Nauroz Khan who died in prison years later. No other political offers or efforts to redress grievances were carried out, and following the surrender, the security forces were alleged to have committed "widespread depredations" against the tribal groups in the surrounding area, killing as many as 1,000 Baloch civilians.1 94 193 194 Syed Iqbal Ahmad, 1992, p. 169. Dhar, 2008. 177 3.1.3. Baloch IH (1963 - 69)19s: Enfeeblement Background. Almost immediately after the Pakistan Army withdrew its major forces, the Baloch People's Liberation Front, a Marxist Baloch nationalist group, began organizing a new resistance base, first in the northern Marri and Bugti tribal region to counter Punjabi settlements, and soon this rebellion spread down to Kalat and Khuzdar districts. Sher Muhammad Marri led another Baloch rebellion by ambushing and raiding Army and Frontier Corps (FC) convoys, sniping at engineers and signals corps troops, and bombing infrastructure in the area. In the northern area alone there were an estimated 22 base camps and 400 full-time guerillas to train recruits and sustain a low-level insurgency with classic guerilla tactics. Violence. The Pakistani state response remained largely offensive, enemy-centric, and brutal towards civilians. The years of 1964 - 1965 witnessed large clashes but for the most part, the state relied on indiscriminate airpower to destroy the livelihoods of Baloch peasants reliant on subsistence agriculture. For a brief period, former Chief of Army Staff (COAS) Gen. Musa Khan, who was from Balochistan, was appointed governor and sought a different approach by releasing 1,300 rebels from custody. But distrusting this conciliatory gesture was only tactical, the rebels resumed their insurgent activities. Soon after, Maj. Gen. Tikka Khan was put in command of counterinsurgent forces and earned the nickname "the Butcher of Balochistan" during this time.196 Khan quickly abandoned Musa's gesture, sought to depose all ruling Sardars, and launched an offensive in 1968 on guerilla base camps. The offensive supported by the air force was alleged to have used napalm and served to further unite the Baloch rebel groups. The Account drawn from Harrison, 1981, pp. 29-33; Bansal, 2010, pp. 40-41; Bennett Jones, 2002, p. 67; Syed Iqbal Ahmad, 1992, pp. 174-177; Dhar, 2008; Awan, 1985, pp. 227-233; Cohen, 2004, p. 220; Titus and Swidler, 2000, pp. 55-58; Lionel Baixas, "Thematic Chronology of Mass Violence in Pakistan, 1947 - 2007," Online Encyclopedia ofMass Violence, June 24, 2008, p. 10 of 33; Human Rights Commission of Pakistan (HRCP), Annual Report, 1992 to 2006, Conflict in Balochistan:HRCP Fact-FindingMissions, December 2005-January 2006. 196 Mehtab Ali Shah, The ForeignPolicy ofPakistan:Ethnic Impacts on Diplomacy 1971-1999, London: I.B. Tauris, 1997, p. 96. 195 178 "Butcher of Balochistan" was not an outlier in the military. In fact, Pakistan's leader at the time, Field Marshall Ayub Khan, threatened the Baloch with "extinction."1. 97 Effort. Despite the brutality, the state deployed anemic efforts, relying on scant numbers of Frontier Corps, intelligence operatives, local police, other paramilitary militias, and air power. 198 The regular army did not begin to seriously engage in the campaign until the final year and seemed more focused on violence management rather than victory. After a change of leadership from one military dictator to another in 1969, a unilateral ceasefire and withdrawal was declared by Pakistan without any actual resolution of the conflict. The state again abruptly withdrew leaving guerilla formations not only in tact but in "a state of combat readiness."' 9 9 The insurgents continued to expand beyond their 900-strong force and retained "de facto authority of Marri-Bugti area." 2 00 One Pakistani officer writing on these campaigns prior to 1970 in the 2002 PakistanArmy Green Book explains, "In all three insurgencies, the army was very effective in controlling it but never managed to finish it." 201 State perceptions that the region was remote and the rebellion inconsequential continually facilitated such limited responses. This perception was shared by keen outside observers. When visiting Pakistan on a mission from President Kennedy in 1962, Henry Kissinger was asked to comment on the situation by a local journalist, and stated, "I wouldn't recognize the Baluchistan problem if it hit me in the face."2 0 2 Tahir Amin, 1988, p. 96. The exact number of forces is not really known but it one author estimates 37 platoons of militia. 199 Harrison, 1981, p. 34. 200 Syed Iqbal Ahmad, 1992, p. 177. 201 Brig, Syed Ithar Hussain Shah, "How to Cope with LIC at National Level," PakistanArmy Green Book 2002: Low Intensity Conflict, Rawalpindi, Pakistan Army General Headquarters, 2002, p. 180. 202 Selig S. Harrison, "Baluch Nationalism and Superpower Rivalry," InternationalSecurity, 5:3, Winter 1980-81, p.152. 197 198 179 - 3.1.4. Baloch IV (1973 -77) 203 Attrition Background. In the fourth Baloch rebellion, the GOP enhanced its efforts to counter the insurgency with a strategy much closer to attrition. The conflict began when Baloch irregular provincial forces allegedly attacked Punjabi settlers and federal employees in Las Bela. Soon after the Baloch were alleged to have been receiving arms shipments through the Iraqi embassy. However, most scholars agree the newly elected civilian leader Zulfikar Ali Bhutto was looking for a pretext to undermine the newly elected Baloch nationalist provincial government to further empower his own Pakistan People's Party. As the Baloch chief minister was trying to raise the presence of ethnic Baloch in the state security forces, Bhutto may have orchestrated both the Las Bela and Iraq embassy incidents to precipitate military action, consolidate political authority over the province, and distract the chief institutional rival of the military all in one fell swoop. 2 04 This Baloch uprising would spread to even more territory and involve the three largest Baloch tribes-the Marris, Mengals, and Bugtis-as well as lesser nationalist tribes. The GoP was convinced the Afghan government was responsible for "causing and sustaining rebellion in Baluchistan" backed by other powerful forces, a "Moscow-Kabul-Delhi axis." 205 Reported sizes of the insurgency vary by an order of magnitude, but this appears to have been the largest Balochistan insurgency to date in terms of expanse, duration, and scale. Some have estimated the insurgency involved as many as 11,500 - 20,000 rebels, but a more realistic estimate based on Accounts drawn from Dhar, 2008; Syed Iqbal Ahmad, 1992, pp. 184-86; Mazari, 1999, pp. 310-368; Titus and Swidler, 2000, pp. 58-63; Bansal, 2010, pp. 42-46; Harrison, 1981, pp. 33-40; Talbot, 1998, pp. 224-27; Awan, 1985, pp. 278-302; Kennedy, 2003; White Paperon Baluchistan, Rawalpindi: Government of Pakistan, October 19, 1974, p. 27; Alexandre Dastarac and Robert Dersen, "Baloutchistan: La Guerre Oubliee," Le Monde Diplomatique, August 1976, pp. 38-39. 204 Syed Iqbal Ahmad, 1992: 178-181; Titus and Swidler, 2000: 60-61; Nawaz, 2008, p. 335; Farhan Hanif Siddiqi, The Politics of Ethnicity in Pakistan: The Baloch, Sindhi and MohajirEthnic Movements, New York: Routledge, 2012, pp. 65-67. Advisor Rafi Raza claims Bhutto had intended to dissolve the government and send troops into Balochistan regardless of any provocation, See Rafi Raza, Zulfikar Ali Bhutto and Pakistan, 1967-1977, Karachi: Oxford University Press, 1997 p. 269. 205 Pervaiz Iqbal Cheema, The Armed Forces ofPakistan,New York: New York University Press, 2002, pp. 128, 131. 203 180 demographics and past rebellions is that it only began with a few hundred and may have escalated to 3,000 - 5,000 insurgents.2 06 This time, insurgents not only took to the mountains to ambush and sabotage state forces, but also conducted attacks and planted bombs in a wider area including the urban center of Quetta and other parts of northeast Balochistan where Pashtuns resided, as well as spreading to in the edges of Sindh and Punjab.2 0 7 Effort. The Pakistani campaign constituted a major escalation in effort compared to previous campaigns with the army deploying between two to four army divisions, 2 08 though over a larger contested space.209 Additionally, the force presence was sustained, remaining in the region at the behest of a civilian government years past when it claimed to have defeated the insurgency. At the same time contradictory accounts indicate army units avoided direct confrontation, rarely venturing into the mountains to minimize encounters with insurgents, potentially due to a casualty-averse force that had just taken a beating in East Pakistan. In almost three years of counterinsurgency, only 178 encounters were reported, almost half exclusively in the Marri region with "sporadic fights" and "few large battles."2 10 Estimates of troops lost were only 317 men from 1973 through mid-1976 (though they may have lost an entire unit of 300 men to an ambush in late 1976).211 This all suggests some degree of sticking to the playbook of the 206 Titus and Swidler, 2000: 61; Mazari says a "few thousand" while Dhar says 20,000 and Harrison 55,000, of which 11,500 were hardcore. Haqqani, 2005 also says 10-15,000 were trained in Afghanistan. The fact that the government declared victory in October 1974 after alleging to have only killed 385 and capture 5,501 Baloch suggests the estimates of a "few thousand" is more likely. See Mazari, 1999, pp. 318, 357, 368; Dhar, 2008, Harrison, 1981. 207 It spread to Dadu and DG Khan. See Titus and Swidler, 2000, pp. 61-63. 208 Some report as high as six divisions. 209 Mazari, 1999, p. 301. most likely the Army deployed no more than 3-4 divisions as another 30-35,000 troops (presumably Frontier Corps) were already stationed in Balochistan administering as much as 85-95% of the province. See "Justice Minister Denies Army Attack on Balochistan," Agence France-Presse,October 3, 1974. FBIS DAILY REPORT. Middle East & North Africa, FBIS-MEA-74-194 on 1974-10-04 p. TI. Accessed via Newsbank on February 19, 2014. 2 Dastarac and Dersen, 1976. The quotations are approximate as this article was translated by the author. Titus and Swidler (2000, p. 61) also describe it as "sporadic encounters" rather than "heavy fighting." 21 Titus and Swidler, 2000, p. 62. For encounters, see Harrison, 1981, p. 39. For casualties, the White Paper(1974) estimates 142 killed 1973-74 and Dastarac and Dersen (1976) estimate 175 from 1975-76. 181 previous three Baloch campaigns and containing conflict below a certain threshold. At least after 1974, the government forces seemed to relied on local tribal militias to police order though they may have colluded with the rebels.2 12 The GOP's larger investment of effort may have been driven by the spread of violence to Pashtun areas and outside the province, but it also appears Bhutto pushed for this extensive and extended deployment, largely to serve his own political ambitions rather than mitigate any heightened threat perception. Bhutto wanted abolish the Sardar system, install a submissive provincial government that would clear the way for his political agenda, and also occupy the military and much more expansive and intensive campaign served all these purposes. In the midst of the conflict, the government released a white paper in October 1974 declaring that despite the ongoing insurgency, the conflict had "virtually ended" and that the military operation had "achieved the results at which it was aimed" even though its writ extended to only 0.1% of the territory and forces would remain there another two and a half years. Violence. The state's strategy of violence intended to once again "pacify Balochistan by brute force."2 14 The military relied on a tremendous imbalance of firepower, indiscriminate shelling, aerial bombing and strafing missions involving thirty new Cobra helicopters with 20mm cannons against rebels with self-loading rifles. In a sparsely populated province, an estimated 6,000 - 15,000 civilians were killed.215 The state appeared to shift to more of a "war of attrition," trying to grind up insurgent manpower and its base of support, but also made periodic use of "shock action" like the massive air attack in Chamalang reported to have killed hundreds 212 Titus and Swidler, 2000, p. 62. 213 White Paper, 1974, pp. 3, 44. 214 Mazari, 1999, p. 301. Dhar, 2008 says 15,000. One former Pakistani official pointed out to me that despite these casualties being listed as insurgents, nearly all of them were civilians killed in collateral damage. Author Interview, Lahore, Pakistan, October 3, 2011, (#20). 215 182 of civilian villagers, as well as punitive attacks on villages in retaliation for ambushes.2 1 6 Civilians were also reportedly placed in the equivalent of concentration camps. 217 Non-kinetic tactics were deployed in name only. The government initiated some development projects but these seemed oriented to surveillance, repression, and extraction of resources rather than serious efforts to win over the population. One outside observer who traveled on a delegation to the Marri region noted that much of this development work was a sham for publicity but chronic poverty persisted. He also noted that negotiation attempts with tribal leaders were "half-baked" and the military forces actively deprived tribesmen of food supplies.2 1 8 In 1977, the military launched a coup against the civilian government and soon folded up operations against the Baloch insurgents with a declaration of amnesty. The Baloch tribal leaders, exhausted from the campaign, relented or retreated to Afghanistan. 2 19 However, the martial law administrator appointed to Balochistan began to flood the region with other ethnic migrants to water down the Baloch base of support and limit its capacity for rising again. What is curious is that while faced with a similar challenge or opportunity in a region composed of a different ethnic bloc, the same Pakistani leader chose a very different path, suggesting group identity might have played a role. Alongside the Baloch campaign, Bhutto also accused the KPK government (also outside his party's control) of colluding with the proMoscow Daoud government in Afghanistan on a Pashtun separatist agenda. While Bhutto used Syed Iqbal Ahmad, 1992, pp. 184-85; Mazari, 1999, p. 317; and Titus and Swidler, 2000, p. 61 all use the same language. 2 17 Mazari, 1999, p. 366. 218 Mazari, 1999: pp. 311, 316, 349. 219 For instance Marri retreated to Afghanistan with 3,000 fighters, whereas Mengal left the country for the UK. 22 0 MK Dhar, 2008. 216 183 this as a pre-text to dismiss their government, he deployed little violence to the region and instead worked to counter Pashtun nationalism through appeasement of the Islamist clergy.22 1 3.1.5. Baloch V (2004 -present).: Enfeeblement Background. After nearly 30 years without major conflict, Baloch insurgents re-emerged in late 2004 in resistance to an inequitable distribution of resources, administrative restructuring that would have undercut tribal leaders' authority, and state encroachment in Baloch territory to establish of military cantonments and a naval base at Gwader. The actual spark for Baloch violence was lit by the assault of a female Baloch doctor, allegedly by federal workers and they began attacking infrastructure in the region, particularly the Sui gas fields and pipelines in 2004 - 2005. Though there was little indication that this rebellion had separatist intent, the Pakistani state immediately saw it in these terms and redeployed troops from FATA to Balochistan because it saw this limited number of poorly armed or skilled Baloch rebels as a "bigger security problem."223 This insurgency was confined initially to the tribal militias and estimated at 1,000 insurgents, mostly targeting infrastructure and government security forces, but it eventually escalated and spread across the province into semi-urban areas and amongst the middle class with an estimated size of 6,000 - 7,500.224 Praveen Swami, India, Pakistan and the Secret Jihad: The Covert War in Kashmir, 194 7-2004, New York: Routledge, 2007, pp. 127-128. 222 Account drawn from Wirsing, 2008; Bansal, 2008; Watts et. al., 2014, pp. 138-143; Mansoor Akbar Kundi, "Insurgency Factors in Balochistan," The Nation, Oct 20, 2004; Frederic Grare, "Balochistan: The State Versus the Nation," Carnegie Paper, April, 2013; Human Rights Commission of Pakistan (HRCP), "Conflict in Balochistan" Report of Human Rights Commission of Pakistan Fact-Finding Mission, December 2005 - January 2006; Malik Siraj Akbar, Refined Dimensions ofBaloch NationalistMovement, Xlibris, 2011; International Crisis Group, Pakistan: The Worsening Conflict in Balochstan, Asia Report No. 119, Islamabad/Brussels: ICG, September 14, 2006; Center for Research and Security Studies, Balochistan'sMaze of Violence, Islamabad: Center for Research and Security Studies, 2011. 223 Zahid Hussain, The Scorpion's Tail: The Relentless Rise of Islamic Militants in Pakistan-AndHow It Threatens America, New York: Free Press, 2010, p. 87. 2 Estimates based on sources from previous footnote as well as "Pashto-Language TV Station [Khyber TV] Interviews Pakistan's President Musharraf," BBC MonitoringSouth Asia, May 21, 2006; Saeed Shah, 2012; Zahid 2 184 (source: hconomist) Effort. In response to the new outbreak of violence, the state initially deployed limited number of army troops-between a brigade and a division-followed by a number of frontier corps units to quickly crush the tribal insurgency and decapitate the leadership. After a brief lull, the insurgents resumed activities and the state increased deployments to an estimated 50,000 frontier corps troops (but withdrew army forces) backed by another 32,000 local police and levies, which was a relatively small force to manage the entire western Baloch region.225 The state was relying on poorly trained and equipped forces, and instead of investing in and improving them to conduct effective counterinsurgency activities, it also turned to unaccountable irregular assets handled by intelligence operatives.22 6 Operations were never dispersed or sustained and the state was reportedly limited to conducting periodic punishment and reprisals Hussain, "It's War Now," Newsweek, January 15, 2006; "2006: Who's Who in Balochistan," Dawn, May 28, 2011; Non-state actor dataset from Cunningham et. al., 2009. 225 Saeed Shah, 2012; M Ilyas Khan, "Balochistan 'Troops Numbers Cut,"' BBC News, May 14, 2008; Siraj Ahmed, "Police Reform in Balochistan," StabilizingPakistan Through PoliceReform, Report by the Independent Commission on Pakistan Police Reform, Ed. Hassan Abbas, New York: Asia Society, July 2012, pp. 113-119. 226 Author Interview, Former Pakistani official, Tysons, VA, August 15, 2013, (#139). 185 ' but could not effectively target insurgents or control territory.227 In recent years, the state has adapted slightly by choosing actions that garner less attention but it still has not developed a sophisticated method of intelligence collection and targeting. This is perhaps why one Pakistani elected representative described the state strategy as "not serious" without a plan for how to end the insurgency. At present, the GOP appears content with containment, "confident it can digest the current amount of violence that is taking place." 228 Violence. When President Musharraf dispatched state forces in 2005 to combat the insurgents, he boasted that they "won't know what hit them." 22 9 One scholar described the limited campaign as indiscriminate in its "unfettered brutality" by with bombing, shelling, mass displacement, prohibition of NGO relief, and the assassination of the Bugti and Marri tribal leaders.230 The objective was punishment to disperse insurgents, deter popular support, and disrupt insurgent leadership capacity. Despite these activities and the claim that the insurgency had been "crushed" with surgical actions, or perhaps because of them, the insurgency expanded and spread beyond its tribal roots, gaining support from broader sectors of society including the middle class and student organizations.2 3 Repressive operations nevertheless continued but in a less direct and visible guise, entering a period described by many as a "dirty war." This strategy, privately acknowledged by some in the Pakistani security establishment and reportedly waged by irregular assets and intelligence agents, has involved arbitrary arrests, abductions, enforced disappearances, and a Author Interview, Pakistani journalist, Washington, DC, May 23, 2013 (#122). Malik Siraj Akbar, "The Ifs and Buts of Negotiating with Insurgents in Balochistan," The Baloch Hal, August 1, 2013. 229 Naveed Ahmad, "Trouble in Pakistan's Energy-rich Balochistan," ISN Security Watch, January 30, 2006; Khurram Shahzad, "What Ails Balochistan?" The Nation, Feb. 14, 2005. 230 Quotation from Wirsing 2008, pp. 29-30. Force numbers range from 4,500 to 70,000. See Bansal, 2008; HRCP, 2005/06; Akbar, Refined Dimensions, 2011, p. 54; Tim McGirk, "Code of the Frontier," Time, January 30, 2005; International Crisis Group, Pakistan: The Forgotten conflict in Balochistan, ICG Asia Briefing No. 69, Brussels/Islamabad: ICG, October 22, 2007, p. 6. 231 Dhar, 2008; Grare, 2013; Mavish Ahmad, "Balochistan: Middle Class Rebellion," Dawn, June 5, 2012. 227 228 186 "kill and dump" policy that leaves mutilated bodies on the roadside.2 32 One former brigadier likened the treatment of the Baloch to that of the Bengalis.m Meanwhile civilian casualties are not accidental but actively created along with ruthlessly targeting of any and all insurgents. There is no such thing as a surrender and rehab policy the way there has been in other regions of Pakistan and, despite some modest changes to the distribution of funds to the provinces, substantively little has been done to improve public welfare or address the political demands of the insurgent base. 3.2. Counterinsurgency East Pakistan East Pakistan (1971): Attrition Background. The most well-known failed counterinsurgency campaign is the case of the Bengali secession from Pakistan, which led to the formation of the new state of Bangladesh. The Pakistani government had treated East Bengal as a colony with callous disregard for the welfare of Bengalis, uneven economic distribution, limited access to state institutions, and inequitable political representation. This second-class treatment combined with the nullification of an election results that would have brought Bengalis to power in Pakistan precipitated a nationalist political mobilization that escalated from civil disobedience to political violence and terrorism, to finally full-fledged guerilla insurgency led by the Mukti Bahini that broke out at the end of March 1971 and involved more than 100,000 trained insurgents. 23 4 The Pakistani state conducted 232 International Crisis Group, The Forgottenconflict in Balochistan, 2007, p 5; Grare, 2013; Walsh, 2011; "We Only Receive Back the Bodies," The Economist, 2012; Matthew Green, "Hidden War Embodies Pakistan's Struggle," FinancialTimes, May 25, 2012; "Baloch Missing Persons: 621 or 23,000? A Matter of Contested Figures," Express Tribune, March 9 2013; Sohail Khan, "Disband Agencies 'Death Squads', Stop Mly Action in Balochistan: SC," The News Sept 12, 2012. 233 Author Interview with retired Brigadier General, Islamabad, Pakistan, October 15, 2011, (#37). 234 The defection of the East Pakistan Rifles and civil armed forces in late March created 20-30,000 insurgents overnight and Swami estimates the Indian government trained 83,000 insurgents by the end of November 1971. See 187 a campaign of attrition that killed an estimated 100,000 to one million Bengalis, most of them civilians, and drove out millions of refugees. These actions led to a conventional inter-state war with India resulting in Pakistan's humiliation. In losing the counterinsurgency campaign and the conventional war, Pakistan also lost more than half its population, one-sixth of its territory that had produced nearly half its GDP, and more than 90,000 troops held captive by Indian forces. 235 Figure 4.7: East Pakistan Map "G"Mn' GiIAMSTA jyO PA'r> 'AB Effort. The Pakistani state expended considerable effort in the Bengal counterinsurgency campaign, though perhaps not as much as it could have. Despite the gravity of the insurgency, the military retained the bulk of its forces in West Pakistan fearing Indian maneuvers to exploit this vulnerability, but it still managed to muster one of the largest counterinsurgency forces it had ever deployed. It expended tremendous effort in the logistical feat of rapidly airlifting two Swami 2007, cited in C. Christine Fair, Fightingto the End: The PakistanArmy's Way of War, New York: Oxford University Press, 2014, p. 148. 2 Akbar Ahmed, 1997, p. 261; Zaheer, 1994, p. 216; Naseem, 1975, pp. 33-52; 25 Years ofPakistanStatistics, 1972, p. 194. 188 infantry divisions on long-haul flights around the subcontinent to reinforce existing divisions.2 3 6 It also recruited local irregular forces, managed to reconstitute the equivalent of a division to replace forces that had defected, and built up its total counterinsurgent force to an estimated at 150,000.237 The GOP operations initially cleared and secured urban areas and then sought to regain near total control of rural regions as well as the borders. The troops were also forced to operate against their conventional training by dispersing into smaller units to cover a larger area 238 and defend the borders, positions they would sustain until the Indian invasion.28 Overall, Pakistani forces did not simply attempt to pacify or disrupt but rather to defeat the insurgency and "transform" the region while completely wiping out insurgent combat power to prevent future uprisings.2 39 Violence. There's little doubt that along with high-effort, the GOP employed an extraordinarily brutal and violent counterinsurgency strategy during this campaign. The objective was to crush insurgent combat power and bludgeon Bengali civilians into submission. Initial shock actions of Operation Searchlight launched on March 26, 1971 sought to principally regain full control of urban centers. 2 4 0 This was followed by months of "search and sweep" and "clearing" operations, "punitive actions" and "kill and bum missions" 24 1 to regain ostensible control over the rest of the province and rural areas. The state unleashed the conventional firepower of tanks and mortars in urban and rural areas to both "break the backbone" the Haqqani, 2005, p. 78; Anthony Mascarenhas, The Rape ofBangladesh, New Delhi: Vikas Publications, 1971, p. 132. 237 Figures counted from Haqqani, 2005, p. 87; Lt. Gen. (Ret.) Amir Abdullah Khan Niazi, The Betrayal ofEast Pakistan, Karachi: Oxford University Press, 1999, pp. 52, 75-79; Siddique Salik, Witness to Surrender,Karachi: Oxford University Press, 1977, pp. 105, 123, 127. 238 Gul Hassan Khan, 1993. 2 39 Mascarenhas, 1971, p. 121; Abdul R. Siddiqi, EastPakistan the Endgame: An Onlooker's Journal1969-1971, Oxford University Press, 2004, p. 111. Karachi: 240 Jahan Dad Khan, 1999, p. 116. 241 Anthony Mascarenhas, "Genocide," The Sunday Times, June 13, 1971; Mascarenhas, 1971, p. 117. 236 189 insurgency 242 and "terrorize [the] population into submission." 24 3 In addition to the hundreds of thousands killed, the state's use of violence created 6 - 10 million refugees fleeing to India, precipitating a crisis and grounds for an Indian conventional intervention that defeated the Pakistan Army. The scope of the campaign and the violence was obviously much larger than anything that had been conducted in Balochistan up to this point as the state sought to brutally "cleanse" the province.244 However, the scale of effort had created significant externalities that had led India to intervene and balance Pakistan on behalf of the Bengalis, forfeiting all the territorial gains Pakistan's counterinsurgency efforts had accrued. 3.3. Counterinsurgency in Sindh Sindh has been the site of two insurgencies each led by its two largest ethnic groups. In 1983, ethnic Sindhis in rural Sindh led a civil disobedience movement that erupted into a rebellion and elicited a swift but brutal crackdown by the military. In urban Sindh, ethnic Mohajirs reasserting their power in the ethnic turf wars of Karachi in the late 1980s directly challenged the state, which led to a military-led counterinsurgency campaign starting in the summer of 1992, specifically targeting the MQM. In both cases, the state believed both ethnically based insurgencies possessed separatist goals. In both cases, the state expended high effort given the value of Sindh to the state but varied in their use of violence. I argue that in rural Sindh, the state employed a high-effort/high-violence campaign of attrition for only a short 1994, p. 141; Hamid Hussain, 2002; Salik, 1977, p. 92; Nawaz, 2008, pp. 266, 269. Hamid Hussain, 2002; Quotation from Consul General Archer Blood, Confidential Telegram 1007, U.S. Consulate (Dacca) Cable, Extent of Casualtiesin Dacca, March 31, 1971, Confidential, Source: Record Group 59, Subject Numeric File 1970-73, Pol and Def, Box 2530. Accessed via National Security Archive Electronic Briefing Book No. 79, Dec. 16, 2002. 244 Mascarenhas, 1971: 121; Abdul R. Siddiqi, East Pakistanthe Endgame: An Onlooker's Journal1969-1971, Karachi: Oxford University Press, 2004, p. 111. 242Zaheer, 190 period that successfully cowed the Sindhis into submission. By contrast, in Karachi, the state employed a high-effort/low-violence strategy of population control. 3.3.1. Sindh - MRD (1983): Attrition Background. In 1983 the Movement to Restore Democracy (MRD) organized major protests across the country against the military dictatorship of General Zia, but in Sindh where the movement had its strongest adherents, the national civil disobedience movement escalated into a violent uprising. After a few hints of unrest and violence, the Pakistani state perceived it to be a separatist insurgency and unleashed the military in what some have charged was a premature campaign of attrition against what the state claimed were "dacoits" and "criminals" with little consideration of negotiations or accommodation of the Sindhi base. 2 45 Constituted primarily of ethnic Sindhis, the movement was most intense in rural Sindh, particularly in the districts of Dadu and Nawabshah, and gave the impression to central elites that it was a motivated by Sindhi grievances and potential territorial aims.246 In fact popular support for Sindhi separation was incredibly high.24 7 The civil disobedience movement turned into an "open revolt" likened to the Palestinian intifada with "militant, insurrectionary actions from the start" targeting police stations, prisons, railways, banks, and other government establishments and killing a number of police and paramilitary personnel.2 48 Villagers were reported to be 245 Maleeha Lodhi, "A Nation At Risk," South: The Third World Magazine, London, December 1983, p. 22. 246 One sympathizer with the MRD movement notes that it had turned into more of a Sindhi nationalist movement by the time violence broke out. See Mazari, 1999, p. 551. Even supporters from the PPP acknowledged the moveent had aggravated "separatist feelings in Sindh." See "Demonstrations, Riots, Clashes Reported," Agence FrancePresse, August 21, 1983. FBIS DAILY REPORT. South Asia, FBIS-SAS-83-163 on 1983-08-22, p Fl (Accessed via Newsbank on Feb 19, 2014). 247 One journalist reported 95.8% of Sindhis would prefer to opt out of Pakistan. Suranjan Das, Kashmir and Sindh: Nation-Building, Ethnicity, and RegionalPolitics in South Asia, London: Anthem Press, 2001, p. 127. 24'Aijaz Ahmad, "The Rebellion of 1983," 1984, p. 42; Feroz Ahmed, 1998: 123. The violence was so intense that it was explicitly described by some scholars as a civil war. See Rahman, 1997: 127 cited from Nichola Khan, Mohajir Militancy in Pakistan: Violence and Transformation in the Karachi Conflict, London: Routledge, 2010. 191 "challenging often violently any image of martial authority they could find." 2 4 9 The Army was sent in to occupy Sindh and conduct counterinsurgency measures against the uprising and the power-brokers of the rural Sindh feudal system, the waderas (large landowners) and Pirs (saints) who had joined the fight against state forces,25 as well as well-entrenched defensive positions to punish and round up thousands amongst the armed and unarmed parts of the movement. 251 Effort. The state seems to have massively overreacted by deploying a three-division strong army and paramilitary force totaling as many as 40,000 - 60,000 troops along with the use of combat air support. 2 5 2 It is unclear whether the divisions came from the Army's V Corps in Karachi or was brought in from elsewhere but this was an unusually substantial commitment and costly decision. At the time, Pakistan was deeply embroiled in supporting an insurgency in Afghanistan against Soviet advances and facing an increasingly aggressive India on its Eastern border that would launch offensives in Siachen in early 1984 and Operation Brasstacks two years later. Pakistan also continued to sustain its forces in Sindh to patrol and police the rural areas for at least five months, well past the height of the violence, to ensure the movement was soundly crushed. Violence. Pakistan's deployment of the Army to deal with protesters-turned-violent appeared to be an extreme measure, but its bulldozing of houses, burning of crops, and use of Suranjan Das, 2001, p. 127. Harrison, 1992, p. 242. 251 Amir Ali Chandio, Mughis Ahmad, and Fouzia Naseem, "Struggle for Democracy in Sindh: A Case Study of Movement for Restoration of Democracy (1983)," Berkeley JournalofSocial Sciences, 1 (1), Jan 2011; Nichola Khan, 2010, p. 35; Jahan Dad Khan, 1999, p. 173 (who specifically served as Martial Law Administrator and Governor in Sindh in 1984). 2 s2 Tahir Amin (1988, p. 182) says 2 divisions totaling 40,000 troops. Talbot (1998, p. 254) says 3 divisions (which could be as high as 60,000). Harrison (1992, p. 242) says 45,000 troops plus an 8,000 man Hur force though he claims they remained aloof. Chandio et al, 2011 state that Hurs, Rangers, and Paramilitary were also activated; Roedad Khan says frontier corps and frontier constabulary were involved. See Roedad Khan, Pakistan-ADream Gone Sour, Karachi: Oxford University Press, 1997, p. 89. 249 250 192 airpower including planes dropping petrol bombs indiscriminately was certainly overkill.25 The military behaved with "extraordinary brutality" in rural areas and seemed to have little interest in a political approach, "having determined already in favor of a military one."254 The tactics involved arbitrary arrests, reprisal attacks against villages, aggressive use of torture, "search and destroy missions" targeting rebel locations, and "mopping up operations" with helicopter gunships.2 5 One journalist with close ties to the military wrote that the campaign involved "a ferocity reminiscent of the army's action in the former East Pakistan in 1971 .,,256 Within a threemonth period, as many as 600 - 800 Sindhis, mostly civilians, were killed (and perhaps 1,500 over the year) and 5,000 - 10,000 arrested, a remarkable figure given the fairly weak to almost non-existent insurgent capacity. 257 It is noteworthy that while the MRD agitation and resistance had stretched beyond Sindh to Punjab, a much more "measured, more prudent" approach was taken against MRD activists in Punjab, suggesting ethnicity played a major role in the response.25 8 3.3.2. Karachi - MQM (1992 - 96): Population Security Background Less than a decade later, Sindh was on fire again, this time in its urban hub of Karachi where the Mohajir Quami Movement (MQM) had began to both politically and violently assert itself. The Mohajirs felt a sense of status reversal after reforms under the PPP government in the 1970s imposed quota systems and language policies to benefit Sindhis at the 253 Chandio et al, 2011; Lodhi, 1983, p. 22. 24 Aijaz Ahmad, "The Rebellion of 1983," 1984, p. 42; This is confirmed by a government spokesman who stated that "the government had no intention of opening a dialogue with the opposition." See "Demonstrations, Riots, Clashes Reported," Agence FrancePresse, August 21, 1983. 255 Aijaz Ahmad, "The Rebellion of 1983," 1984, p. 42. 256 Lodhi, 1983, p. 22. 217 Das, 2001, p. 127; Tahir Amin, 1988, p. 197 estimates 600-800 with 800-2000 injured. A total estimate of 1500 killed is offered by Anatol Lieven, 2011, p. 314. Lodhi (1983) estimated 6,000 had been arrested by December. 25 Aijaz Ahmad, 1984, p. 42; Lodhi, 1983 p. 22. 193 expense of Mohajirs.2 59 Trying to forestall relative decline in state and local government influence, Mohajirs under the MQM leadership instigated an ethno-political turf war that ultimately turned violent against the state apparatus, 260 and may have flirted with territorial ambitions, seeking an independent country or "Hong-Kong-like status" for Karachi to be named "Jinnahpur." 26 1 While ethno-political violence in Karachi had been mounting since the late 1980s, in the early 1990s the MQM pivoted to an "insurrection," 2 62 began to target attacks against other political parties, civilians, and organs of the state, and even threatened to escalate to separatism.2 63 Clashes periodically occurred with the police, and eventually the army was called in for a period to quell this "mini-insurgency," patrol the curfewed streets of Karachi, and prevent MQM retaliations.2 64 In May of 1992, the military initiated a strategy of population control with Operation Clean Up-a 30-month military operation-led by the V Corps under the direction of Gen. Asif Nawaz.2 65 Effort. The deployment of the Army throughout Sindh but also to urban areas was itself both unprecedented and a costly choice for the state, made to restore some order to its key megalopolis. Along with the Army divisions from the V Corps, the state poured in additional material resources drawing on the Pakistan rangers, a 30,000 - 40,000 paramilitary border force commanded by Army officers, which worked with regular army forces to guard check posts with 2 59 Jahan Dad Khan, 1999, pp. 277-79. Both the President and Prime Minister described this as a rebellion or insurgency. See Talbot, 1998, p. 305; Azhar Hassan Nadeem, Pakistan: The PoliticalEconomy ofLawlessness, Oxford University Press, 2002, p. 129. 261 Malik, 1997: 225, 251, 254; Cheema, 2002, p. 133; C. Christine Fair, Urban Battle Fields ofSouth Asia: Lessons Learnedfrom Sri Lanka, India, and Pakistan, Santa Monica: RAND, 2004, p. 114; Christophe Jaffrelot, A History of Pakistanand Its Origins,London: Anthem Press, 2002, p. 35; Jahan Dad Khan, 1999, p. 288. These suspicions have since proven to be false. See, Ansar Abbasi, "Where PPP, PML-N and MQM stood on Jinnahpur in 1992," The News International,September 3, 2009. 262 Waseem, 1998, p. 89. 263 Malik, 1997, p. 226. On separatism, see p. 68 of Yunas Samad, "In and Out of Power but not Down and Out: Mohajir Identity Politics," Pakistan:Nationalism Without a Nation, Ed. Christophe Jaffrelot, New York: Zed Books, 2002, pp. 63-84. 26 Malik, 1997, p. 242; Jahan Dad Khan, 1999, p. 288. 265 Talbot, 1998, p. 324. 260 194 local police, protect the city, win local support, and gather actionable intelligence for more effective targeting of MQM insurgents.2 66 The government also expanded military presence with six new military cantonments in 1993 to limit infiltration and weapons smuggling. 2 67 In close coordination with the civilian administration, Army troops were forced to adapt to "police work" and tasked with "search and arrest" of suspects involved in violence against the state. 2 68 Local intelligence and policing activities were a major shift for officers unaccustomed to "patrolling in civic jungle" and who feared it would detract from their core expertise of focusing on kinetic anti-militant operations. 269 Likewise, coordination of intelligence across multiple agencies was uncharacteristic of the state.2 70 Finally, amidst a highly volatile political environment, the Army committed to continuous adaptation and shuffling of commanders to ensure professionalism and results.2 71 Even after the military drew down at the end of 1994, the state continued its sustained commitment to the campaign by maintaining at least 7,000 rangers patrolling within the city,272 expanding police forces by almost 20% (to 32,000) and upgrading them with special equipment, firepower, and transport vehicles, and subjecting them to an organizational and leadership overhauls to improve professionalism and generate results, a rarity in policing in Pakistan. 273274 By the late 1990s, violence continued but it remained exclusively criminal in nature. The threat Fair, 2004, p. 124; Quatrina Hosain, "Paramilitary troops to be increased in Karachi," Agence FrancePresse, June 18, 1995; Haider A. H. Mullick, Pakistan'sSecurity Paradox:Counteringand FomentingInsurgencies, Hurlburt Field, FL: Joint Special Operations University, December 2009, p. 21. 267 "Six New Military Bases to Be Set Up in Sindh," The Pakistan Observer, March 10, 1993, p. 1. 268 Nawaz, 2008, pp. 454-55; Brian Cloughley, A History of the PakistanArmy: Wars and Insurrections, New York: Oxford University Press, 2006, pp. 289-90. 269 Mullick, Pakistan'sSecurity Paradox, 2009, p. 21; Syed Ithar Hussain Shah, "How to Cope with LIC at National Level," 2002, pp. 181-82; Major Ata Ur Rehman, "Responsibilities of Pakistan Army in LIC Environment," PakistanArmy Green Book 2002: Low Intensity Conflict, Rawalpindi, Pakistan Army General Headquarters, 2002, .,0202. 7Fair, 2004, pp. 124-25. . 266 272 Anthony Davis, "Pakistan: State of Unrest," Jane'sIntelligenceReview, 11 (1), January1 1999, p. 33. 2 274 Fair, 2004, p. 124. John Stackhouse, "Karachi Pays High Price for Peace," The Globe and Mail, Oct. 25, 1996, p. A16. 195 to the state was considered to be defeated by 1996 and the MQM were brought into a political coalition in the central government in 1997.275 Figure 4.8: Map of Karachi, Sindh Sindh Province (Source: CNN & Wiki Commons) Violence. The campaign was also surprisingly restrained and limited civilian casualties considering state security forces faced a disciplined and organized insurgent group estimated at 1,500 - 4,000 capable of pitched battles with security forces and armed with automatic weapons and rocket launchers. 2 7 6 First, the transformation of the Army's aggressive practices in rural Sindh to patrolling, collecting intelligence, and making arrests was a stark contrast. 277 This was reflected in the much higher ratio of insurgents arrested to killed for an Army accustomed to shooting-first rather than making arrests.27 8 Even cordon and searches, decried as terribly indiscriminate methods that allowed for score-settling, 2 79 were still far more selective and less violent than any previous operations conducted in rural Sindh or Balochistan. Here the state was Cheema, 2002, p. 133; Nichola Khan 2010, p. 45; Anthony Davis, "Pakistan: State of Unrest," Jane's Intelligence Review, 11 (1), January 1, 1999, p. 33. 276 Waseem, 1998, p. 89; Fair, 2004, p. 119. 277 Cloughley, 2006, pp. 289-90; Nawaz, 2008, pp. 454-55. 278 Lt. Col. Hamid Hussain Naqvi, "Army and Sindh's Internal Security," PakistanArmy Journal, 35 (3), Autumn 1994, p. 60; Ikrarn Seghal, "The Ostrich Syndrome," Ikram Seghal Blog, December 22, 1994. 279 Fair, 2004, p. 125; Nawaz, 2008, p. 456. 275 196 - actively trying to distinguish and sort insurgents from Karachi's general population of "law abiding and patriotic citizens" 2 0 in order to minimize casualties. After the Army retreated to the background and the police took over, they also conducted operations primarily based on intelligence collection rather than indiscriminate operations, though there were accusations of some extra-judicial means in dealing with militants. 281 However, the use of the police meant that for one of the first times, state counterinsurgent forces were uniquely accountable and subject to disciplinary measures. This resulted in the reprimanding of 2,000 police officers, the dismissal of 800, and the jailing of 200.282 Even the Army subjected its forces to disciplinary actions for excess force against civilians.283 Furthermore, in addition to the relative restraint it displayed, Operation Cleanup also involved a political and social programs to improve social welfare, create employment, and provide public goods to complement the military's kinetic activities. 2 84 The state also sought to minimize excess force and pursue violent negotiation. Even when violence escalated in 1995 96, the Army sent in the lighter paramilitary rangers because "it did not want to be seen as fighting a civil war against the Mohajireen" scenes. 2 86 285 and in fact pushed for a settlement behind the Even the state's attempt to split and prop up a moderate faction, the MQM-H, sought to segregate and rehabilitate a part of the insurgents with the goal of reducing violence and enabling bargaining over brute force, though it has been criticized for backfiring and escalating violence.2 87 Ultimately this violent negotiation paid off as the MQM was brought into a political 28 0 Nawaz, 2008, p. 456. Waseem, 1998, p. 89; Talbot, 1998, p. 344. 282 Fair, 2004, p. 125; Stackhouse, 1996. 283 Cloughley, 2006, p. 291. 2 1 Nawaz, 2008, p. 455; Naqvi, 1994, p. 61. 285 Malik, 1997, pp. 250-25 1. 286 Talbot, 1998, p. 343. 287 Siddiqi, 2012, pp. 107-08; Fair, 2004, p. 121. 281 197 coalition with the ruling party in 1997 and then made a central component of the military's ruling strategy after General Musharraf took over in a coup in 1999. Even as they fought, there appeared to be restraint and close ties between elements of the state and the MQM. For an Army built like a hammer, which had never found an internal security problem that didn't look like a nail, Army leadership was surprisingly hamstrung in this case, "unable to provide or implement any solution that may involve the greater use of force," 2 88 partly due to the identity of the combatants. Reports suggest Mohajir Army chief of staff Aslam Beg was reluctant to deploy the Army, which is why operations did not begin until the summer of 1992 when Asif Nawaz took the helm.2 89 A restrained Army operation "had no wish to create martyrs," 290 or desire to "[fight] against their fellow citizens," 2 9 1 sometimes to a fault, allowing "major culprits [to remain] at large." 292 Intelligence officials who have observed the Mohajirs's transformation to more disciplined and militant politics betray a begrudging admiration for the group's strength and resiliance, similar to how the state has often viewed Pashtuns.2 93 There was also evidence the intelligence agencies may have even supported the MQM and Mohajir militants for a time. 294 The bureaucracy, who controlled local police forces, also helped restrain state actions. One former Army officer who served as governor of Sindh describes the empathy accorded to the Mohajirs: "it was obvious that by then most of the Mohajir bureaucrats sympathized with MQM and were not prepared to take any step which was detrimental to its interests."2 95 288 289 290 291 292 293 294 291 Malik, 1997, p. 253. Cloughley, 1999, p. 288. Malik, 1997, p. 248. Malik, 1997, p. 253. Malik, 1997, pp. 248-249. Lieven, 2011, p. 326. Cohen, 2004, p. 355 note 21; Khalid Hassan, "Zia Formed the MQM: General Beg," Daily Times, May 27, 2007. Jahan Dad Khan, 1999, p. 281. 198 & 3.4. Counterinsurgency in Khyber-Pakhtunkhwa 194296. 3.4.1. Malakand I (1994) Mitigation Background. In 1994, Pakistan faced an Islamic insurrection when the Tehrik-e-Nifaz-eShariat-e-Mohammadi(TNSM), an offshoot of Jamaat-e-Islami,began to agitate in large numbers for Sharialaw and against a federal decision eliminating tribal courts. These agitations in the Malakand region of KPK, predominantly populated by ethnic Pashtuns, turned violent in the spring and by the fall, had escalated to a rebellion. The TNSM had been mobilizing for months in the spring of 1994 and blocking off access to government officials declaring they had established "a state within a state" and even began restricting access to the regions borders.297 In November, the 2,500 - 5,000-strong armed wing of the TNSM with an estimated support base of 40,000 declared jihad, denied the legitimacy of the state, and took over government buildings, an airport, and major roadways. 2 98 In response, the Pakistani government deployed a mitigation strategy involving very limited locally based paramilitary forces to do just enough to quell the situation with minimal bloodshed. Effort. The GOP hesitated for months to get involved and deploy force, tacitly accepting the militants' expansion, arms buildup, and political demands in order to avoid conflict, but after the actions in November, the state finally deployed a 4,000-strong force composed of local levies International Crisis Group, Pakistan: CounteringMilitancy in PATA, 2013; Asad Munir, "Narratives on the TTP," The News International,July 6, 2013; Javed Aziz Khan, "Soldier Who Led Operation Against TNSM Passes Away," The News International,April 20, 2009; "Country's 'Biggest-Ever' Drug Seizure Reported," The News, January 29, 1995, p. 1. FBIS Daily Report. Near East & South Asia, FBIS-NES-95-019 on 1995-01-30, p. 71.(Accessed via Newsbank, February 19, 2014). 297 "The Showdown in Malakand," The FrontierPost, November 7, 1994, p. 10 (FBIS Daily Report. Near East South Asia, FBIS-NES-94-216 on 1994-11-08, p. 70 (Accessed via Newsbank, February 19, 2014); Author interview with a former U.S. official based in Pakistan during this period, Washington DC, July 11, 2013 (#134). 298 Robert LaPorte, Jr., "Pakistan in 1995: The Continuing Crisis," Asian Survey, 36 (2), Feb 1996, p. 179-89; Lubna Abid Ali and Naveed Iqbal Khan, "The Rise of the Tehreek-e-Nifaz-e-Shariat-e-Mohammadi in Malakand Division, NWFP: A Case Study of the Process of 'State Inversion,"' Pakistan Vision, 11 (1), 2010, pp. 89-119; Charles Lindholm, "Justice and Tyranny: Law and the State in the Middle East," Journalof the Royal Asiatic Society, 9 (3), Nov 1999, p. 386; Beena Sarwar, "Pakistan-Politics: Radical Muslims Ignite Northern Hills," IPS-Inter Press Service, November 7, 1994; Tahir Amin, "Pakistan in 1994: The Politics of Confrontation," Asian Survey, 35 (2), Feb 1995, pp. 140-146. 296 199 and frontier corps troops to deal with the rebellion. State forces swiftly conducted operations to eject militants from government buildings and free government employees who had been held hostage and then retreated after one month. The state made no comprehensive efforts to improve the local administrative or policing capacities that had allowed this uprising to emerge or address the economic and legal grievances that had precipitated such a rebellion. Instead, political promises were made to appease the militant base, but nothing was ever enforced to implement substance change in the environment. Figure 4.9: Malakand Division of Khyber-Pakhtunkhwa (KPK) (Source: Stratfor) Violence. The state strategy employed a highly restrained combination of "force and persuasion"299 confining the scope of the counterinsurgency operation to just those rebels engaged in direct combat against government forces and institutions. Paramilitary forces limited 299 Amin, 1995: 143. 200 violence by only targeting of insurgents, avoiding civilians, and trying to ultimately co-opt TNSM leadership with a negotiated deal. The head of operations, IG Frontier Corps Maj. Gen. Fazal Ghafoor managed to flip the leader of the TNSM, Sufi Mohammad, and worked with him to convince TNSM forces in fortified positions around arterial roads and hilltops to stand down and minimize the actual clashes. Within a month the state managed to restore order with minimal casualties, alongside political concessions for Sharia law (though they were never fully implemented) and a great deal of latitude or deferral to TNSM as an armed political force. 3.4.2. Malakand H (2007 - 13) 30: Population Security Background. The TNSM would remain an active political force in the Malakand region and Swat Valley but would not turn to political violence against the state until 2007. After TNSM leader Sufi Muhammad was jailed in 2001, his son-in-law Maulana Fazlullah assumed leadership of TNSM. Fazlullah managed to revive the organization, promote violently zealous enforcement of Sharia, intimidate locals, and disrupt development measures like vaccinations and the education of girls. However, following the military's raid on Islamabad's Lal Masjid (Red Mosque) in July 2007, the TNSM, like other Islamist militant groups, declared jihad against the government in the summer of 2007 and began to attack security forces and government buildings, effectively seizing control of the region by October 2007. After some initial fumbling, the state would exert the necessary combination of sustain and intensive material and 300 Shuja Nawaz, "Pakistan's Summer of Chaos," ForeignPolicy, June 17, 2009; Mullick, "Lions and Jackals," 2009; McKelvey, 2011; Watts et. al., 2014, pp. 132-138; Haider Ali Hussein Mullick, "Holding Pakistan: The Second Phase of Pakistan's Counterinsurgency Operations," ForeignAffairs (online), March 24, 2010; Reza Jan, "Paradise Regained: Swat One Year On," Critical Threats, May 25, 2010; Inter-Services Public Relations (Pakistan), "COAS Key Note Address - National Seminar Deradicalization at Mingora Swat," ISPR No PR157/2011, July 6, 2011. 201 organizational resources, restraint and non-kinetic substitutes for a strategy of population control that would restore order to Malakand region. Effort. The government initially sent in more than an army division to dislodge the TNSM, but appearing "hesitant to take them on"30 1 began to conduct a mix of conventional operations alongside attempts at deals with militants. After the TNSM broke a third deal and began encroaching into other districts that threatened the capital, Islamabad and the Karakoram highway, the state dramatically upgraded its efforts and increased its deployment to 52,000 troops to the region area for major military operations to dislodge the insurgents. The state also upgraded the training and capabilities of frontier corps troops, recruited new local policing forces, and sustained their posture long after the state declared victory. Violence. This strategy went to extensive lengths to not only defeat insurgents but also minimize collateral damage by evacuation of as many as 2 million people (along with their rapid repatriation within months after initial operations) at great expense to the state. The state then and then remained to protect the residents of Malakand after "clearing" operations, and converted to a much more defensive "presence-oriented strategy" and sustained posture of holding the territory, taxing the military's blood and treasure. Efforts to demobilize insurgents were initiated with the establishment of de-radicalization centers for former insurgents while the security forces worked with local civilians to restore public services such as schools and public infrastructure. Despite an enduring force presence, political authority continued to be devolved with the creation of a new local police force, tribal peace councils, and the implementation of new judicial reforms. 301 Imtiaz Gul, 2011, p. 118. 202 3.5. Counterinsurgency in the Federally Administered Tribal Areas (FATA) As the earlier description of FATA establishes, this peripheral frontier experienced frequent uprisings and political violence against state authority even predating the British Raj such that most of this violent resistance failed to attract much historical attention because they were treated as common, everyday occurrences. During the early years of Pakistan's independence, the state faced a number of these uprisings amidst the backdrop of Pashtun nationalism, but the cases that garnered interest were those that also drew in the Afghan state into border disputes.3 02 Aside from a rebellion in 1975, the region would not be seriously viewed through the lens of rebellion until 2002. 3.5.1. South Waziristan (1975)303: Mitigation Background. Both the British Raj and the Pakistani state were long accustomed to the periodic raids and inter-tribal conflicts of the Pashtun-dominated Federally Administered Tribal Areas (FATA) that sometimes turned against state forces. One dispute arose between the Wazir and Mehsud tribes in South Waziristan Agency in 1975 when Mullah Noor Muhammad mobilized 2,000 armed Wazirs to take over parts of the agency, launch guerilla raids on security posts, and even flirt with secession and a Pashtun nationalist alliance with Afghanistan.304 The risk of secession was quite real given that the Mullah's rebellion coincided with increasing Afghan material support secessionists and the creation of a Pashtunistan. The Daoud government 302 For some descriptions of these conflicts See Jacob Bercovitch and Richard Jackson, InternationalConflict: A ChronologicalEncyclopediaof Conflicts and Their Management, 1945-1995, Washington, DC: Congressional Quarterly, 1997, p. 65, 67, 99. 303 Most of this account is from Akbar S. Ahmed, 2004, but it is also referenced in Mansur Khan Mahsud, "The Taliban of South Waziristan," Talibanistan:Negotiating the Borders Between Terror, Politics, andReligion, Ed. Peter 304 Bergen and Katherine Tiedemann, New York: Oxford University Press, 2013, p. 164. Ahmed, 2004, p. 75. 203 in Afghanistan had specifically stepped up financial support and guerilla training for Pashtuns, 3 05 and the Afghan state had a long history of inciting Pashtun separatist movements in the 1950s and 1960s and intervening with Afghan irregulars. 306 Nevertheless, the state employed a mitigation strategy of limited effort and violence to simply restore the balance of tribal order. Effort. The Mullah's "empire [had] become an obvious threat to the authority of the Government," 307 and so it sent in the paramilitary Waziristan Scouts, backed by the threat of tanks and air support. The limited force was used primarily for leverage and only a short period of limited operations and limited targets. State security operations narrowly focused on clearing militants from the road network, degrading some insurgent capabilities, and capturing the Mullah. They operated with the belief that rebellion would quickly collapse thereafter, the normal political order of the tribal areas would be restored, and paramilitary forces could withdraw. No attempt was made to alter the political order or address the root causesperceptions of favoritism to the Mehsud tribe-that had sparked the rebellion. Tribes were unruly and this was simply the cost of doing business. Violence. Instead of using decisive kinetic force, the state tried to leverage its deployments to broker a deal and co-opt the Mullah. After the third attempt to settle the dispute, the state resorted to more sever actions but even then, it made extensive use of arrests and expulsion rather than deadly force, minimizing tribal fatalities to roughly 30 insurgents and few if any noncombatants. 308 When the GOP did eventually resort to major kinetic force, it used airpower to target and demolish economic assets. Demolition of houses, buildings, and economic 305 Syed Iqbal Ahmad, 1992, p. 183; Boggs, 2012; Haqqani, 2005, p. 174; Mohan Guruswamy, "Call for Pashtun Nation is Not Far Away," Rediff.com, July 28, 2009; Iqbal Khattak, "Kabul Trained Baloch, Pushtoon Youth in 1970s: Ex-ANP Leader," Daily Times, February 15, 2010. 306 Syed Iqbal Ahmad, 1992, p. 150. 3 07 Ahmed, 2004, p. 75. 308 Account from Ahmed, 2004, pp. 70-85. 204 assets was a deliberate decision by the state in the tribal areas to degrade an insurgency but avoid killing so as to not trigger blood feuds and minimize escalation.3 09 The Mullah was eventually captured and imprisoned, some of his property was confiscated, and his family expelled from the agency, a comparatively light and narrowly targeted punishment to deter other miscreants. 3.5.2. FA TA I (2002 - 07): Mitigation Background. The 2001 U.S. invasion of Afghanistan and outpouring of foreign fighters precipitated similar types of agitations in the tribal regions of Waziristan and other parts of FATA. In combination with local militant commanders indigenous to the tribal regions, these foreign fighters began to violently assert their influence, challenging local political orders and attacking Pakistani state security forces. In response, the Pakistan military conducted operations as early as 2002 to mop-up foreign fighters and al-Qaeda members but relied heavily on local paramilitaries, tribal militias and warlords for support. In 2004, the government expanded operations to target local tribal militants who hosted foreign fighters and fought against the state. This mix of foreign fighter and local militant hosts constituted the inchoate beginnings of the Tehrik-e-Taliban Pakistan (TTP) that would eventually expand attacks beyond the tribal regions to other parts of the country. 311 Although the TTP was rooted in Islamist principles, given its Pashtun tribal base and opposition to specific government policies, some analysts and officials feared it was driven by nationalist sentiments 309 31 Lieven, 2011, pp. 384-85. Ahmed, 2004, pp. 70-76. This account is based on Gul, 2011; Hasan Askari Rizvi, "Understanding the Insurgency," Daily Times, October 5, 2008; Shuja Nawaz, FATA-A Most DangerousPlace: Meeting the Challenge of Militancy and Terrorin the FederallyAdministered TribalAreas ofPakistan, Washington, D.C.: Center for Strategic and International Studies Press, January 2009; Syed Saleem Shahzad, Inside Al-Qaeda and the Taliban:Beyond Bin Laden and 9/11, London: 311 Pluto Press, 2011; Hassan Abbas, "A Profile of the Tehrik -i-Taliban," CTC Sentinel, 1 (2), January 2008, pp. 1-4; Carey Schofield, Inside the PakistanArmy, London: Biteback Publishing, 2011. 205 and could potentially escalate to a separatist insurgency. 3 12 By 2004, the GOP acknowledged the militant activities to constitute an open rebellion, but the Pakistani state viewed it through the same lens as the 1975 Mullah Noor Muhammad rebellion-internal tribal rivalries mixed with Pashtun nationalism had simply been redirected against the state. Consequently the state deployed the same strategy of mitigation with minimal levels of effort and violence. Effort. Believing it had exacerbated the problem by sending in regular military forces in 2002 that had not been inducted into the region since 1947, the Pakistani state was "reluctant to maintain a presence" or exert "any effort to control the area fully." 3 13 The state began to seriously pursue the cooptation strategy through a repeated cycle of negotiations and side deals. While the military claimed troops in the region totaled as high as 80,000 by mid-2007, most of these forces were already based in the area with the Army's XI Corps. Moreover, the vast majority of these forces did not actively participate in operations or patrols but instead manned static posts on the border or remained on their bases.31 One senior officer described the campaign as a "sitzkrieg" 315 and when they did conduct operations, they were "sporadic," 3 16 "not sustained over time." 3 17 Furthermore, the main forces utilized in these operations were poorly trained and armed frontier corps that had little experience with the level of high-intensity combat they were thrown in to, and took high casualties. The state believed it was "less expensive to keep them at arms length with special laws and different kind of administration" and treated the M Ilyas Khan, "Taleban Spread Wings in Pakistan," BBC News, March 5, 2007; Cheema, 2008; Lieven, 2011, p. 372-73. 13 Mullick, "Lions and Jackals," 2009. 314 Author phone interview, former U.S. official previously based in Pakistan, June 14, 2013 (#129); Ahmed Rashid, "Pakistan's Continued Failure to Adopt a Counterinsurgency Strategy," CTC Sentinel, 2 (3), March 2009, pp. 7-10. 315 Shuj a Nawaz, LearningBy Doing: The PakistanArmy's Experiencewith Counterinsurgency,Washington DC: Atlantic Council, February, 2011, p. 10. 316 Haider, 2014. 317 Fair and Jones, 2009/2010, p. 172. 312 206 distinctive militant groups in FATA as a "segmented problem" apart from the rest of the country. 318 Violence. Though the state occasionally deployed combat forces to fight high-intensity battles with well-armed insurgents in fortified positions, it made efforts to narrow its targeting to clusters of foreign fighters so as to avoid engagements with tribal militants or noncombatants. The state also heavily relied on deals with militant warlords and cooptation of tribal militias to manage and minimize violence. Major deals were signed in 2004, 2005, 2006 and early 2008 and a number of informal deals were believed to exist as well. 3 19 The deals sought to "[co-opt] influential tribal elders" on the condition that they "generally commit to provide stability in the area." 32 0 Pakistani officials believed they could manage militants, use them to stem foreign fighters thought to be the main source of instability, and return the region to the pre-2001 system of "very peaceful" indirect governance run by tribal elders and political agents.321 Accepting that political violence would continue, the objective of the deals was to minimize direct confrontation and bloodshed. In exchange for greater latitude, local warlords would mitigate violence and protect the area from foreign infiltration.32 2 Paired with these deals, military actions appeared to treat the insurgent threat as something "to be contained, not defeated."3 23 Author interview, former Pakistani official, Lahore, Pakistan, October 3, 2011 (#20). Daud Khattak, "Reviewing Pakistan's Peace Deals with the Taliban," CTC Sentinel, 5 (9), September, 2012, pp. 11-14. 320 Kamran Rasool, "Pakistan's Perspective on the 'War on Terror,"' Military Technology, 32 (11), November 2008, 17. Schofield, 2011, p. 136. 322 Gul, 2011, pp. 55-59; Ejaz Haider, "Intelligent Intelligence," Daily Times, July 26, 2007; Ismail Khan, "Militant Leader Threatens to Attack Forces, Dawn, July 30, 2005; Khattak, 2012; Interview with a former U.S. government official, Washington, DC, June 13, 2013 (#127). 323 Weinbaum 2009, p. 75. The term "containment" is echoed in Ijaz Khan, 2008, p. 14; Brian Cloughley, War, Coups, and Terror: Pakistan'sArmy in Years of Turmoil, New York: Skyhorse, 2008, p. 189; Jerry Meyerle, Unconventional Warfare and Counterinsurgencyin Pakistan:A BriefHistory, Alexandria, VA: Center for Naval Analysis, November 2012, p. 29. 318 3 19 207 Figure 4.10: FATA, North & South Waziristan, and TTP/Mehsud Insurgency Base NWFP (North-west Administied Kabut AFGHANISTAN Not i Rawalpindi P U N J A 8 JS T A 3...FTAMe," 8prsnt)Mie N A (Source: Economist) 3.5.3. FA TA H (2008 -present): ML-xed Background. The conflict in FATA took a rapid turn for the worse after July 2007 when the Pakistani government directed a military siege and stormed Islamabad's Red Mosque, where Islamist militants with ties to the TTP had holed themselves up. The Red Mosque incident provided a major impetus to TTP cohesion and escalation of violent operations. The incident triggered a wave of attacks and suicide bombings beyond FATA against military targets and civilian population centers throughout the country including urban centers. With Pakistan's core under fire from the TTP, the state was forced to adjust its strategy. However the strategy it employed toward FATA was segmented. The state escalated violence 208 against the TTP and its affiliates but retained a low effort strategy of enfeeblement for most of the region. In the TTP stronghold of the Mehsud tribal region in South Waziristan, the GOP launched an attrition campaign to completely dismantle the insurgent infrastructure, but in North Waziristan where the state's quasi- allies were concentrated, the GOP continued a strategy of mitigation. Mehsud Tribal region - South Waziristan Agency (SWA): Attrition. The state decided to decisively dismantle the hardcore TTP insurgent organization, concentrated in the Mehsud-tribal region of eastern South Waziristan, not because the territory itself was valued but because the insurgents had been able to project high levels of violence throughout valuable parts of urban Pakistan. Since that organization was based in a specific section of South Waziristan, the military adopted a unrestrained attrition campaign to degrade and destroy the TTP base of combat power, lines of communication, and command and control with the use of massive firepower, indiscriminate bombing runs by the air force, and three divisions to clear, occupy, and control the territory.32 4 The operation produced high collateral damage, displaced most of the Mehsud tribal population, and made few attempts to repatriate or rehabilitate them such that one government official remarked, "all that's left in South Waziristan is the army and dogs."3 25 North WaziristanAgency (NWA): Mitigation. By contrast, the state held to its cooptation strategy and ceasefire deals with a number of tribal warlords in North Waziristan and the western half of South Waziristan despite repeated attacks on the state outposts and institutions emanating from the region. Though it maintained forces and periodic operations in the region, it had 324 Nawaz, 2011; Sameer Lalwani, "Pakistan's Counterinsurgency Strategy," Talibanistan:Negotiatingthe Borders Between Terror, Politics, andReligion, Ed. Peter Bergen and Katherine Tiedemann, New York: Oxford University Press, 2013; Jamal Hussain, "PAF in the Countrinsurgency Role in the Tribal Areas of Pakistan, Defence Journal (Karachi), November 2012; Frederick W. Kagan, Reza Jan, and Charlie Szrom, "The War in Waziristan: Analysis of Operation Rah-e-Nijat," CriticalThreats, November 18, 2009. 325 Reza Jan, "Trickling Home to South Waziristan," CriticalThreats, December 13, 2010. 209 effectively outsourced control to warlords like Mullah Nazir and Hafiz Gul Bahadur (and informally to Jalaluddin Haqqani) under a deal that generally held since 2007. 326 Other FA TA Agencies: Enfeeblement. For the rest of FATA, Pakistan pursued an inconsistent, intermittent, and non-decisive "enemy centric" approach heavily reliant on airpower and artillery to mop up the remnants of the TTP that had infiltrated into other agencies from Waziristan and resumed fighting the state. 327 The state had dispelled its illusions about managing unruly tribesmen with minimal force now that the TTP had unrelentingly targeted and killed thousands of civilians and state officials, and security forces no longer felt constrained in the use of force and firepower. The state approach more closely resembled enfeeblement and an intentional "whack-a-mole" approach that U.S. special envoy Richard Holbrooke described as "dispersing" rather than "destroying" insurgents. 3 2 8 Difficult terrain alone could not explain this choice of strategy as the state had been willing to invest significant effort in Malakand, which in 329 The some ways posed greater challenges in terms of mountainous and forested terrain. difference was that the stakes in Malakand were greater and the state stood to lose more. Instead of expending the same effort as the population control strategy in Malakand or the attrition campaign in South Waziristan, the Pakistani state fought TTP elements while accepting periodic 326 Nazir and Gul Bahadur's groups were considered the Pakistani government's allies even though they may have provided tacit support to "bad Taliban" elements. See Anand Gopal, Mansur Khan Mahsud, and Brian Fishman, "The Taliban in North Waziristan," Talibanistan:Negotiatingthe Borders Between Terror, Politics, andReligion, Ed. Peter Bergen and Katherine Tiedemann, New York: Oxford University Press, 2013, pp. 128-163; Mansur Khan Mahsud, 2013. 327 The current deputy national security advisor, Anthony Blinken was reported to have told Pakistani Ambassador Hussain Haqqani that "there is appeasement one day, confrontation another day and direction a third day." See Woodward, 2010, p. 286; David J. Kilcullen, "Terrain, Tribes, and Terrorists: Pakistan, 2006-2008," Brookings Institution Counterinsurgency and Pakistan Paper Series, No. 3, September 10, 2009, p. 10-12; "PAF Conducted 5,500 Bombing Runs in FATA since 2008," Express Tribune, November 14, 2011. 3 Jane Perlez, "Pressure from U.S. Strains Ties with Pakistan," New York Times, October 26, 2009. 329 "Q&A: Pakistan's Waziristan Challenge," BBC News, October 19, 2009. 210 resurgences-like those that cropped up in Khyber in 2011, and the Tirah Valley and Mohmand in 2013-so long as they could eventually be contained with more limited military actions.330 3.6. Summing up Pakistan's Counterinsurgency Cases The Pakistani state expended far greater effort fighting over East Pakistan (East Bengal), Sindh, and the Swat/Malakand region of KPK than it did during most conflicts in Balochistan or FATA. These efforts were evidenced not simply by manpower but also in other ways including: expenditures of money and materiel (heavily consumed in East Pakistan and Malakand); the level of difficulty (airlifting multiple divisions to East Pakistan and heli-borne troop insertions in Malakand); organizational adaptation (re-tasking troops away from conventional doctrine in Karachi and Malakand to population protection and rebuilding physical and social infrastructure); strategic tradeoffs (re-deploying troops away from the existential threat theater in Sindh and Malakand); and in all cases, high levels of sustained activity and prolonged deployments in the field. Likewise, the Pakistani state employed considerably more force against the Bengalis and Baloch insurgencies (even during very limited engagements) with not only disregard for civilian casualties but in some cases purposeful targeting and disproportional violence. Though the level of violence perpetrated by the state forces against Bengalis may have been exaggerated even by an order of magnitude, it would still be one of the most intensely violent counterinsurgency campaigns since WWII. 3 3 1 Balochistan campaigns also consistently relied on brutality and high 330 Reza Jan, "Pakistani Military Operations in Orakzai," CriticalThreats, May 13, 2010a; Reza Jan and Sam Worby, "Limited Goals, Limited Gains: The Pakistan Army's Operation in Kurram, CriticalThreats, September 6, 2011; Tayyab Ali Shah, "Pakistan's Challenges in Orakzai Agency," CTC Sentinel, 3 (7), July 2010; Daud Khattak, "'Clearing' Kurram," ForeignPolicy's AfPak Channel, August 25, 2011. 331 Sarmila Bose, DeadReckoning: Memories of the 1971 Bangladesh War, London: Hurst and Co., 2011; Ulfelder and Valentino, 2008. 211 violence even if the body count is not comparable. One former general described early operations in Balochistan as "trying to kill a fly with a sledge hammer."3 32 Today, direct military confrontations in Balochistan have ceased but a "dirty war," conducted by state security forces, 333 In the case of Sindh, the state came intelligence services, and death squads, has taken its place. down extremely hard to defeat what it saw as a nascent insurgency, though it may have more accurately been described as inchoate, opportunist criminal violence amidst an ethnically based political movement. The state was motivated to deploy this extremely disproportional response (particularly in comparison to Malakand I) both to safeguard critical territory, but also based on heightened suspicions of all ethnic Sindhis' loyalties. 3.7 Measuring Strategy Based on the accounts provided above and qualitative measures of the dependent variable, I can generate a distribution of Pakistani strategies for each campaign. As described in chapter 2, I assign a value for the strategy, based on qualitative assessments of a particular campaign's effort (in terms of resources, organizational investments, sustainment, and adaptive learning) and violence (in terms of collateral damage, enemy-centricity, accommodations, and public welfare), and plot it on and XY graph. Based on these qualitative codings, the Pakistan state deployed strategies of attritionin East Bengal, Balochistan IV, Sindh, and FATA IISWA/Mehsud, populationsecurity in Karachi and Malakand II, enfeeblement in Balochistan (I- III, V) and FATA II-Other, and mitigationin South Waziristan, Malakand I, FATA I, and FATA II-NWA. Akbar Khan, 1960. 333 Grare, 2013; "We Only Receive Back the Bodies," 2012; Green, "Hidden War Embodies Pakistan's Struggle," 2012; Walsh, "Pakistan's Secret Dirty War," 2011; Sohail Khan, "Disband Agencies 'Death Squads'," 2012. 11 212 Figure 4.11: Qualitative Distribution of the Dependent Variable 8 East Pakistan * 4 Baloch V Baloch IV 4 4 * * .2 Sindh Baloch II Baloch III Baloch I 4 FATA II - Other * FATA II - NWA FATA II - SWA/ Mehsud 4 * FATA I * 4 Waziristan * Karachi Malakand II Malakand I 0 0 4 Effort 8 Figure 4.12: Quantitative Distribution of the Dependent Variable * * Baloch II (1958-59) East Pakistan (1971) Baloch 1 (1948) Baloch IV (1973-77) FATA II - SWA/ Mehsud (2008$13) Baloch III (1963-69) ----- ----* * indh(1983) Baloch V (2005-13k FATA II -NWA (2008-13) FATA I (2002-07) Malakand II (2007-13) Waziristan (1975-76) -5 -5-~4 -1.5 FATA II - Other (2008-13) Malakand I (1994) -- Karachi (1992-96) -0.25 Effort 213 1 Quantitative measures of deployed manpower to measure effort and state-inflicted 33 4 civilian casualties to measure violence (while controlling for area, population, and duration) help confirm values assigned by the qualitative measures with some slight departures. In Balochistan V, state-inflicted fatalities are likely underestimated by the quantitative measures 335 due to the tight restrictions and difficulties of accessing information in the region. The state is suspected to have authorized a large number of disappearances-to eliminate both insurgents and political activists-but bodies do not always turn up, or are disguised to look like ordinary crime or sectarian violence. Additionally, the FATA II cases appear to all involve high effort due to high troop deployments but in fact, the previous description of the FATA II reveals these numbers were likely inflated and their activities quite limited. Forces that were largely parked in the FATA region to appease international and American demands only conducted sporadic engagements if at all, or mostly left kinetic actions to the air force, except in the Mehsud tribal region of South Waziristan. This is confirmed in fatality data, which shows that even though more troops were deployed to FATA, Pakistan's security forces absorbed far more casualties and fatalities in KPK (both in offensive engagements and overall) than in North Waziristan or the rest of FATA. The quantitative data is certainly fraught with problems as any estimation of conflict data is subject to, and should not be taken literally. However, it does offer insights on the distribution and clustering of campaigns and confirmation of the relative distance of these clusters. Given this evidence, there should be little doubt that the state employed altogether different strategies of violence in Malakand and Karachi than in Balochistan and East Bengal. Similarly, force 334 Note the actual measure of effort is the Log[Manpower/Area of operation] and the measure of violence is Log[(state-inflicted civilian casualties/yr)/Area population]. 33' Hamid Mir, "The Tragedy of Journalists in Balochistan," The News, October 2, 2012; Sumit Galhotra, "Blacklisted in Balochistan" Committee to ProtectJournalists,March 28, 2014; Karlos Zurutuza, "A Black Hole for Medis in Balochistan," Al Jazeera, February 5, 2014. 214 deployments reveal the stark difference in effort between Malakand I and Malakand II and for that matter, most of the Baloch campaigns compared to operations in East Bengal and Sindh. The distribution of a strategy's effort is further substantiated by alternative quantitative measures for recent campaigns where data is available, such as the ratio of offensive to defensive engagements to measure violence and the ratio of state activity to insurgent activity to measure effort. These also help to substantiate the qualitative measures and locate the different campaigns since 2001 in relation to each other in terms of effort and violence (figure 4.13). Other estimates of violence, kinetic activity, and collateral damage, include data on state-produced civilian fatalities, the use of airpower, and repatriation of internally displaced persons (table 4.6). Meanwhile loss-exchange measures help estimate the degree of effort as investments in quality personnel reduce losses. After confirming values of counterinsurgency strategy in each of the 15 Pakistan campaigns, we can now begin to evaluate how the core-periphery dynamics of territory and identity fare in explaining this distribution of strategies. Figure 4.13: Ratio of Activity (Effort) and Ratio of Offense to Defense (Violence) Balochistan V (09-12) C FATA I- Othc FATA (09-12) 4 C U SC FATA II - North Waziristan (09-12) 0 Malakand II * C o FATA II - South Waziristan (09-12) (09-12) Swat Valley (09-12) 0 0.5 1 State/Non-State Actions Ratio (Effort) (Source: Pakistan Institute for Peace Studies, HRDP) 215 & Table 4.6: Comparisons of State Violence in Different Regions Balochistan FATA Swat/ Malakand 252 370 358 Civilians killed in Offensive Operations (2009-12) 0.057 0.089 0.041 Civilians killed per 1,000 people N/A 1.832 0.707 Offensive Clashes/ Defensive Clashes .2 million .75-1.07 million 2-2.5 million Estimated Displacement 0% 40-50% 90-100% % Repatriated (Source: PIPS; Gishkori, 2013; Hussain, 2012; Express Tribune, 2011; FATA DMA; UNHCR; IDMC; The News) Table 4.7: Comparative Measures of Effort in Recent Campaigns (bold denotes data that departs from expected predictions) FATA II South North Malakand II Balochistan Waziristan Waziristan Other FATA Date 2007-2013 2005-2013 2008-2013 2008-2013 2008-2013 Strategy Expectation Population Control High Enfeeblement Low Attrition High Mitigation Low Enfeeblement Low 0.58 0.05 0.61 0.27 0.56 33.9 10.3 20.7 0.3 36.5 0.06 0.11 0.06 0.01 0.02 2-2.5 million .2 million Offensive Ops Defensive Clashes/Insurgent Attacks Troop Loss in Ops/Yr Milt Loss-Exchange (SF per insurgent killed) in Ops Estimated Displacement % of displaced Repatriated 90-100% by 2013 _ (Source: PIPS and other data sources) 1 .75-1.07 million 0% 40-50% _ Table 4.8: Comparative Measures of Violence in Recent Campaigns Date FATA II North Other FATA Waziristan Malakand II Balochistan South Waziristan 2007-2013 2005-2013 2008-2013 2008-2013 2008-2013 Control Low Enfeeblement High Attrition High Mitigation Low Enfeeblement High 1.05 11.7 0.19 22.6 1.66 63.1 1.33 10.4 1.95 57.1 10.8 36.7 14.5 14.2 83.3 N/A 131 N/A 10,600 Population Strategy Expectation Offensive Ops/Defensive Clashes NonSt killed/yr/IOOK pop Civs killed in State cios 100,000 tat Actions/ 100,000 capita Airpower: Avg. Sorties/Month (during major operations) Airpower: Bombs dropped I 1,700 II 216 1 18.5 IV. Evaluating the Theory 4.1 Correct Predictions After examining each of the Pakistan counterinsurgency campaigns to describe and measure the state's strategy, we can now begin to evaluate the predictions generated by the core periphery theory in section 2.3 (figure 4.4). The medium-N analysis allows me to maintain concept validity while examining the entire set of Indian counterinsurgency campaign observations to identify how variation of the dependent variables of effort and violence correlates with the hypothesized independent variables of territory and identity. Not only does this allow us to estimate how many of the cases can be explained with the theory, the process of data collection and coding the variables for the medium-N analysis allows me to identify cases worth further exploration through intensive and focused comparisons in the next chapter. Table 4.10 shows that the values of the dependent variable (strategy) are, for the most part, well predicted by my theory's independent variables. High effort strategies were in fact primarily deployed in the more "settled" and valued regions of Sindh, East Bengal, and Malakand (in the past recent decade) while restrained, low violence strategies were primarily deployed when facing rebels with a Pashtun or Mohajir ethnic base. Based on the coding procedures for the dependent variable, the theory completely explains 12 of the 15 cases. Specifically, territory correctly predicts the level of effort in 14 of the 15 cases and identity correctly predicts the degree of violence in 13 of the 15 cases. Since the theory's predictions are probabilistic, this is bound to happen from time to time. A close examination of the incorrectly predicted cases does not actually disconfirm the theory but simply reveals the importance of some other factors at play. 217 Table 4.9: Detailed Predictions of Strategy IV2: Identity IV1: Territory No. Observation 1 Baloch I 2 Baloch II 3 Baloch III 4 5 East Pakistan Baloch IV Region/ Province Kalat, Insurgent Group(s) Baloch Rebels Balochistan (Prince Karim) Khuzdar, Baloch rebels Balochistan (Nauroz Khan) Kohlu/Jhalwan, Baloch rebels Balochistan (BPLF) East Bengal Bengali rebels (Mukti Bahini) Baloch Rebels Balochistan Projection Low Worth Low Embedded Low Predicted Strategy Enfeeblement Low Low Enfeeblement Date(s) 1948 Incentives Low Cost High 1958-1959 Low High Low 1963-1969 Low High Low Low Low Enfeeblement 1971 High High Low Low Attrition 1973-1977 Low Moderate -High High Moderate Low Low Enfeeblement 1975-1976 Low High Low High High Mitigation 1983 High Moderate High Low Low Attrition Population Control Mitigation ____ (BPLF) 6 7 8 9 Waziristan Sindh-MRD KarachiMQM Malakand I South Wazirs (Mullah Waziristan, FATA Noor Muhammad) Rural Sindh MRD/Sindhi (Nawabshah, Dadu) political groups Karachi, Sindh MQM 1992-1996 High Low High Moderate Malakand Division/ Swat TNSM 1994 Low Moderate Low High ModerateHigh High AQ/TTP 2002-2007 Low High Low High Moderate Mitigation Baloch 2005-2013 Low High Low Low Low Enfeeblement TNSM 2007-2013 High LowModerate High High High Population Control 00 Valley, KPK 10 FATA I FATA 11 Baloch V Balochistan Nationalists 12 Malakand II 13 FATA II SWA FATA II Other FATA FATA II NWA Malakand Division/ Swat Valley, KPK 14 15 S. Waziristan, FATA Other FATA TTP/Mehsud 2008-2013 Low High High Low Moderate Attrition TTP/Other 2008-2013 Low High Low Moderate Moderate Mitigation N. Waziristan, FATA TTP/Gul Bahadur, 2008-2013 Low High Low High Moderate Mitigation Haggani, Nazir Table 4.10: Evaluating Predictions Dependent Variable No. Observation 1 Baloch I 2 Baloch II 3 Baloch III 4 East Pakistan 5 Baloch IV 6 Waziristan 7 Sindh-MRD 8 Karachi-MQM Predicted Strategy Enfeeblement Enfeeblement Enfeeblement Attrition Enfeeblement Mitigation Attrition Population Effort Low Low Low High High Low High High Correct Prediction? X Violence High High High High High Low High Low Correct Prediction? Control Control 12 Malakand I FATA I Baloch V Malakand II 13 14 FATA II - SWA FATA II - Other FATA 15 FATA II - NWA 9 10 11 Actual Strategy Enfeeblement Enfeeblement Enfeeblement Attrition Attrition Mitigation Attrition Population Mitigation Mitigation Enfeeblement Population Control Attrition Mitigation Low Low Low High Low Low High Low High Low High High X Mitigation Mitigation Enfeeblement Population Control ~Attrition -/? Enfeeblement Mitigation Low Low ' Mitigation ci 4.2 Under-predicted Cases 4.2.1. Missed Predictions of Effort The core-periphery theory fails to accurately predict the Baloch IV (1973 - 77) case in terms of the expected level of state effort. The first reason for this simply may be that the strategy in fact straddled the line between enfeeblement and attrition. The Baloch IV strategy of effort deployed was very difficult to cleanly code because despite a large force presence and intensive initial operations, in 1974 the military scaled down its operational activity and mostly took up static positions. And while many features of the case tilt it towards attrition, its political aim still seemed closer to administrative occupation rather than usurpation and integration of the region. The second and more compelling reason for the theory's failed prediction in Baloch IV is that leader agency and preferences simply overruled the strategic incentives described by the core-periphery theory. It appears the main reason for this high effort campaign was the political machinations of the newly elected civilian leader Zulfikar Ali Bhutto. Prime Minister Bhutto pushed for the high effort, activity and manpower in Balochistan to undermine his Baloch rivals, occupy the military and consolidate his power. Not only does the level of manpower and violence seem to have been excessive and spearheaded by Bhutto, the conflict itself appeared to be contrived by him. In any other circumstances, the military likely would have objected to this unnecessary deployment but in its weakened state after its devastating 1971 military defeat to India in East Pakistan, it had no capacity for resistance. Bhutto had fired the previous Army chief of staff and appointed a supplicant in Gen. Tikka Khan knowing he would remain loyal. Leaders' 220 preferences undeniably command substantial influence over state choices and strategy and in this case, there were no capable institutions to challenge or encumber Bhutto's preferences. 3 36 Despite the theory's weakness in this case of Balochistan IV, all the theory's main competitors equally fail. The GOP launched a campaign of high effort despite its state strength and capacity having been severely diminished by the loss of East Pakistan. It launched a campaign of high violence despite the government being firmly in the hands of a democratically elected civilian leader who had the upper hand over the military for a period of time. And it persisted with yet another costly attrition campaign despite having just established its reputation for toughness three years earlier. 4.2.2. Missed Predictions of Identity For the most part, the level of violence was well predicted by the level of group worth and embeddedness, but the two exceptions are the most recent cases of FATA II-Mehsud/SWA and FATA 11-Other. When the FATA II campaign is disaggregated, it becomes clear that the Pakistani state spared mostly Wazir insurgents in FATA II - NWA and deliberately increased the level of violence and repressive measures against one sub-group of Pashtuns, the Mehsuddominated TTP. The TTP was initially based in the Mehsud tribal region of northeastern South Waziristan, but after military operations targeted them in FATA II, they dispersed to other agencies like Orakzai, Khyber, and Kurram. The Mehsuds were targeted because they composed the brunt of the Pakistani Taliban (roughly 80% according to some estimates) 3 37 and conducted most of the suicide bombing campaign against urban targets in Punjab and Islamabad from 2007 - 10 (see figure 8). In this case, it appears the dynamics of civil war endogenously reshaped the Daniel L. Byman and Kenneth M. Pollack, "Let Us Now Praise Great Men: Bringing the Statesman Back In" InternationalSecurity, 25 (4), Spring 2001, pp. 107-146. 337 Author interviews with Pakistani analysts and former officials. 336 221 once esteemed identity group and state perceptions of the Mehsud/TTP network through a number of mechanisms, transforming it from an "us," like most other Pashtuns, to a disloyal "other." 3 38 In four previous cases of Pashtun rebellion, the state responded with great restraint. But after the TTP sub-group targeted the very core of state identity, it revealed itself to be totally disloyal and irreconcilable, warranting the full wrath of the state. First, the state was able to associate the Mehsud tribe with foreign elements and externalize an internal threat as it had done before with Bengalis, Baloch, and Sindhis. The Mehsuds had brought in Uzbek militants as had many other tribes in FATA but after the foreign militants began targeting Pakistan and the state turned its sites in particular on these Uzbek militants, only the Mehsud tribe continued to offer them shelter. The state attempted to condition each peace deal with Mehsud on giving up the Uzbek militants but they continued to go back on their word. As a result, Uzbeks became a loyal force to Baitullah Mehsud, but in doing so, gave the state at least the pretext that he was aligned with foreign interests. Survey data from FATA supports this contention as about 20% of those surveyed believed that the Pakistani Taliban were 3 39 composed of foreign fighters and 43% stated the military should force them to leave. Second, the identity of the faction of the Mehsud clan led by Baitullah Mehsud had morphed somewhat during the fight. The state had frequently sought to parlay with the Mehsud militants but the rebels broke their agreements with the state in 2005 and again in early 2008, choosing to directly target the military and Pakistan's core territory with a vengeance. Mehsud had not only struck against the Pakistani urban core that had gotten both the attention of the public and national elites and imposed enough costs to cause them to expend more effort, the This is similar to the identity hardening endogenous to conflict. See Kalyvas, 2008, p. 1046; Chaim Kaufnann, "Possible and Impossible Solutions to Ethnic Civil Wars," InternationalSecurity 20 (4), Spring 1996: 140-44. 339 Naveed Ahmad Shinwari, UnderstandingFATA: Attitudes Towards Governance, Religion & Society in Pakistan'sFederallyAdministered Tribal Areas, Vol IV, Islamabad: Community Appraisal and Motivation Program, 2010, pp. 69-71. 338 222 targeting of his efforts was also important. Under Mehsud's leadership, the TTP had deliberately dispatched suicide attacks not only against frontier army outposts, but also against soldiers' homes, their families and their places of worship in the Pakistani "mainland." They had struck an emotional chord within the army that quieted the early reservations of a number of Pashtun commanders and soldiers within the Army apparatus. The state made the deliberate choice for greater violence against a sub-group, once close to the "core," whose actions had driven a wedge between and expanded the social distance between themselves and the idea of Pakistan: A very senior general, commanding a corps... said that as the insurgency picked up and terrorists began knocking out targets in Islamabad, Rawalpindi, and Lahore, the corps commanders favored an all-out campaign against the groups involved. A consensus was reached that all militants must be dealt with indiscriminately-however close they may have been to our institutions,he said. 340 No longer were these Mehsud militants considered misguided, overzealous, and in need of rehabilitation like the other tribal groups including Wazirs. Empathy had been shown and alternatives attempted to no avail. Their actions and targeting revealed themselves to the Pakistan army to be a true enemy of the state and no institutional ties or representation within the institution could restrain the Leviathan. Public opinion reflected this shift in perceptions of identity of the TTP. Not only had the general public come out in support of the military operations by the summer of 2009,41 Pakistani Pashtuns specifically began to perceive the TTP as a distinct and foreign entity. Pashtuns expressed heightened fears of the TTP (previously absent in earlier polls) and even higher than average support for Pakistani military operations rather than dialogue and negotiations (which 340 Gul, 2011, p. 136 (emphasis added). Based on survey data released by Pew Global Attitudes, Gallup, International Republican Institute, Terror Free Tomorrow, Gilani Research, and World Public Opinion. 341 223 3 42 In other words, the had previously been the longstanding preferred choice of respondents). public and Pashtuns in particular were getting behind the use of force against the TTP, whowith their actions-had dislodged themselves from being part of the state's in-group that might be reconcilable. Even a large majority of the Pashtuns in FATA came out in support of military offensives in FATA in late 2009.343 FATA Pashtuns' support for the military offensive in the Malakand region in 2009 amounted to 70% and their support for military operations in FATA totaled 67%. Only in South Waziristan did opponents (46.8%) of the military offensive outstrip slightly supporters (43.6%). If this was the case for the general public and Pashtuns within the tribal areas then Pashtuns in the armed forces were likely even more willing to unleash force on this small Pashtun faction that had turned their guns on the Pakistani Army. Third, the state was also able to cultivate other Mehsud tribal elements and warlords to challenge the TTP and Baitullah and Hakimullah's leadership including Qari Zainuddin344 prior to the launch of major operations in the fall of 2009, and after they began the operations, with Abdullah Mehsud and Turkistan Bhittani.34 5 In doing so, it revealed the state was still attentive to managing the perception that this was a fight against all Pashtuns or Mehsuds and they took significant actions to minimize these perceptions for the Pashtun regions and the country. 4.3. Evaluating Competing Explanations For the purposes of my larger study, I have controlled for state-level structural variables like strategic/organizational culture, civil-military relations, doctrine, as well as state strength For Pashtun views, one can look to the NWFP responses in the survey analysis disaggregated by province. See C. Christine Fair, "Islamist Militancy in Pakistan: A View from the Provinces," World Public Opinion, July 10, 2009. 143 Shinwari, 2010, pp. 57-58. 344 Saeed Shah, "Pakistan Taliban Leader Faces Threat form Fellow Tribesman," McClatchy Newspapers, June 13, 2009. 345 Mahsud, Talibanistan,2013, p. 192. 342 224 and regime type in order to explore the sub-national variation within a country. Strategic culture and civil-military relations generally do not temporally vary much at all but some structural conditions can change in theory. Pakistan is relatively constant in its macro-conditions as a weak, non-democratic state. 346 Its political institutions have consistently been rated as non-democratic and not free, based on Polity and Freedom House scores, but there are some minor fluctuations, particularly in the past two decades. And while it has exhibited spurts of economic growth during periods of international and American aid (in the 1980s and the 2000s) this has not fundamentally altered its state strength the way it did for the East Asian NICS as it still remains in the bottom third of development in terms of GDP/capita (even PPP) as well as measures of political extraction. Therefore, in the Pakistan cases these structural conditions are relatively constant, but they can be relaxed slightly to explore whether some more fine-grained variation can explain patterns of the dependent variable. 4.3.1. State Capacity Pakistan has generally been a weak state throughout its history but it could be argued Pakistan was relatively stronger during periods of external support (primarily through U.S. military and economic assistance), which coincided with periods of higher economic growth,34 7 and this might account for variations in effort. Since these exogenous injections of aid largely coincided with military rule and involved a substantial degree of military assistance, it would bolster expectations that this increased capacity would directly translate into military effort. In fact, as the earlier examination of cases demonstrates, the timing of these exogenous shocks fails to correlate with high effort strategies. The first major input of U.S. aid began in 346 Watts et al, 2014, pp. 116-119. See also Jalal, 1990; Haqqani, 2005; Siddiqa, 2007. 347 Epstein and Kronstadt, 2012, p. 40. 225 1954 when the two states signed a mutual defense assistance agreement. Despite this windfall of resources, the Pakistani state continued low effort strategies in Baluchistan in 1958 and 1963. Yet it was only after the US attenuation of aid beginning in 1965 following the Indo-Pakistan war 34 8 that Pakistan exerted its greatest efforts yet to counter the East Pakistan insurgency. It then followed this with another high effort campaign in Balochistan in 1973, not only with limited external aid, but also following on the heels of losing Bengal, which had provided half its GDP and its largest source of foreign exchange. Figure 4.14: Pakistan State Strength/Capacity, 1947 - 2008: GDP/capita and CINC 0.014 2500 0.012 2000 0.01 0.016 000Q;0 0.004 500 0.002 0 V 1947 n0 1957 1967 - 1977 GDP/capita 1987 - 1997 2007 CINC The second major round of U.S. aid to Pakistan began during the Afghan jihad in the 1980s. In 1982 resources began to increase after President Regan signed an aid package, which did coincide with the attrition response to the Sindhi rebellion in 1983. Nevertheless, aid dropped precipitously after 1990 due to the Pressler amendment and U.S. aid to Pakistan was at its lowest E. Wayne Nafziger and William L. Richter, "Biafraand Bangladesh: The Political Economy of Secessionist E Conflict," JournalofPeaceResearch, 13 (2), 1976, p. 96. 226 ,- ebb throughout the 1990s. 349 Yet Pakistan still mustered the resources for a protracted, higheffort population control campaign in Karachi since valued territory was at stake. Figure 4.15: U.S. Assistance to Pakistan, 1948 - 2010 Greenbook data 1948-2000 4000 (a) 3500 appropriations 2001-2010 (b)(c) (d) (e) $ 3000 2500 . E 2000 15001 S1000 500 0 FIscal Years Economic Assistance, Total MISecurity Assistance, Total (Source: Susan B. Epstein and K. Alan Kronstadt, Pakistan: U.S. Foreign Assistance, R41856, Washington DC: Congressional Research Service, October 4, 2012, p. 40) Figure 4.16: Pakistan State Strength/Capacity, 1960-2007: Relative Political Extraction (RPE) and Reach (RPR)) 1.2 1 0.8 0.6-AN 0.4 0.2 0 t- M MR M(Agr I' -RPE (Agr) N C-DRP M 0 CGD '- -- -RPE (GDP) 349 Epstein and Kronstadt, 2012, p. 40. 227 -RPR Finally, the third major round of aid began in 2001 with direct cash transfers as the US enlisted Pakistani support for its military campaign in Afghanistan. Despite this new capacity, Pakistan invested very little effort during the first campaign in FATA and in Balochistan. Though the result of this decade-long stream of aid may have assisted the Pakistani state to make a decisive turn in KPK against the TNSM and against the TTP in one sub-section of South Waziristan, its low-effort strategies in the rest of FATA and Balochistan remained unchanged suggesting at the very least that the state capacity hypothesis remains incomplete. Capacity alone was not sufficient to motivate Pakistan's high effort strategies both historically and at present. 4.3.2. Regime Type/Civilian Rule This paper has treated Pakistan generally as a non-democracy given that for the vast 35 0 the military majority of its history, it has generally not met the criteria for democracy, continues to possess "latitude,"35 and civilian authority has been circumscribed regardless of formal institutions. 3 2 However, democratic governments were nominally in charge during a number of counterinsurgency campaigns warranting a test of whether this restrained the use of state violence as some theories suggest. The track record of civilian rule and influence is not compelling. Civilian Prime Minister Liquat Ali Khan, who was a founding father of the country and arguably one of the most powerful figures at the time, presided over the first high violence campaign in Balochistan against the Kalat rebellion.35 3 By many accounts, Bhutto, along with other civilian political leaders, supported the ruthless and aggressive counterinsurgency campaign against Bengalis in Pakistan scores a 7 or above in the Polity index for 17 of 64 years and is coded as free by Freedom House in only 9 of 41 years. 351 Author Interview, former Pakistani official, Lahore, Pakistan, October 3, 2011 (#20). 352 Jalal, 1990; Siddiqa, 2007. 353 Syed Iqbal Ahmad, 1992, 150. 350 228 - 1971 .14 And civilian rule since 2008 has not fundamentally changed the high-violence strategy in Balochistan and in fact coincided with a shift to a high violence campaign against the TTP in different parts of FATA. Even in periods when civilians ruled during low violence campaigns, they did not seem to actively play any real restraining role in military-led operations in Karachi in the early 1990s, and instead were consumed with domestic political intrigues. All this evidence contradicting the regime hypothesis might be dismissed by those who argue the military still dominated security decision-making, if not politics altogether. At minimum, this civilian impotence would reveal the weakness of the regime theory-formal institutions and norms of political freedom are not sufficient to constrain states. However, even a more favorable test reveals that democratic, civilian regimes could be just as ruthless as military regimes. Figure 4.17: Pakistan Regime Type, 1950 - 2012: Polity and Freedom House 10 8 6 14 ... .................. . ......... -- ---- - -- - --- - - ..... - .......... 12 10 4 8 S2 0 6 1 -4 -4 2 -6 -I . -r- 'r=' I 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 - Polity score -......"Democracy" threshold Freedom House score --- 3 54 "Free" threshold Hamid Hussain 2002; Akbar Ahmed 1997; Gul Hassan Khan, 1993, p. 272; Ashok Kapur, Pakistan in Crisis, New York: Routledge, 1991, p. 179 note 15 referencing Oriana Fallaci's interview with Bhutto; Srinath Raghavan, 1971: A Global History of the Creation ofBangladesh, Cambridge: Harvard University Press, 2013, p. 50. 229 The Balochistan IV campaign provides the easiest test for the regime type theory as it occurred during the period when civilian power was at its highest relative to a military, broken and discredited after its exploits in East Pakistan. However, all evidence suggests the state conducted another brutal counterinsurgency campaign in Balochistan, not in spite of but at the behest of civilian leadership led by Bhutto. 355 This was not the first time Bhutto had called for unnecessary force, ordering the military to fire on striking police with artillery within the urban center of Peshawar.3 56 Even more revealing, as soon as the military took over the country in the 1977 coup, it abruptly "abandoned" military operations in Balochistan and reconciled with the insurgent leaders.3 57 Former Amb. Hussain Haqqani writes, "Had the rebellion been the real threat to Pakistani security it was made out to be, army action against it would not have been so readily terminated." 358 Realizing that "Bhutto was using the army for his own political purposes,"359 the army commanders likely retreated at this point because they didn't have a serious stake in the fight beyond containment. 4.3.3. Reputation Theories of internal deterrence or reputation do a good job explaining why states choose to fight despite low stakes, but they offer less traction explaining how they fight. If we are to take the theory's predictions to their conclusion, we should expect to see high levels of effort and violence early on in order to signal to other self-determination groups the price of attempting 35 One of Bhutto's own advisors acknowledges that Bhutto had intended to dissolve the government and send troops into Balochistan regardless of any provocation. See Raza, 1997, p. 269. 356 Gul Hassan Khan, 1993, pp. 361-62. 35 Author Interview, former Pakistan government official, Islamabad, Pakistan, October 12, 2011; Rizvi, 1987, p. 227. Talbot 1998, p. 226 notes that Bhutto claims the Army overrode his efforts to withdraw the military from Balochistan. However, it appears they had no problems ending the counterinsurgency campaign in 1977 and even before that, appeared to be deployed in Balochistan but not necessarily doing anything except limited raiding and infrastructural development. 358 Haqqani, 2005, p. 170. 359 Nawaz, 2008, p. 335. 230 separatism. Pakistan certainly fended off early insurgencies and used high violence during these enfeeblement campaigns but their inconsistent and limited approach, lack of publicity, the absence of increasing accommodations over time, and unwillingness to seriously address challenges by Pashtuns significantly weakens the reputation argument in these cases. The state was terribly inconsistent in cultivating a reputation for toughness. Though it went in to Balochistan and routinely doled out high levels of violence, the lack of effort invested in this whack-a-mole strategy did not impose significant enough costs on the Baloch rebels, allowing them to reorganize and fight again a few years later. This signaled to other more capable challengers that the state would not follow through on counterinsurgency operations. Another clear inconsistency in the reputation logic is that when the state made a significant "concession to East Pakistan" with the 1956 Constitution in order to accommodate the Bengali opposition very early on. 3 60 Though it had not built up a reputation for toughness, the Pakistani state had signaled its willingness to walk away from fights with insurgent organizations still intact and accommodate self-determination movements with concessions. Moreover, the state did not bother to publicize their actions in order to deter others. A state concerned with its reputation for how it fights would have been deeply troubled by the repeated resumption of Baloch rebellions, especially so early in its statehood, as it sent dangerous signals to the restive Bengalis and Pashtun communities that displayed clear risks of secession. After its East Pakistan campaign, the GOP managed to bury for thirty years a secret government-commissioned report detailing the lengths the state had gone to fight separatism and hold on to Bengal. Their actions had been costly, but the state was doing its utmost to muffle the 360 Kamal Azfar, "Constitutional Dilemmas in Pakistan," Pakistan Under the Military: Eleven Years of Zia ul-Haq, Ed. Shahid Javed Burki and Craig Baxter, Boulder, CO: Westview Press, 1991, pp. 49-85. Quotation from p. 57. 231 signal and compromise any deterrence value it might have had.361 Three years later they were fighting another Baloch separatist insurgency. Yet Pakistan remained unconcerned, routinely walking away from Baloch insurgencies without defeating them, believing the Baloch to always be manageable with minimal containment efforts. Third, reputation theories also predict that eventually, a state will start to make accommodations after it has sufficiently signaled its toughness in a number of rebellions. However, after Pakistan had emitted one of the costliest signals of toughness in modem history during its counterinsurgency campaign in East Bengal, it again pursued the same strategy two years later in its fourth Baloch campaign. It would continue to use high violence strategies in the Sindh, Balochistan V, and FATA II-TTP campaigns, a recurring habit the reputation theory has a hard time explaining. The rationalist utility of violence in Balochistan has likely run its course but there is no sign of accommodation as the state persists with this high-violence strategy, 362 motivated in part because of the identity of the rebels. Finally, the biggest weakness of the reputation theory in explaining the Pakistani state's behavior is that it cannot account for why the state consistently passed on early opportunities to signal their toughness against Pashtun challengers. The Faqir of Ipi, who led a tribal rebellion against the British and drove them mad in a manhunt for him, continued to promote his brand of Pashtun separatism after the British withdrew from South Asia. Even after the formation of Pakistan had undercut some of his support, he continued his "Free Pashtunistan" movement with his tribal Mehsuds to fight the Pakistani military and pose a threat to the settled areas. The Faqir's continued resistance coincided with numerous efforts by Kabul to stoke the Pashtunistan The Hamoodur Commission Report conducted by Chief Justice Hamoodur Rahman was officially completed in 1972 and then a revised version was completed in 1974, but the report was not publicly available until it was leaked in 2000. 362 Wirsing, 2008; Grare, 2013; Akbar, 2011. 361 232 issue including direct material support and intervention with irregular forces in collusion with local tribals in Pakistan's FATA, particularly between 1950 - 51. In Ramzak, North Waziristan, the Faqir would declare himself the President of Pashtunistan, the Afridi tribe in the Tirah Valley had also lent its support'to the Pashtunistan idea, and in Dir and Bajaur, local strongmen sought to wage war against the state with the help of Afghan irregulars. Whether it was the Faqir's continued rebel activity, his declarations along with other tribes of Pashtun independence, or the repeated incursions of Afghan irregulars in collusion with FATA tribals, Pakistan passed on numerous opportunities to signal its toughness and slap down these challenges to its sovereignty with serious force to demonstrate the costs of such actions before they spread elsewhere. The state repelled Afghan incursions and attempts to stir up tribal fueds and violence but took great pains not to punish, let alone blame the Pashtuns of FATA.3 6 3 Instead of rising to the occasion to crush these movements early and often as the reputation theory would expect, the state maintained confidence in Pashtun loyalty, tried to co-opt Pashtun tribals with an "Islam over tribe" mantra, and redirect their energies to other activities like the 1947 Kashmir campaign and greater inclusion within the state.3 64 Decades later, the government continued to downplay the risk of separatism during the post-2001 rise of the Taliban. Though the FATA insurgency from 2002-07 evoked concerns of a movement that had broader Pashtun nationalist intentions, even expressed by some military officers, because of the group's reach and ties to the, the government generally remained unalarmed and even responded with what many decried as an appeasement policy. 365 363 Cheema, 2002, p. 131 notes how Afghan agents infiltrated the tribal areas to stir up tensions. 364 Analysis based on Brig. A.R. Siddiqui, "Faqir of Ipi's Cross Border Nexus," Khyber.org, September 16, 2005, Accessed August 4, 2014; Brig. Asad Munir, "The Faqir of Ipi of North Waziritan" Express Tribune, Nov. 15, 2010; "Pashtunistan - 1947-1955," Global Security. Accessed August 4, 2014; Vahid Brown and Don Rassler, FountainheadofJihad: The HaqqaniNexus, 1973-2012, New York: Oxford University Press, 2013, p. 31-36. M. Ilyas Khan, "Taleban Spread Wings in Pakistan," 2007; Selig S. Harrison, "The Pashtun Time Bomb," New York Times, August 1, 2007. 233 The missed opportunities to signal toughness to the early Pashtun nationalists may have proven costly. Afghan incursions and continued support for a militant Pashtunistan insurgency would continue to be a thorn in Pakistan's side through the 1970s. More importantly, the willingness to tolerate Pashtun secessionist activities could have empowered other groups to consider rebellion against the state as well. Some analysts even linked the growth of post-2001 Taliban violence with Pashtun religious nationalism due to Pakistan's past tolerance of tribal rebellion. 366 Nevertheless, the Pakistani state continued to filter these strategic threats through the lens of identity. 4.3.4. Learning/Organizational Culture One variable debated in the literature is the effect of learning during and between campaigns and whether the military organization fosters a learning culture.3 6 7 The GOP's learning from mistakes does not appear especially powerful between campaigns. It repeated the same strategy against the first three Baloch rebellions despite repeated renewal of insurgency. Pakistan also repeated the same attrition strategy in Balochistan after suffering a catastrophic loss in East Bengal less than two years prior. While there is certainly learning within counterinsurgency campaigns, especially those with a much longer duration, lessons learned can also be bounded or trumped by the state's selection of strategy. In the case of Malakand and the TNSM, Pakistan quickly moved from a strategy of cooptation to one of population control within a year of launching a major campaign as it had learned from mistakes made here and in FATA and overhauled unit training and 366 367 Nadir Hassan, "The Faqir of Ipi and the Taliban," Express Tribune, April 18, 2012. Nagl (2002) argues learning cultures are more likely to move towards a particular type of campaign while Jackson (2008) finds there to be learning during campaigns that shape strategy but then regression between campaigns that end up favoring other strategies. 234 preparations for counterinsurgency operations in 2008.368 Yet even after it profited from these lessons in Malakand, Pakistan deployed a mixed strategy in FATA, applying massive force in the Mehsud region of South Waziristan, a cooptation strategy in North Waziristan and the border regions of South Waziristan controlled by Mullah Nazir, and an enfeeblement strategy in much of the rest of FATA. Meanwhile, in Balochistan, the Pakistan state pulled back from overt displays of force but pursued the same high violence/low effort strategy with more shadowy "dirty war" methods to advance the same goals. While the military was arguably more tactically competent in all areas and soldiers' survival rates improved in operations, their improved competence was selectively distributed and serviced different strategies in different regions.3 69 4.3.5. External Threat Recent scholarship has pointed to factors in the international environment that might motivate states to fight harder in counterinsurgencies, specifically links between insurgents and an external rival that render expectations of a future external threat environment all the more threatening. 37 The difficulty in using this to explain sub-national variation in Pakistan case is that to some degree, external collusion with insurgents, or the state's suspicion of it, is present in nearly every observation of insurgency (whether from Afghanistan or India), making it difficult to explain variation on the dependent variable. 37 ' The Pakistani military and state elite operates with the fundamental belief that insurgency is not possible without external support, that India 368 Haider, 2014; Haider Ali Hussein Mullick, "Recalibrating U.S.-Pakistan Relations," The Washington Quarterly, 35 (3), Summer 2012, pp. 93-107. 369 For instance, Pakistan's Frontier Corps (FC) in the northwest went through a process of learning and was upgraded, retrained, and better outfitted for tactical operations. But these lessons and retraining were not deployed to FC units in Balochistan, which persisted with its enfeeblement strategy. 370 Butt, 2012. 371 See Table 4.11 235 Table 4.11: Suspected External Support for Pakistan's Insurgencies Conflict Tye Country Source Baloch I Political support, safe haven marginal political support planning as early as 1965; Political support (to formulate the 6 Pt plan); Training of Guerillas Afghan reportedly trained 10-15K fighters, arms, safe haven India, Afghanistan Soviets India Lt Col Syed Iqbal Ahmad, Balochistan and its Strategic Importance, Karachi: Royal Book Company, 1992, p. 302 Lt Col Syed Iqbal Ahmad, Balochistan and its Strategic Importance, 1992, p. 212 Lt Gen Jahan Dad Khan, Pakistan Leadership Challenges," Karachi: Oxford University Press, 1999: 118; ; Shireen Mazari, India's Unconventional Strategy, Defence Journal, January, 1999 (http://www.defencejournal.com/jan99/rawfacts.htm); Lt Gen Gul Hassan Khan, Memoirs, 1993: 285-86; Fair, 2014 on P. Swami, 2007 Lt Col Syed Iqbal Ahmad, Balochistan and its Strategic Importance, Karachi: Royal Book Company, 1992, p. 178, 180, 181, 183; Brig Asif Harnoon Raja, "Seething Lava of Balochistan," The Muslim/DailyMuslims.com, March 4, 2012; Shireen Mazari, 1999; Haqqani, Pakistan: Between Mosque and Military, p. 174; Guruswamy, 'Call for Pashtun 425 Nation is Not Far Away', p. 2; Iqbal Khattak, 'Kabul Trained Baloch, Pushtoon Youth in 1970s: Ex-ANP Leader', Daily Times, February 15, 2010; Titus and Swindler, Baloch II East Pakistan Baloch IV India, Russia, Iraq, Afghanistan 2000: 63 Waziristan Sindh-MRD Karachi MQM RAW sought to nurture the Pakhtunistan issue in the 1970s; training of Pashtun militants in camps Training camps by RAW; support for Jeay Sindh movement Information; Covert material Support for violence from RAW; funds, weapons (believed to have continued into 1999) FATA I Baloch V training camps, Arms India India India, Afghanistan India Shireen Mazari,1999; Boggs, 2012; Haqqani, Pakistan: Between Mosque and Military, 2005 p. 174; Guruswamy, 'Call for Pashtun 425 Nation is Not Far Away', p. 2; Iqbal Khattak, 'Kabul Trained Baloch, Pushtoon Youth in 1970s: Ex-ANP Leader', Daily Times, February 15, 2010 Suranjan Das, 2001: 144; Shireen Mazari, India's Unconventional Strategy, Defence Journal, January, 1999 (http://www.defencejournal.com/jan99/rawfacts.htm); Anthony Davis, "Pakistan: State of Unrest," Jane's Intelligence Review, 11 (1), January 1, 1999: 33 Suranjan Das, 2001: 146; Lt Gen Jahan Dad Khan, Pakistan Leadership Challenges," Karachi: Oxford University Press, 1999: 281; Lieven, 2011: 315; I Malik, 1997: 246; Shireen Mazari, India's Unconventional Strategy, Defence Journal, January, 1999 (http://www.defencejournal.com/jan99/rawfacts.htm); "Six New Military Bases to Be Set Up in Sindh," The Pakistan Observer, March 10, 1993, p. 1; Anthony Davis, "Pakistan: State of Unrest," Jane's Intelligence Review, 11 (1), January 1, 1999, p. 33 MK Dhar, 2008, quoting lots of people including Mushahid Hussain interview, Outlook, April 24, 2006 (http://www.outlookindia.com/article/RAW-Is-Training-600-Balochis-InAfghanistan/231032); Brig Asif Harnoon Raja, "Seething Lava of Balochistan," The Muslim/DailyMuslims.com, March 4, 2012 Afghanistan Swat 11 FATA II India Suspected Baitullah Mehsud an Indian agent India Mazetti, 2013: 227 C~I "is always on the look out for reaping advantage from any misfortune that befalls Pakistan,"3 72 and consequently, that India "has always supported separatist elements in Balochistan, Sindh, and NWFP." 3 7 3 This is mirrored by Indian impressions of the repeated involvement of Pakistan's ISI in numerous Indian insurgencies. Since India and Pakistan have been locked in a rivalry since 1947, it is perfectly reasonable that each side regularly exploits the others' fissures by abetting ongoing insurgencies. If one is to measure degrees of collusion, then India's support of insurgents would seem most threatening and motivate the greatest effort. However India's links to and longtime support of Pashtun nationalism in Pakistan's northwest never triggered the same effort or violence in the TNSM I or Noor Muhammad cases. While Pakistani military and civilians have made a much bigger deal about India' support for the current Baloch insurgency, by most measures they have fought much harder in South Waziristan and Malakand, primarily because these threatened the core while the Baloch insurgency, even with stronger Indian support, did not. The comparisons of the Sindh-MRD and Karachi-MQM cases also reveal the limitations of external collusion in explaining violence. India's support for the MRD in Sindh was mostly rhetorical while it provided material support to the MQM in Karachi. Nevertheless, for such little external support, the state used much more violence against the MRD movement's Sindhi base than the Mohajirs in Karachi. It seems possible that identity and external collusion may interact in some way to produce more violent state responses. There may be evidence or suspicions of Indian involvement with the Taliban or MQM, but the perception of whether these links are merely tactical or whether they signal a groups' total disloyalty and insurgent alignment with Pakistan's chief rival and Author interview, Pakistan military official, Washington DC, July 1, 2011 (#13); Gul Hassan Khan, 1993, p. 285. 373 Syed Iqbal Ahmad, 1992: 181. 372 237 existential threat may turn on the identity of the insurgent base, group status, and representation within the state. At a glance, it seems the Pakistani state was willing to believe the worst of the Bengalis, Baloch, and Sindhi insurgents, and that their external links ran much deeper, demanding more violent responses. 4.3.6. Sons of the Soil Similar to the rebel aims variable, there is a notion that a certain type of rebellion, a "sons of the soil" rebellion where the rebelling group fears a loss of local/regional power through demographic encroachment of mainland migrants, can explain the dynamics of a conflict, particularly its duration, and also be extended to explain strategies of low effort. 7 Rebellions in Karachi, Sindh, Bengal and Balochistan have all exhibited "sons of the soil" elements with insurgents motivated by perceived encroachment and or exploitation of their homeland. Nevertheless the state has employed distinct strategies in these cases with only half corresponding to the "sons of the soil" predictions of low effort. Meanwhile, low effort strategies were employed in four FATA cases and one from Malakand, none of which had anything to do with settler encroachment or exploitation. In sum, this theory has little explanatory power for effort in the Pakistan cases, while the value of the territory fares much better. 4.3.7. Insurgent Threat There are a number of variables hypothesized to capture the level of insurgent threatconceptualized and measured by Gleditsch et al-but these do not offer traction on the Pakistan cases. More often, the territory threatened and the identity of combatants filters the state's threat perceptions. 3 74 Fearon, 2004; Fearon and Laitin, 2010; Mukherjee, 2014. 238 Rebel Size. If insurgent size is supposed to explain levels of violence, 3 7 5 then it has a hard time grappling with the high levels of violence deployed against the much smaller Baloch insurgencies throughout Pakistan's history. In the Baloch V campaign, far more violence was doled out to Baloch insurgents and their ethnic base than in Malakand II even though the TNSM rebel size was either the same or double that of Balochistan.37 6 For that matter, the state used mitigation strategies in Malakand I and FATA I though they were also roughly the same size as the Baloch V insurgency. Even after the state upgraded its FATA strategy with more violence and in some cases more effort, it still conducted minimalist activities against the militant organizations ranging from 5,000 - 20,000 in strength in North Waziristan. 377 Even in the case of Bengal, rebel size cannot easily explain the level of effort or mass violence. Though insurgent forces did eventually swell to between 80,000 - 100,000 by the end of November 1971, this mostly took place at the end as India stepped up its guerilla training in preparation for an invasion. Through June 1971, the Bengalis were livid with the Indians for slow-rolling them and only training 6,000 guerillas, yet it is precisely in these early months from March to June when the Pakistan security forces committed the greatest acts of violence.3 78 Rebel Strength. The skill and fighting capacity of a rebel group is though to be distinct from its size but also fails to explain most Pakistan cases. The state has used all strategies except mitigation against "much weaker" Mohajir and Baloch insurgencies while relying on mitigation for a number of TTP-affiliated insurgents judged to be at "parity" with the state. Moreover, the Pakistani state deployed much greater violence against the general population when fighting an inchoate rabble of "miscreants" in Sindh but was relatively cautious and discriminate in its use of 375 Valentino et al, 2004. 376 Baloch insurgents were estimated between 1,000-5,000 while in Swat they were estimate at 6,000-10,000. 377 Brig. Javed Hussain, "The Hornet's Nest," The News, January 19, 2011. Raghavan, 2013, p. 211; Gary J. Bass, The Blood Telegram: Nixon, Kissinger, anda Forgotten Genocide, Knopf, 2013, pp. 203-204. 378 239 force against a highly discipline Mohajir insurgency in Karachi. Here again, group worth and embeddedness help to explain these biases. Furthermore these variables are often highly endogenous to the state strategy and lack the predictive power prior to the outbreak of rebellion that the core-periphery theory's identity variable offers. Neither rebel size nor strength can explain the state's actions in early 1971, even before there was an organized Bengali insurgency composed of defecting battalions of the East Pakistan Rifles or guerillas trained by the Indians, when the armed opposition was just disparate acts of unorganized, political violence. The state initiated its attrition campaign with Operation Searchlight, an operation planned months in advance, to preemptively crush the beginnings of any insurgency and, through that, it in all liklihood galvanized the rebellion instead. Despite relatively matched levels in terms of insurgent strength between the total Mehsud/TTP forces in South Waziristan and total Baloch nationalist forces around 2006, one former state official judged that the Baloch "do not pose a serious military threat" or a threaten Pakistan's "territorial integrity." 3 7 9 Despite the high number of Baloch attacks, the difference is that the TTP/Mehsuds were able to project violence outside of the frontier region on to core territories making them a qualitatively different threat to the state because of where they were able to exert strength (figure 4.18). CentralizedControl. Insurgent organizations characterized with centralized control did not consistently elicit any more force or violence than those that lacked it. Both the Baloch and MQM movements had centralized control but elicited state strategies of enfeeblement and population control respectively, varying in both violence and effort. Similarly, both the TTP and TNSM lacked centralized control yet all four types of strategies were deployed against these organizations. Additionally, Insurgent centralized control is typically not coded well in macro379 Author interview, Pakistani former official, Lahore, Pakistan, October 3, 2011 (#20). 240 ire 4.18: Force Projection by TTP and Baloch nationalists through 2012 241 level datasets as it also may vary within the same insurgency across groups or over time and be endogenous to the state counterinsurgency, weakening its predictive power relative to the variables specified in the core-periphery theory. Arms Procurement. The Pakistani state has cracked down violently on groups lacking serious arms procurement capacity (e.g. the Baloch) and been much more careful with groups that have possessed it (e.g. the MQM and TNSM). It has also exerted high effort in a number of cases where the insurgents lacked basic weapons such as in Sindh and Balochistan IV. Bengali insurgents would eventually gain better training and arms from the Indians by the summer of 1971, but the Pakistani state actually initiated a high-effort (and intensely violent) crackdown in March and April long before they separatist movement received any arms and was only capable of mass protests and violent, unruly mobs. TerritorialControl. The effect of insurgent territorial control still rests on the value of the territory. Certainly the control of territory by the Mukti Bahini in Bengal and the TNSM in Malakand motivated high effort but the defacto control of territory insurgents in Balochistan and FATA generally did not elicit the same effort. The problem this variable raises provides further validity to my theory. Insurgent seizure of territorial control is only easily identifiable in cases where they wrest it from a political entity that previously had firm control. However, as my theory contends, if insurgents hold peripheral territory not previously controlled by the state, this will go unmeasured. What the variable is then actually capturing is insurgent encroachment into valued territory rather than actual insurgent strength. Insurgent Goals/Aims. Some theories purport to explain conflict processes such as conflict recurrence or duration based on insurgent aims.380 However, there are not clear predictions for what type of strategy this should produce. In the cases described above, Pakistan 380 Stanton, 2009; Licklider, 1995. 242 ' has launched high effort campaign against separatist insurgencies (as in the case of East Pakistan) as well as low effort campaigns against the separatist Baloch insurgents. Centerseeking rebels have also been met by both high effort responses like the case of Malakand II or the Mehsud/TTP in South Waziristan as well as low effort one like the first Malakand rebellion in the early 1990s or campaign in FATA prior to 2007 and for the most parts of it, even to the present day. Insurgent aims don't seem to have much traction on violence either with declared separatists being met with high amounts of violence as the cases of the Baloch and Bengali, rebellions, as well as by low levels of force like the case of the Mohajirs during the 1992 counterinsurgency campaign in Karachi. The bigger problem with this theory is that it is very difficult for states to gauge insurgent aims are but they can infer the worst based on their identity base. Some state officials believed that rising violence in FATA and Malakand post-2001 was potentially a case of rising Pashtun nationalist separatism, particularly given the history of Pashtunistan national movements.3 8 Nevertheless, the state still approached it with caution and low violence. Similarly, the GOP viewed the center-seeking populist MRD in 1983 as a separatist movement given its ethnic Sindhi base, but in this case, the state cracked down on it early and brutally in an attrition campaign. Meanwhile, most experts have regarded the Baloch insurgencies as separatist in aims 3 82 but some Pakistani analysts and officers suggest the previous campaigns were simply violent tribal bargaining rather than separatism. 3 83 Nevertheless, the state has consistently used brute force against the Baloch. 381 382 M Ilyas Khan, "Taleban Spread Wings in Pakistan," 2007; Cheema, 2008. Wirsing, 2008; Harrison, 1981. Author interviews with former Pakistan officials and analysts, Dec 1, 2010; Sept-Oct, 2011; May 23, 2013 (#3, 122). 383 243 VIII. Conclusion This medium-N analysis of Pakistan's counterinsurgency campaigns confirms the main predictions offered by the core-periphery theory on the role of territory and identity in shaping incumbent strategies. In nearly all cases, the Pakistani government deployed high effort strategies to defeat rebellions in core territories (e.g. Sindh, East Bengal) and regain full control of the territory while it deployed low-effort strategies to contain insurgencies in peripheral territories (Balochistan, and FATA). As Khyber-Pakhtunkhwa transitioned from a peripheral to a core region by the late 1990s, the state also transitioned to a high-effort strategy of control. An insurgent organization that was deemed containable in the 1990s had to be dealt with in a more decisive an enduring manner two decades later because the territory it threatened had become much more consequential for the Pakistani state. The costly lengths the state went to regain control were not simply demonstrated in terms of manpower, but also in terms of a much wider array of material resources, political commitments, organizational and tactical adaptations all for a broader project of transformation or rehabilitation of the contested region. The next chapter details in depth the efforts undertaken in some of Pakistan's campaigns. When the territorial stakes were high, the state had no qualms deploying all types of resources, implementing costly decisions and organizational adaptations, and sustaining deployments and high operational-tempo for extended periods of time. Pakistan's high-effort strategies for these core territories held even when resources were in short supply and state capacity at its lowest ebb, as evidenced by their efforts in East Pakistan (1971) and Karachi (1992-96) when U.S. assistance had declined or been entirely cut off. Even though the case of Balochistan IV (1973-77) departs from my theory's predictions, it demonstrates that even when 244 Pakistan lost half its productive resources, state strategy was not dictated by its diminished capacity, but rather in this case by Prime Minister Bhutto's political agenda. The Pakistan cases also confirm that rebels from core identity groups valued and embedded within the state (e.g. Pashtuns and Mohajirs) generally faced far less violent counterinsurgency strategies by a state restrained by the identity of its adversary. In these cases, the concern for excess collateral damage against "our people," exceeded concerns about the insurgents' violent, subversive (often separatist) challenges to state authority. By contrast, rebels with social base much more distant from the state's core identity groups (e.g. Baloch, Bengalis, Sindhis) did not benefit from the same restraint and instead met strategies of brute force and indiscriminate violence. In this latter set of cases, the Pakistani state exhibited no restraint, and by groups of low worth, with suspect loyalties, and very little embededness within the state. The mechanisms for how this identity distance translated into higher violence will also be explored in depth in the following chapters. The actual identification of variations in strategy, not just theoretically but empirically with clear and replicable measurement, also provides a meaningful contribution. As demonstrated by the numerous empirical observations from Pakistan, the state may be the resident leviathan but when fighting internal threats, it can selectively choose when to deploy more or less violence, and more importantly, effort. Given the number of cases that populate the enfeeblement and population control quadrants, "fighting hard" is not linear and in fact moves unevenly along two different dimensions of violence and effort, a concept generally underappreciated by the civil conflict literature. Therefore, a viable theoretical and empirical category of strategy in fighting insurgency can involve varying levels of violence, but without the objective of defeating insurgents outright. This approach may involve overlapping tactics 245 - with containment, 3 84 violence management 3 85 or rational appeasement, 386 and still be relatively effective at reducing the threat to the state. Despite very different strategies employed against the TTP in North and South Waziristan, violence was brought down to roughly the same levels in both (see figure 4.19). Figure 4.19: Insurgent Incidents in North (NWA) and South Waziristan Agency (SWA) 350 300 2501 200 _ 150 02001 - 2002 2003 NWA Incidents - 2004 2005 2006 2007 NWA Fatalities -- 2008 2009 SWA Incidents - 2010 2011 2012 -SWA Fatalities (Source: PIPS data) Finally, disaggregating Pakistani's political violence into a larger set of observations and cases reveals much more variation that cannot be explained with macro-level structures and requires a meso-level explanation of how state incumbents perceive threats and calculate the utility of different strategies. A state has plenty of motives, including reputational ones, to fight a rebellion that challenges its authority, but unproductive or non-strategic regions do not provide strong incentives for a costly, decisive campaign. Combined with the demands of sustaining a Watts et. al., 2014. 385 Staniland, 2012. 386 Treisman, 2004. 384 246 campaign and state-building in a frontier region with little physical or social infrastructure, Pakistan has balked at such strategies, conserved its resources (even without multiple simultaneous civil conflicts), and deployed minimalist strategies that merely contain conflict below a certain threshold, sometimes yielding a steady-state of chronic, low-level warfare and ungoverned spaces. A weak state like Pakistan needs to husband those precious resources for a variety of other security, political, and developmental contingencies. After all, "Pakistan is always managing risk and violence," 387 and subsequent chapters will demonstrate other statesmore democratic and stronger-do the same. This novel, yet disturbing finding belies the commonly held assumption that war is the more costly enterprise states seek to avoid or end, when in fact monopolization of violence is the far more daunting task. 387 Author interview, Pakistani civilian analyst, Lahore, Pakistan, October 1, 2011 (#18). 247 Chapter 5 Core-Periphery Mechanisms in East Pakistan, FATA, and Malakand I. Introduction For a country that has always emphasized its severe external threat environment and articulated a military doctrine around this geopolitical feature, the Government of Pakistan (GOP) has engaged in a striking number of counterinsurgency campaigns for more than half its period of independence.' Chapter 4 established a strong relationship between the GOP's value of the territory and identity groups and corresponding strategies it deployed in the 15 counterinsurgency campaigns it has fought. More often than not, insurgent challenges on valued territory elicited strategies of high effort and insurgents derived from valued identity groups encountered strategies of low violence and restraint. This chapter seeks to build on these findings and provide more in-depth examinations of the specific evidence linking the dependent variable of strategy to the independent variables of territory and identity. Specifically, I try to look for evidence over the course of Pakistan's formulation and conduct of counterinsurgency that supports my core-periphery theory's expected observations. This chapter then focuses on tracing territory and identity mechanisms that structure incentives, constraints, and opportunities for the expenditure of effort and violence in a particular counterinsurgency campaign 1.1. Empirical Strategy and Case Selection I selected three significant cases of Pakistani counterinsurgency -- East Pakistan, FATA I, and Malakand II to be closely examined in this chapter. These counterinsurgency cases were partially selected for the richness of historical analysis and high quality data available, which I estimate that Pakistan has been fighting counterinsurgency campaigns in 34 out of 66 years since its independence. 248 makes for better theorization and empirical tests. More importantly, however, these cases offered opportunities to closely examine three different strategies and make comparisons on both dimensions of effort and violence. In East Pakistan the state chose a strategy of attrition, in FATA it chose a strategy of mitigation, and in Malakand, a strategy ofpopulation control. These three cases are compared across a set of territorial incentives and identity mechanisms to evaluate what role they played in observable effort, violence, and the state's overall strategy. The tightly structured comparisons by the same incumbent (and in some cases over the same period of time) are able to control for confounding structural variables, but more importantly, they focus these comparisons on key observable implications and mechanisms that reveal strategic choices-that is, whether and how the state was expending serious effort or committing serious violence-as well as illuminate the motives behind them. My argument is highly contingent on the quality of accessible information, which admittedly is often quite limited, particularly on the national security decision making of a highly suspicious state that closely guards its secrets. Nevertheless, I attempt to compensate for this with the extensive data collection, fieldwork, and methods described in chapter 3 in order to make a contribution on the motives and processes that undergird the observable strategic choices often taken as given by most scholars of these conflicts. 1.2. Mechanisms In this chapter, I find that a host of specific mechanisms help to confirm in rich detail the relationship established in chapter 4 between territory and effort and between identity and violence in Pakistan's counterinsurgency strategies. 249 Territory Mechanisms. Observations of various types of investments - material, political, organizational, and time-can reveal beyond a reasonable doubt that the state was not merely throwing resources at the problem or even feinting effort, but actually seriously overcoming tactical and operational challenges in order to take its best shot at victory. These types of effort are revealed in quality as well as quantity of deployments, in tradeoffs the state was willing to make and risks it was willing to absorb, in the ambitions and scope of the counterinsurgency campaign revealed in speech evidence and operational scale, and in innovations and adaptations the state undertook to correct itself when containment was too dire to contemplate. In both East Pakistan and Malakand, the GOP's exertion of effort was clearly demonstrated with the investments made and risks incurred. Closer microanalysis of the cases can also help us evaluate whether other structural or strategic factors motivated these efforts or whether the territorial stakes instead drove it. In FATA for instance, the state resisted taking on such costly efforts even with a boost in capacity from outside assistance and economic growth because they knew the territorial stakes were not worth it. Identity Mechanisms. The more important move this chapter makes is to illuminate the ways in which observed employment of violence and non-kinetic substitutes was shaped by affective and strategic mechanisms triggered by the identity of the insurgents. Working backwards from detailed observations of the counterinsurgency campaign's properties, if we can observe evidence of the intervening identity mechanisms hypothesized in chapter 3, that would offer convincing evidence that the state formulated and employed a strategy with consideration of the insurgents' identity group. Foremost of these identity mechanisms, if observed state restraint and low violence is deliberate, we ought to observe evidence of empathy in word and deed prior to and during 250 conflict, as well as Pakistan's attempts to comprehend the grievance of the insurgents. This could provide convincing evidence that choices of restraint, selective targeting, and proposals for nonkinetic substitute approaches were deliberate and purposeful in their design, as will be demonstrated in the cases of FATA and Malakand where soldiers and statesmen express sympathy for the insurgent base with whom they has emotional and familial ties. If there was high violence, we should observe a total lack of empathy of outright expressions of contempt and disdain. For instance, this was evident in the speech evidence of Pakistan's leaders and commanders during the East Pakistan campaign vocally attributing the roots of the insurgency to the Bengali character and arguing that only significant violence could hope to alter this. Similarly, if strategies choices were deliberately formulated to minimize violence against the insurgent group or their identity base, we also might observe evidence of vulnerability to the identity group's power or political influence within the state-whether through backlash and criticism (voice) or even mutiny and defection (exit)-through internal debates, dissent, and reevaluation. Early in the FATA campaign, the Pakistani Army began observing these risks in limited form and made deliberate choices to avoid escalations of violence against Pashtun insurgents in FATA and Malakand. In strategies of high violence, the lack of any debate or course corrections and persistence with a brutal strategy could suggest insensitivity to any strategic risks. In East Pakistan, the GOP never felt the need to adjust because there were not enough (if any) Bengalis within the high command or state leadership to voice criticism or threaten to leave. Additionally, if the observation of low violence and selective targeting resulted from a co-identity (or proximate identity) advantage the Pakistani state possessed, we ought to be able to observe this in the state's capacity to gather and leverage strategic information on the 251 insurgents' motives, leadership, and behavior, as well tactical information for location, identification, and targeting. Pakistan demonstrated their ability to work with local tribes in FATA in order to separate out rebels whose demands could be appeased or co-opted and specifically target foreign fighters. In Malakand, the Army took great pains to resurrect the information systems of local surveillance like the police and peace militias working in conjunction with a rapidly repatriated local population, all to selectively identify and discretely target rebels. Conversely in cases of high violence, we should observe the total absence of this strategic and tactical information and the inability to gather intelligence, make distinctions in the commitment levels of insurgents, and delineate priority targets to punish the more hardcore and leadership elements. In East Pakistan, the near total absence of Bengalis in their regular or even irregular ranks rendered the GOP incapable of gathering or operating on any human intelligence and had no choice but to lump 75% of the roughly 75 million population base into a suspect category. Related to information, if the state was able to successfully make rebel surrender and rehabilitation, rebel cooptation, devolution of political authority, or the provision of public goods a large part of the strategy, this would suggest there was a degree of trust the state held in elements of the insurgent group and vice versa. Even as the GOP fought insurgents in FATA, it made significant use cooptation and violent negotiation, and in Malakand, it demobilized insurgents through non-kinetic means, provided public goods that were welcomed by the population (most notably security), and even devolved some political authority with judicial reforms. By contrast, the GOP would not even negotiate in good faith with Bengali nationalists prior to conflict and made little use of such trust-laden tactics during the conflict. When it did 252 eventually make some half-hearted attempt at a surrender and amnesty, it evoked no trust form the insurgent to amount to anything more than a failed public relations stunt. 1.3. Roadmap This chapter contains five sections including this introduction to examine the East Pakistan, FATA, and Malakand campaigns in order to evaluate the causal argument of the theory. In the second section, I examine the high effort, high violence attrition strategy in East Pakistan where there was strong evidence of the territorial interests motivating effort but a distinct absence of identity mechanisms restraining violence. Section three turns to the case of the low effort, low violence mitigation strategy in FATA, which exhibits just the opposite dimensions. While there was little material incentive to motivate the state past a containment strategy, the identity of the rebels compelled significant restraint by the government of Pakistan, almost to a fault. The fourth section turns to Malakand and examines a case where violence and effort specifically do not co-vary, as the state felt highly threatened and motivated to expend significant effort and force, but still managed to restrain its use of violence and utilize a wide range of defensive and non-kinetic tactics. The fifth section concludes with implications for the Pakistan cases and for a broader understanding of counterinsurgency strategy. II. Ravenous Leviathans: Counterinsurgency in East Pakistan Insurgency, 1971 In 1971, the Pakistani military launched an attrition campaign right out of the gate against a nascent Bengali insurgency in East Pakistan. East Pakistan's significant material, strategic and ideational incentives motivated the GOP to commit to a high effort strategy even before a real insurgency had emerged, though it soon grew into a massive one. Simultaneously, the Pakistani 253 state regarded Bengalis with low group worth, even contempt, while Bengalis had scant embeddedness in state institutions, both of which combined to unleash the state's use of violence. The GOP's brute force tried to degrade all forms of insurgent combat power and any potential base of support, committed indiscriminate violence against civilians due poor strategic or tactical information, and dismissed consideration of non-kinetic substitutes or bargaining almost completely. 2.1. Territory and Effort With immense value placed on East Pakistan, second only to Punjab, due to its productivity and contributions to the Pakistani economy, the GOP would exert tremendous effort to control this province. Bengal was too valuable to lose on material grounds alone but with the strategic assets it possessed, and its centrality to the "idea of Pakistan" as the Muslim state in the subcontinent at stake, the government had strong incentives to invest in a strategy of attrition. While some operational costs would be high, primarily the distance, the physical, and human terrain was slightly more manageable. The GOP therefore took great pains to invest in a campaign that involved high material resources, organizational changes and adaptation, costly strategic tradeoffs, a sustained and comprehensive deployment, and a politically ambitious agenda. 2.1.1. Costly Material Investments Despite claims to the contrary, the state invested significant resources in East Pakistan over a very short period of time to grip the territory tightly, even if it may have been too little, too late. The Pakistani state has been accused of caring little for East Bengal because prior to the 254 insurgency, it maintained only one infantry division as well as 20,000 paramilitary, but Pakistani defense doctrine during this period relied on defending the Western half and launching a counteroffensive to bail out the East. This did not mean they did not value East Pakistan but valued its center, Punjab, more. After all, the same doctrine only allotted one infantry brigade and artillery regiment to the defense of Sindh as the bulk of the Western frontline and reinforcements were concentrated in Punjab.2 And at the first whiff of violent instability, the state escalated quickly. In 1971, following mounting Bengali disenfranchisement, the Pakistan Army was sent in to East Pakistan to quell a nascent "Awami League rebellion"-a well-organized nationalist mobilization for an independent Bangladesh involving the establishment of a parallel government and province-wide strikes, protests, violence, and terrorism. 3 Military leaders were convinced that violent agitations, which began in March after the postponement of the National Assembly and crippled the province, were a prelude to a larger separatist insurgency being planned in part by a retired Colonel Osmany. 4 The GOP deployed the military on March 2 to manage the unruly political agitations that were escalating to spats of mob violence led by student militias and inflicted heavy casualties on these groups. It then demonstrated its commitment to East Pakistan by tripling the size of its Eastern Command regular forces over the course of a few weeks, and eventually tripling its irregular and paramilitary forces as well. The first the two divisions of reinforcements began arriving in late February and a brigade was airlifted on March 2, more than 3 weeks before the outbreak of the insurgency. 6The military requisitioned the Pakistan International Airlines' seven Boeings and moved two Gul Hassan Khan, 1993, pp. 174, 290-94. 3 Salik, 1977, pp. 228; Talbot, 1998, pp. 206-08. 4 Gul Hassan Khan, 1993, pp. 270-7 1. 5 Gul Hassan Khan, 1993, pp. 263. 6 Gul Hassan Khan, 1993, pp. 261 2 255 divisions worth of troops continuously for 14 days from Karachi to East Pakistan by flying long haul routes around India via Sri Lanka.7 Such movement of manpower and equipment by airlift and ferry were "real feats of logistics" and with such determination, by May, "the army was in full physical command and control of the province." 8 The GOP then initially deployed 34,000 regular army troops and 11,000 Civil Armed Forces (CAF) composed of the East Pakistan Civil Armed Forces (EPCAF), the Vulnerable Points Force, the Industrial Security Force, the EPCAF Special Force, and the police. Pakistan continually expanded its CAF, which may have peaked at 62,000. In addition, the state recruited and armed between 35,000 to 50,000 Razkars (militias drawn largely from the ranks of Biharis).9 Therefore its possible the state fielded a maximum of 146,000 security forces,' 0 well short of the 250,000 - 375,000 troops some in Pakistan believed they needed (based a T.E. Lawrence formula of 20 men per four square miles)," but certainly one of the largest counterinsurgency efforts Pakistan ever mounted. Because of the distance for logistics and supply, sustaining this large a force was a feat of power projection. 2.1.2. Comprehensive and Sustained Pakistan's comprehensive coverage of territory and sustainment of this posture also exerted tremendous effort. As soon as the military launched Operation Searchlight on the night of March 7 2 5 th and took over the urban centers like Dhaka within a few days, the commanding Mascarenhas, 1971, pp. 132; Haqqani, 2005, pp. 78. 8 Siddiqi, 2004, p. 132. 9Figures counted from Salik, 1977, p. 105 and Niazi, 1999, pp. 52, 75-79. 1 Excluding the Razkar numbers, we know that 1300 Pakistani military were killed in the war with India in Nov/Dec 1971 and another 79,700 were captured (Haqqani, 2005, p. 87) so there were at minimum 81,000 military/paramilitary (not counting the Razkars) in theater prior to the inter-state phase of the conflict. Prior to the fight with India, Salik (1977, p. 123) estimates Pakistan had the 3 divisions (of 34,000 troops) plus the paramilitary and Razkars totaling 73,000. With five additional battalions arriving in the end of November (Salik, 1977, p. 127), the peak force may have been 111,500. " Salik, 1977, p. 101 256 General Tikka Khan dispatched troops across the province to take over and hold major towns and communication centers, seal the borders, and patrol and dominate the spaces in between "to dislodge any bases the rebels may establish" in what was expected to be a long campaign.' 2 Army units were also dispersed within the province to provide support and leadership to irregular Razkar units guarding vital infrastructure points.1 3 By august, senior military officials expressed confidence that they had regained full control of the region.1 4 Even after "the restoration of 'complete normalcy'," no one questioned why the military remained in place because the expectation was that they had to maintain control with constant patrols of roads, urban centers, and parts of the countryside. 5 16 2.1.3. Costly Tradeoffs Throughout this campaign, commanders stuck with the deliberate but costly choice of dispersing rather than concentrating troops. The move was a significant tactical adjustment for the troops as well as for senior officers coordinating the campaign between General Head and the theater command. This decision for dispersal involved a significant tradeoff. Border outposts manned by platoons served the counterinsurgency objective of denying the rebels any maneuver space or territory to control, both for tactical and political reasons, but in doing so, this dispersal made units much more vulnerable to aggressive guerilla attacks and even conventional probes or an all out Indian invasion. Furthermore, it left critical nodes like Dhaka much less defensible, even when the state had concluded an Indian invasion was inevitable.1 7 The Pakistan military Gul Hassan Khan, 1993, pp. 275, 277 13 Gul Hassan Khan, 1993, p. 313. 14 Gul Hassan Khan, 1993, pp. 275, 277. " Siddiqi, 2004, p. 156. 16 Mascarenhas, 1971, p. 126. 17 Gul Hassan Khan, 1993, pp. 275, 277, 302-304, 314. 12 257 ' was unwilling to concede any separatist gains and in so doing, may have made itself more vulnerable to conventional war. 2.1.4. Politically Ambitious In addition to the military choices, the state pursued a highly ambitious political goal. Pakistan would not be satisfied with containment and was seeking a radical transformation in political authority, one of usurpation. One Brig. general who served as the head of ISPR at the time wrote, "The army had moved in not just to control an immediate law and order situation but also to brainwash the people.. .The army, it seemed, had come to stay. It was not just a temporary mission of pacification but one of conversion and transformation..."18 A reporter who traveled through Bengal as part of a press delegation described phrases he repeatedly encountered from military officers - "cleansing process" and "rehabilitation effort" 19 which distinguished this type of campaign from containment strategies that relied on limited punitive actions. Part of this may have involved a degree of religious cleansing by purposefully driving out Hindus to flee as refugees,2 0 and the "crusading zeal" of the theater commander, Gen. Niazi, to "convert 'Hinduized' Bengalis into good Muslims." 2 The East Pakistan campaign was an attempt to achieve total victory and control through a remaking of society with the blunt instrument of force. Knocking the insurgents around for a bit like the state periodically did every decade in Balochistan would not prove sufficient here; instead the state tried to comprehensively destroy their capacity and will for generating combat power to return the territory to a peaceful state. Though the elected civilian leader who would 18 Siddiqi, 2004, p. 111. 19 Mascarenhas, 1971, p. 121. 20 Bass, 2013; Raghavan, 2013. 21 Siddiqi, 2004, p. 170. 258 pick up the pieces of Pakistan after 1971, Zulfikar Ali Bhutto, blamed the atrocities of the campaign on military dictator "Yahya Khan and his gang of illiterate psychopaths," he would go on to defend the principal of using brute force by invoking the examples of other brutal nationbuilders. Bhutto stated: You can't build without destroying. To build a country, Stalin was obliged to use force and kill. Mao Tse-tung was obliged to use force and kill. To mention only two recent cases, without raking over the whole history of the world. Yes there are circumstances where bloody suppression is justifiable and justified. Officers down the ranks may have unwittingly agreed that this campaign was about nation building. One officer confided that the benefit of the campaign might have helped to make soldiers out of "docile and chicken-hearted" Bengalis and hoped that Bengal would have a new generation of warriors "as a result of the rapes."23 2.2. Identity and Violence Pakistan's contempt and disdain for Bengalis due to deep-seated racial prejudice along with the fact that the group was tremendously underrepresented in the state and military owing to historical contingencies enabled the state to fight in a completely unrestrained manner. The previous chapter demonstrated that even prior to considerations of rebellion, Bengalis had no group worth and hardly embedded in state institutions. My theory expects that in the event of Bengali rebellion, we should expect to observe evidence of contempt driving violence, limited strategic and tactical information driving indiscriminate targeting, nonchalance over any blowback effects, and distrust foreclosing on alternative tactics or bargaining. A close examination provides evidence to support these expectations. 22 23 Oriana Fallaci, Interviews with History and Conversationswith Power, New York: Rizzoli, 2011: pp. 286-88. Siddiqi, 2004, p. 166. 259 Simply put, the Pakistani leviathan was unleashed due to the lack of empathy, vulnerability, information, or trust. First, the Pakistani government viewed Bengalis as disloyal, with contempt, and of low worth, generating no empathy or restraint and in fact a strong incentives for brute force against any and all Bengalis potentially supporting the separatist movement. Second there were few to no Bengalis at high positions within the state to impose costs with their dissent or defection, inoculating Pakistan from the strategic or political costs of its actions to which other incumbents might have been vulnerable. Third, there were no Bengalis left within the state or even civil servants or officers with the ability to inform state strategy with strategic guidance or tactical intelligence for better decision-making and more selective targeting. Finally, the state's perceptions of Bengali disloyalty and continued distrust between the two parties inhibited any interest or prospects of non-kinetic substitutes or bargaining alongside ongoing violence. 2.2.1. Empathy The Bengal case offers an illustration of how ethnic contempt translated into an excessively violent counterinsurgency campaign. Being the most underrepresented in state institutions and reviled by most West Pakistani leaders that ran the state, we should expect to observe the most unrestrained violence practiced in this counterinsurgency campaign. The ethnic stereotypes combined with an absence of any Bengalis within the state to exhibit empathy or advise caution or alternative methods led to a strategy that called for excessive violence to quell the rebellion, resulting in significant casualties, escalation, and later inter-state war. 260 ' Contempt Prior to the conflict, Bengalis were the victims by "unspeakable outpourings of racial hatreds and stereotyping." 24 West Pakistanis had inherited British stereotypes of Bengalis as part of the "feminine races" that were "feeble and spineless."2 5 Field Marshall Ayub Khan's peculiar perceptions of Bengalis (described above) were reflected by West Pakistani ethnic exclusion and chauvinism, which identified Bengalis with "a very different ethnicity, history, tradition, language, and culture." 26 Intelligence chief Maj. Gen. Akbar Khan referred to Bengalis as "bastards," and President Yahya Khan referred to them as "Bingos." 2 7 The belief that "certain citizens are more equal than others,"28 perceptions of Bengali inferiority, and disregard for their lives were evident in day-to-day state policy long prior to the counterinsurgency. While revenue generated from East Pakistan was high, reinvestment back into it was low, and per capita expenditures of West Pakistan averaged 87% more than those in East Pakistan. 29 Furthermore, the military doctrine of "defence of East Pakistan from West Pakistan," that concentrated forces in the West Pakistani core and sacrificed the East signaled to Bengalis "you are worthless and indefensible." 30 Disdain became even more palpable with a very poor government response following a cyclone in the Bay of Bengal that killed an estimated 200,000 Bengalis and a drunken Gen. Yahya Khan, publicly making light of the situation.3 24 Talbot, 1998, p. 212. Barbara D. Metcalf and Thomas R. Metcalf, A ConciseHistory ofIndia, p. 111. Talbot, 19 9 8 , p. 186. 27 Salik, 1977, p. 107; Ali Hassan, Pakistan: Generals aur Siyasat, Urdu (Pakistan: Generals and Politics), Lahore: Vanguard Books, 1991, p. 167. 28 Jaffrelot, 2002, p. 18. 29 For all data, refer to pages 4,11-12, and 266 of Twenty-Five Years ofPakistanin Statistics, 1947-72, Central Statistical Office, Economic Affairs Division, Ministry of Finance, Planning and Development, Govt. of Pakistan, Karachi, 1972. 30 Hamid Hussain, 2002. 31 Nawaz, 2008: 261. 25 26 261 Furthermore, Bengali distance bred beliefs about disloyalty and little regard was paid to whether Bengali grievances bore some legitimacy. Theater commander Gen. Niazi for instance feared that Bengalis were closet Hindus who would betray the state.32 Chief of General Staff Gen. Gul Hasan Khan expressed what many senior officials believed, that the landowning Hindu classes had "brainwashed" the people of East Pakistan.3 3 Bengalis as a whole had long been considered "anti-state," "anti-Islam" and "dogs let loose on the soil of Pakistan." These types of statements revealed the "background of general mistrust and prejudice against Bengalis" held across the board as "almost all civilian and military leadership was of the same view" that a healthy dose of violence was the only method to "sort them out."34 Brute Force Pakistan's total lack of empathy would permit and in fact demand an intensely violent attrition campaign to crush all resistance as quickly as possible. Operation Searchlight was launched on March 25, 1971 to reassert state authority through martial law, tanks, and mortars in a political crackdown of students, journalists, political leaders, and local Bengali regiments whose allegiances were suspect and would compose the bulk of insurgents. This "scorched earth policy" involved indiscriminate violence and civilian victimization to terrorize the population into submission was actively encouraged by the commanding generals.3 6 3 7 Leaders including the military head of state, Gen. Yahya Khan, the Information secretary, and other civil servants, expected that a little collective punishment-"a taste of the danda (big stick)," "a whiff 32 Niazni, 1999, p. 34. 3 Gul Hassan Khan, 1993, pp. 244-45. 3 Hamid Hussain, 2002. 35 Hamis Hussain, 2002. 36 Niazi, 1999, p. 46; Blood, 1971; Nawaz, 2008, p. 269. 37 Niazi, 1999, p. 46. 262 of grapeshot," and "some fear of god"-would quickly dispense the insurgency and "cow down" Bengalis into submission.3 8 The utter lack of empathy, or more apt, the contempt bordering on hatred for the Bengalis resulted in what one UN official estimated was between 5,000-25,000 civilians killed within the first two days of the crackdown. 39 The campaign would then expand across the province for another month retaking and seizing authority from locals, mostly aligned with Bengali nationalists, to directly control it from West Pakistan. Most Bengalis seeking political rights appeared disloyal, irreconcilable, and therefore viable targets, necessitating a brutal counterinsurgency campaign. The commanding officer, Lt. Gen Tikka Khan pursued a conventional attrition strategy to "break the backbone" of the Bengali insurgency by rapidly depleting their numbers, 40 and utilizing irregular Bihari militias (Razkars) as death squads. 4 1 He treated Bengalis like a foreign enemy, and entirely complicit, rather than "dealing with his own misguided and misled people," stating "I want the land and not the people" 42 and vowed "to kill all the traitors, and if necessary, raze Dacca to the ground. There will be no one to rule; there will be nothing to rule."43 . Tikka's right hand, Maj. Gen Farman Ali wrote in his diary the "green land of East Pakistan will be painted red." 44 In a retrospective analysis, one Brigadier writes that the first month of the campaign was purely about reclaiming territory with little regard for the people: The typical military mindset had regained land at all costs, but the people inhabiting the land had been lost very cheaply.. .the ruthless efficiency with which phase 1 of counterinsurgency was completed alienated the whole population."45 38 Zaheer, 1994, p. 141; Siddiqi, 2004, p. 195; Iqbal Akhund, Memoirs ofa Bystander: A Life in Diplomacy, Karachi: Oxford University Press, 1997, p. 211. 39 Raghavan, 2013, p. 149. 40 Salik, 1977, p. 92. 42 Haqqani, 2005, p. 79. Niazi, 1999, p. 46. 43 Salik, 1977, p. 53. 41 All quotes from Niazi, 1999, p. 46. 45 Shah, "How to Cope with LIC at National Level," 2002, p. 179. 44 263 What motivated high state violence was not the strength of the insurgency but Bengali identity itself. Senior officers professed Bengalis "only [knew] the language of force" and were "Hindu at heart." 46 Officials talked of "cleansing" East Pakistan, "purify[ing] the Bengali race" and bragged that they had "sorted out" Bengalis "well and proper - for at least a generation." 47 The dehumanization of Bengalis was carried down the ranks to younger officers and the enlisted, translating into a complete lack of restraint. Another Pakistani reporter who witnessed "kill and burn missions" was given a sense of the distance to which Pakistani security forces would go for control. When visiting the 1 &"division headquarters, officers repeatedly told him, "we are determined to cleanse East Pakistan once and for all of the threat of secession, even if it means killing off two million people and ruling the province as a colony for 30 years." 48 The counterinsurgency campaign may well have achieved something close to this. Though some revisionist scholarship place total fatalities at 100,000 and Bengali separatists claim 3 million were killed, estimates of 300,000 to 1 million killed at the hands of state forces may be plausible given private Pakistani estimates. 49 Regardless, such levels are extreme for a campaign that lasted a little over 8 months and is the reason why many scholars regard as genocide for the level of viciousness displayed. 50 Believed to be the root cause of the instability, Hindus were also specifically targeted en masse. Observers noted the high degree of "racial prejudice" and "racial hatred" evident in the counterinsurgency operations as well as that high Salik, 1977, p. 78; David Loshak, Pakistan Crisis, New York: McGraw Hills Book Company, 1971, p. 112. 47 Salik, 1977, p. 78; Hussain, 2002; M. Tazeen Murshid, "A House Divided: The Muslim Intelligentsia of Bengal," Ed. A. Donald Low, The PoliticalInheritanceofPakistan, London: MacMillan, 1991, p. 147. 48 Mascarenhas, June 13, 1971, p. 118. 49 Bose, DeadReckoning, 2011; Ulfelder and Valentino, 2008. Former commanders Tikka Khan and Farman Ali Khan estimated the numbers of Bengalis killed to be 34,000 and 50,000 respectively. By late April, in country estimates were on the order of 100,000 and by July, West Pakistani officials privately estimated 250,000 people had been killed in the conflict. See Mascarenhas, 1971, p. 118; Bass, 2013, p. 251. 50 Bass, 2013. 46 264 level officers directed specific but indiscriminate targeting of Bengali Hindus. While its unclear what percentage of the casualties they constituted, by the end of the campaign both the CIA and Indian government estimated that Bengali Hindus made up roughly 80% of the refugees. 2 2.2.2. Information StrategicInsight At the strategic level, due to the lack of any Bengalis within state leadership circles, state decision makers-civilian and military alike-were clouded by their own prejudices and lacked guidance or insight to accurately diagnose the cause and motives of rebellion, or the efficacy of brute force. ... The only commodity in abundance was ignorance mixed with raw emotions and rhetoric. The military elite failed to understand the dynamics of their own society. They embarked on an ambitious but un-realistic goal of higher level of national cohesion in the absence of genuine participation. In March 1971, they took the fatal decision on the basis of their thoughts that the conflict is an artificial one and they will control it by attacking it directly and with brutal force. In this assumption, all the civilians of West Pakistan (including civilian bureaucracy and all political parties) were in agreement with the military's point of view. A dysfunctional information feedback loop emerged. The absence of Bengali input in the decision making cycle rendered the regime "blind," "suffering from delusions," and "ignoran[t] of ground realities," unable to make accurate evaluations of strategy and course correct, and instead led to decisions based on "deep-seated prejudice".5 4 Ironically the only real resistance to this attrition strategy came from the few West Pakistan commanders who had actual previous experience on the ground in East Pakistan. Had there been more like them, or more Bengalis in 51 Bass, 2013, pp. 81, 83. 52 Bass, 2013, p. 236; Raghavan, 2013, p. 209. 53 Hamid Hussain, 2002. 54 Hamid Hussain, 2002. 55 Zaheer, 1994, p. 159. 265 the high command, the Pakistani state's military strategy might have been more restrained or at least sensitive to losses and mistakes, and potentially more adaptive. Commanders dispatched to East Pakistan did not possess special knowledge or insight or refused to listen to any that came their way. Though out of power, in his diary Ayub Khan remarked on his surprise at the selection of Lt Gen. Tikka Khan being dispatched to East Pakistan to take control of the region. Ayub wrote, "Why Tikka is difficult to understand. He has no experience of that province and knows nothing about it. Tikka is a nice and a simple man, but comes from the Punjab." He went on to forecast that this would sow distrust with Bengalis "and will be used as a proof of Punjabi domination." 56 One officer writes that even when Maj. Gen. Niazi, who took over from Tikka, still had some access to Bengali advisors who could help guide his actions, he refused to heed their advice.57 Intelligence Early on the campaign, the Army chief of General Staff Lt. Gen. Gul Hassan Khan assessed that the military was effectively flying blind as "the intelligence system had broken down completely" due to the few Bengalis within the military having mostly defected or covertly working for the insurgency to feed misinformation. Aside from getting tactical field reports from troops who were equally hamstrung, Pakistan possessed no strategic intelligence "no information that could aid us in constructing an overall scenario on which we could build our plans."58 The callousness and indifference of West Pakistanis towards Bengalis not only produced indifference to civilian victimization but also limited the state's access to vital information even if it wanted to be more selective in its approach. Civil servants posted to Bengal had mostly Baxter, 2008, p. 460. 57 Gul Hassan Khan, 1993, p. 326. 58 Gul Hassan Khan, 1993, p. 301. 56 266 behaved like distant colonizers and distanced themselves from the "natives," refusing to engaged with the people or learn the language and consequently affording the state little knowledge of the human terrain.59 Pakistan had also foreclosed on the only real local sources of information and intelligence they potentially had when they had tried to disarm the EPR and EBR, causing them to mutiny and defect. As a result there was little knowledge within the military or bureaucracy about how to navigate the human terrain, collect intelligence or identify informants or potential Bengali collaborators. It's difficult to distinguish whether the lack of empathy or information generate more indiscriminate violence as there is evidence for both. The military conducted "search and sweep" and "clearing" operations, sometimes due to bad intelligence, sometimes out of a desire for retribution, but in both cases, "no witnesses were left." 60 The military lumped all Bengalis into three categories - white, grey, and black - and with the 1 million ethnic Biharis and only 20-25% of ethnic Bengalis being considered "whites," nearly "all Bengalis were potential suspects."61 Since the government struggled to separate insurgents from noncombatants, this allowed security forces to indiscriminately identify anyone as a "miscreant" (essentially an insurgent) and "dispatch them to Bangladesh" (i.e. kill them).62 A strategy that made little distinction between civilian and combatant did even less to separate out types of committed and uncommitted insurgents. Consequently every bit of resistance had to be stamped out with brute force. One officer told a field reporter, "It took me 59 Gul Hassan Khan, 1993, pp. 340-41. For instance, see Siddiq Salik, 1977, pp. 92-96; Siddiqi, 2004, p. 130. Siddiqi, 2004, p. 132. 62 Siddiqi 2004, pp. 132, 148. 60 61 267 five days to get control of this area. We killed everyone who came in our way. We never bothered to count bodies." 63 2.2.3. Trust As violence began to escalate between the Bengali nationalist opposition and government forces, the ethnic divides and contempt the state harbored for the Bengali people contributed to deep trust, which prevented attempted negotiations from de-escalating the conflict during the run up to insurgency, the early stages of inchoate violent resistance, and months into the counterinsurgency campaign. The state signaled its distrust of Bengalis early. By refusing to seat the National Assembly, it nullified Bengali victory in the December 1970 elections. Tensions mounted as Bengali nationalists began to mobilize strikes, demonstrations, and protests in the province of East Pakistan. Escalation may have been avoided during this mass mobilization if local forces rather than Army units had been tasked to handle the situation, but the central government had no confidence in the local paramilitary and police run by the state government and refused to turn over control to them. 64 In the early days of the nascent insurgency when mobs clashed with the armed forces in early March 1971, there may have been ample opportunity to cut a deal between the Awami League and the GOP, but this opportunity was constrained by deep "mutual mistrust" between Bengali leader Mujibur Rahman and PPP leader Bhutto, with the GOP President Gen. Yahya Khan trying to serving as an independent broker. Talks between Yahya and Rahman in midJanuary followed by Bhutto-Rahman talks at the end of that month yielded little progress. 63 64 Loshak, 1971, p. 112. Gul Hassan Khan, 1993, p. 261. 268 Harmless actions were viewed with intensifying suspicion. A meeting between Bhutto and Yahya was perceived by Bengalis as collusion and Rahman's unwillingness to travel to West Pakistan for negotiations in February was seen as confirmation of secessionist ambitions. In midmarch when Yahya met Rahman in Dhaka, "the level of mistrust between Yahya and Mujib was so great" that they met in a bathroom of the Presidential House because Rahman feared the meeting rooms were bugged.65 Bhutto's meetings with Rahman never achieved resolution, in part because he was negotiating in bad faith, unwilling to countenance serious negotiations, and tried to "browbeat" Rahman into a compromise. 66 Some accounts place the President as trying to facilitate negotiations between Bhutto and Rahman but after multiple attempts and Bhutto rejecting all of them, he then suggested "a short and sharp military action" to change the bargaining table. 67 The military had in fact been considering this option since February if not earlier. 68 Outside observers suggest that there was never any good faith bargaining over autonomy in the early months of 1971 because the state had always viewed their Bengali counterparts as disloyal and that "negotiations were carried on and protracted chiefly in order to enable the Army to build up its strength in East Pakistan." 69 One contemporary defense analyst writes: In soldier's mind when conclusion was reached that Bengalis are traitors then the next line of action was quite obvious. You don 't negotiate with traitors,you finish them off to save the country from their ravages. Soldier was now ready, mentally prepared to deal with a group who was seen as coward and only to be dealt with force."70 Instead of leaving any room for bargaining alongside counterinsurgency operations, the state launched a massive attrition campaign with Operation Searchlight on the night of March Talbot, 1998, p. 202. Gul Hassan Khan, 1993, p. 283. 67 Gul Hassan Khan, 1993, p. 272. 68 Talbot, 1998, p. 206. 69 K. Wolff of the UNDP, quoted by Raghavan, 2013, p. 148. 70 Hamid Hussain, 2002. 65 66 269 25 h by killing or capturing as many political leaders of the Bengal movement as they could find. Afterwards Bhutto exclaimed to reporters, "By the Grace of Almighty God, Pakistan has at last been saved," believing military action alone would be sufficient to end the political impasse. 72 Months into the brutal counterinsurgency campaign, attempts at political outreach and resolution were not seriously attempted or entertained given the degree of violence that had ensued. For the most part, the Chief of General Staff (CGS) did not sense any effort underway to initiate broad political outreach, which was why his visits to the East Pakistan theater were constrained to "purely military matters." 73 After months of slaughter, the Pakistani state initiated some conciliatory gestures for international propaganda, but they underestimated the degree of Bengali disaffection that prevented these gestures from de-escalating the conflict. In September, the GOP sought to appease Bengalis by appointing a civilian Bengali governor (though clearly under the thumb of the military) and even announced an amnesty for rebels, though it specifically excluded the top leadership, which "took the wind out of the sail of the general amnesty." 74 The CGS was disappointed to find that these "momentous political concessions" were ignored and "did not elicit even a lukewarm response." He attributes it to the indoctrination and restrictions placed on the separatists by the Indians.75 Evidently the Pakistan military discounted how their actions had widened the gulf of distrust. On top of the state's attrition campaign that perpetrated mass killings, just a month before they made these political gesture, the GOP had published a white Gul Hassan Khan, 1993, p. 273; Gary Bass, 2013. Raghavan, 2013, p. 51; Gul Hassan Khan, 1993, p. 283. 73 Gul Hassan Khan, 1993, pp. 301, 337. 74 Siddiqi, 2004, p. 163. 75 Gul Hassan Khan, 1993, pp. 282-83. 71 72 270 paper laying all the blame and of the violence on Bengali political leaders and then announced Rahman would be tried for treason.76 Brig. Siddiqi, head of the ISPR dismissed the amnesty as "ill-time, inadequate, and illconceived", a "crude joke," and "neither general nor real," that had zero effect. Had the GOP considered this in May, it might have had some traction but by September, extreme violence had eliminated any prospects for politics due to mutual hatred. As evidence of this bad faith, just as the amnesty was announced, GOP forces began a massive manhunt to round up Bengalis and pre-date their capture so as not to come under the purview of the amnesty. 7 7 Even the CGS recognized some of these actions were simply not genuine, like the local elections to fill vacant seats, which he describes as a "complete farce." Ultimately, all these efforts at negotiation after utter ethnic contempt had led the state to slaughter hundreds of thousands of Bengalis were "a case of too little, too fragmentary, and too late." 78 On the other side, as late as November, the Indian government sought to convince U.S. leadership to pressure Pakistan into negotiations but the military was too dug in with contempt for anything to emerge.79 2.2.4. Vulnerability From 2002 onwards, the Pakistani state would feel an acute sense of vulnerability turning its guns on Pashtuns in its counterinsurgency efforts in the tribal areas, but in Bengal it suffered no real costs to massive indiscriminate violence against Bengalis, or at least remained to immune to them. Though there was a sizeable contingent of Bengalis in uniform serving in East Pakistan, 76 Raghavan, 2013, p. 208. 77 Siddiqi, 2004, pp. 163-170. Gul Hassan Khan, 1993, pp. 282-83. 79 Bass, 2013, p. 253. 78 271 - they were considered suspect and expendable, and their defection imposed no serious cost. Along with Operation Searchlight, the military sought to quickly disarm the East Bengal Regiment and East Pakistan Rifles, though the effort was botched and the officers and troops mutinied and fled to set up liberated zones as the vanguard of the insurgency.8 0 While not ideal, the state did not see this as a real cost because the disloyalty of Bengali units was inevitable "Bengali soldiers.. .could hardly be trusted." 8' A GOP spokesman explained to reporters that these units were loyal to Rahman and when the state initiated a military crackdown "as expected, they had initiated the mutiny within minutes of the military action." 82 The military did express some concern about Bengali defections, but primarily from a tactical standpoint. Bengali defection would quickly strengthen the insurgency by providing intelligence to the rebels on Pakistani Army assets, capabilities, and positions, 83 but without real standing within the broader military apparatus or the state, this would have no ripple or backlash effect that would compromise the center from within. Instead of exposing the state or military to vulnerabilities and prompting a course correction as it would with the Pashtuns, military decision makers doubled down on their strategy of violence and preemptively relieve Bengalis from their posts. 8 4 Bengalis had a slightly larger role in the Pakistan Air Force but they too were soon grounded as all "the loyalties of Bengali personnel became suspect" because "Yahya saw a 'Mujib' in every Bengali."8 5 80 81 82 83 84 85 Raghavan, 2013, p. 52. Siddiqi, 2004, p. 90. Siddiqi, 2004, p. 121. Gul Hassan Khan, 1993, pp. 277-78. Gul Hassan Khan, 1993, pp. 338-39. Siddiqi, 2004, p. 162. 272 2.3. Alternative Explanations A significant amount of scholarship has been devoted to perhaps this most brutal of counterinsurgency campaigns in the modem era, but three alternative explanations for the nature of the conflict and the violence are offered. First, some argue this was willful genocide and had nothing to do with the nature of the rebellion. Others suggest quite the opposite that the scale of violence was not deliberately directed by the principal but agents gone rogue and committing private acts of violence. A third camp suggests that the campaign might have been conducted very different had civilians been at the helm of the state. Each of these alternative explanations is considered and refuted below. 2.3.1. Willful Genocide? Charges of genocide have been frequently leveled on the case of East Pakistan, but there are plenty of reasons why this was not a deliberate attempt to exterminate an entire race of people. While undoubtedly a case of mass killing that involved ethnic cleansing, the motives for this level of violence, while perhaps not purely rational, had a strategic, though shortsighted, "logic of appropriateness" grounded in the identity politics and prejudices of Pakistan as my theory argues. First, the charge of genocide was made by U.S. Ambassador Archer Blood in a memo to the state department, but was initially done for shock value and the actual charge was disputed by others in the embassy like the USAID chief.86 There can be little doubt this was a case of mass killing based on extreme ethnic prejudice, but the Pakistan military was reacting to mass mobilization and rebellion. If it had wanted to pursue a genocidal policy, it would have initiated the campaign long before the crisis had escalated with popular mobilization, perhaps as early as 86 Bass, 2013, p. 81. 273 December 1970-January 1971 (or for that matter, years earlier since Bengali opposition had been a problem for some time). The military had operational plans long before it initiated them but continued to pursue some meager attempts at negotiated solutions for a few months. Killing Bengalis was cheap but it was not costless and did not afford the state a major advantage in its own right-in fact it was a disadvantage given how much the state was able to extract from Bengali production. The state found utility in killing Bengalis primarily to deter or defeat violent challenges to its authority. The Pakistani military was also preoccupied with too many other activities that departed from a systematic genocide of Bengalis. Dispersing forces to guard borders rather than spend all their time massacring civilians in villages actually worked against a systematic killing campaign since it did not mass firepower and made platoon outposts more vulnerable to counterattack. The reason things turned so brutal in the countryside was that rebels practiced guerilla warfare on a very large scale, which is a key predictor of the use of mass killing.87 Second, there was theory of victory behind the mass killing that was rooted in ethnic prejudices. Ethnic cleansing specifically targeted Bengali Hindus who were believed to be the root cause of the insurgency. The main form of this came not by systematically killing all Hindus but sparking enough violence and fear to drive most of them to flee. 8 The strategy appeared to work as roughly 8 million Hindus had fled in total by the end of the conflict. Moreover, the GOP disputed figures on Bengali refugees citing only the two million Muslim Bengalis who had fled, a sign it had no intention of allowing them to return. 9 The use of cleansing in the counterinsurgency campaign was founded on a highly warped racial theory that Hindu presence and influence provided the source for Bengali weakness, Valentino, 2004; Valentino et. al., 2004. Bass, 2013, pp. 82-83, 169. 89 Raghavan, 2013, pp. 206, 209. 87 88 274 assertions of distinctiveness, and separatist inclinations. The GOP held warped notions of the "enslavement" of Bengal by Hindus, 90 and officers and politicians repeatedly suggested that Bengali ethnic separatist "plot" was driven by the influence of Bengali Hindus and landlords. 9 1 It therefore seemed appropriate (though perhaps not rational) to target them specifically in order to root out the driver of the insurgency. Simultaneously, Pakistan state elites believed the ethnic character of Bengalis was inherently weak-willed and that that exposure to high levels of violence and "shock action" would induce them to back down. 92 Third, an undefined but substantial amount of the violence resembled inter-ethnic violence of partition such as the violence reprisals between Bengalis and 1 million Biharis who had settled in Bengal after partition, whose loyalties were generally aligned with the GOP. Though state department dissidents led by Archer Blood blamed most of the killing on the Pakistan's initiation of violence, they still estimated at least one third of the deaths were nonBengalis, a far cry from one-sided violence that typifies genocide.9 3 Others have placed the level of private, inter-ethnic violence to be much higher owing to brutal actions by insurgents, religious differences, ethnic insecurity, and score settling that took place during the conflict. 94 Finally, the GOP judged high intensity violence necessary early in the conflict to deter India from predation specifically in this case, and later in the conflict to gamble for resurrection. The military was "desperate" to firmly resolve the conflict with brute force so as to prevent prospects for separatism as well as Indian predation and intervention, 95 which had been 90 Bass, 2013, p. 82. 91 Bass, 2013, p. 169. Hamid Hussain, 2002. 93 Bass, 2013, p. 85. 92 94 9' Siddiqi, 2004; Bose, 2011. Siddiqi, 2004, p. 149. 275 anticipated months if not years before it actually began. 96 This would always remain a vulnerability of Pakistan and thinking this specific insurgency was orchestrated by the Indian agents in collusion with Bengali Hindus to test their resolve, they had to act firmly and quickly to regain control before the clear military superiority of India could weigh in to dislodge East Bengal from the Pakistan's sovereign control.97 In this case, the external threat uniquely interacted with the rebel identity base to escalate the state's use of violence. 2.3.2. Indiscipline? Just the opposite of the genocide argument, some have advanced that the intensity of violence was not a strategy but either acts of a desperate state completely caught off guard by the rebellion or a principal-agent problem of soldiers gone rogue and pursuing private violence based on fear, anger, resentment, or ethnic hatreds. However, there is clear evidence that high violence and collective punishment was not simply a result of spiraling escalation action and bloodlust but part of a pre-meditated strategy due its anticipated efficacy. First, documents regarding a secret, premeditated contingency plan, "Operation Blitz" that called for aggressive action had been prepared and circulated even before the disputed 1970s election. The plan called for the declaration of a state of emergency and authorized "the greatest vigour and determination to create an unmistakable impact and remove any doubts regarding the type of martial law, which is being imposed in contra-distinction to the deliberately watereddown martial law to which people have become conditioned. Shock action would therefore be imperative...There should be no hesitation in using force for effect." 98 Gul Hassan Khan, 1993, pp. 285-86; Jahan Dad Khan, 1999, p. 118; Shireen Mazari, "India's Unconventional Strategy," Defence Journal, January, 1999. 97 Gul Hassan Khan (1993, pp. 290-293) talks about fears of increasing Indian power. 98 Nawaz, 2008, pp. 265-66. 96 276 Additionally, state leaders appointed a theater commander on the basis of his ability to reliably carry out a strategy of brute force. Lt. Gen. Tikka Khan (known as the "Butcher of Balochistan" for his previous campaigns) had been dispatched to take charge of the Eastern Command with the expectation that he would "get the job done in short order." 99 President Yahya Khan had replaced a squeamish general in the Eastern Command because he was skeptical of a military solution and in his place, and charged Tikka with the responsibility because he was "unclouded by strategic thinking or complicated vocabulary."' 00 In other words, he was the right man to carry out the state's plan of brute force with grim efficiency. Though there is certainly evidence to suggest that indiscipline had broken out in the ranks over the course of the campaign, this was not only countenanced by some officers up the chain of command but also actively encouraged as part of the overall strategy.10 1 Gen Niazi was believed to have tacitly or sometimes actively encouraged sexual violence and one officer writes that this strategy was directed from the top -- "the soldiers were under orders not to show any mercy to subversive, anti-state elements."' 0 2 An officer recounts one general reportedly asked his men 'How many Bengalis have you shot?"1 03 2.3.3. Would Civilians Be Different? Some scholars argue that had civilians been in charge of the Pakistani state during this period, the counterinsurgency campaign might have been different. Bhutto himself stated "I would have done it with more intelligence, more scientifically, less brutally."1 04 99 Nawaz, 100 2008, p. 266. Nawaz, 2008, p. 266. 101 Nawaz, 2008, p. 277; Haqqani, 2005, p. 75. 102 Siddiqi, 2004, p. 167. 103 Bass, 2013, p. 70. 104 Fallaci, 2011. 277 However, this theory has significant holes. First this claim discounts how much military action was encouraged by civilian leaders, as described in the details above. Accounts by various civil servants, military officers, and politicians at the time all suggest there was strong support within both the military chain of command and the West Pakistan-dominated civilian bureaucracy and political class for an aggressive counterinsurgency campaign to knock "sense" into the Bengalis.10 5 In fact this was in keeping with Bhutto's approach in particular, which was to push for military aggression for populist reasons and then disavow the aftermath, as he did with the 1965 war.106 Second, the earlier quotation by Bhutto suggested he knew exactly what type of force was necessary to effectively build a state, and while he was happy to outsource it to the military, he effectively stood by it. Moreover, despite being the leading civilian in West Pakistan and playing and having a significant voice in Yahya's political circle, 107 he never urged restraint in any of the Army's actions until after they had been defeated by India. Even after the Hamood Dar Commission he appointed condemned the military for its actions, Bhutto had no qualms appointing Tikka Khan, the theater commander who had played a decisive role within the Bengal counterinsurgency campaign, as Chief of the Army Staff. He later defended the choice in an interview that Tikka was "a soldier doing a soldier's job." 0 Third, considering the counterfactual, it is unlikely that had a civilian been in charge and force was authorized that civilian leaders would have then been able to alter the course of the campaign strategy and leash the military because the fundamental incentives that restrain both 15 Talbot, 1998, p. 206; Hamid Hussain, 2002 (see notes 33-38); Akbar Ahmed, 1997; Gul Hassan Khan, 1993, p. 272; Raghavan, 2013, p. 50; Fallaci, 2011. Even Haqqani who critiques military rule acknowledges civilian support of a harsh response. See Haqqani, 2005, p. 70. 106 Ashok Kapur, Pakistanin Crisis, 1991, p. 111. 107 Ashok Kapur, Pakistan in Crisis, 1991, p. 79. 10 Fallaci, 2011. 278 militaries and states, was absent. Civilians maintained the same ethnic contempt for Bengalis and were unwilling to brook political concessions to them. 109 They also considered the rebellion a supply-side problem, possessed the same insensitivities to violence against Bengalis, and lacked the strategic and tactical information to guide the campaign more selectively. Even if civilians had the political strength to intervene and overrule military choices, the identity composition of the state lacked the desire (home team discount) or the tools (home team advantage) to implement a strategic change. Finally, even if regime type offers some traction on the case, at a minimum, the type of regime interacted with the lack of group worth and embeddedness to intensify conflict. Military rule meant Bengali underrepresentation in the institution would escalate tensions. One Pakistani analyst specifically links Bengali underrepresentation in the security forces combined with military dominance, to the intensifying violence: In a multi-ethnic society like Pakistan, where all ethnic groups are not represented in the institution of armed forces ... The military's 'seizure of power [has] the effect of ethnicization of areas of politics not formerly ethnically salient and/or intensifying ethnic awareness where it already exists... a prelude to a violent showdown between the state and the aggrieved ethnic group.110 III. Reluctant Leviathans: Counterinsurgency in FATA I & II, 2002-13 3.1. Background The FATA campaigns present a case of the very opposite strategy used in East Pakistan. Here the Pakistani state was neither motivated to deploy serious effort, nor significant levels of violence, instead settling on a mitigation strategy for what was a very formidable threat. 109 110 Talbot, 1998, p. 206. Hamid Hussain, 2002. " This account is drawn from a number of sources including Haider, 2014; Brian Cloughley, 2008; Nawaz, FA TA, 2009; Schofield, 2011; Gul, 2011; Lalwani, 2009; Lalwani, 2013; Shahzad, 2011; Ahmed Rashid, Descent into Chaos, London: Penguin Books, 2008; Seth G. Jones and C. Christine Fair, Counterinsurgencyin Pakistan, Santa Monica, CA: Rand Corporation, 2010. 279 Following the September 11, 2001 attacks and U.S. operations in Afghanistan, Pakistan soon began to experience a rise in militancy and terrorism on its Western border, which eventually developed into an insurgency as a result of two simultaneous factors: the infiltration of foreign fighters formerly based in Afghanistan into Pakistan's Federally Administered Tribal Areas (FATA) and parts of the North West Frontier Province (renamed Khyber Pakhtunkhwa (KPK) in 2010) and, the induction of Pakistani regular forces into the region. A mix of Al Qaeda members, foreign fighters, and Afghan Taliban escaped into the region and some eventually began to integrate with certain tribal networks and affiliated militias, which were a regular feature of FATA and to some extent KPK. With the help of coercion and inducements, Pakistan committed to supporting U.S. operations in Afghanistan and by April of 2002, it had deployed two regular infantry brigades from the Army's XI corps to accompany its Frontier Corps wings to FATA.1 1 2 Though elements of the Pakistan regular army had moved into the FATA for short durations, this was significant for being the first time ever that the Army had seriously deployed in the area since the country's independence in 1947.113 After the initial round of cooperation, the Pakistani military conducted operations to capture the al-Qaeda remnants and foreign fighters scattered around FATA. 14 These militants had begun to assert their dominance over local tribal power structures, challenge the local political order, and in conjunction with local militants, wage war against the GOP, mostly targeting its outposts and installations around the tribal regions. Some local tribal militias saw opportunities to ally with foreign elements for motives ranging from greed, grievance, Markey reports an Army division was deployed to Khyber and Kurram in June 2002 to block militant escape routes from Afghanistan. See Daniel Markey, Securing Pakistan's TribalBelt, Council Special Report No. 36, Washington DC: Council on Foreign Relations, August 2008, p. 11. 1 3 Nawaz, FATA, 2009, p. 9; Cloughley, 2008, p. 186; Some accounts state that the military had entered the tribal areas before but in a very minor way - both in scale and duration - so this was at least a significant departure in terms of the scale of force employed and remaining based in the area. 114 Schofield, 2011, p. 133; Fair and Jones, 2010, pp. 46-56. 12 280 identity, social norms, and enhancing their status and position in local political and power configurations,"5 and the conflicted had escalated to a full-scale insurgency by 2004,116 which a number in the Pakistani government feared had nationalist and separatist goals.' 17 In response to Pakistan military operations targeting al-Qaeda after 2001 and more robust campaign around 2004, an inchoate and nascent insurgency eventually formed the basis for the Tehrik-i-Taliban Pakistan (TTP) 3.2. Territory and Effort The GOP judged that the low value of territory in FATA combined with prohibitive costs presented little appeal and warranted only a modest low effort campaign. Described by some in the Pakistani government as the "wild west" and known as known as "yaghestan," the land of unrest, 11 FATA encompassed harsh terrain, poor infrastructure, and a history of unruly resistance to the state "where the government's writ meant little" 19 This led analysts and military leaders wanted no part in a campaign demanding excessive blood and treasure for a region that would require "hundreds of years to tame."12 0 As one journalist described it "the game hadn't been worth the candle."121 " 5Nawaz, FA TA, 2009; Gul, 2011, p. 31. 116 Kilcullen ("Terrain, Tribes, and Terrorists ," 2009) writes, "By the end of 2004 the tribes were in a full, though undeclared, frontier war against the government."; "Imran Khan's Statement on Acts of Violence," Pakistan Press InternationalInformation Services, Aug 6, 2004 (Factiva). 117 Harrison, "The Pashtun Time Bomb," 2007. 118 Lieven, 2014, p. 193. 2013, p. 36. 119 Mazetti, 10 Author Interview, Pakistani former official, Lahore, Pakistan, October 3, 2011 (#20). 12 Mazetti, 2013, p. 105. 281 3.2.1. Territory Incentives Historically, FATA had no capacity for "surplus rent extraction."1 22 The land could not sustain large-scale agricultural activity and with no revenue source, there was no "magnet for the central government to assert its control more aggressively." 123 FATA's economy had depended on government allowances to tribal elders, NGO aid, trading in illicit products, and migrant worker remittances, none of it really extractable by the state or worth controlling. It was essentially dead space but for the fact that it was territory and potentially served as a buffer as it had during the great game days of British colonialism.12 4 Since then this strategic value had declined. Furthermore, FATA offered little ideational value since it had been ruled indirectly outside the normal authority of the state. After Pakistan's independence, it continued the British system of indirect rule. The state relied on political agents to enforce order by managing relations with a set of tribal elders or maliks, utilizing the frontier crimes regulation for judicial authority, and the frontier corps for punishment or enforcement if necessary. This was not necessarily an institutionally weak system and appeared to work well through the 1970s but it was certainly a different one.125 Consequently the people of FATA were perceived to doubt the notion that a monopoly of violence was a public good and "never believe[d] they [we]re better off with the state." 26 FATA's historic peripherality, isolation, and treatment as a colony, 2 7 had acquired a 122 Adnan Naseemullah, "Shades of Sovereignty: Explaining Political Order and Disorder in Pakistan's Northwest," Unpublished Paper, 2013. 123 Hamid Hussain, 2007. 124 Amina Khan, 2011; Naseemullah, 2013. 12 Akbar S. Ahmed, 2004. 126 Author interview, Pakistani analyst, Lahore, Pakistan, October 6, 2011, (#23). 127 Ibid. 282 path-dependence due to increasingly prohibitive costs of development and a set of vested interests in the semi-autonomous nature of these ungoverned spaces. 128 The physical challenges to state reach in FATA were visible with just one look at the terrain and limited physical infrastructure but weak social infrastructure lowered expectations of what policymakers though was feasible and acceptable to the people of FATA. Measures of human development -particularly education, literacy, and health were the lowest here than any other region in the country and only half the country's average if not lower.1 29 As a result, the state came to believe that whatever changes it implemented by force "would not take" and this made the state less ambitious in its approach to the region. 3 0 Even when the GOP promised time and again to implement more development program to "corrupt them with progress" they expected this would be a decades-long venture that required proceeding very gingerly. 13 1 Management, Not Control Given the costs and incentive structure, the region of FATA was always thought of a place with periodic rebellions that could be only be contained. American analysts wanted to believe that Pakistan's "fundamental strategic aim is different" from the British policy of maintaining FATA as an ungoverned space when this was in fact distinctly not true. 3 2 A former chief secretary for KPK stated, "Waziristan is a management issue, a containment issue, even during British [colonial] times. It has never been controlled... Any commander with any sense Ijaz Khan, 2008, p. 20; Amina Khan, 2011. Measures of human development presented in Shakeel Kakakhel, "Integration of FATA into NWFP: An Analysis," Forum of Federations Project, Feb. 2010, p. 7. Also see International Crisis Group, Pakistan'sTribal Areas: Appeasing the Militants, Asia Report No. 125, Islamabad/Brussels: ICG, Dec. 11, 2006, p. 9-10; Nawaz, FATA, 2009, p. 8; Arshad Ali, "Internal instability in Pakistan - Ideological and Socioeconomic Perspectives," Strategic Studies, (ISSI), 31 (1-2), Spring 2011, p. 97. 130 Author Interviews with U.S. and Pakistani analysts and officials, Washington, DC, May-July, 2013. (#123, 127129, 133). 131 Author interview with Pakistan Army official, Washington, DC July 11, 2013 (#133). 132 Kilcullen, "Terrain, Tribes, and Terrorists ," 2009, p. 10. 128 129 283 would not use ground troops - you'd be entrapped. Waziristan can swallow six or seven [army] divisions."1 33 The concept of violence management was practically enshrined in Pakistani doctrine. So long as violence remained below a certain threshold and did not threaten survival, the state could tolerate these challenges. This approach is consistent with rationale expressed in Pakistani Army doctrinal writings. In the 2002 PakistanArmy Green Book, one officer writes: The type of aggression encountered in [low intensity conflict] is not as blatant as that in war. Subversion, sabotage, assassination and guerilla operations.. .may pose a threat to national interests, but the threat to national survival may be neither imminent nor obvious.. .The inevitable ambiguity of the proper employment of force demands that weight be given to other considerations. Limited nationalinterests...and lack offeasibility help explain apparenttolerance ofsome undesirablesituations. On the other hand, non-involvement accepts the piecemeal degradation of security interests and tolerates unnecessary human suffering, both of which might be prevented or alleviated by a more active, if necessarily, selective approach.134 Even Gen. Kayani who ran the ISI from 2004 onwards and would ascend to the rank of Army Chief of Staff from 2007-2013 understood the importance of violence management. Kayani's master's thesis at Ft. Leavenworth on the Afghan Mujahedeen emphasized the difficulty and prohibitive costs of a decisive military defeat of tribal insurgents in the border region of Afghanistan, which would entail massive, and perhaps intolerable personnel losses, and economic and political cost." 135 Khalid Aziz, a former chief Secretary of KPK, quoted in Saeed Shah, "Pakistan Plans Air Assault in Waziristan," The Globe and Mail, July 13, 2009. 134 Rehman, 2002, p. 203. 135 Mazetti, 2013, pp. 112-113. 133 284 3.2.2. Effort Because the territory was both worthless and operationally difficult, the state chose to "contain rather than eliminate militancy" in FATA, 13 6 and deployed limited effort in terms of material investments, organizational investments, sustainment, or comprehensiveness. The state's objective was not to transform upgrade political authority but to restore the stability of the old system, a form of nonincorporation that abdicated authority to local, informal power structures. Limited and Poor Quality MaterialInvestments The GOP deployed limited and poor quality material investments in FATA. Despite official troop numbers, during this period from 2002-07, after adjusting for troops actually in operation (the estimated "real" force numbers) and force quality, it is evident the GOP did not invest in military superiority in the FATA. Though it officially deployed as many as 70-80,000 troops and broadcasted these figures to counter accusations by US and NATO allies that they were not doing enough, few took into account the types of forces employed, whether they were already garrisoned there, and how they were actually utilized. In fact the GOP primarily used poor quality forces, limited operations to much smaller scales, and left most of their forces sitting in the region but dormant. Save for some operations in South Waziristan, for the most part Pakistan seemed to rely heavily on the paramilitary Frontier Corps (FC) "managing militant threats" 137 despite being developed primarily for border patrol and smuggling interdiction operations rather than counterinsurgency. As a result, they were ill 136 Ijaz Khan, 2008, p. 14. Cloughley echoes this in stating "the army was doing all it could to contain violence." See Cloughley, 2008, p. 189. Others echo this. Weinbaum (2009, p. 75) argues this was the case until early 2009. Meyerle (2012, p. 29) states the military believed the insurgency could be contained to the tribal areas prior to 2007. 137 Markey, 2008, p. 13. 285 equipped, poorly trained, at times unwilling, and generally much less capable at this time than regular Army force.1 3 8 FC forces may have composed as much as 60% of the deployment. 13 9 Two regular army divisions were already based in the region around Peshawar and were simply moved into FATA. Additionally, the descriptions of most major operations during this campaign only seemed to involve 10,000 or fewer Army troops with the real operations led by SSG forces, 140 leaving the most of the 35-40,000 Army force remained garrisoned in their bases or tied down to static posts. When GOP forces did engage in periodic strikes, despite the clear asymmetry in power and capability, it incurred casualties at nearly the same rate as the insurgents, further evidence of the low quality forces being used. By the end of 2004, the GOP had captured 656 militants and killed 302 but GOP security forces had incurred 221 fatalities and 482 casualties/injuries.141 142 Even when Frontier Corps forces were sent in on an operation, their limited capabilities rendered them vulnerable, forcing the army to send in heavy-handed reinforcements and close air support to bail them out.1 4 3 138 Lalwani 2009, pp. 37-38; Fair and Jones, 2010, pp. 38-39; Hassan Abbas, "Transforming Pakistan's Frontier Corps," Terrorism Monitor, 5 (6), March 30, 2007; Tariq Mahmud Ashraf, "The Pakistan Frontier Corps in the War on Terrorism- Part I," Terrorism Monitor, 6 (15), July 25, 2008; Arthur Keller, "Propaganda and Peace Deals: The Taliban's Information War in Pakistan," CTC Sentinel, 1 (8), July 2008. 139 In September of 2008, Pakistan's Secretary of Defence stated that of the 90,000 troops deployed in the war on terror at the time, about 50 paramilitary/Frontier Corps personnel were deployed to each of the 1100 check posts along the 2430 km Afghan-Pakistan border. Assuming this balance held in years prior to this period, this means that at least 60% of the armed forces Pakistan had claimed to commit were FC and of much lower quality than the regular army. See Rasool, 2008. 140 Schofield, 2011, in particular details a number of these. The reliance on the SSG is echoed in Haider, 2014 and Author Interview with Pakistan Military official, Washington, DC, July 2013 (#133). 141 Tariq Mahmud Ashraf, "Pakistan's Frontier Corps and the War Against Terrorism - Part Two," Terrorism Monitor, 6 (16), Aug 11, 2008. 142 Gul, 2011, p. 31. 143 Schofield, 2011; Gul, 2011; Jones and Fair, 2010, pp. 49-50. 286 The state was simply unwilling to allocate resources to the FATA theater, even as tensions with India began to abate from 2003-2007, reducing demands on its eastern border.144 The military was not unaware of other models of counterinsurgency that called for more comprehensive approaches that also addressed the demand-side drivers of insurgency,1 45 but they chose not to apply them. Even after another counterinsurgency campaign prompted the state to upgrade some deficient capabilities,1 46 they still selected more limited campaigns for most of FATA.1 47 Limited Operations With No OrganizationalChanges Counterinsurgency experts noted that Pakistani forces maintained a very limited operational tempo, static deployment pattern, and narrow scope of targeting that hardly adapted to the new challenges or environment.1 48 Other journalistic sources suggest that for the most part, the Pakistan military was hunkered down in its bases and doing very little patrolling or operations.149 Based on his meetings with officers, one Pakistani analyst described the approach in FATA: In other parts of FATA, the military presence did not lead to active battles. South Waziristan remained largely the domain of the TTP, as it used its tribal Mehsud area as a staging post. In North Waziristan, the land of the Ahmedzai Wazir and the Daur tribes, and home also to the 144 See Steve Coll, "The Back Channel: India and Pakistan's Secret Kashmir Talks," The New Yorker, March 2, 2009; Author interview with a senior Indian intelligence officer, December 5, 2012 (#101). 145 Many of the articles from the 2002 Pakistan Army Green Book recognize the importance of political and economic prongs to address insurgency. See also Commander Adnan Mazher Khan (PN), "Tribal Areas - Need for Integration into National Mainstream - Recommendations, PakistanNavy War College Review, 2005-06. 146 Daniel S. Markey, No Exitfrom Pakistan:America's TorturedRelationship with Islamabad, New York: Cambridge University Press, 2013, p. 128. 147 For investments, improvements and successes see Mullick, "Lions and Jackals," 2009; Nawaz, 2011; General David Petraeus, "Striking a Balance: A New American Security," Transcript of Keynote Address, Center for New American Security, Washington DC, June 11, 2009; Frederick W. Kagan, "The Two-Front War; Pakistan is Finally Doing its Part. Now We Need to Do Ours," The Weekly Standard, 15 (8), November 9, 2009. For continued limitations see Reza Jan, "Trickling Home to South Waziristan," 2010; Jan and Worby, 2011. 148 Kilcullen, "Terrain, Tribes, and Terrorists," 2009, pp. 10-12 149 Ahmed Rashid, 2009. Also verified through author interview with former U.S. intelligence official, June 14, 2013 (#129). 287 Haqqani group of the Afghan Taliban, the army adopted what a senior military officer derisively called a policy of "sitzkrieg"-meaning, sitting in camps without any aggressive actions. The army described the policy as "dominating space."150 A CIA officer deployed to the region observed that the militants controlled the road networks and that during this period, the Pakistan military deployed to the area were conducting "few patrols and spending nearly all of their days inside protected barracks.""5 When the GOP did summon greater resolve and effort, it still relied on limited conventional tactics. Two dozen major operations conducted "search and destroy missions"15 2 with tactics including "sweeps, searches, and occasional bloody battles,"1 53 and the occasional pounding of suspected militant hideouts with helicopter gunships and fighter aircraft. 5 4 Roughly a hundred minor operations conducted rapid and short, almost surgical, operations to strike militant safe houses where foreign fighters were believed to be based.1 55 156 In both cases, the objective was narrow - to round up any foreign fighters but for the most part avoid area domination or any conflict with local militants even if they had already turned against the state.1 57 The same CIA officer observed "no elements of the Pakistani military [would] slug it out with the Taliban in ground combat." 5 8 ' Nawaz, 2011, p. 10. 1 51 Mazetti, 2013, pp. 156, 158. 52 ' Markey, 2008, p. 11. 153 1 Fair and Jones, 2009/10, pp. 168, 172. 4 Gul, 2011, p. 26. Haider, 2014; Jamal Hussain, 2012. 156 See bar graph on p. 132 of Zahid Ali Khan, "Military Operations in FATA and PATA: Implications for Pakistan," Strategic Studies (Islamabad), 31/32 (4/1), Winter 2011 and Spring, 2012, pp. 129-146. 157 Fair and Jones, 2009/10, p. 172. 15 Arthur Keller, 2008. 155 288 Lack of Sustainment The campaign in FATA was most criticized for its lack of sustainment. Operations were "sporadic"1 5 9 and "not sustained over time,"16 0 "short-lived" 161 and conducted "within a much narrower prism." 162 Though military tried to "degrade leadership and, where possible, kill Taliban fighters,"163 most observers and even Pakistani officials acknowledged retrospectively that military actions were "incomplete, inconclusive, and at times, appeared insincere." 164 Following quick clearing operations, the army was "reluctant to maintain a presence in them afterwards." 16 5 This inattention to holding territory allowed insurgents to repeatedly return to the area and resume their activities. Selective Rather than Comprehensive Pakistan also demonstrated limited effort in its FATA I campaign by choosing a highly selective and segmented strategy rather than a comprehensive one, both in terms of targeting and deal making, and maintained this approach even after 2007. After being pummeled by TTP attacks in Pakistan's core, the GOP launched a new campaign, FATA II, beginning in 2008 to crush and dismantle TTP/Mehsud faction in South Waziristan. Nevertheless, it would continue to narrowly target and only contain other TTP-affiliates and militant organizations throughout FATA that had fought against state forces, aligned with TTP's goals, and still posed an armed threat to the state's authority in the tribal areas. 159 Haider, 2014. 160 Fair and Jones, 2009/10, p. 172. Weinbaum, 2009, p. 76. Shahzad, 2011, p. 184. 163 Haider, 2014. 164 Weinbaum, 2009, p. 76. 165 Mullick, "Lions and Jackals," 2009. 161 162 289 Initially, Pakistani operations very narrowly targeted foreign fighters and al-Qaeda operatives along with some of their local supporters.1 66 As the state pursued these fighters in the Waziristan region, their limited incursions faced resistance from Wazir and Mehsud tribals that sparked a broader insurgency. To manage them, the state frequently indulged in deal making with all tribal groups even after they had fought the state and solely concentrated on the foreign fighters and al-Qaeda operations. This pattern persisted until roughly 2008-09. Eventually, due to the TTP projection of violence on Pakistan's core, the state shifted its strategy in FATA II to aggressively attrite a particular group fighting the state, though it kept the scope relatively narrow. The GOP disproportionately focused its targeting on the Mehsud camp (which later became synonymous with the TTP) and effectively gave a pass to other major militant tribal organizations led by commanders Mullah Nazir and Hafiz Gul Bahadur. All the major tribal groups in Waziristan had at one point fought the state since 2002 and continued to conduct activities that undermined it, but only some were deemed real enemies. The Mehsud tribe (led by Abdullah, then Baitullah, then Hakimullah), which would eventually be labeled as "bad Taliban," appeared to put up more a more sustained and consistent resistance to the state, conducted hit and run attacks against the Pakistan military between 2004-07 and later formed the core of the TTP.1 67 Additionally, the Mehsud tribe began launching attacks outside of FATA very early on in the conflict - as early as Dec 2004.168 Those tribal warlords dubbed "good Taliban" would go largely untargeted even though they also had participated in fighting the state. Gul Bahadur, who led the Uthmanzai Wazirs in Some note that little attention was paid to the Taliban, in part because at least until 2005, the US did not prioritize Pakistan's policy towards the Taliban in talks and may not have asked them to target the Taliban. See Vali Nasr, The DispensableNation: American ForeignPolicy in Retreat, New York: Doubleday, 2013, p. 69; Markey, 2013, pp. 113, 127. 167 Mazetti, 2013, pp. 156-57. 168 Meyerle, 2012, p. 22. 166 290 North Waziristan, fought Pakistani forces in 2006 and 2008 and at one point assumed the role of deputy commander in the TTP. Even when he opposed fighting the Pakistani state, 169 he still agreed to an alliance with Mehsud, harbored anti-state militants fleeing South Waziristan, and maintained close ties to Hakimullah Mehsud, Baitullah's successor. 7 0 Mullah Nazir, who led the Ahmadzai Wazirs of South Waziristan, also fought against the state under the leadership of Nek Mohammad in 2004, who had raided Frontier Corps posts and disrupted operations between 2002-04, and his predecessors had aligned the Ahmadzai Wazirs with the TTP in late 2006. Nazir had also harbored and aligned with al-Qaeda foreign fighters even after they were declared an enemy Pakistan in 2002.171 Moreover, Nazir and Gul Bahadur's forces periodically clashed with state security forces and even launched suicide attacks on military positions in North Waziristan, suggesting, "the military's control over the two commanders was limited and that their cooperation was fragile and conditional." 172 While Nazir's Wazir group was mostly hostile to the Mehsuds, in part due to historical Wazir-Mehsud tribal rivalries, at times Nazir seemed to warm to the Mehsud-led anti-state TTP group like in the spring of 2009 alongside with Gul Bahadur, and in early 2012 towards Hakimullah Mehsud. 173 Despite clear evidence that Nazir and Gul Bahadur potentially threatened the GOP given their past history, harboring of Pakistan's enemies of Pakistan, and latent future threats, the military never launched a major offensive against them and treated them differently from the 169 170 Gopal et. al., 2013, p. 147; Mahsud, 2013, p. 169. Gopal et. al., 2013, p. 148. Meyerle, 2012, p. 24; Mazetti, 2013, pp. 104-05; Syed Saleem Shahzad, "Taliban and al-Qaeda: Friends in Arms," Asia Times Online, May 5, 2011; "Profile: The 'Good' Taliban Leader," Express Tribune, January 4, 2013. 172 Meyerle, 2012, pp. 37-38. 173 Meyerle, 2012, pp. 30, 37, 42; Mahsud, 2013, p. 169. 171 291 Mehsud-TTP faction. A militant outfit would "have to be pretty blatantly and consistently antigovernment for the Pakistanis to take a concerted effort, otherwise they could give it a pass." 174 No Real Effort/No Real Plan Overall, the absence of these inputs of effort led some to describe the strategy as a lack of a "coherent plan." The strategy seemed to be defined as much in what the state chose not to do. There was "no attempt to develop a comprehensive COIN policy," and "there was no policy to hold territory and exploit or conduct sustained operations." 75 Despite the rise in violence committed by FATA militants against the people and state of Pakistan, the military relied on a strategy much closer to counter-terrorism-what one Pakistani military officer described as "cost-effective"' 76 -because "it did not involve any effort to control the area fully" leave alone adopt any political or economic changes.1 77 With limited objectives in the fight in FATA through much of 2007, the scope of operations and effort would remain limited. The government believed that a combination of short military operations, negotiated ceasefires, propping up of local warlords to control areas, and financial rewards would be "more than adequate."1 78 Even as late as 2009, GOP officials described their strategy in the FATA as principally focused on political dialogue with tribes, socio-economic development, and the occasional use of force when necessary as a deterrent.' 7 9 Author interview with former U.S. intelligence official, June 14, 2013 (#129). Haider, 2014. 176 Author Interview with Pakistan Army official, July 11, 2013 (#133). 177 Haider, 2014. 178 Hamid Hussain, 2007; Weinbaum, 2009, p. 76. 179 "Editorial: 3Ds Dialogue, Development, Deterrence?" Daily Times, February 3, 2009. 14 175 292 3.3. Identity and Violence In general terms chapter 4 found that the Pakistani state generally employed low-violence strategies against Pashtun-based rebellions in due to the state's esteem and value of the group. Pashtuns held group worth in the eyes of the dominant Punjabis and were deeply embedded within state institutions, especially the military, dating back to the period of British rule of the subcontinent. My theory then expects that in the event of rebellion, the position of Pashtuns would constrain the state's use of violence and incent a greater role for non-kinetic approaches. While in FATA, the Pakistani state continually hesitated to use significant force, something it had no qualms with in most of its previous counterinsurgency campaigns (nor in a simultaneous campaign in Balochistan beginning in 2005). In Swat/Malakand, the state was forced to adapt to a more comprehensive but costly strategy to hold the area and protect the population (involving greater risk exposure for security forces). The reasons for both these choices can be traced back to Pashtun co-identity representation in the state, and particularly in the military, triggering mechanisms of empathy, strategic sensitivity, information, and trust. This section examines the role of my four mechanisms in shaping strategy against the Pashtuns in campaigns in FATA and Malakand. First, state elites felt a high degree of empathy for the Pashtun rebels grievances as well as Pashtun civilians liable to be killed in a high-violence campaign. Decision makers and troops from the Pashtun community felt particularly morally torn over the use of violence against their own kin and attributed insurgent actions to geopolitical circumstances or grievances rather than the inherent loyalty of the group. This made the state narrow the definition of an insurgent, selective in its targeting, willing to surrender and rehabilitate militants, and seriously consider political accommodations. Second, the heavy presence of Pashtuns within the military made the 293 - state vulnerable to backlash or defection should they become unhappy with state choices. Pashtuns not possessed a voice within the state to shape state policy and strategy but a compelling threat of exit. This propelled the state to be more judicious in its use of force and rely heavily on negotiations and accommodations. Third, with a strong contingent of Pashtuns within the military, the state had a strategic advantage of personnel who had special insights on the human terrain and socio-political structures as well as troops on the ground with situational awareness, cultural, and language skills. This enabled the state to access better strategic insights and judgments as to what actors were reconcilable or cooperative, better tactical information for more selective targeting, and more fruitful interactions with the population they sought to protect. Finally, the number of ties between military, civil servants, and insurgent leaders enabled channels of dialogue and trust. This facilitated the state's wider use of non-kinetic tactics to deplete the insurgency by addressing demand-side of rebellion with side deals, negotiations, public goods, and political accommodations. 3.3.1. Empathy First, empathy acted as a major restrainer on state strategy as Pashtuns, were truly perceived to be Pakistan's own loyal people. Unlike some other ethnic groups, Pakistan had, as one officer writes, "social, moral, economic and religious links with the people of FATA." 80 Pashtuns in the state institutions, who possessed a reasonably large voice in the state strategy, held greater empathy for those Pashtuns-both combatants and non-combatants-that might have been victims of broad-based military operations. Some of them sought to dissuade other state leaders that violence in FATA was just nationalist expressions of anger rather than direct Commander Adnan Mazher Khan (PN), "Tribal Areas - Need for Integration into National Mainstream Recommendations," PakistanNavy War College Review, 2005-06, pp. 63-68. 180 294 challenges to state authority. 18 Senior officers sought to make clear distinctions between the population and the insurgency, and expressed confidence that 80% of the population wanted nothing to do with the insurgency, fifteen percent could be won over and only five percent were "hard core" who had to be dealt with by force.18 2 The Punjabi majority both within the government institutions as well as amongst the broader public also expressed this empathy with the Pashtuns and opposition to heavy force.'s3 Surveys conducted during this period showed widespread (and majority) opposition to using the army against Taliban in Pakistan's tribal areas. 184 Officers were particularly hesitant to endorse or participate in such extensive military action, because they could identify with insurgents. One former senior police officer pointed out how difficult it was that, "this is a war with Muslims" 185 (despite the fact the state was brutally fighting Baloch Muslims without any qualms). But state and military officials felt they "should not be warring against our own people-our own kith and kin."18 6 One well-regarded analyst and former official explained how the social ties of Pashtun soldiers on the ground transmitted moral qualms up the chain of command and restrained the use of force: Although there is no such thing as a Pashtun grouping, their opinion certainly matters. ... When they go on leave, they interact with the people of their area and are influenced by local opinions. What happens in the NWFP and the tribal areas directly affects their views to which the military hierarchy is not oblivious. 87 181 182 Rashid, 2008, p. 277 - described of Jan Orakzai Paul McGeough, "Inside the Taliban's Heart of Darkness: The Violent Frontier," Sydney Morning Herald, May 22, 2007. 183 184 Commander Khan, PakistanNavy War College Review, 2005-06, p. 64. Fair and Jones, 2010; Cloughley, 2008; Ijaz Gilani, Voice of the People: Public Opinion in Pakistan, 2007-2008, Karachi: Oxford University Press, 2010. See also surveys conducted by the International Republican Institute (IRI) between 2006-2009. 185 Author Interview, former Pakistani police Inspector General, Lahore, Pakistan, October 2, 2011, (#19). 186 Author Interview, former Pakistani official, Lahore, Pakistan, October 2, 2011, (#20). 187 Abbas, "Musharraf Contends with the Pashtun Element in the Pakistani Army," 2006. 295 One senior general responsible for forces in the region admitted that on a few occasions, the central command prevented him from conducting operations fearing the potential blowback.188 This empathy also extended to the tribal regions. Recruitment of army officers from FATA, particularly Wazirs and Mehsuds tribes at the center of the insurgency had increased substantially (some estimate tripled) since the 1970s, particularly during the Afghan jihad. 189 A CIA analyst deployed in FATA during this time argued the problem was that the high degree of co-identity between the state forces and insurgents enabled the Taliban's propaganda to resonate within these institutions and demoralize soldiers.' 90 Levels of Violence This empathy found its manifestation early in puzzling levels of restraint and a bumbling set of operations by what was ordinarily considered a highly professional military. While the military certainly did conduct offensive operations, kill and capture militants, and cause some collateral damage with some clunky, "sporadic" military tactics like the use of air power,191 the army seemed to overly emphasize restraint and operations revolved around reaching negotiated micro-settlements in order to withdraw rather than further degradation or disruption of insurgent combat power. The policy of "sitzkrieg" was meant to avoid aggressive engagements and real on these deals, and at times, the army was accused of being restrained to a folly resulting in operational mistakes.1 92 Missions were also incredibly selective, many of them narrowly focused on killing or capturing a handful of individuals rather taking the fight to militant forces. The state believed " Author Interview, former Senior Pakistani Army Officer, Arlington, VA, Dec 2012 (#117). 2011, p. 4; Gul, 2011, pp. 36-37. 190 Keller, 2008. 189 Nawaz, 191 Haider, 2014. 192 Schofield, 2011, pp. 141-42; Nawaz, 2011, p. 10. 296 that foreign fighters who had infiltrated the region were the root cause of the problem and instability would be quelled if they were dealt with.1 93 Part of the reason state forces incurred high casualties and terrible exchange ratios was that the state was generally dispatching its less capable frontier corps rather than regular military assets that would have inflicted higher kills but also higher collateral damage.1 94 Civilian casualties were modest (though data is not readily available on this) partly due to restraint and partly due to the GOP forces unwillingness to fight.1 95 In fact, conscious attempts to avoid collateral damage actually exposed security force operations to serious risk and danger.1 96 Some of the early tactics involved leveraging the collective punishment capacity of the century old Frontier Crimes Regulation and "siege" to areas and using punitive measures like arresting tribesmen and destroying or impounding the assets of tribal elites to compel them to cooperate and give up foreigners and Al Qaeda suspects, or those harboring them.197 As much as these were criticized by outside observers, these were locally accepted mechanisms that generally minimized bloodshed so as to not trigger blood feuds. This frontier management repertoire of limited punishment also set up bargaining with tribal leaders to co-opt them and outsource the GOP's war.1 98 3.3.2. Vulnerability Second, early into the campaign, Pakistani state leaders became acutely aware that it was vulnerable to backlash within its institutions if it pushed its use of violence too far. Even when 193 Rasool, 2008 194 Schofield, 2011; Gul, 2011; Nawaz, 2011 2008. Schofield, 2011, p. 141. 197 "Tribesmen May Seek more Time," Dawn, October 11, 2003. 198 See chapter 3 of Marten, 2012, pp 31-63. 195 Keller, 196 297 the state did muster the stomach and deploy the forces to seriously combat the Taliban on occasion, they began to face some real internal strategic risks that alarmed them. The frontier corps troops who were nearly all Pashtuns and even regular army troops and some officers (one fifth of whom were estimated to be Pashtuns) protested or refused to participate in operations as they were reluctant to fight their own brethren. The military began to see a rising number of surrenders, desertions, suicides, discharge requests, early retirements, and court martials of midranking officers, all plummeting morale, triggering serious cohesion problems for the army, and bringing soldiers "near-mutiny".19 9 Pashtun troops were "surrendering en masse" to insurgents, 2 00 and fears arose that the visible indisipline and continued operations "could split the army, which was clearly unhappy at the prospect of fighting their own people," particularly at the behest of an outside agent.201 A state like Pakistan, which believed it faced a perpetual existential threat from its neighbor, could ill-afford morale and cohesion problems, leave alone the risk of an actual mutiny. One reason for continuing a strategy of mitigation in FATA was to prevent dissention within the Army's officer corps. President Musharraf had built his reputation as Army Chief of Staff on building bridges between Pashtun and Punjabi officers who dominated the military and might have been motivated by a desire to "keep his constituency happy and united." 2 02 If Pashtun influence ran from Peshawar to 'Pindi then a backlash but military personnel and officers would 199 Quotation from Shahzad, 2011: 43. See also Haider, 2014; Cloughley, 2008, p. 187; Abbas, "Musharraf Contends...,"2006; Hamid Hussain, 2007. 200 Lieven, 2014, p. 193. 201 Zahid Hussain, The Scorpion's Tail: The Relentless Rise of Islamic Militants in Pakistan-And How It Threatens America, New York: Free Press, 2010, p. 90. For more on this risk, also see Syed Saleem Shahzad, "Warning shots for Musharraf ," Asia Times Online, June 12, 2004; Syed Saleem Shahzad, "Musharrafs moment draws near," Asia Times Online, June 9, 2004. 202 Lieven, 2014, p. 196; Abbas, "Musharraf Contends..." 2006, references a statement made by Ahmed Rashid at a conference in Kabul in October of 2006. 298 not just threaten the tribal regions but could directly disrupt or harm the state's center of power. Musharraf then sought to pre-empt this backlash and "pacify his Pashtun generals." 203 For the most part, the Army sought to avoid significant offensives. Casualty data shows that in 2007, insurgent fatalities from attacks they initiated were double what the Army doled out in offensives. 2 04 Instead of a heavy doses of military operations, the state switched gears to emphasize governance reforms as early as 2004 and a "comprehensive approach" that emphasized development alongside security in 2006. One general argued that "patience, time, services and economic development are fundamental" to the strategy. 2 05 Of course, these were not well integrated nor effective plans given the state's general lack of effort. 206 3.3.3. Information A third reason for Pakistan's more restrained, selective approach was that Pashtun commanders with heavy responsibilities in KPK region and tribal areas asserted they had better information and special knowledge of tribal networks, and were given charge of the region. Most notable of these was Lt. Gen Jan Aurakzai, XI Corps Commander in Peshawar who was not only a Pashtun but specifically from the tribal areas rather than settled regions. Aurakzai was the commander responsible for FATA from 2001-04 and was later appointed governor of KP from 2006-08, was considered the architect of what others labeled an "appeasement" strategy. 2 07 He believed that in Pashtun regions, military operations would begat further militancy rather than 203 204 205 206 Harrison, 2007. This ratio of fatalities was 393 to 185. Based on data from the Pakistan Institute for Peace Studies (PIPS). McGeough, 2007. Markey, 2008, p. 12; Joshua T. White and Shuja Ali Malik, "Governance Reforms in Pakistan's Tribal Areas: The Long Road to Nowhere?" Peace Brief No. 135, United States Institute of Peace, October 15, 2012, p. 4. 207 Schofield, 2011, pp. 141-42, 151; International Crisis Group, Pakistan's TribalAreas, 2006; Khaled Ahmed, "Zahid Hussain's counter-narrative," Express Tribune, Dec 11, 2010; Zahid Hussain, "The Perils of Appeasement," The Scorpion's Tail, 2010, pp. 63-91. 299 stem it, 2 08 but through this approach, with numerous carrots and an occasional stick, the Pashtun tribes could be cultivated as strategic assets and effectively deputized to regulate violence in the area.209 Aurakzai had proven his loyalty to Musharraf years before and was a trusted confidant. He was warning Musharraf that significant military intervention could actually instigate a broader insurgency.2 0 President Musharraf took this seriously and frequently made public statements separating the greater Pashtun community from the Taliban in order to promote selective targeting by both his own officers as well as his new allies of the U.S. and NATO.m In addition to Aurakzai, the state's other key interlocutors, Maj. Gen. Hamid Khan commanding the Frontier Corps, Maj Gen. Niaz Mohammad Khattak, and KPK Gov. Lt. Gen. Iftikhar Hussain Shah were both Pashtuns who possessed a strategic understanding of the regional tribes and could communicate directly with them and their leaders in Pashto. Insiders suggest that their identity played a critical role allowing them to identify the right tribal leaders or maliks with whom they could coordinate the military's deployment and build support for their targeting of specific militants. This close relationship generated the intelligence necessary for discriminate and narrow targeting of foreign fighters and al Qaeda operatives, which the state effectively pursued during this period. Insurgents in the area soon wised up to the state had assets and information flows and they began to specifically target these tribal maliks. Hundreds of maliks were killed within a few years,2 13 and the government lost critical intelligence assets essential to its low-violence strategy of managing the region. At this point, the state began to heavily resort to cutting deals in the area 20 8 Ali Mohammad Jan Orakzai, "Situation in FATA: Causes, Consequences and the Way Forward," Policy Perspectives (published by Institute for Policy Studies), 6 (1), Jan-June, 2009. 209 Another Pashtun officer serving in a senior position in the region reaffirmed some of these calculations and conclusions. Author Interview, Dec. 19, 2012, (#117). Mazetti, 2013, p. 37. Schofield, 2011, p. 145. 212 Hamid Hussain, 2007; Gul, 2011, pp. 32-33. 2 13 Nawaz, FATA, 2009. 210 211 300 "to give time to the government to rebuild its local intelligence network."2 14 Even after retiring from the military, Aurakzai was appointed Governor of KPK and became the primary advocate of utilizing peace deals in North Waziristan. 2 11 3.3.4. Trust The final mechanism of trust, due to the identity of the rebel base, limited the state's use of violence and afforded it other non-kinetic avenues for demobilizing rebels and containing the insurgency. In a strategy involving violent negotiation, trust was a critical mechanism that helped the state bargain effectively with local tribes and even co-opt some militants to serve as proxies for the state in battling insurgents. The state still held confidence in the partial loyalty of insurgents to negotiate in good faith and cooperate with the state. Many of these deals fell apart after affording the state an important tactical pause, but some endured for years and helped Pakistani manage violence in the near to medium term, As part of its strategy of containment, the GOP cycled between limited engagements and hastily developed peace agreements-what many detractors labeled as appeasement. 2 16 Defence Secretary Rasool promoted the strategy of "co-opting influential tribal elders" on the condition that they "generally commit to provide stability in the area."217 The GOP engaged in deal making soon after recognizing its own vulnerabilities, reluctance to expend effort or force, and desire for a tactical pause. Though the GOP recognized the risks and tenuousness of these deals, one 214 211 Hamid Hussain, April 2007. Mazetti, 2013, p. 171 Rashid, 2008, p. 388; International Crisis Group, Pakistan's TribalAreas, 2006. On Pakistan military acknowledgement of this, see Schofield, 2011, pp. 151, 176-77. On Pakistan NSC acknowledgment of this, see Ismail Khan, "Talibanisation Imperils Security, NSC Warned: Immediate Action Urged," Dawn, June 22, 2007. 216 217 Rasool, 2008, p. 17. 301 former US official gathered they still believed they had sufficient information and leverage to manage the insurgents by activating other tribal militias against them. The military used these deals to deputize some local militant commanders-some the military had been fighting just days earlier-as the state's own warlords to maintain some stability in exchange for a truce. The military promised withdrawal to its cantonments and freedom of movement for warlords and militias on the condition that they cease activities against the state and police the area of foreign militants who were the root cause of the disturbances. 2 19 Simultaneously, the state tried to bolster the century-old political agent and malik (tribal leader) system of indirect governance under which FATA was believed to be "very peaceful." 220 At times this approach appeared to be fruitful. The GOP partnered with Mullah Nazir, a declared Taliban affiliate and known al-Qaeda supporter aligned with Hizbul-i-Islami Gulbuddin who emerged in South Waziristan in 2006.22 1The state enlisted Nazir to help crack down on foreign Uzbek fighters, who had been attacking the state and were harbored in Waziristan. Alongside local militias, embedded army commandos, and army intelligence and artillery, Nazir led the charge on behalf of the state and eliminated upwards of 250 Uzbek militants while the rest (estimated around 1,000-2,00022) relocated to Mehsud territory of South Waziristan. 2 In exchange, Nazir was essentially given a free pass, "allowed to run his own local government," and in this de facto alliance, functioned as the state's warlord in their broader cooptation strategy. 218 Author Interview with former US official, June 13, 2013. (#127) 219 See for instance, Rasool 2008; "Pashto-Language TV Station [Khyber TV] Interviews Pakistan's President Musharraf," 2006. 220 Schofield, 2011, p. 136. 221 Meyerle, 2012, p. 24. 222 Markey, 2008, p. 16. 223 Gul, 2011, pp. 55-59. 224 Haider, "Intelligent Intelligence," 2007. 302 Day to day enforcement of these deals was absent and the military would often turn a blind eye to violations of these informal deals, so long as they stayed within certain boundaries. Of course once a certain threshold was crossed, the military would return in force again to conduct some more limited operations and then seal them with another deal. Some of the deals held for an exceptionally long period of time. Both Nazir's group and Gul Bahadur for the most part maintained their pacts with the state. The military still had to come in and occasionally knock heads or conduct a brief strike but for the most part state forces deferred to these local militias. In the grand scheme of things, these deals may have been mistakes and backfired, as some within the Pakistani intelligence warned,2 25 but nevertheless the state's frequent reliance on these tools certainly signaled the state's trust that the Pashtun social base was not inherently disloyal. From 2002-07 survey data revealed the broader Pakistani public supported these deals. Finally, there was high support for negotiations and peace deals. Meanwhile a majority in Pakistan generally supported peace deals or negotiations with the Taliban.2 2 6 As chapter 4 indicates, in 2008, Pakistan began to change its strategy in some parts of FATA due the extreme intensity of the conflict and direct targeting by certain TTP factions, but once violence was brought down by 2011 with the dismantlement of the TNSM and TTP/Mehsud base in South Waziristan, the Pakistani public and many elected leaders resumed their long term support for negotiations and peace deals with the Islamist militants above military operations, something informed experts suggest extended to the Pakistani military's rank and file. 22 7 2 226 2 Mazetti, 2013, pp. 107-108. IRI Survey, May 11, 2009; Terror Free Tomorrow poll, June 2008; World Public Opinion, September 2007. Lieven, 2014, pp. 194-95. 303 3.4. Alternative Explanations: An Inchoate Taliban? One objection to this line of argument might be that the Taliban insurgency's inchoate state in Pakistan could explain the low levels of effort and violence. While the insurgent organization of the TTP certainly strengthened over time, there are a number of reasons why the argument does not. First, even when the Taliban was most cohesive and at the peak of its strength, the Pakistani state purposefully still limited its expansion of operations to one small region of South Waziristan to dismantle the part of the TTP that was conducting long-range attacks on Pakistan's urban centers. For the rest of FATA, the state continued its low-effort containment strategies. For instance, operations in 2011 involving narrow targeting were described as having "limited goals, limited gains" yielding tactical improvements but "[did] not seem to have been a significant contribution to the overall fight against militancy in the FATA." Second, the Pakistani state had never shied away from high-levels of violence just because a rebel organization was weak and inchoate. Pakistan's history with Balochistan involved continuous use of high-violence strategies to contain these mostly weak insurgencies. Even in 2005, the state repeated this by launching a short operation of bombing and artillery to degrade a very weak and poorly organized BLA force. Finally, the size of Taliban forces should have provided the GOP every reason to be alarmed by insurgency in the tribal areas very early on. Militants had been attacking frontier posts and GOP security forces, disrupting the balance of tribal power in FATA, and had even launched some significant attacks into urban areas including assassination attempts on President Musharraf. By early 2006, one leading analyst estimates the local Pakistani Taliban alone composed 35,000 fighters not counting the 3,000 Afghans and 2,000 foreign fighters based in 228 Jan and Worby, 2011. 304 Waziristan.229 Even if the core strength of "full-time" or "hardcore" fighters was an order of magnitude lower as some insurgency experts tend to estimate,23 0 and has been reported by other journalists, 23 1this was still a substantially sized insurgency. Additionally insurgents were not dormant but had conducted as many 254 attacks in 2005 and 675 in 2006, totaling over 1,000 deaths.23 2 Still, the state betrayed a curious lack of alarm at the situation and believed the insurgency "did not in itself threaten the existence of Pakistan." 233 This low threat perception was shared by the general public, which opposed early counterinsurgency operations in the tribal areas.23 4 IV. Restrained Leviathans: Counterinsurgency in Malakand II, 2007-13 4.1. Background The case of Malakand II, centered on the TNSM insurgency in the Swat Valley, offers a tight comparison to the FATA case. While insurgencies in both FATA and Malakand were composed of a militant Islamist Pashtun social base, the state acted more decisively and with greater urgency in Malakand because the real estate was more consequential. This manifested itself in more comprehensive efforts to both rid the region of the insurgency as well as rehabilitate this "settled" territory and the Pashtun popular base of the insurgency. The GOP's efforts in Malakand exhibited the same impulses of restraint as in the FATA campaigns, but The the TTP alliance in Dec 2007 declared it was 40,000 strong. See Shahzad, 2011, p. 29. This seems to have some support by other estimates like Javed Hussain, 2011. 229 230 David Kilcullen, The Accidental Guerrilla:FightingSmall Wars in the Midst of a Big One, New York: Oxford University Press, 2009, pp. 48-49. Zahid Hussain reports that in 2004, Baitullah Mehsud had about 4,000 armed men under his command. Hussain, The Scorpion's Tail, 2010, p. 76. 232 Reported in a PIPS 2008 report cited by Sartaj Aziz, 2010. 231 233 Lieven, 2014, p. 193. In surveys conducted in March-June of 2006, the plurality of respondents opposed military operations in Waziristan. See Gilani, 2010, p. 46. 234 305 instead of containment sought a more comprehensive and stable outcome offered by a strategy of population control. Around 2004, Maulana Fazlullah took over control of the TNSM from his father-in-law Sufi Mohammad, who had led the group in an insurgency in 1994 but had been imprisoned in 2001. Fazlullah gained a following in the Malakand region with his sermons broadcast over an illegal radio station and stoked increasing radicalization. Fazlullah's virulent propaganda led to some violent and overzealous enforcement of Sharia in Swat, the terrorizing and intimidation of locals, disruption of development measures like polio vaccinations and girls' education, and the attraction of foreign fighters into the region.235 Though the provincial government tried to cut deals with the TNSM, it escalated violent against state and civilian targets after the Lal Masjid incident in the summer of 2007.236 In November of 2007, the military launched its first major campaign in Swat and would cycle between these and peace deals until 2009 when it would escalate its efforts significantly to defeat the TNSM and decisively reclaim control of Malakand. 4.2. Territory and Effort Chapter 4 described the province of KPK's rise to significance by the end of the 1990s and the turn of the century. The substantially province rose in terms of its overall value to the Pakistani state ever since the economy had grown consistently, the physical and human terrain had been upgraded and become more "legible," and most importantly its ideational value had fully transformed. Government officials and elites talked of the region as "settled area," a distinction that gained greater prominence as the violence in FATA began to encroach on KPK's Christine Fair, "Pakistan Loses Swat to Local Taliban, Terrorism Focus, 4 (37), November 13, 2007; Afzal Khan, "Revolt in Pakistan's NWFP: A Profile of Maulana Fazlullah of Swat," TerrorismFocus, 4 (38), November 20, 2007. 236 Accounts drawn from Schofield 2011, p. 178; Gul, 2011, pp. 112-119; Khattak, 2010. 235 306 Malakand region and Swat Valley. When the TNSM, an extremist group with deep social roots and a history of violent activism in the region, launched an insurgency in this settled region, it deeply alarmed Pakistani state and elites. The state fumbled initially for a few months as nationwide political upheaval transitioned the country back from military dictatorship to civilian rule, but by 2008, the Pakistani state and military had embarked upon on an intensive and exhaustive counterinsurgency campaign that would be made evident in its 2009 Operation titled Rah-e-Rast (Righteous Path). Relying on the model of Pakistan's strategy in FATA, interested analysts and advisors to the United States were skeptical that the Pakistanis would seriously fight, endure, or succeed, but they would soon be proven wrong by a very different sort of "presence-oriented strategy" 238 The campaign against the TNSM in Malakand II, would expend tremendous blood and treasure to defeat the insurgency, regain control of the region, and rehabilitate it to the normalcy of a settled region. In contrast to FATA, the material stakes in Malakand were too high for simple containment and compelled Pakistan to employ a high-effort strategy. At the same time, because the social identity base of the rebels was Pashtun, and specifically settled Pashtuns, this restrained the state from unleashing the type of brute force it had in Sindh and Bengal. Instead, the state organically developed a strategy of population control involving significant restraint, accommodations, and non-violent substitutes. The state deployed significant material, political, and organizational resources to Malakand and the Swat Valley, sustained them for an indefinite period, and adapted to new roles in the region. David Kilcullen in particular voiced deep skepticism of what Pakistan could and would accomplish. See James Blitz, James Lamont, and Farhan Bokhari, "Swat Outlook 'Pretty Bleak' for Pakistan," FinancialTimes, May 13, 2009; "Pakistan and the Taliban: A Real Offensive or a Phoney War?" The Economist, April 30, 2009. 237 238 Mullick, "Lions and Jackals," 2009. 307 ' 4.2.1. Material Investments First, far more troops were deployed for active operations in the Malakand region than anywhere in FATA. At its peak, this amounted to 52,000 troops deployed at the peak of its operations in Swat. Although the TNSM forces had grown to between 5,000-10,000 insurgents this augmentation of forces was not a function of matching troop ratios. Taliban forces in FATA were believed to have grown from 20,000 to 40,000 around the same period but except for in one pocket of South Waziristan, the same force ratio was not mirrored here.239 The significant escalation in troops arrived in April 2009 when the TNSM broke the peace agreement it had signed with the government, began expanding into a neighboring province of Buner, and officials feared they might threatens strikes on the capital of Islamabad or be able to sever lines of communication by gaining control of part of the Karakoram highway.2 40 Second, rather than exclusively relying on less capable frontier corps troops, the state dispatched some of its best units to Swat. In addition to divisions from the 1 It Corps in Peshawar and the Frontier Corps, which had conducted the majority of previous operations against the Taliban, the Pakistan Army moved in units from elite divisions in critical areas on its Eastern border including units from its I Corps, a strike corps, its 30 Corps based in Gujranwala less than 50 miles from the Indian border, and units from the l0t Corps used to secure the Line of Control with India.2 4 2 39 Mullick, Pakistan'sSecurity Paradox, 2009 estimates 6,000; Mullick, 2010 estimates 10,000; Khattak, 2010 estimates 5-6,000; and Hussain and Rosenberg estimate 6,000-8,000. See Zahid Hussain and Matthew Rosenberg, "Pakistani Peace Deal Gives new Clout to Taliban Rebels," Wall Street Journal,April 14, 2009. When adding up the figures of total militants killed in Swat, Dir, Malakand, and Buner between 2007-12 (-4587 based on PIPS data) plus the number in prison, in rehabilitation centers or release from rehab programs (-3,500; see Gandhi, 2012) and acknowledging that many militants may have escaped, the figure of close to 10,000 seems more plausible. See Rohit Gandhi, "Deradicalization: A Tall Order," Asia Pacific Defense Forum, July 1, 2012. 240 Author Interviews including Pakistan security scholar, Islamabad, Pakistan, October 11, 2011 (#30). 241 Brian Cloughley, "Insurrection, Terrorism and the Pakistan Army," Pakistan'sQuagmire: Security, Strategy, and the Future of the Islamic-nuclearNation, Ed. Usama Butt and N. Elahi, New York: Continuum International Publishing, 2010, p. 108; Petraeus, 2009. 308 Third, because of what was at stake, the state invested incredibly high levels of financial resources into defeating the TNSM, regaining control of Malakand and the Swat Valley, and resettling and rehabilitating the area. The state incurred such high costs that it sought a special round of support from donors to defray the estimated the campaign's $2.5 billion price tag for a country whose entire military budget in 2008-09 was $3.6 billion. 242 4.2.2. Organizational Investments Besides improvements in the deployment of material assets, the Pakistani military sought to change its methods of force employment. It began to make some departures from organizational pathologies that typically plague militaries in counterinsurgency and sought to adapt its doctrine, training, and tactics. At the highest level, a new doctrine was constructed in 2008 that began to be implemented in 2009. Junior officers, who accumulated experience with counterinsurgency in recent years, offered advice that was incorporated into the planning and decision making loop. 243 At the ground level, unlike during the campaign in FATA, army units were sent to one of three newly established Battle Inoculation Training Centers for pre-induction training modules for a period of 4-12 weeks to improve their tactical and situational awareness skills and retrain for a counterinsurgency rather than conventional fight before being sent into battle.24 4 Meanwhile, in other parts of KPK, the state was proactively building up and retraining an elite police force to inoculate the rest of the region from further Taliban infiltration. 245 Woodward, 2010, p. 287; Briefing slides from the Pakistan Army's Inter-Service Public Relaitons (ISPR). Mullick, "Lions and Jackals," 2009; Mullick, Pakistan'sSecurity Paradox,2009; Mullick, 2010. 244 Haider, 2014;; Mullick, 2012, p. 97; Nawaz, 2011, p. 19; Shuja Nawaz, "The Pakistan Army and Its Role in 242 243 FATA," CTC Sentinel, 2 (1), January 2009, pp. 19-21; Author interview with Pakistan Army official, July 11, 2013 (#133). 245 Lieven, 2011, p. 454. 309 Pakistan's military also made a bold organizational and political decision to allow U.S. and British trainers to be sent to Northwest Pakistan to help train and upgrade Pakistan's Frontier Corps troops for counterinsurgency. While portrayed by some as a no-brainer, this was not a costless transaction for the military. Both the Americans and the British were extremely unpopular at the time amongst the Pakistani public and within the ranks of the military. The wider discovery of this training could hurt the state's standing amongst the population and galvanize greater support for insurgent forces, and force it could compromise the morale of the military's rank and file. Pakistan was already fending off the impression that it was acting at the behest of the United States, which had also hampered support for its campaigns, and so to actively embrace training by that country exposed the state to high-risks.2 4 6 Additionally, few professional military organizations are pleased to cede their prerogative to other state organs but the Pakistan Army allowed in foreign military trainers because the threat to what it most valued exceeded its organizational interests. Eventually, the Army would also allow in U.S. psychological operations trainers to help the Pakistan military effectively engaged in counter-propaganda. 247 4.2.3. Sustainment In sharp contrast to its operations in FATA that had featured rapid withdrawals and an outsourcing of control to local warlords, Pakistan's military sustained its force presence in the Malakand region to control it years after it dislodged insurgents. In July 2011 when the military officially declared an end to operations, it continued to maintain two divisions in the region even The fact that this program had to be shut down after the Raymond Davis episode in February 2011 and the Bin Laden raid in May 2011 demonstrates the state felt exposed. 247 Author Interviews with officials involved in the training program, April 2013. 246 310 while it had rebuilt a police force to replace it in security duties, and the military forces remained into 2014 with the promise of a brigade cantonment to be permanently stationed.2 4 8 The state's willingness to sustain the operation was also demonstrated in the losses it sustained. While offensive operations did not incur significant losses, two years of defensive operations required the military to absorb very high costs. 249 The Army chief of staff reported that after two years, Army casualties in Swat/Malakand region reached 985 including an officer to troop ratio of 1:8, which he claimed was one of the highest in history. 250 4.3. Identity and Violence Similar to FATA, the Pashtun identity base of the TNSM rendered the GOP much more cautious and hesitant about high violence in Malakand, 25 1 and it initially interspersed peace deals (in May 2007, May 21, 2008, and April 2009), with quick, short military operations and the arming of local lashkarsspecifically to "reduce civilian casualties."252 Soon after realizing this strategy would not be sufficient and the TNSM threat was expending in KPK, it initiated a massive military operation. The Mechanisms of restraint played as much a role in Malakand as they did in FATA, if not more so, but in this case, the campaign was more comprehensive. Instead of managing militants and containing violence, the state employed a strategy of population control to decisively defeat the insurgency and restore its full political authority to the region. "Pakistani PM Announces Permanent Base in Swat," RFE/RL, January 16, 2014. One expert writes, "Pakistan measures its costs not only in the movement of troops and loss of public support for its actions, but also in the deaths of.. .soldiers." See Nawaz, "The Pakistan Army and its Role in FATA," 2009. 250 Inter-Services Public Relations (Pakistan), "COAS Key Note Address - National Seminar Deradicalization at Mingora Swat," ISPR No. PR157/201 1, July 6,2011. 251 Schofield, 2011, p. 181; Gul, 2011, p. 118. 2 52 Nawaz, , "The Pakistan Army and its Role in FATA," 2009 quotes Gen. Janjua who commanded the operation. 248 249 311 4.3.1. Empathy Minimizing CollateralDamage Even as it expanded efforts in Malakand, the empathy the Pakistan army maintained for the Pashtun base of the TNSM compelled it to take costly actions to minimize harm to civilians in Malakand. A number of former servicemen had joined the TNSM organization since its founding, rendering the state strategically sensitive to using excessive force against the group. 2 53 The most visible demonstrations of empathy and restraint began in its new operations in 2009. The GOP employed a highly unconventional and unprecedented method to minimize civilian casualties while dislodging against insurgents, who had dug-in to fortified defenses in semiurban centers and throughout the valley. In May 2009, the state began to rapidly evacuate 2.5 million people out in the area and provide temporary camps and allowances for about 25% of them. Though widely criticized early on, this move turned out to be quite successful and different from other population displacements Pakistan had created and neglected in Balochistan and FATA. In Malakand, the state would actively support the repatriation of residents, and within a few months, 90% of the residents of Swat had moved back to their homes.25 4 By contrast, the state had no qualms about displacing 85% of the Baloch population during 2005 operations in Dera Bugti region where there had been little effort to resettle displaced Baloch civilians and even proactive state actions to block relief efforts. 25 5 The population relocation move had a surprising amount of support. One retired officer explained that many people in Swat were willing participants in their own displacement. "The people of Swat and Buner were begging for the government intervention to expel the militants," Hassan Abbas, "The Black Turbaned Brigade: The Rise of the TNSM in Pakistan," Terrorism Monitor, 4 (23), November 30, 2006. 254 Mullick, 2010 reports 1.8 million people by September; Jan, "Trickling Home," 2010 says 90% by July, 2009. 255 International Crisis Group, Pakistan: The Forgotten Conflict in Balochistan, 2007; "Government was Blocking Access to Balochistan IDPs," Dawn, January 17, 2007. 253 312 he stated. "People of Swat even offered to migrate and vacate their homes to enable the military to single-mindedly root out the militant evil once and for all... Most of them who have suffered at the hands of the militants have accepted their dislocation as a price for getting rid of militancy."256 Public support, discriminate use of force, significant and rapid repatriation efforts, and paid compensation and reconstruction for homes damaged from operations indicate this population displacement in Swat demonstrated significant effort as well as restraint.257 SustainedProtection A sustained Army presence served not only to target militants but also to protect the population. During this campaign the ratio of defensive to offensive engagements was 2:1 indicating the military was much more focused on protecting locals than hunting insurgents. Army forces dispersed units, conducted regular patrols, manned numerous checkpoints, and fully inculcated the lesson that "helping civilians is as important as fighting and killing insurgents." 258 Two years after the initial offensive, half the army forces remained 25 9 and continued to man roughly 40 checkpoints throughout the valley and a permanent cantonment was built years later to sustain some forces indefinitely. 260 Sustainment of the military presence provided one public good of security but also provided the conditions for other public goods. Whereas the absence of a durable security 256 Jeremy R. Hammond "The Situation in Swat: An Interview with Shahid R. Siddiqi," ForeignPolicy Journal, July 9, 2009. Jan, "Paradise Regained," 2010; Jan, "Trickling Home," 2010; McKelvey, 2011; Omar Warraich, "Refugees Head Home After Army Scatters the Taliban," The Independent, Aug 3, 2009. 25 2 59 Nawaz, 2011, p. 10. Ahmed Rashid, Pakistanon the Brink: The Future ofAmerica, Pakistan, and Afghanistan, New York: Viking, 2012, p. 144. There were still two Army divisions in place through 2011 along with paramilitary and police forces. Author Interview, Pakistan analyst, Wahsington, DC, September 19, 2012 (#63). 260 "Pakistan Army Strategy in Question After Attacks," Dawn, October 22, 2012. 257 313 presence had compromised most of the international development efforts in FATA,2 6 ' the sustained Army presence in Malakand created confidence and sufficient local security for local NGOs backed by international donors to operate safely and rebuild infrastructure like schools.2 6 2 One NGO head described the continued army presence as "a blessing." 263 Because insecurity had prevented judges from adjudicating cases in Swat, sustained force presence also facilitated the conditions for improvements in the justice system, which had been a core grievance of the TNSM and the people of Swat. 2 64 Security also underpinned the general resumption of normal life. Marketplaces were crowded, schools for girls were reopened, and the staples of a once thriving economy--agriculture, tourism, fashion, and the cosmetics-were resurrected. 2 65 2 66 Even as the military began to transfer authority back to the civilian government in certain districts and towns, it pledged to remain in place to supervise and assist. 2 67 Public Goods In addition to facilitating NGO operations, the Pakistan state and military actively participated in the provision of public goods. Immediately after operations were completed, the Army partnered with civilian commissioners to work with their Engineering Corps and rapidly provide services such as electricity, water, and sanitation. 2 68 By the spring of 2010, it had spent 515 million Rs to complete 435 projects and was slated to spend an additional 420 million Rs on Author Interview with former USAID official based in Pakistan, July 12, 2013 (#135). Some argue that a stable security environment is a critical prerequisite to effective development efforts in Pakistan. Its absence in FATA has undermined aid to the region and development programs. See Hassan Abbas and Shehzad H. Qazi, "Rebellion, Development and Security in Pakistan's Tribal Areas, CTC Sentinel, June 25, 2013. 262 Rina Saeed Khan, "Swat Valley NGO Finds a Solution to the Assault on Education," The Guardian,May 10, 2011. 263 Emmanuel Duparcq, "Swat Rebuilds Year After Pakistan Floods," Agence FrancePresse, July 24, 2011. 264 Fazal Khaliq, "Reclaiming Swat: Influential Peace Jirga Member Escapes Attack," Express Tribune, Nov 14, 2011. 265 Author Interview with a Pakistan analyst, June 17, 2013. (#130) 266 "Pakistan Army Uses Bullets and Classrooms to Fight Militancy," Reuters, April 25, 2012. 267 Delwar Jan, "Malakand Division on Road to Civilian Control," The News International,Aug 8, 2012. 268 Mullick, 2012, p. 98 261 314 another 995 projects. Additionally another 50 billion Rs (or 2% of the federal budget) was allocated for the rehabilitation of IDPs from Malakand. 2 69 The military also supported infrastructure investments in a Swat Expressway, the construction of new bridges, the reopening of an airport, (all of which helped to reinvigorate Swat's tourism industry), the creation of vocational centers for women, and employment creation through special recruitment of Malakand youths into the army.2 70 4.3.2. Information With a heavy presence of Pashtuns within its ranks and the trust of the local population, the military possessed the capacity to set up local structures and leverage information for a more selective violence and targeted operations against insurgents. All it needed was to invest in formal and informal local structures to more effectively interface with the population. The state's rapid repatriation of the population and the Army's sustainment of ground forces allowed a "presence-oriented strategy" to take shape and engage with local actors to secure the region and the population. 2 71 The Pashtun population of Swat were highly appreciate and cooperative with state forces. One Pakistani analyst pointed out, "One of the reasons Swat worked was because the people who lived there felt they were equally involved in the operations."27 2 Despite what might have been seen as an intrusion in other parts, most of the Swat inhabitants wanted the 269 Ahmad Hassan, "Operation in Malakand, Waziristan is No Solution," Dawn, Nov 8, 2009. "Countering Terror: 'Militants Are trying to Regroup in Swat,"' Express Tribune, Jan 16, 2013; Fazal Khaliq, "New Beginnings: After Six Years of Closure, Saidu Sharif Airport Ready to Become Operational," Express Tribune, July 12, 2012. 270 271 Mullick, "Lions and Jackals," 2009. 272 McKelvey, 2011. 315 army to remain in place as they perceived them to be delivering security and actively cooperated with them.2 73 First, Army forces themselves functioned for a time as policing units, manning checkpoints, patrolling the area, and even setting up a low grade version of population tracking through colored lanyards and identification cards that had to be worn at all times. Second, the government focused on maintaining the gains they had won by bolstering informal local policing and information collection to serve as the eyes and ears of the state rather than relying on blunt force or indiscriminate cordons. In addition, the security forces coordinated with locals and village councils, developed and utilized informant networks to monitor and prevent TNSM and Taliban infiltration, armed and coordinated about 30,000 lashkars(militias) to also police the area, and empowered peacefirgas and their village defense committees to maintain security in the area and prevent re-infiltration from Afghanistan through Upper Dir.2 74 Malakand's involvement of local militias and peace jirgasdeparted from the more aloof cooptation efforts in FATA. In this case, the military led the operations and used militias as an auxiliary force rather than deferring to them. The state also worked to organize and network the militias to improve their effectiveness and more aggressively backstopped them throughout the campaign. 275 Finally the GOP rebuilt and expanded a more capable civilian police force with stronger local ties. The previous Swat police force had only been roughly 1700 strong and had been depleted and stripped of its morale resulting in about 40% deserting and another 30% killed or 27 3Khattak, 2010, p. 11; Author interview with Pakistan analysts, Washington DC, September 19, 2012 (#5) and 18, 2013 (#4). April Jan, "Trickling Home," 2010; Mullick, "Lions and Jackals," 2009; Mullick, 2010; Inter-Services Public Relations (Pakistan), 2011; "Pakistan Army Strategy in Question After Attacks," Dawn, Oct. 22, 2012; Author interview with a Pakistan analyst, September 19, 2012 (#5).; Fazal Khaliq, "Peace Maintainable Only Through Public's Commitment," Express Tribune, Dec 29, 2012. 275 Ejaz Haider, "Time for Plan B," Daily Times, April 21, 2009. 316 refusing to leave their homes. 27 6 Addressing this problem began with rebuilding and strengthening the entirety of the KPK police force. Manpower in the KPK force, salaries and benefits, and procurement of new equipment and better weapons were all dramatically increased between 2007 and 2010, almost doubling the size of the police force (78.5%). New special units were also created including an elite force of 7,500 that trained with the army to better confront militants, and a second 7,000-strong special police (recruited more directly from the local population), quick response forces, and bomb disposal units. Additionally, the KPK government hired as many as 2500-4,000 retired Pashtun ex-military service members from the region to serve in the KPK police force. 277 855 of these were specifically for Swat police force, which was expanded to 3,200 plus an additional 500 person Levies Force. 27 8 4.3.3. Trust The state' trust in the Pashtun population of Malakand had allowed it to accept a number of peace deals between 2007-09, but even as they collapsed and the state was forced to escalate kinetic force, it still maintained confidence that the population was mostly loyal. This assurance, stemming from the high degree of Pashtun embeddedness within the state, allowed the Pakistani government to continue attempting other non-kinetic substitutes to violence including the public goods it provided, surrender and rehabilitation programs for former insurgents, the active persuasion of the population, and even some devolution of political authority. Even prior to the outbreak of the insurgency, the government recognized the instability in the neighboring FATA might spillover and sought to reach out to the TNSM with two rounds of Lalwani, 2009, p. 38. Hasan Asad Alvi, "Police and Counterterrorism in Khyber Pakhtunkhwa," Stabilizing PakistanThrough Police Reform, Report by the Independent Commission on Pakistan Police Reform, Ed. Hassan Abbas, New York: Asia Society, July 2012, pp. 104-112. 278 Delwar Jan, 2012; Alvi, 2012; Author Interview with Pakistan analyst, Washington DC, January 30, 2012, (#60). 276 27 317 agreements in 2006 and 2007, believing they could help contribute to the stability in the region. After the outbreak of an unforeseen insurgency, alongside military operations, the Pakistani state still made three attempts to negotiate with TNSM leaders Mullah Fazlullah and Sufi Muhammad as they had some ties to the ISI and were considered manageable, semi-loyal, and even potential allies to the state in the past.2 7 9 Rehabilitation In the aftermath of the 2009 offensives, Pakistani security officials-believing that many of the Pashtun TNSM insurgents were still of moral character and loyal at heart-initiated a program to de-radicalize and reintegrate former TNSM and Taliban members back into society. Unlike in Balochistan or even parts of South Waziristan, here the state actively tried to separate hardcore TNSM/Taliban foot soldiers from "accidental guerillas"-drawn in through tribal networks, coercion, grievances, or the lure of social prestige or financial incentives-who could be demobilized through non-kinetic means and even rehabilitated. The military departed from its previous counterinsurgency efforts with an innovative approach of "socio-psychological" deradicalization to reconcile with several thousand insurgent recruits and ex-combatants, many of whom had been trained for suicide bombing missions. The Army treated these "misguided youth" as "miscreants" who were not hardened militants but impressionable youth that needed to be shown the correct path.280 Roughly half of the TNSM fighters captured during operations in Swat were given a choice of standing trial for treason or 279 Lieven, 2011, p. 465; Shahzad, 2011, p. 171; International Crisis Group, Pakistan: CounteringMilitancy in PA TA, 2013. COAS General Kiyani specifically used this term to described killed and captured combatants in Swat, See COAS Keynote Address, ISPR, PRI 57, July 6, 2011; DIG of Police Muhammad Azeem, "The Softer Option," Dawn, March 25, 2013; Dr Fawad Kaiser, "Swat Deradicalisation Project - II," Daily Times, February 3, 2014. 280 318 enrolling in a deradicalization center where they would also learn vocational training. The Army's two rehabilitation centers employed psychologists, anthropologists, educators, and clerics to re-educate these former Taliban and teach them basic vocational skills.2 82 By the end of 2012, over a thousand had been rehabilitated, released from these institutions, and successfully reintegrated back into Pakistani society.283 As a preventative measure, one program began to reach out to another 5,000 youth in Swat Valley vulnerable to the predations of the Taliban and al Qaeda.2 8 4 Dissuasion While rehabilitation demobilized former militants, dissuasion was a critical component of inoculating Malakand from future insurgent uprisings. In addition to continuing to forcefully shut down illegal radio stations propagating extremism and militancy, 285 the Army sent personnel to be trained in psychological operations and learn how to effectively counter the propaganda of the Taliban and TNSM forces trying to sway the public in Malakand. With the help of advisors, Pakistani units conducted information operations by setting up their own competing local Pashto radio stations through both Army and civilian channels, to challenge and The numbers are difficult to estimate but there are believed to be 850-1000 who have been released and reintegrated, "hundreds" currently enrolled, and 2,000 still imprisoned of whom about 700 are eligible for deradicalization. Assuming the total figure is something like 3,500, it seems that roughly 2,000 will undergo this process. For more detail, see Rohit Gandhi, "Deradicalization: A Tall Order," 2012. 282 Muhammad Amir Rana, "Swat Deradicalization Model..." Conflict and Peace Studies (PIPS journal), 4 (2), April-June, 2011, pp. 5-12; Dina Temple-Raston, "Pakistan's Ambitious Program to Re-Educate Militants," NPR, April 1, 2013; Taha Siddiqui, "Deradicalizing Boys in Pakistan, ChristianScience Monitor, May 25, 2012; Cathy Scott-Clark and Adrian Levy, "How to Defuse a Human Bomb: Rescuing the Taliban's Teenage Recruits," The Guardian,October 15, 2010. 283 Fazal Khaliq, "Much Changed for Better Last Year, More Still to Come: GOC," Express Tribune, January 18, 2013. 284 Kristen Seymour, "De-radicalisation: Psychologists' War Against Militants," Express Tribune, July 17, 2011. 285 "Pakistani Authorities Close Three Radio Stations for Pro-Taliban Broadcasts," Associated Press, November 1, 2009. 281 319 discredit the radio sermons of extremist Mullahs radio and promoting the virtues of civic service. Internal surveys suggested this began to have an effect on public opinion.2 86 Devolution and PoliticalAuthority: JudicialReforms In tandem with military operations, the state sought to address the core grievance that insurgents had leveraged to mobilize popular support: justice. The GOP undertook a number of actions towards judicial reform to improve the performance and speed of the system ofjustice in Swat. 2 87 The people of Swat and Malakand had been frustrated by the dysfunctional and extortionist legal and judicial institutions in the region since the early 1990s, and the TNSM had built its following on this issue. First, the government quickly instituted initiatives to speed up the legal system by hiring more judges, establishing new courts including a circuit bench in the urban center of Mingora, creating mobile courts to more rapidly adjudicate minor cases and disputes, and imposing duration limits on cases. All of these moves led to a reduction in the backlog of cases from 18,000 down to 2,300.288 Second, the GOP satiated public demand for the implementation of the Nizam-e-Adl Regulation (Order of Justice System), which it had controversially approved back in 2009 prior to military operations by establishing an appellate court run on the principles of the Sharia to expedite cases. While detractors criticized the move for creating a parallel justice system, supporters-including the Supreme Court Chief Justice Iftikhar Muhammad Chaudhrypraised it both for co-opting the issue from the TNSM and for the effectiveness in dispensing 286 McKelvey, 2011; Tayyeb Afridi, "Pakistani Province Reaches Out Through Radio," KUT News, July 8, 2011. Ajaz Maher, "Pakistan's Army Steps Up Radio Wars," BBC, August 14, 2012; Author Interview with U.S. advisor involved in the program, April 2013 (#120). 287 Osama Siddique, Pakistan'sExperience with FormalLaw: An Alien Justice, New York: Cambridge University Press, 2013, pp. 385-391. Reza Sayah, " Can the Return of Justic Halt the Taliban?" CNN, Nov 14, 2009; "Judicial Reform in Swat," Dawn, March 25, 2010; Rania Abouzeid, "Taliban Gone, Pakistan Area Still Wants Islamic Justice," Time, April 25, 2010; Fazal Khaliq, "Speedy Justice: Mobile Courts for Swat Announced," Express Tribune, Nov. 26, 2011. 288 320 justice in a record number of cases, while in accordance with the constitution.289 Finally, the government also embraced auxiliary reconciliation committees as a third form of dispute settlement to ease the burden on police and the formal judicial system and these have shown promise as well with a rapid and high rate of resolution. 29 0 Mutual Trust Pashtuns continued to maintain trust in the state as well. Though hesitant at first, after some time, local Pashtuns began to overwhelmingly support military operations and presence in the Malakand/Swat region, and survey data found that even tribal Pashtuns of FATA supported the military's endeavors. 29 1Critics of the campaign even acknowledged that it had secured the confidence of the people of Swat. 2 92 This trust of the Pashtun population would provide the essential support for all the states tactical substitutes for violence including the civilian evacuation and repatriation, security presence, reconstruction and reform efforts. In Swat, the Pakistani state appeared to depart from past approaches in FATA, and outside observers took notice. President Obama remarked that Pakistan had "ramped up their [counter-terrorism] cooperation in a way that is significant" 293 while General David Petraeus noted how the Pakistani campaign had "carried out really quite impressive counter-insurgency operations" mirroring U.S. doctrine, stating, "it's not just clear and leave, it's clear hold, build, 289 Fazal Khaliq, "Nizam-e-Adl Regulation: Top Sharia Court Set Up in Swat," Express Tribune, January 19, 2011; Fazal Khaliq, "'Terrorists Behind Swat Unrest to Face Justice,"' Express Tribune, Nov 27, 2011. 290 Fazal Khaliq, "Timely Justice: Community Body Gives Quick Solutions for Small Disputes," March 13, 2013. Express Tribune, Fair, "Islamist Militancy in Pakistan," 2009. Also, see also polling and survey data from Pakistan Gallup, IRI, Terror Free Tomorrow, World Public Opinion, and Pew Global Attitudes data after May 2009. For instance, over 70% of co-ethnic Pashtuns in the tribal areas of FATA (who are generally less supportive of the state) continued to support military offensives in the Malakand region through the summer of 2010 (while only 16% opposed it). See Naveed Ahmad Shinwari, UnderstandingFATA 2011: Attitudes Towards Governance, Religion & Society in Pakistan'sFederally Administered TribalAreas, Vol. V, Islamabad: Community Appraisal and Motivation Program, 291 2012. 2 92 Khattak, who'se work seems quite critical acknowledges this, Khattak, 2010, p. 11. 293 Woodward, 2010, p. 369. 321 and even transition." 294 One noted jouralist who had criticized the Pakistani military in 2009 for its "failure to prepare for counterinsurgency warfare,"2 95 lauded its reinvigorated strategy in Malakand by 2012: The war in Swat was the first, and so far the only, time the Pakistan Army successfully completed a counterinsurgency campaign according to the book: the militants were killed, captured, or driven out, the area was secured, the displaced population returned, their homes were rebuilt, and the civic administration was revived. The army had finally learned the principles of "clear, hold, build, and transfer"-the mantra of General Petraeus's counterinsurgency strategy in Iraq and Afghanistan-and could carry them out when it had the will to do so. 2 96 4.4. Alternative Explanations: U.S. Pressure? One alternative explanation for Pakistan's escalation of effort in Malakand was that US support and pressure had helped guide it toward the ideal model of counterinsurgency. This seems to be contradicted by much of the empirical evidence leading up to and after the initiation of the Malakand campaign. First, accounts supporting this claim tend to rest on GOP civilian leaders publicly acceding to US demands when in fact the military was in charge. President Zardari may have promised US leaders that aid would allow for a change in Pakistani strategy but he had no real control of the strategic decisions. 2 97 Second, while General-turned-President Musharraf proved to be relatively pliant wit US demands, at least between 2002-04 when the Americans had substantial leverage in the early years of Pakistan's campaign, the Army chief of Staff from 2007-13, General Kayani, was did not easily succumb to pressure.29 8 As a result those the U.S. pressure/support explanation cannot account for the stubborn refusal to accede to U.S. demands to expand targeting to "good Petraeus, 2010. 2009. Rashid, 2012, pp. 143-44 297 Woodward, 2010; Rashid, 2012. 298 Mazetti, 2013. 294 295 Rashid, 296 322 Taliban" and expand operations to North Waziristan, both of which went unheeded between 2009-2014. Third, there are a number of scholars and practitioners who argue that public U.S. pressure actually postponed operations in Malakand because the GOP could not afford to be seen acting at the behest of the United States, which was wildly unpopular with the Pakistani general public.2 9 9 In fact, the state was able to embark on a decisive and comprehensive campaign in Malakand specifically because the public had voiced strong support for the campaign, with analysts explaining that the overwhelming support was because insurgency in settled areas was simply intolerable. 30 0 V. Conclusion This chapter has provided detailed evidence from Pakistan's campaigns against the East Pakistan, FATA, and Malakand insurgencies validating the mechanisms of the core-periphery theory, specifically the incentivizing territorial mechanisms of costs and incentives as well as the restraining identity mechanisms of empathy, vulnerability, information, and trust. The more finegrained evaluation of tactical and operational decisions and actions during the counterinsurgency campaigns allows me to demonstrate some of the unique contributions of my theory, even if alternative theories offer some purchase on the case. East Pakistan and Malakand, both campaigns which contested highly value territory, present opportunities to observe detailed evidence of high effort, such as costly deployments, organizational changes, sustained commitment, and continuous adaptations to improve the prospects for success. By contrast, FATA presents a case where one would have expected to also A number of Pakistani scholars and former officials expressed this sentiment including one former diplomat. Author Interview, Islamabad, Pakistan, October 23, 2011 (#52). 300 Lalwani, 2013; Author Interviews in Islamabad 2011 and Washington DC, 2013. 299 323 observe high effort given other conditions-a fairly strong Taliban insurgency and a Pakistani state capacity bolstered by international support-but instead we observe distinctly weak efforts made to only contain the insurgency. Ultimately, the campaign in Swat/Malakand was more comprehensive because the TNSM insurgency threatened more valuable real estate in these "settled" areas, pre-existing state capacity (and expectations) were higher, and physical and social infrastructure only needed to be rebuilt rather than built from scratch. This was not the case for Balochistan or FATA. A significant amount of scholarship has been devoted to detailing the incredible violence and brutality the Pakistani state perpetrated in its counterinsurgency campaign in East Pakistan, but I demonstrate how identity distance, ethnic contempt, and exclusion from state institutions can specifically structure the state's choice of extreme violence. In this case, the Pakistani state felt no empathy or vulnerability to morally or strategically restrain it, possessed no useful sources of information to target violence more selectively against hardcore insurgents, and neither held nor commanded trust with the rebel base to enable alternative methods for demobilizing insurgents or bargaining for an end to the conflict. The state had few options then but to pursue a brutal, high violence attrition strategy The FATA and Malakand cases allow me to examine whether ethnic Pashtuns position in Pakistani society and the state institutions effectively constrained the Pakistani government's strategic choice of violence, even after more than a decade of war. Once again, the identity mechanisms help me to explain how restraint stemmed from the rebel's social base rather than structural, normative, or strategic factors, and also why a state capable of such brutality against Bengalis was hamstrung against very dangerous and capable insurgencies in the Pashtun tribal and settled areas of Pakistan's northwest. 324 The broader case studies also help to review competing alternative explanations for the cases. Its difficult to deny that the brutality in East Pakistan was the product of authoritarian tendencies, but this seems to have more as much to do with ethnicity and a prejudiced theory of victory. Furthermore, it is unclear whether civilians would have done much different had they been in charge since Bhutto executed a very similar campaign three years later. The charge of genocide too is less convincing of an explanation given the state seemed to really believe ethnic cleansing by driving out Hindus would effectively dismantle the insurgency. Pakistan's relatively sluggish and nonchalant efforts in FATA seem to be less motivated by an inchoate TTP since the insurgent numbers indicate a formidable threat. Instead, what concerned the GOP was what the TTP would do with their insurgent capacity. When they began targeting core territory, Pakistan took decisive action, but only against the TTP/Mehsud group capable of power projection, while it continued to accept containment as a viable strategy against other TTP-affiliates and fellow travelers. Hardly the model of quiescent state militant assets, these various militant groups still targeted state authority and security forces, but in a wild frontier where the GOP was constantly managing violence, keeping things to a low boil was the best Pakistani leaders believed they could hope for. Finally, this theory specifically helps shed light on a puzzle policymakers grappled with over the past decade-the baffling behavior of the Pakistani state. The United States escalated its efforts and troop commitments in Afghanistan, in part with the expectation that Pakistan would change its calculus, deploy a control strategy, and monopolize violence in the FATA region,30 1 and failed to comprehend the Pakistani state's minimalist and half-hearted approach in the tribal regions. This misunderstanding of Pakistani strategy stemmed from the assumption that violence and effort would be correlated and the unfamiliarity with containment strategies frequently used 301 Woodward, 2010, p. 187-88. 325 by Pakistan (and indeed many other weak states) throughout its history, given the limited value they placed on full sovereignty in undervalued, peripheral regions. Though certainly other factors were at work including the tumultuous U.S.-Pakistan relationship, this empirical analysis offers a more robust picture of what may have been structuring Pakistani strategic incentives and choices during this period. 326 Chapter 6 A Medium-N Analysis India's Counterinsurgency Campaigns I. Introduction Like many post-colonial states, India has encountered its fair share of conflict since its independence in 1947 but in the ten years between 1985-1995, despite near macro-structural conditions, its responses to these different challenges varied tremendously. In one case, it marshaled its military resources in a campaign of "ruthless mailed-fist repression"' with "brutality... institutionalized by the security forces" 2 while in another case it deployed significant assets to target combatants in "surgical" 3 operations and "protect" the population from insurgent violence.4 In one of the conflicts, the Indian state fought a "dirty war"5 relying on shadowy tactics and extra-judicial killings and still in others, despite deploying a militarized response, the Indian state seemed to neglect as "out of sight, out of mind,"6 with "liberal," and "ostrich-like" approaches 7 that appeared to tolerate instability and insurgent activity below a certain threshold. While the insurgencies themselves were not cut copies of each other, most had strikingly similar properties as ethnic separatist insurgencies with roughly 4,000-6,000 armed 1Sumantra Bose, Kashmir: Roots of Conflict, Paths to Peace, Cambridge, MA: Harvard University Press, 2003, p. 111. 2 Robert G. Wirsing, India, Pakistan, and the Kashmir Dispute: On Regional Conflict and its Resolution, New York: St. Marten's Press, 1994, p. 158. 3 Arjun Chowdhury and Ronald R. Krebs, "Making and Mobilizing Moderates: Rhetorical Strategy, Political Networks, and Counterterrorism," Security Studies, 18 (3), 2009 p. 391. 4 KPS Gill quoted in Rahul Chandan, KPS Gill: The ParamountCop, Noida: Maple Press, 2013, pp. 153-54; 17072. 5 Andrew Buncombe, "Victims of India's Dirty War in Manipur Poised for Justice," The Independent, August 5, 2013; Bruce Rich, To Uphold the World: The Message of Ashoka and Kautilyafor the 21st Century, New Delhi: Viking, 2008, p. 137. 6 For instance, one international journalistic account of Assam was titled "Northeast Erupts, But New Delhi is Not Panicking," (by Praful Bidwai, Inter-PressService, January 27, 1994). The opening paragraph begins ""Out of sight, out of mind" seems to be the Indian government's blinkered policy on the northeast, a simmering cauldron of endemic strife and militancy." The same phrase used by Chakravarti regarding the Naxal insurgency. See Sudeep Chakravarti, Red Sun: Travels Through Naxalite Country, New Delhi: Viking, 2008. 7 Former DG of the CRPF describes the vacillating policies in Andhra Pradesh and of its neighbors. Trinath Mishra, Barrelof the Gun: The Maoist ChallengeandIndian Democracy, New Delhi: Sheriden Book Company, 2007, pp. 112-116. 327 insurgents.8 Undoubtedly, the Indian state leviathan responded to all these violent challenges to its authority with force of arms, but in some cases it exhibited greater reluctance to mobilize its strength and in others, greater restraint in the use of force.9 With incumbent and insurgent properties largely held constant, theories of civil war dynamics are at a loss to explain this variation in strategy. The core-periphery theory laid out in chapter 3 can offer some purchase in explaining these puzzles. 1.1. Literature on Indian Conflict There is a large and growing body of literature on Indian civil conflict but studies of state strategy are scarce or unsystematic. Incumbent responses to the vast number of Indian rebellions dating back prior to the British Raj10 and particularly the 1857 "Great Rebellion" has been the subject of deep historical inquiry but other than the Kashmir conflict, primarily studied through the lens of inter-state relations, India's modern civil conflicts are largely neglected. Recent work on Indian internal conflict has begun to explain patterns of conflict onset, spread, and 8PWG in Andhra estimated at 5,000 in 1988 (Clodfelter, 2008, p. 641; "Naxalism: Y2K Problem Sans Solution?" The Hindustan Times, January 29, 2000 [Accessed on Factiva on July 16, 2014]); ULFA estimated at 5,000 (Sanjoy Hazarika, "Insurgents in India's Northeast Call for Peace Talks," New York Times, February 20, 1991); Sikh insurgents were estimated at 6,000 at their peak (Hamish Telford, "Counter-Insurgency in India: Observations from Punjab and Kashmir," The Journalof Conflict Studies, 21 (1), 2001); Kashmiri insurgents were estimated at 5,0006,000 (Lt. Col. Iqbal Singh Singha, "Combating LIC in J&K," Pratividrohi, 1995, p. 34; Roger Howard, "Evolving Rather than Receding, the Killing in Kashmir Continues," Jane's IntelligenceReview, 11 (1), Jan 1, 1999, p. 39); Manipur estimated at between 4,000-10,000 by two sources each (Kaushik Roy, The Armed Forces ofIndependent India, 1947-2006, New Delhi: Manohar Publishers, 2010; Lawrence E. Cline, "The Insurgency Environment in Northeast India," Small Wars & Insurgencies, 17 (2), 2006, pp. 126-147; Bibhu Prasad Routray, "Manipur: The State Abdicates," South Asia Intelligence Review, 6 (43), May 5, 2008; Prakash Singh, India'sNortheast: The Frontierin Ferment, Hurlbert Field, FL: Joint Special Operations University, 2008, pp. 38-39; Wasbir Hussain, "Manipur: Impending Collapse of Governance," South Asia Intelligence Review, 3 (4), August 9, 2004); Nagaland is reported to have 5,000 insurgents between the NSCN-K and NSCN-IM. (See Cline, 2006; Prakash Katoch (Lt Gen), "Why Maoist Insurgency Will Flourish in India," Indian Defence Review, June 13, 2013). 9 Two impartial scholars estimating data on state-caused civilian casualties estimate an order of magnitude difference between the Punjab and Kashmir campaigns. See Valentino and Uldfelder, 2008. My own analysis of casualty data draws similar conclusions. 10 Lilach Gilady and Joseph MacKay, "Bringing the Insurgents Back In: Early Wars in British India," Terrorismand PoliticalViolence, 26, 2014, pp. 1-2 1. 328 reduction," insurgent mobilization and fragmentation, and patterns and responses to communal violence. 13 (Though casualties hardly serve as the sole basis for weighing political violence, it is noteworthy that the level of casualties generated in India's insurgencies far outstrips the casualty levels of India's communal violence since 195014 and potentially even communal violence from the subcontinent's partition between 1946-48.15) While offering significant advances in the study of civil conflict and India, few of these even pay much attention to state responses to rebellion. A recent wave of Indian civil conflict scholarship has surveyed the state's counterinsurgency responses, but while it has richly described the variation of these state strategic responses, it has largely neglected to explain it.1 6 Many of these studies tend to default to and coalesce around one of two explanations: 1) Each case of rebellion and the state approach " Bethany Lacina, "How Governments Shape the Risk of Civil Violence: India's Federal Reorganization, 1950-56," American Journal ofPoliticalScience, 58 (3), July 2014, pp. 720-738; Shivaji Mukherjee, ""Econometric analysis of spatial variation in Maoist insurgency in India, " Paper Presented at American Political Science Association Annual Conference, Chicago, IL, August 2013; Devesh Kapur, Kishore Gawande, and Shanker Satyanath, "Renewable Resource Shocks and Conflict in India's Maoist Belt," Center for Global Development Working Paper #302, August 2012; Joseph Flavian Gomes, "The Political Economy of the Maoist Conflict in India: An Empirical Analysis" Unpublished Paper, 2011; Ajay Verghese and Emmanuel Teitelbaum, "Colonialism and Armed Conflict in the Indian Countryside," Working Paper, SSRN, July 18, 2014; Gaurav Khanna and Laura Zimmerman data, "Guns and Butter? Fighting Violence with the Promise of Development," Working Paper, SSRN, April 1 2014. 12 Staniland, 2014. 13 Ashutosh Varshney, Ethnic Conflict and Civic Life: Hindus and Muslims in India, New Haven: Yale University Peess, 2002; Steven I. Wilkinson, Votes and Violence: Electoral Competition and Ethnic Riots in India, New York: Cambridge University Press, 2004; Saumitra Jha and and Steven Wilkinson, Does Combat Experience Foster Organizational Skill? Evidence from Ethnic Cleansing during the Partition of South Asia," American Political Science Review, 106 (4), November 2012, pp. 883-907. 14 Wilkinson estimates 10,000 (2004, p. 12) since 1950. 15 Wilkinson estimates 200,000 (2004, p. 13) but a quick back of the envelope calculation shows Indian insurgencies eclipse this between 1948-2013. 16 Sumit Ganguly and David P. Fidler, India and Counterinsurgency: Lessons Learned, New York: Routledge, 2009; E.N. Rammohan, CounteringInsurgencies in India:An Insider's View, New Delhi: United Service Institution of India, 2011; Ved Marwah, India in Turmoil: Jammu & Kashmir, the Northeast andLeft Extremism, New Delhi: Rupa, 2009; Kaushik Roy, The Armed Forces of Independent India, 1947-2006, New York: Manohar, 2010; Sankaran Kalyanraman, "The Indian Way in Counterinsurgency," Democraciesand Small Wars, Ed. Efraim Inbar, Portland: Frank Cass and Company, 2003, pp. 85-100; Durga Madhab (John) Mitra, UnderstandingIndian Insurgencies:Implicationsfor CounterinsurgencyOperations in the Third World, Carlisle, PA: Strategic Studies Institute, February 2007; Anil A. Athale, CounterInsurgency & Questfor Peace: A Comparative Study, New Delhi: Vij Books, 2012; Vivek Chadha, Low Intensity Conflicts in India: An Analysis, New Delhi: Sage, 2005; Cline, 2006; Prakash Singh, 2008; Anit Mukherjee, "India's Experience with Insurgency and Counterinsurgency," The Routledge HandbookofAsian Security Studies, Ed. Sumit Ganguly, Andrew Scobell and Joseph Liow, New York: Routledge, 2009. 329 to counter it is sui generis, having its "own peculiarity," 17 often due to mysterious factors like "leadership," or sometimes "political will," 18 or 2) Due to its macro-structural features, Indian state responses have displayed a certain "consistency" 9 involving "a particular mix of coercive and accommodation-oriented strategies." 2 0 This second approach is divided into two camps with one emphasizing India's brutal and repressive approach to counterinsurgency 21 and the second emphasizing India's "deliberate restraint,"2 2 professionalism, and democratic "self-correcting mechanisms."2 Nevertheless, all these approaches discount meaningful variation in Indian strategy across different campaigns. The literature on Indian counterinsurgency, which advances macro explanations like strategic culture, regime type, and state strength, or micro-explanations like leadership remains "maddeningly incomplete." 24 More importantly, much of this literature suffers from the lack precise measures, a tight comparative approach, or serious attention to research design to isolate the factors that may be driving variation in strategies. In this chapter, I intend to further demonstrate that factors of territory and identity-that operate between the micro-level individual choices of India's leaders and the macro-level structures of the Indian state-shape strategy. Chadha 2005, p. 17. In my numerous interviews with Indian practitioners and analysts, these are typically the two factors emphasized in explaining variation. 19 Rajesh Rajagopalan, "Insurgency and Counterinsurgency," India Seminar, 599, July 2009. 20 Bidisha Biswas, Managing Conflicts in India: Policiesof Coercion andAccommodation, Lanham, MD: Lexington Books, 2014, p. 13. 21 For instance, Nandani Sundar emphasizes this aspect in much of her work on the Maoists, Kashmir, and the Northeast. 17 18 22 23 24 Kalyanraman, 2003, p. 97. Anit Mukherjee, 2009, p. 140. Borrowing a phrase from Jackson, 2008. 330 Table 6.1: List of 14 Indian Counterinsurgency Campaigns Date Start 1948 Date End 1948 Inclusion in other Datasets Meets Fearon Lyall /Laitin Lyall (10OObd) ID # IND-01 Insurgent Razkars led by Qasim Rizvi War Hyderabad IND-02 IND-03 IND-04 CPI Nagas/Federals NSCN - IM, NSCN-K Telangana Naga I Rebellion Naga II 1948 1955 1976 1951 1975 ongoing IND-05 IND-06 Mizo, MNF Naxalite (CPI-ML) 1966 1967 1986 1971 IND-07* Naxalite (CPI-ML) 1967 1971 IND-08 1982 ongoing IND-09 PLA, UNLF, KCP, KNF, PREPAK Sikh Militant Groups -Panthic Committees I, II & III - KCF, KLF, BTKF, BKI, etc Mizo Revolt India v. Naxalite I -all/Andhra Pradesh India v. Naxalite I - West Bengal Manipur Insurgency India-Sikh Insurgency 1983 1993 IND-10 PWG Naxal II - Andhra 1987 2004? IND-Il JKLF, HM, HuA, LeT, JeM. And other Kashmiri groups Kashmir 1990 2003 1 IND-12 ULFA Assam Insurgency 1990 ongoing ~I IND-13 CPI-M Naxal III 2004 ongoing IND-14* JKLF, HM, HuA, LeT, JeM. And other Kashmiri groups Kashmir 2004 ongoing 1 ~ 1 I Proof of battle death criteria Sunderlal Report, Noorani Book; Sundarayya 1 9 SATP data PRIO 'I 1 ~ MHA data; SATP data 9 9 1 1 SATP data (Sahni article) SATP data MHA data; SATP data 4 MHA data; SATP data 9 SATP data 1 split into two distinct observations the Naxalite I case (due to different strategies by region/state) and the Kashmir Case (sue to an ostensible change in strategy around 2004). 1.2. Recap of the Theory and Empirical Strategy Instead of fighting hard and violently against all insurgencies, the core-periphery theory expects that in the event of a rebellion, state incumbents will employ minimalist, low effort strategies like enfeeblement or mitigation to contain insurgencies threatening peripheral territory with few incentives and high costs, unless rebels manage to project violence on core territory. Furthermore, it expects affective and strategic motives will restrain state incumbents to low violence strategies like mitigation or population control if the rebel identity group is perceived to have high moral worth or is highly embedded in state institutions. To this end, India provides a valuable opportunity to test the core-periphery theory because of its tremendous sub-national variation while controlling for confounding factors. The Indian state has presided over an extremely diverse and contentious ethnic and political landscape carved out of British India in 1947. It has fought 14 counterinsurgency campaigns2 against a multiplicity of ethnic, social, caste, and religious groups throughout the country, and at certain points in time it has prosecuted multiple counterinsurgency campaigns simultaneously allowing for unique tests and structured, tightly controlled comparisons. The dynamics of India's different civil conflicts also vary, potentially offering greater generalizability. Some of counterinsurgency campaigns have been quite short (Hyderabad, Telangana, Naxalite I) or firmly resolved (Punjab, Mizoram) while others last for decades with no end in sight (Nagaland, Naxalites, Manipur). In some campaigns, the Army has been the lead counterinsurgent force (Kashmir) while in others it has been the local police (Punjab) or paramilitary (Naxals, Northeast cases). Money appears to be no object in some cases as resources are poured in (Punjab, Kashmir) while other campaigns appear starved of them (Northeast, These campaigns are identified with PRIO criteria but include a number of cases missed by the Armed Conflict Dataset that are defined by Pakistani academic and government sources as counterinsurgency campaigns. 25 332 Naxalites). And in some counterinsurgencies, the state has offered significant political accommodations (Punjab) while in others, comparatively little (Kashmir). India's structural conditions have been largely constant since independence. India has been a weak but democratic state (by most standardized measures), fielded a very large and capable professional military dominated by civilians, and fought four major wars with Pakistan, its external threat and rival since its birth. These features allow this medium-N analysis to hold constant country-level variables like regime type, state strength, strength relative to insurgents, and threat environment, and examine how the core-periphery relations theory fares in explaining sub-national variation. Later, even these controls can be relaxed to compare periods of relative difference in India's structural conditions. Figure 6.1 Timeline of Indian Counterinsurgency Campaigns Naxal III Assam Kashmir I & I Naxal 11 - Andhra Manipur Naxal I- AP &WB Mizoram Nagaland I&I T na Uderabad 1945 1955 1965 1975 1985 (Source: Start dates based on PRIO/UCDP data) 333 1995 2005 2015 1.3. Roadmap The argument made in this chapter seeks first to convincingly establish a relationship between patterns of counterinsurgency strategy and the hypothesized independent variables of territory and identity. The second section draws on qualitative and quantitative data to identify the specific values of the theory's independent variables-the importance of various territories and identity sub-groups-across India's regions and decades since independence in order to generate predictions. The third section identifies the strategy employed in each of the 14 334 counterinsurgency campaigns26 using quantitative and qualitative measures of the dependent variable as well as descriptive accounts of the effort, violence, tactics, and political authority. Finally, the fourth section evaluates the performance of the core-periphery theory in explaining variation in strategy across different campaigns and weighs it against alternative explanations, and the fifth section concludes. The following chapter will then pick up on the arguments made here and proceed to more closely examine certain Indian cases for evidence of hypothesized mechanisms linking territory to effort and identity to violence. II. Measuring the Independent Variables: The Value of Territory and Identity 2.1. The Incentives and Operational Costs of India's Regions India's disparate geography and uneven development have produced regions of varying material, strategic, and ideational worth and imposed varying costs on the state's reach and access. India's conventional history chronicles events of the Indo-Gangetic plain, the Deccan plateau, and some coastlands, but neglects much of its other territories. Historically, India's northern half, the Hindi heartland, has been the site of its wealth and political power (Delhi, Haryana, Punjab, and Uttar Pradesh 2 8) and increasingly its Western half has been the site of Though often included in the analysis of Northeast insurgencies, I exclude the case of Tripura because it largely appears to be about intragroup violence rather than meeting the definition of an insurgency and posing a challenge to the state itself. This is in accord with other scholars. See Biswas, 2014. However, some accounts label this an insurgency even though the target of their violence is other sub-groups and not the state See A.S. Abraham, "Optimism on the North-East Some Signs of Pacification," Times ofIndia, March 29, 1985, p. 8 27B.G. Verghese, India's Northeast Resurgent: Ethnicity, Insurgency, Governance, Development, Delhi: Konark, 1996, p. 280 28 Candice Gianetti, Gaye Facer, Andy McCord, Smita Patel, and Vikram Singh, "North Central India," India, Ed. Diane Mehta, New York: Fodors LLC, 2005, p. 122. It is striking how central Uttar Pradesh is (likely due to its sheer population size) to Indian politics. Not only has it been the home state and political base for the Gandhi family and Congress Party but it was also the base for a number of non-Congress Prime Ministers including Charan Singh (Janata), V.P. Singh (Janata Dal), Chandra Shekhar (Janata Dal), and Vajpayee (BJP). See also Jyoti Malhotra, "The 26 Road to Delhi Runs through Uttar Pradesh," Foreign Policy's South Asia Channel, May 6, 2014. 335 Indian economic growth (Maharashtra, Gujarat, Karnataka, Tamil Nadu). 29 Of the regions contested by rebellions, the territories of considerably greater material, strategic, and ideational importance were Punjab and Kashmir, while those considered of less or little importance were most regions in the Northeast and the patchwork sections of central-east India, referred to as the "Red Corridor" or "Naxal belt."30 In order to identify whether a particular territory was core or periphery, I drew on qualitative and quantitative data to systematically measure the incentives and costs of projecting force in a particular region. To estimate the relative material importance of each region, historical data was collected to measure material incentives based on production (GDP) and wealth (GDP/capita). Strategic value was primarily estimated based on expert qualitative judgments of the region, and ideational value was based on content analysis/frequency of mentions in Indian news headlines. To measure the relative operational costs of projecting force on to each region, I collected data on the distance from the capital, physical terrain or infrastructure (road and electrification density), and human terrain or social infrastructure (human development index scores). Since states will actively invest in the physical and social infrastructure of regions that are materially or strategically valuable, thereby reducing future operational costs, these provide further indicators of the territory's importance. Based on these measures, Punjab, Kashmir, and West Bengal are clearly core territories, while the Northeast and the central Indian "Naxal belt" or "red corridor" are distinctly peripheral territories. Some of the quantitative measures do not adequately capture the importance of territory and subsequent qualitative descriptions of these regions based on expert and practitioner Sanghamitra Bandyopadhyay "Why is Economic Growth Across Indian States Uneven?" London School of Economics Blog February 18, 2013. 30 Jonathan Kennedy and Lawrence King, "Adivasis, Maoists and Insurgency in the Central Indian Tribal Belt," EuropeanJournalofSociology, 54 (1), April 2013, pp. 1-32; Chakravarti, 2008, pp. 11, 309. 29 336 ) accounts help substantiate and confirm the initial measures. While a small population and production capacity limit Jammu & Kashmir's material value, it possessed great ideational value for Indian leaders and strategic value for defense planners. In fact, both Kashmir and Punjab were the main grounds for the Army's ability to launch counter-offensives against India's chief rival, Pakistan, should the need arise.3 ' On the other hand, the Naxal belt's low value and limited accessibility are understated by state-level data and would be even more pronounced if calculated with district-level data to isolate the "Naxal Belt" as sub-regions within the large states of central India (Andhra Pradesh, Bihar, Chhattisgarh, Jharkhand, Maharashtra, Orissa, and West Bengal32 ). Figure 6.3: Economic Wealth by State >~1965 N_--) 1995 ,p (J ~ ~~~1 47 -4, Indian states: Rich (pink), Poor (blue) Source: Sanghamitra Bandyopadhyay, Convergence clubs in incomes across Indian states: Is there evidence of a neighbours' effect?, Economics Letters, Volume 116, Issue 3, September 2012, Pages 565-570 V. Krishna Rao, In the Service of the Nation: Reminiscences, New Delhi: Viking, 2001, p. 475. Some eastern districts of Uttar Pradesh are sometimes included and prior to 2000, the regions of Chhattisgarh and Jharkhand were part of Madhya Pradesh and Bihar respectively. 31K. 337 Table 6.2: Cost/Incentives and Predictions of Territory Costs* Incentives* Cost-Benefit/ Expected Return Strategic Ideational/ Psych. Distance Physical Human high high high moderate moderate moderate 3 (High) High Telangana high high moderate moderate moderate moderate 2 (High) High Nagaland I low moderate low high high moderate -4 (Very Low) Low Mizoram Mizoram low moderate low high high low -3 (Low) Low Andhra Naxal I-AP high low moderate moderate moderate moderate 0 (Low) Low West Bengal Nagalim Manipur Naxal IWB Nagaland II Manipur high high high moderate high moderate 2 (High) High low low low low low low high high moderate high moderate low -4 (Very Low) -4 (Very Low) Low Low Punjab Andhra Pradesh Assam Jammu & Kashmir Central Punjab Naxal IIAP Assam Kashmir I high high high low high moderate low moderate low moderate low moderate 6 (Very High) 0 (Low) High Low low high Moderate high moderate low -2 (Low) 2 (High) Low High low moderate moderate low moderate/ high moderate high moderate Naxal III low moderate / low low high -3 (Low) Low Kashmir II moderate high high low moderate moderate 3 (High) High State/ Region Conflict Material Hyderabad Hyderabad Telangana Nagalim Predicted Effort (greater Nagaland) Pradesh India (7 most affected Naxal states) Jammu & Kashmir *Value assigned to each incentive/cost based on the following: high=2, moderate=1, low=O; Cost-benefit therefore ranges from -6 to 6 00 Figure 6.4a: Economic Assets by District: % Households with TV, Computer, Phone, and Vehicle (darker), 2011 les than 7% hoshod 6'14-21% hw,.dd E Mom th.,21%ho,.,d,,Id Source: Avinash Celestine, "We Are the 5%" Data Stories, Data Stories, April 1, 2013 (http://datastories.in/blog/2013/04/01/indias-5-percent/) (Based on 2011 census data) Figure 6.4b: Economic Assets by District: % Households with No Assets (darkeer), 2011 Les thon 16% households 16-32% households 3248% household& More th. 48% households Source: Avinash Celestine, "What Would a Poverty Map of India Look Like?" DataStories, April 9, 2013 (http://datastories.in/blog/2013/04/09/what-would-a-poverty-map-of-india-look-like/) (Based on 2011 census data) 339 2.1.1. The North East The Indian Northeast is a region that generally poses few incentives but plenty of operational costs. Surrounded by Bhutan, Burma, China, and what is now Bangladesh (formerly East Pakistan), the Northeast is barely connected to the Indian mainland by the narrow Siliguri Corridor also known as the "chicken's neck." One journalist has described the Northeast as "the periphery of the periphery" and if proposes that if one "...Read[s] 'Indian' history as it is taught... you will scarcely know the Northeast exists." 33 The region has been described and treated as a "wilderness,"34 a colonial "frontier," "emotionally" remote,3 6 the part of India that can be most easily broken away or cut off,3 7 and this peripheral treatment has only deepened even as India's economy has grown.3 8 This peripheral treatment has deep historical-institutional roots. While the British annexed this territory, they designated much of it as the North East Frontier Agency and accorded it "almost total autonomy as the process of subjugation was too expensive" and "isolate[d] them from the plains politics." 39 The region remained "either un-administered or lightly administered" 40 because, in James Scott's words, it was "illegible space" replete with slash and burn agriculture, mobile populations, and wars over people and not territory. 3 Verghese, 34 1996, p. 280. Nirmal Nibedon, Nagaland, The Night of the Guerillas,New Delhi: Lancer Publishers, 1978, p. 53. 3 Sanjib Baruah, DurableDisorder: Understandingthe PoliticsofNortheast India, New Delhi: Oxford University Press, 2005, pp. 5, 24-25, 40, 78. 36 Marwah, 2009, p. 162. 37 Author Interview with an Indian national security analyst, New Delhi, India, Dec 20, 2013 (#152). An example of this dysfunction and poor planning is the fact that the entire Northeast can be effectively cut off from the mainland during the monsoon season (4 months out of the year) in the all-too-frequent event of a landslide blocking the one narrow highway that connects the two regions. Anecdote shared by another Indian scholar, New Delhi, India, November 15, 2012 (#74). 31 Marwah, 2009: 8 39 Quote from page 56 in Brig JR Mukherjee, "Perspective of the North East," Pratividrohi,7 (2), September 1995, pp. 52-60. Baruah, Durable Disorder,2005, pp. 8, 39. 340 This colonial treatment has rendered the northeast "psychologically too distant," 41 and the disinterest of political elites 42 and an "out of sight, out of mind" mentality shapes Indian policy towards the region.43 One anecdote underscores the degree of neglect. The only state map of Nagaland submitted to the Home Ministry in 1979 had apparently gone missing with no copies could be found. The disappearance of such a critical document reveals how casually the region was treated--"From an official point of view it might as well not exist." 44 The Northeast is so absent from the national imagination that one civil servant contends most Indians could not identify the individual states on a map. While the national anthem mentions all of India's major rivers, Assam's major river-the Brahmaputra-is conspicuously left out. 45 An editorial in a local Manipur paper lamented the "prolonged neglect" and "the step motherly attitude of the centre" arguing, "those in the helm of affairs in Delhi... have never taken the region very seriously." 46 Limited governance in the region is then both a contributor and byproduct of poor economic development, generating few material incentives. 47 Today, the Northeast remains neglected with underdeveloped economies suffering from high unemployment, poor Patricia Mukhim, "Where is this North-east?" India InternationalCentre Quarterly, 32 (2/3), Winter 2005, p. 187. Baruah, Durable Disorder,2005, p. 51. 43 Praful Bidwai, "Northeast Erupts, But New Delhi is Not Panicking," Inter Press Service, January 27, 1994, Factiva accessed 3/5/14 44 Jonathan Glancey, Nagaland:A Journey to India's Forgotten Frontier,London: Faber & Faber, 2011, p. 243. 45 This was made clear by a number of people in interviews including a retired senior Indian civil servant, New Delhi, November 23, 2012, (#83), and an Indian scholar, New Delhi, India, November 15, 2012 (#74). Quoted in N. Somorendro Singh, "Integration and Development in North-east India: An Assessment," Indian JournalofPoliticalScience, 67 (2), April-June 2006, p. 338. 47 An interactive graphic was posted by The Economist likening every Indian state to its nearest "country equivalent" in terms of GDP and GDP/capita, revealing the sorry state of the Northeast. In GDP, the Northeast states' closest equivalents were Bolivia, Zimbabwe, Belize, Liberia, Somalia, and Central African Republic while in GDP/capita they were closest to Tajikistan, Laos, Mauritania, Lesotho, Sudan, and Cambodia. By contrast, some richer states like Punjab, Haryana, and Maharashtra were closer to Tunisia, Serbia, and Singapore in GDP and Fiji, Armenia, and Sri Lanka in GDP/capita. See "An Indian Summary: Comparing Indian States and Territories with Countries" Economist, 2011. 41 42 341 industrialization, infrastructure, or human development,48 poverty rates above the national average, 49 and despite decades of development assistance and dependence on the center, significant debt liability and non-integration with the "mainland."5 0 This limited development and integration underpins a steady state of "durable disorder." 1 Even the state of Assam, which "harbours the most potential for economic development in the region",5 was treated as a "hinterland" under British rule, and this status solidified under the post-colonial Indian state. 53 Despite its neighbors, the Northeast is not seen as strategically critical. In the partition of the sub-continent, two founding fathers of the Indian state, Nehru and Sardar Patel, were believed to have so little regard for the region that they were willing to cede it to Pakistan.54 Even after it was absorbed into India, some leaders' public statements revealed they placed little strategic value on the Northeast. 5 Though borderlands, particularly those bordering great or rival powers, are often treated as strategic territory, historically this was treated as one early intelligence director described it as inactive and "dead frontier." 56 India's western border with Pakistan has always held far more strategic value for the survival of the country. 48 Anil Bhat, Assam: Terrorism and the DemographicChallenge, New Delhi: Centre for Land Warfare Studies, 2009, p. 2; Krishna Chaitanya Vadlamannati, "Why Indian Men Rebel? Explaining Armed Rebellion in the Northeastern States of India, 1970-2007," JournalofPeace Research, 48 (5), 2011, p. 607; Marwah, 2009, p. 162. 49 N. Somorendro Singh, "Integration and Development in North-East India: An Assessment," Indian Journal of Political Science, 67 (2), April-June 2006, p. 334. (Note the poverty figure is for all Northeast states except Mizoram). 50 Singh, "Integration and Development," 2006, p. 334; Baruah, DurableDisorder,2005, p. 42; Marwah, 2009, 174; KV Krishna Rao, 2001, pp. 378-379. 51 Baruah, Durable Disorder,2005, p. 25. 52 Anil Bhat, 2009, p. 9. 5 Sanj ib Baruah, Beyond Counter-Insurgency:Breaking the Impasse in NortheastIndia, New Delhi: Oxford University Press, 2009, p. 151. 54 Maloy Krishna Dhar, Open Secrets: India's Intelligence Unveiled, New Delhi: Manas Publications, 2005, p. 398. 5 Former governor and national security advisor Ved Marwah argues that Nehru's remarks after the 1962 Chinese invasion of India's Northeast gave the impression "as if the rest of the country was reconciled to its conquest by the and "not too bothered about losing Assam" (2009, p. 9). 5Chinese" 6 B.N. Mullik, My Years with Nehru: 1948-1964, Bombay: Allied Publishers, 1972, p. 169. 342 2.1.2. Central India (The "NaxalBelt") A large chunk of central India described as Naxal belt or Red Corridor is another ideationally distant region and has been treated as much like a "peripheral zone,"" frontier area, 58 or a "boondocks" 59 inside India. This region cuts across the mountainous, hilly, forested terrain of the Indian states of Andhra Pradesh, Bihar, Chhattisgarh, Jharkhand, Madhya Pradesh, Orissa, Uttar Pradesh and West Bengal. The human and physical terrain is poorly understood and illegible, extending across "a huge forest unsurveyed and unmapped even in size, decades after Independence," with extensive poverty, illiteracy, and human development, 60 limited road 62 connectivity,61 treated as "virtually a 'no-go' area for security forces and administration." Furthermore, "the state's ability to project its disciplinary power remain[s] limited," since the peoples residing in these region were "not politically or socio-economically dominated by centralized state-governed societies on the plains." 63 Similar to the Northeast, these largely tribal regions in central India provide little material incentive to the state having been treated apart from other territories in part for their expected economic potential, or "assumptions about their geographical concentration and isolation, and lack of economic differentiation or 'modern' economic sensibilities."6 4 Though in the heart of 57 This was discussed by Nandani Sundar in a panel "Spatial Issues and Maoist Insurgency in India," Observer Research Foundation, New Delhi, February 24, 2012. 58 Author interview, with a retired senior Indian civil servant, New Delhi, India, November 23, 2012 (#83). 59 Saumitra Mohan, "Naxalism: The Enemy Within," Institute for Peace and Conflict Studies, #2284, May 7, 2007. 60 Palash Ghosh, "Blood Red Corridor," InternationalBusiness Times, Feb. 28, 2014. 61 Deepak Aneel Boyini, "Explaining Success and Failure: Counterinsurgency in Malaya and India," Unpublished masters thesis, Naval Postgraduate School, December 2010, p. 4. 62 Quote from page 25 of Col. Purushottam Singh, "To Combat Left Wing Extremism, Tackle its Roots Rather than Merely Treating Symptoms," Pratividrohi,1 (3), March 2011, pp. 25-29. 63 Jonathan Kennedy and Sunil Purushotham, "Beyond Naxalbari: A Comparative Analysis of Maoist Insurgency and Counterinsurgency in Independent India," ComparativeStudies in Society and History, 54 (4), 2012, p. 852. 64 Louise Tillin, "United in Diversity? Asymmetry in Indian Federalism," Publius, 37 (1), Winter 2007, p. 56; Stuart Corbridge, "Competing inequalities: The Scheduled Tribes and the Reservations System in Jharkhand, India," Jharkhand:Environment, Development, Ethnicity, Ed. S. Corbridge, S. Jewitt, and S. Kumar. New Delhi: Oxford University Press, 2004, p. 176 343 central India, the region is effectively an "outland" and "a neglected, poorer, lawless place" 65 with poor governance, infrastructure, and human development. The area is "generally out of sight and, therefore, out of mind"66 and has remained "beyond the reach of any development projects, social welfare schemes and agencies of administration" and marked by "extremely distressing socio-economic conditions." 67 Theoretically, the region possess substantial potential value with extractable resources and mineral wealth but much of this is yet unexploited. Still, "the State has failed to ensure for the local population access to basic services, such as education, health, water and sanitation, employment and human development in general."6 8 Resource "cursed" countries are evidence that natural resource wealth is certainly distinct from development, as extraction requires little by way of physical or social infrastructure. In both the Northeast and the Red Corridor, the "Indian state is quite comfortable living with under governance" 69 Finally, the geography of the region poses little strategic incentive because it is not central to the defensibility of the Indian state from an external, conventional power. 70 West Bengal is the one exception to the rest of Central India, at least in the first three decades post-independence. Bengal had been one of the crown jewels of the British Raj, a primary site of agricultural surplus extraction and therefore administration, and the seat of colonial power.7 ' Bengali early adoption of English made them useful administrators in the British system and subsequently the leading intellectuals of India.72 In the 1960s it was one of 65 66 67 "India's Ungoverned Spaces: Out of the Trees," The Economist, June 1, 2013. Sudeep Chakravarti, 2008, p. 129. Bidyut Chakrabarty and Rajat Kumar Kujur, Maoism in India: Reincarnationof Ultra-Left Wing Extremism in the Twenty-First Century, New York: Routledge, 2010, p. 27. Siddharth Chatterjee, "Flaws in India's Strategy to Counter the Maoist Insurgency," Global Observatory, International Peace Institute, August 27, 2012. 69 Author Interview with an Indian journalist, New Delhi, India, Dec 17, 2013, (#147). 70 Based on Author interview with a retired senior Indian Army officer, Gurgaon, India, November 9, 2012 (#67). 71 Naseemullah, 2011; Anderson 2012. 72 Mark Magnier, "In India, Bengalis Seek to Recapture Their Glory as Intellectuals," Los Angeles Times, September 8, 2012. 68 344 the richest states, buoyed in large part by one of the largest urban and manufacturing centers, the port city of Calcutta.73 Eventually the state would go into economic stagnation in the mid 1970s, but it is the one region that was of great ideational and material import to the state during the Naxal uprisings. 2.1.3. Punjab In contrast to the Northeast and Central India, the North and West of India have always been at the core of India's economic and political identity, and one of the leaders has always been Punjab.74 "It is relatively small in terms of territory and population size but significant in terms of its economic, political, and strategic importance in the Indian federation."7 5 Historically, the region of Punjab was of immense strategic importance under the British empire. It provided the bulwark against the Russian threat from the Northwest requiring the creation of lines of supply and communication, the building of cantonments, and an administrative machinery resulting in "a significant network of communications and markets for Punjab."7 6 Two districts of Punjab served as an essential bridge for armed forces in mainland India to access Kashmir, and loss of their control could delink Kashmir and leave it vulnerable to Pakistan.7 7 Punjab was also economically critical to India's survival. Since independence, the state was referred to as India's Iowa because 60% of the grain was produced here after the agricultural boom of the green revolution and provided "India's only cushion between enough-to-go-around 7 Amartya Lahiri and Kei-Mu Yi, "A Tale of Two States: Maharashtra and West Bengal," Review ofEconomic Dynamics, 12 (3), July, 2009, pp. 523-542. 74 The centrality of Punjab to the national identity was underscored in various interviews. 7 Pritam Singh, Federalism, Nationalism andDevelopment: India and the Punjab Economy, New York: Routledge, 2008, p. 3. Rajit K. Mazumder, The Indian Army and the Making ofPunjab, Delhi: Permanent Black, 2003, p. 47. Kumar, "Sikh Ethnic Uprising in India and Involvement of Foreign Powers," Faultlines, 18, January 2007. 77Suneel 345 and famine, between self-sufficiency and foreign dependency." 78 Overall, Punjab was one of the fastest-growing and richest states with one of the lowest poverty rates, and in the 1980s, it was also the fifth most industrialized state in India. The growth of its agricultural sector spilled over to develop other aspects of the state such as road density, electricity, and irrigation and entrepreneurship, all of which would help lower operational costs in that terrain.79 Moreover, Punjab was in close proximity to other valuable regions in the more developed Western half of India8 0 including Haryana, Chandigarh, and most importantly, the capital of New Delhi. 2.1.4. Jammu & Kashmir (J&K) Kashmir is sometimes described as a part of India's periphery given its location as a border state, distance from Delhi, and dense mountainous terrain. Despite its similarity to the Northeast in beingphysically located on the extremities of British India and post-independence India, "Jammu and Kashmir has been more central to the national imaginary of India than the North-East."" It is why "the level of central government intervention in, and manipulation of, the state's politics has been greater than in any other state since independence." 8 2 Though economically Jammu & Kashmir (J&K) may have only had moderate value and its physical terrain posed significant challenges, J&K was strategically and psychologically very important both to Indian leaders and elites in Delhi.83 Rone Tempest, "Sikh Unrest: Roots are not Persecution," Los Angeles Times, August 4, 1985. 79 Swaminathan S. Anklesaria Aiyar "Why Punjab Has Suffered Long, Steady Decline," Indian Express, Nov 8, 2012 80 Sanghamitra Bandyopadhyay, 2013. 81 Tillin, 2007, p. 58. 82 Tillin, 2007, p. 62. 83 Author Interview with an Indian national security analyst, New Delhi, India, December 20, 2013 (#152). 78 346 J&K was fabled to have great economic value and "prosperity in the state due to its vast natural resources," and "its picturesque tourist attraction"8 4 Nehru had argued before the UN that economically "Kashmir is intimately related to India" based on historical trade routes.8 5 After the conflict was militarized, it took on a strategic importance and gave way to "territorial obsession." 86 Given its geography, some contend, "The security of Kashmir is vital to the security of India"87 Its mountainous terrain and surroundings withan average elevation of 16,000 feet certainly limit physical infrastructure and pose major operational costs, but this is offset by the security this feature affords the rest of the country. Srinagar and the Kashmir valley had strategic value to India and its rivals due to strategic communication links and ingresses in the 1947 and 1962 wars as it guarded vital entryways into the very Punjab and Haryana plains and even Delhi. 8 Therefore, it continued to be valued by the Indian military as a strategically vital region.89 Finally, in terms of the political identity, India values the region of Kashmir as "the cornerstone of its identity as an inclusive, secular state and the focal point of its bitter enmity with Pakistan" 90 Even prior to any conflict in the region, it is well recognized that India's first leader, Nehru, held a deep emotional attachment to his ancestral home of Kashmir 9l and asserted, "Anything that happens in Kashmir has a certainimportancefor the rest of India." This has led Singha, 1995, p. 35. J. Bandyopadhyaya, The Making ofIndia's ForeignPolicy, New Delhi: Allied Publishers, 1970, p. 225. 86 Mona Bhan, Counterinsurgency,Democracy, and the Politics ofIdentity in India, New York: Routledge, 2013, p. 12. 87 White Paper on Jammu & Kashmir, New Delhi: Government of India, 1948 pp. 45-46 drawn from Rajpal Budania, India'sNational Security Dilemma: The PakistanFactorand India'sPolicy, New Delhi: Indus Publishing Company, 2001, p. 92. 8 Stephen Cohen, India:EmergingPower, Washington DC: Brookings Institution Press, 2002, pp. 212-13. 89 Capt. Yudhishthir Moza, "A Perspective on Jammu and Kashmir," Pratividrohi,Autumn 1996, p. 54. 90 Sumantra Bose, 2003, p. 111. 91 Christopher Snedden, The Untold Story of the People ofAzad Kashmir, New York: Columbia University Press, 2012, p. 32-33; Dhar, 2005, p. 379. 84 8 347 some historians to assess that "the fact that [Nehru] was the prime minister made Kashmir a state different from the 500 other princely states." 92 Since then, Foreign Minister Jaswant Singh has stated that Jammu and Kashmir is at "the core of Indian nationhood." 93 Others have also argued out Kashmir's importance has less to do with Pakistan but rather "what it [means] for Hindu-Muslim relations at the all India level," 94 and because it "affects the existence of the Indian Republic as a democratic, multi-lingual, multireligious and multi-ethnic pluralistic state." 95 The quantitative and qualitative evidence on the four regions suggests that Punjab, Kashmir, and West Bengal were core regions due to their economic and strategic value along with their lower operational maintenance costs, while the Northeast and Central India were generally undervalued and treated as peripheral zones. If threatened by rebellion, the former would elicit much greater alarm. Varshney confirms this, arguing: Rightly or wrongly, the small Northeastern states of Nagaland and Mizoram do not-and did not-affect politics in India's political heartland. Punjab and Kashmir, where insurgencies erupted in the 1980s and 1990s, are different. ... A Sikh insurgency in the 1980s was profoundly shocking, generating great anxieties about whether Punjab would stay in India. ... Being a Muslim majority state, Kashmir was even more critical, especially for the rise of Hindu nationalism.% Consequently, as the regions of Punjab, Kashmir, and West Bengal are contested by rebellion, I expect the Indian state to exert a higher level of effort in these more valued regions and strategies that seek to defeat rather than merely contain insurgencies such as attrition or population security. Claude Arpi, "The Blunder of the Pandit," Rediff com, June 16, 2004. 93 Stated on Doordarshan television interview with Mrinal Pandey, July 11, 2001 as reported in "Indian Foreign Minister on Summit, Says Kashmir 'Core' of Indian Nationhood," BBC MonitoringSouth Asia, July 12, 2001. 94 Author Interview with an Indian journalist, New Delhi, December 17, 2013 (#147). 95 J.N. Dixit, Anatomy ofa FlawedInheritance:Indo-PakRelations, 1970-94, New Delhi: Konark Publishers, 1995, 200. Varshney, 2002, pp. 74-75. 92 348 2.2. The Moral Worth and Embedded-ness of Indian Identity Groups Like the physical terrain, the landscape of identity groups in India has been varied and uneven with a core and a periphery where some groups dominate positions of power or commanding moral worth (as perceived by the state or state elites), and others are disempowered, marginalized, and regarded with disdain. Without a single dominant cleavage in India (owing to ethno-linguistic, religious, and caste divides even prior to independence), minority groups were not universally subordinated and had opportunities for power and worth, but this varied based on patterns of colonial administration and favoritism as well as pre-existing social hierarchies. Groups that rebel may be partially motivated by grievances but still somewhat empowered. Of the relevant sub-groups that rebelled against the Indian state since 1947, Sikhs were valued with both high group worth and embedded-ness in the state (particularly the security forces). Hindu lower castes and social groups-labeled scheduled castes and tribes (SC/ST) or Dalits (untouchables) and adivasis (original/indigenous peoples)-initially possessed modest group worth and low embedded-ness, but due to state affirmative action policies, this increased over time to a point where they achieved high levels of embedded-ness and worth by roughly the 1990s. By contrast, Indian Muslims (including Deccani and Kashmiri Muslims) were regarded as distinct, of low worth, and had little embedded-ness in state institutions after a status reversal resulting from the partition of the subcontinent. Ethno-linguistic tribal groups from India's Northeast (henceforth "Northeasterners" which includes Nagas, Mizos, and Meiteis/Manipuris), most of whom were Christians and ethnically of Tibeto-Burman origin, were regarded with suspicion and low worth by the "mainland", and were completely un-embedded in state institutions, political or economic networks, owing to their exclusion from the British colonial 349 administrative system. The exception in the case of Northeasterners was the Assamese people of the Bhramaputra Valley who were predominantly Hindu and moderately integrated into the Indian state. Values for each rebel group's worth and embedded-ness can be estimated for each decade using a range of empirical support from scholarly research, government reports, speech evidence, and data on group representation within state institutions. Group worth is largely based on qualitative evidence as well as some data on prejudicial attitudes and discrimination, while group embedded-ness is much more grounded in quantitative data on the group's representation in critical state institutions including the security services (military, paramilitary, police, intelligence), as well as civil service and political institutions. Based off this qualitative and quantitative measures of the position of various groups (summarized in Table 6.4), the coreperiphery theory would then expect strategies of high violence against most Northeast and Muslim groups, as well as SC/ST prior to the 1980s, and it would predict strategies of low violence against Sikhs, Assamese, and SC/STs after the 1990s. Table 6.3: Minority Populations by Decade (%) Scheduled castes (SC) Scheduled tribes (ST) Sikhs Muslims 1961 1971 1981 1991 2001 2011* 14.7 14.6 15.75 16.48 16.2 16.6 6.9 6.9 7.76 8.08 8.2 8.6 1.8 10.7 1.9 11.2 2 11.4 1.99 11.7 1.86 13.4 ? 14.6* 3.83 3.84 3.83 3.72 3.66 3.38 North-easterners --Assam 2.47 2.67 2.64 -- rest of NE 0.83 0.9 0.98 2011 census did not include data on religion. Muslim population 350 2.65 2.59 1.08 1.14 estimated by Pew 2010 2.58 1.14 Table 6.4: Identity Group Position by Decade (High, Mod Low) Worth Hindu Brahmins Other Backward Class Scheduled castes (Dalits) Scheduled tribes (Adivasis) Sikhs Muslims --Kashmiri Muslims Assamese Other Northeasterners High Scheduled castes (Dalits) Scheduled tribes (Adivasis) Sikhs Muslims --Kashmiri Muslims Assamese Other Northeasterners High 1960s Embed Civil Security High High Worth High 1970s Embed Civil Security High High Worth High 1980s Embed Civil Security High High Worth Low Low Low Low Low Low Low Low Low Mod Mod Mod Low Low Low Low Low Low Low Low Low Mod Mod Low Mod High Low Low High Low Low High Low Mod High Low Low High Low Low High Low Mod High Low Low High Low Low High Low High High Low Low High Low Low High Low Low Mod Low Mod Low Mod Low Mod Low Mod Low Mod Low Mod Low Mod Low Mod Low Mod Low Mod Low Mod Low Low Low Low Low Low Low Low Low Low Low Low High 1990s Embed Civil Security High High High High Mod High 2000s Embed Civil Security High High High High High Mod Mod High High High Low Mod High Low Mod High Low Low Mod Low Mod Low Low Worth Hindu Brahmins Other Backward Class 1950s Embed Civil Security High High High 2010s Embed Civil Security High High High High High High High Mod High High Mod High High Low Mod High Low Mod High Low High High Low Mod High Low Mod High Low Low Mod Low Mod Low Mod Low Mod Low Mod Mod Mod Mod Mod Low Low Low Mod Low Low Mod Worth Worth 2.2.1. Sikhs Sikhs have been a critical minority group in India of high moral worth, closely linked to the dominant identity group, and deeply embedded in Indian economic, political, and security institutions, in particular. Group Worth Sikhs have largely enjoyed equal status and lived harmoniously with Hindus, India's dominant religious group.9 7 Historically, the lines between Sikhs and Hindus blurred as they were believed to have "stemmed from the same stock," and conversion and intermarriage were common. 98 Hindus regarded Sikhs as a highly "Indianized" community, practicing "a subset of Hindu traditions," and "fundamentally Hindu," particularly given the significant overlap of religious customs such as festivals, rites, deities, spiritual leaders, and scriptures.9 9 Some went so far as to suggest that "no other two religious communities are bound together by as many inseparable bonds as are Hindus and Sikhs,"100 a claim substantiated by the fact that some Hindu families would volunteer one of their sons to become a Sikh.1 01 This perception was echoed by 97 Gardner Murphy, In the Minds ofMen: The Study ofHuman Behavior and Social Tensions in India, New York: Basic Books, 1953, pp. 43-44. 98 Prakash Tandon, Punjab Century: 1857-1947, Berkeley, CA: University of California Press, 1968, p. 10; Donald L. Horowitz, "Ethnic Identity," Ethnicity: Theory and Experience, Ed. Nathan Glazer, Daniel Patrick Moynihan, Corinne Saposs Schelling, Cambridge, MA: Harvard University Press, 1975, p. 135; Gerald James Larson, India's Agony Over Religion, Albany: State University of New York Press, 1995, p. 24; Ajit Adhopia, "Offspring Feel Threatened by Mother Faith," Toronto Star, February 17, 2001. 99 Quotes from Larson, 1995, p. 24; R.A. Schermerhorn, Ethnic Pluralityin India, Tuscan: University of Arizona Press, 1978, p. 151; Qamar Hasan, Muslims in India: Attitudes, Adjustments and Reactions, New Delhi: Northern Book Centre, 1988, p. 45. On practices, see Eleanor Nesbitt, "Pitfalls in Religious Taxonomy: Hindus and Sikhs, Valmikis and Ravidasis," Religion Today, 6 (1), 1990, pp. 9-12; B.N. Mullik (1972 p. 421) echoes the same ideas. 10 Former Sikh Lt. Gov. of Goa Dr. Gopal Singh, "Hindu-Sikh Fraternity: Don't Negate History of Five Centuries," Times ofIndia, July 14, 1986, p. 8. 10 Murphy, 1953, pp. 43-44. 352 India's first Intelligence Bureau chief stating, "there was little difference between Hindus and Sikhs in Punjab." 02 Additionally, Sikh's were seen as fiercely loyal patriots. 03 After displays of Sikh bravery in India's early wars, "Indians were compelled to admit that Sikhs were the chief defenders of their country."1 04 Even civilian Punjabis demonstrated their loyalty to the nation as Sikh farmers, transport owners, and merchants all helped to provide material support to Army units mobilizing and fighting on the Punjab's border in the 1965 and 1971 wars with Pakistan.1 05 Even in the midst of a turbulent political dispute between Sikh leaders and New Delhi, with the onset of the 1962 war with China, the Sikh people volunteered tremendous support to the central government.1 06 Group Embeddedness Sikhs have been described as a "mobilized minority," "much more rooted in the system"'07 with "influence far exceeding their numbers" who had " won political success that no other community has actually gained."108 Economically Punjabi Sikhs were prosperous,109 held a proportionate share of influence in the New Delhi bureaucracy and the Union Cabinet," 0 and had a disproportionate share of political power in the Punjab state government." Despite being 2% of the population, during the 1980s Sikhs made up more than 8% of the government workforce B.N. Mullik, 1972, p. 421. 2002, p. 74. Kushwant Singh, 1969, p. 159 - form Shermerhorn, 1978, p. 149. 105 Stephen P. Cohen, "The Military and Indian Democracy," India's Democracy: An Analysis of ChangingStateSociety Relations, Ed. Atul Kohli, Princeton, NJ: Princeton University Press, 1988, p. 137. 106 B.N. Mullik, 1972, p. 457. 107 Author phone interview with Human Rights activist, New Delhi, India, November 14, 2012 (#73). 108 Schermerhorn, 1978, pp. 128-129. 109 Varshney, 2002, p. 74. 110 Schermerhorn, 1978, p. 150. "1 Schermerhorn, 1978, pp. 150-151. 102 04 Varshney, 103 1 353 (including the elite civil/administrative service). Many powerful officials at the time were Sikhs, including Indian President, Zail Singh, and one of Prime Minister Rajiv Gandhi's closest advisors, Arun Singh were Sikhs." 2 Even concessions on some Punjab autonomy "led, paradoxically, to more complete incorporation as an unexpected consequence."" 3 In the 1970s, one study based on survey research of Indian political participation and alienation found Sikhs to be deeply entrenched as insiders." 4 Where the Sikhs really punched far above their weight was in the security forces. In the 1980s, they composed the majority of the Punjab state police, but they also composed 7% of total armed forces personnel, 12% of the Indian army (though this had dropped from a high at independence of 25%) and about 20% of the officer corps." 5 In 1962, Sikhs composed roughly 42% of Indian Army flag officers." 6 Sikhs had historically composed a substantial portion of a number of infantry regiments in India including two infantry regiments drawn exclusively from their own community. "7 They were also heavily present in other units including the Army Service Corps, the engineers, artillery, armor, and other specialized branches as well as the air force (estimated to be about one quarter of the pilots) and Indian navy." 8 As officers, Sikhs were highly regarded and had a "brilliant reputation," particularly serving in counterinsurgency 112 Tempest, 1985; also see p. 447 of Paul Wallace, "Countering Terrorist Movements in India: Kashmir and Khalistan," Democracy and Counterterrorism:Lessonsfrom the Past, Ed. Robert J. Art and Louise Richardson, 2007, pp. 425-481. 11 3 Schermerhorn, 1978, p. 152. 114 See page 350 of Ashis Nandy, "Engagement and Alienation in Indian Politics," ComparativePoliticalStudies, 7 (3) 1974, pp. 334-356. 115 Telford, 2001; Tempest, 1985; Barua estimates the Sikhs were 11% of the manpower of the armed forces. See page 125 of Pradeep P. Barua, "Ethnic Conflict in the Military of Developing Nations: A Comparative Analysis of India and Nigeria," Armed Forces& Society, 19 (1), Fall 1992, pp 123-137. 116 Cohen, 1988, p. 134 says six of the Army's Lt. Generals, 13 of the 28 major generals, and 30 of 79 brigadier generals were Sikhs. I estimate estimated to be about 7-8 total Lt. Gen. since the ratio of Lt Generals to Major generals in 2009 was 20 to 75 based on See Army Adopts Promotion Policy for Maj. Generals, Lt. Generals," Indian Express, August 20, 2009. 11 7 P.K. Gautum, p. 47-49; Kundu, 1994, p. 48. 118 Stephen P. Cohen, "The Military and Indian Democracy," India'sDemocracy: An Analysis of ChangingStateSociety Relations, Ed. Atul Kohli, 1988, p. 132. 354 operations in the North East.' 1 Trust in their leadership and loyalty has allowed some to ascend to the heights of intelligence agencies like RA&W and NTIC, and two Sikh generals have even served at the helm of the armed forces as the Army Chief of Staff.12 0 2.2.2. Scheduled Classes and Tribes (SCIST) The positional status of lower caste groups such as scheduled castes (dalits), scheduled tribes (adivasis), and other backwards classes (OBCs) who have both composed the popular base of the various Maoist/Naxal rebellions through modem India's modem history from 1948-the present,12 has changed over India's history. SC/STs, who had been historically marginalized in the pre-independence period under a Hindu dominated caste system, began to be identified by the British policies by the 1930s as the depressed castes and tribes that required reservations to correct this. While there was not much to unite the disparate sets of scheduled castes and tribes, they were unified in their political exclusion from the experience of colonialism and were recognized by the Indian Constitution and accorded a political identity. While STs have been seen as outside Hindu society and SCs subjugated within it, India's Hindu elites developed a degree of empathy for them motivating them to institute affirmative action programs ushering in major socio-economic and political elevation of these groups status and embedded-ness within the state. Meanwhile, middle castes and OBCs who have played a larger role in more recent Naxal insurgencies (II and III) have also advanced socially and politically since the 1980s. 122 119 Cohen, 1988, p. 134. 12 Praveen Swami, "The Colors of Army Green," Frontline,Jan 2, 2004. 121 Though scheduled tribes or adivasis are well-recognized as the political base of the Maoist/Naxalites, scheduled castes have also composed their base. See Satish Kumar, "How Anti-Naxal Forces Weakened in Bihar," IPCS brief, #2821, February 27, 2009; E.N. Rammohan, "Unleash the Good Force," Outlook India, July 16, 2012. 122 For instance, 38% of Naxals interviewed by one sociologist were OBCs. See Chitralekha, "'Committed, Opportunists and Drifters': Revisiting the Naxalite narrative in Jharkhand and Bihar," Contributionsto Indian Sociology, 44 (3), 2010, pp. 299-329. 355 Group Worth Despite their poor socio-economic and political status, in the modem era, both SCs and STs have been recast as possessing significant moral worth. The formal incorporation of SCs and STs was actively led by one of independent India's architects of independence, Mohandas Gandhi, who believed these scheduled castes, whom he termed 'Harijans' (children of God) had to be elevated socially.123 A study conducted amongst a diverse mix of students in 1951 found that each group tended to rank Dalits on the lower end,124 but Dalits have socially ascended since that time. In the 1990s, India's economic takeoff and integration of different castes triggered a number of social changes such as the erosion of discrimination and stigma in inter-personal and inter-caste relations.125 This period has also witnessed increasing social mobility as evidenced by the continuous rise in inter-caste marriages to Dalits or untouchables.1 26 Social attitudes have also shifted as increasing numbers of people have expressed opposition to bans on inter-caste marriage, and a majority (53%) in urban areas indicate a decline in preference for a particular caste in marriage. Recent experimental evidence from the online arranged marriage market in India reveals that a majority of even upper caste respondents are willing to embrace inter-caste marriage.1 Scheduled tribes, referred to as Adivasis (meaning original inhabitants) were isolated from much of mainstream Hindu society and described as "forest and hill peoples" and regarded Cohen, 2002, p. 114. 123 124 Qamar Hasan, 1988, p. 97. Devesh Kapur, Chandra Bhan Prasad, Lant Pritchett, and D. Shyam Babu, "Rethinking Inequality: Dalits in Uttar Pradesh in the Market Reform Era," Economic & PoliticalWeekly, 45 (35), August 28, 2010, p. 39 126 Subodh Ghildiyal and Swati Mathur, "Silent Revolution: Rise in Marriages of Others to Dalits," Times ofIndia, January 31, 2010; Subodh Ghildiyal, "Steady Rise in Inter-Caste Marriages Involving Dalits," Times ofIndia, September 19, 2013. 127 Indian National Election Study in 2004 and a yet to be released N-S dataset in page 4 of Amit Ahuja and Susan L. Ostermann, "Blurred Lines: Intermarriage and the Weakening Caste Cleavage in Urban India," Presented at Annual International Studies Aassociation Conference, Toronto, April 2014. 125 356 as something like aboriginals.1 2 8 These "people of 'Aryavrata,' the Hindi speaking national heartland," that is STs or adivasis, "serve as the stock image of the stereotypical 'Indian."1 29 Many of the scheduled caste are Hindus and even the STs or adivasis have been viewed by some as "backward Hindus,"13 0 or who only require "reconversion."131 The similarities or overlap in physical appearance, cultural practices, beliefs, rituals, and festivals makes for a more intimate link between mainstream Indian (Hindu) society and adivasis.1 3 2 Relative to Dalits, "...adivasis are subject to less stigmatization because their lower (on average) position in society is at least as likely to be attributed to adverse environmental circumstances as to inferior innate abilities." 13 3 While this "valorizing [of] a mythical Other" or the "noble savage" may be essentializing and appropriating as some critical anthropologists contend, 3 4 it still suggests the Indian elite holds these groups in some esteem and admiration. The RSS has made strides in first reaching out to adivasis even before independence,13 5 and seeking to civilize and reform their religious practices. Evidence of adivasi alignment with the Hindutva political project was evidenced by their participation in the communal riots against Muslims in Gujarat in 2002.136 Schermerhorn, 1978, pp. 70-71. Shashi Tharoor, India: From Midnight to the Millennium and Beyond, New York: HarperPerennial, 1998, p. 11 13 0 Kennedy and King, 2013, p. 4. 131 Leena Gita Reghunath, "The Believer," Caravan, February 1, 2014. 132 S.M. Michael, Untouchable: Dalits in Modern India, Boulder, CO: Lynne Rienner, 1999, p. 105-110. 133 Thomas E. Weisskopf, Affirmative Action in the UnitedStates and India:A ComparativePerspective, New York: Routledge, 2004, p. 95. 134 Amita Baviskar, "Indian Indigeneities: Adivasi Engagements with Hindu Nationalism in India," Indigenous Experience Today, Ed. Marisol de la Cadena and Orin Starn, Berg Publishers: Oxford, 2007; Ramachandra Guha, "Radical American Environmentalism and Wilderness Preservation: A Third World Critique." Environmental Ethics 11 (1), 1989, pp. 71-83. 135 Christophe Jaffrelot, Religion, Caste, and Politics in India, New York: Columbia University Press, 2011, p. 252. 136 Baviskar, 2007; Jaffrelot, 2011, pp. 382, 390-93. 128 129 357 Group Embeddedness Both SCs and STs were the beneficiary of an extensive affirmative action and "protective discrimination" program allowing them to gain "as a group"13 7 through reserved parliamentary representation, reservations in the education system and public sector jobs, even the security forces, to accelerate integration, 3 8 and development programs to bridge the economic divide.1 3 9 Initially these quotas in the 1950s-60s were "almost totally ineffective."140 As a result, in a survey in 1969, scheduled castes were judged to have the smallest percentage of "insiders" (7.8%) and highest level of outsiders (31.8%) amongst a range of social groups surveyed (including Hindus, Muslims, Sikhs, and Scheduled Castes). 41 Though the political scene was "profoundly transformed," Dalits were still treated as a Congress party client.1 4 2 While economic standing and representation of STs was quite low early on, by the late 1960s their representation nearly reached half their proportion by the late 1960s.1 4 3 The socioeconomic changes took much longer and hiring quotas only began to have real "catch-up" effects in the 1980-90s.144145 Meanwhile, the implementation of central administration quotas in 1990 for other backwards classes (OBC), who at times have participated and sympathized with Naxal movements, has led to their empowerment in the political and administrative spheres, what one scholar has termed a "silent revolution."1 4 6 137 Schermerhorn, 1978, p. 62. 138 Schermerhorn, 1978, p. 56. 139 Kennedy and King, 2013, pp. 3-4. 140 Steven Ian Wilkinson, "India, Consociational Theory, and Ethnic Violence," Asian Survey, 40 (5), Sept-Oct 2000, p. 777. 141 Nandy, 1974. 142 Jaffrelot, 2011, p. 493. 143 Schermerhorn, 1978, pp. 78-79. 144 Jaffrelot, 2011, pp. 5 17-520. 145 Social, Economic and EducationalStatus of the Muslim Community in India: A Report, New Delhi: Prime Minister's High Level Committee, Cabinet Secretariat, Government of India, November 2006, p. 85. 146 Christophe Jaffrelot, India's Silent Revolution: The Rise of the Lower Castes in North India, New York: Columbia University Press, 2003. 358 Today, a growing Dalit middle class includes many civil servants who were beneficiaries of affirmative action programs. "Caste is losing its grip," argues one Dalit social scientist, and this has been confirmed by studies revealing tremendous growth in the quality of housing and household goods, and entrepreneurship amongst Dalits. 147 Further evidence of socioeconomic mobility comes from the national sample survey data revealing intergenerational education, income, and occupation mobility rates of SC/STs have converged to non-SC/STs between 1983 and 2005.148 Improvements in representation accelerated alongside the movement for advancing Other Backward Castes (OBC), and SC/ST representation in state government employment "shot up during the late 1970s and the 1980s. 149 There is further evidence of a number of "non-material gains" from the affirmative action programs that were introduced to increase SC representation including "cultural respect," "self-dignity" and "symbolic social change" and greater voice in political decision-making. 150 The SC and OBC groups became much more assertive in the 1980s and assumed a share of power by the 1990s,15 1 for instance led by political figures like Mayawati, the "Untouchables Queen," who became the Chief Minister of India's largest state, Uttar Pradesh, in 1995. By the late 1980s, SC/ST composed roughly 28% of district magistrates. In 1997, India even elected K.R. Narayanan, a Dalit, as President. Regardless of the provenance of newly elected Prime Minister, Narendra Modi's status as an OBC, his 147 "I Was One of India's Unclean Dalits... Now I Am a Millionaire," Associated Press, December 27, 2011; SA Aiyar, "The Unexpected Rise of Dalit Millionaires," Swaminomics Blog, Times ofIndia, July 31, 2011. 48 Viktoria Hnatkovska, Amartya Lahiri, and Sourabh B. Paul, "Breaking the Caste Barrier: Intergenerational Mobility in India," Journalof Human Resources, 48 (2), Spring 2013, pp. 435-473. 149 Wilkinson, 2000, p. 782. 150 Simon Chauchard, "Why Representation Matters: The Meaning of Ethnic Quotas in Rural India," Book Manuscript Under Review, p. 24. 1 Cohen, 2002, pp. 114-115. 152 Estimated to be 125 out of 450. See Jaffrelot, 2011, p. 538 (citing a figure stated by Kanshi Ram). 153 PJ Anthony "K.R. Narayanan Dies at 85; 'Untouchable' India President," New York Times, November 10, 2005 359 attempt to cloak himself in that mantle confirms the significant political rise of these lower castes.1 54 Proportionality also extended to the security forces; in the 1980s, SC/ST reservations were implemented and enforced in a number of army and paramilitary units including the Assam Rifles, BSF, CRPF, and Rapid Action Force.155 The government even went so far as to ensure proportional government expenditures earmarked for SC/ST by 1990.156 Today SC/STs are much more embedded in state the institutions such as the political regimes, bureaucracy, and even organs of national security and public order/safety. SC/ST are nearly fully represented in the police force' 57 and though their representation ratio in the defense activities is half of what it should be, this is much higher than Muslims (or Northeasterners). While their representation is largely in lower level positions, when taken as a whole of central government employees, SC and ST representation meets their actual proportion of the country and exceeds the mandated levels.1 58 New civil service appointments in 2010 reflected near proportionality with the SC and ST populations but still not with Muslims.' 59 2.2.3. Indian Muslims New York Times columnist Thomas Friedman recently wrote that Indian Muslims "live in, are the product of, and feel empowered by a democratic and pluralistic society," 60 just a couple years after an exhaustive government-appointed panel report concluded quite the 154 Sheela Bhatt, "Is Narendra Modi Relaly an OBC?" Rediff com, June 5, 2014. 155 Wilkinson, 2000, p. 782. 156 Wilkinson, 2000, p. 783. Based on representation data from the ministry of home affairs as well as Social, Economic and Educational Status of the Muslim Community in India, 2006. p. 102. 158 Subodh Varma, "SC/STs fail to break caste ceiling: No SC in 149 top government officers, 40 pc do menial jobs," Economic Times, September 6, 2012. 159 Data from "Kashmir Muslim boy rules India, tops Civil Services exam," TwoCircles.net, May 6, 2010. 160 Thomas L. Friedman, "No Way, No How, Not Here," New York Times, February 18, 2009. 157 360 opposite. In fact, Indian Muslims have been profoundly disempowered, regarded with low moral worth and disdain, and dis-embedded from nearly all state institutions. While SC/ST may be "victims of indifference," Muslims are the "conspicuous minority" who suffered a major status reversal after India's independence from an elite to a "powerless minority," due to English language policies, property rules, the emigration of most educated Muslims to Pakistan, and the elimination of government reservations for Indian Muslims. 16 1 They have continued on a downward trajectory ever since. Group Worth Anecdotal accounts and social science research through surveys and ethnography continued to confirm perceptions of tremendous social distance, ethnic prejudice (even relative to untouchables or Dalits), disdain, and perceived disloyalty of Muslims. In some of the earliest surveys conducted, even prior to the 1947 partition, Indian Hindus sampled held Muslims in lower regards than Harijans, and this dropped further after partition.1 62 Partition exacerbated communal tensions leading one Muslim professor who was interviewed to point out the conspicuousness of Indian Muslims after 1947: "You're never treated just as a human being; you're treated as a Muslim."163 In another early survey, 84% of Muslims survey perceived some degree of discrimination by Hindus, 58% characterized it as pervasive, and 73% of Hindus sampled viewed Muslims with some degree of unfavorability.1 64 One study of Muslim youth found them to perceive scheduled castes to have greater worth because their job security and Schermerhorn, 1978, pp. 154-55, 162-64 (also citing Hussain, 1965, 131-32 ). S.P. Adinarayan, "Before and After Independence: A Study of Racial and Communal Attitudes in India," British JournalofPsychology, 44 (2), 1953, pp. 108-115. 16' Gardner, 1953, p. 127. 164 Gardner, 1953, pp. 144-45, 148-149. 161 162 361 national loyalty were never in doubt. 165 Other ethnographic work from the late 1970s-early 1980s found Muslims to be a suspect class, subject to communal prejudice, and of lower political status than scheduled castes. 166 A study of Hindu schoolboys conducted in Patna, Bihar in 1950 found intensified social distance and prejudice against Muslims in more formal, institutional, or kinship ties (e.g. intermarriage) with Muslims ranking near the bottom. 167 The study found this prejudice to be a function of Hindu religious mores and stereotypes and noted "the pattern thus emerging seems to be comparable to the pattern of anti-Negro and anti-Jew prejudice in many important respects."16 8 This may have in part derived from Muslims being portrayed in history textbooks as cruel rulers and aggressors.1 69 It also might have derived from an experience in India's first war with Pakistan in 1947, where the Muslims in a mixed the 4t J&K Light Infantry regiment opened fire on their sleeping Hindu comrades and defected to the Pakistani side.17 0 Indian Muslims' loyalty has always been held suspect. Early on in India's history, Sardar Patel warned of disloyalty and a lack of cooperation amongst Muslim bureaucrats.17 1 Muslims were perceived as dogmatic, anti-democratic, aggressive, territorially and demographically expansionist, and disloyal with extra-territorial loyalties (to Pakistan or other Muslims nations). 7 2 These perceptions endured long after partition as revealed by a robust survey in the 165 Nirmala Srinivasan, Identity Crisis ofMuslims: Profiles ofLucknow Youth, New Delhi: Concept Publishing Company, 1980, p. 93. 166 Nirmala Srinavasan, PrisonersofFaith:A View from Within, New Delhi: Sage Publications, 1989. See especially pages 119-20, 135-36, 142-43, 147-49, 157-58, 160-63. 67 Sahib K. Mitra, "Social Prejudice and Adolescent Personality," A Perspective on Psychology in India: Dr. S. M Mohsin Felicitation Volume, Ed. Simaleswar De and Durganand Sinha, Allahbad, India: Sinha, 1977, pp. 256-79. 168 Mitra, 1977, pp. 276-77 169 Yashodhara Dalmia, "History Text-Books Foster Prejudice," Times ofIndia, March 9, 1993, p. 3; Mohammed Hasan, "Communalism in Indian Literature," Man and Development, 1984, 6, pp. 144-49. 170 Lionel Protip Sen, Slender Was the Thread: Kashmir Confrontation, 1947-1948, New Delhi: Orient Longman, 1969, pp. 34-36. 171 Qamar Hasan, 1988, p. 73. 172 Qamar Hasan, "Chapter 5: Extra-Territorial Loyalties and Communal Aspirations of Muslims," 1988, pp. 72-87. 362 1980s. 73 In a sort of early endorsement experiment, on average Hindus estimated that the majority Hindus believed "Muslims were traitors" and "Muslims are more loyal to Muslim countries than they are to India." They also indicated that a near majority (44%) of Hindus believed Muslims sought demographic dominance, exaggerated discrimination to extract undue benefits, that they sought to regain lost power, and were becoming more militant with external support.174 In general, Muslims were perceived as "not full-fledged loyal citizens of the country and need 'Indianizing' to purge them of disloyal tendencies."' 75 Though ostensibly a secular nation, Hindu elites found it difficult to ideationally absorb Muslims.' 76 While staunch adherents to Hinduism and Hindutva found Sikhs, adivasis (ST), Buddhists, and Jains to be proximate to the Hindu identity because they considered India their sacred land, "by the same token Muslims, Christians, Parsis and Jews become non-Hindus and, therefore, non-nationals."1 77 This treatment also underpinned both unconscious' 78 and conscious forms of social discrimination against Muslims, some which even persist today. These included discrimination when applying for jobs, loans, or places of residence. In urban areas of India, Muslims are still secluded in "vast ghettos" and even white collar and middle-upper class Muslims face housing discrimination in urban metropolitan areas with Hindus refusing to rent to them.1 79 The study was survey conducted in 1983-1985 amongst roughly 1800 college students in Uttar Pradesh in a variety of mixed towns with varying degrees of communal disturbance history and Musliam population salience 174 Qamar Hasan, 1988, p. 216 Table 6a, 6b, and 6c. The interpretation of the data is complicated by the way the study was structured. 175 Schermerhorn, 1978, p. 178. 176 Schermerhorn, 1978, p. 180. S.H. Deshpande, "Evaluating Hindutva I - The Muslim Question & Sarvarkar," Times ofIndia, January 4, 1994, p. 14; Varshney, 2002, pp. 65-66 makes the same point. 178 Eric Silver, quoting Salman Kurshid, " India's Other Nation - The Status of the Minority Muslims," The Guardian,August 28, 1986. 179 Gardiner Harris, "For India's Persecuted Muslim Minority, Caution Follows Hindu Party's Victory," New York Times, May 16, 2014. 173 363 Group Embeddedness After partition, Muslims were never embedded in state institutions, particularly security institutions, due to path dependence as well as conscious and unconscious discrimination. Muslim underrepresentation in the state began with post-partition emigration and political changes, like the end of the zamindari (feudal-like) land system, but it was furthered by deliberate policies to remove and screen out Muslims officials as well as informal discrimination in the civil administration. 80 181 Muslims in government were significantly underrepresented by at least half in the 1970s.1 82 Though the Muslim population has risen from 10% to roughly 15% of the country since 1947, at the highest levels of government, only 2% were Muslims (a figure that held steady for almost 50 years18 3 ). At the lowest rungs of the bureaucracy, "outright discrimination is more clearly shown" with Muslims constituting only 0.5%, a problem the government officially did not acknowledge.' 8 4 Today Muslims constitute 3% of the powerful Indian civil services bureaucracy.! 85 Significant underrepresentation in civil service was also mirrored in India's two largest employers -- the armed forces and the railways,186 as well as public sector companies, banks, and the private sector. 871 8 8 The underrepresentation and diminishing status of the Indian Muslim was repeatedly "discovered" and warned of dating back to Nehru's concerns post-partition.' 8 9 Muslim demanded 180 Omar Khalidi, Muslims in IndianEconomy, Gurgaon, India: Three Essays Collective, 2006, p. 39. 181 Khalidi, 2006, p. 1 2 48. Khalidi, 2006, p. 41. 183 The former Indian Chief Justice confirmed these figures in a speech in 1999, from A.G. Noorani, The Muslims of India: A DocumentaryRecord, New Delhi: Oxford University Press, 2003, p. 286. 184 Schermerhorn, 1978, p. 165 (also citing Lelyveld, 1968). 1 Social, Economic and EducationalStatus of the Muslim Community in India, 2006, p. 166. 186 In the 1970s, Muslims were roughly 2% of the Indian railways employees. Khalidi, 2006, p. 50. 187 Noorani, The Muslims ofIndia, 2003, p. 1040. 1" Khalidi, 2006, p. 45. 189 Nehru's Letter to Chief Ministers on 26 April 1954. From Noorani, The Muslims ofIndia, 2003, p. 109. 364 . quotas in the 1970s,190 and findings by a study commissioned by Indira Gandhi in 1980 found "the economic condition of Indian Muslims was worse than that of the Scheduled Castes." 191 But none of these discoveries elicited any changes. In 2006, another government appointed committee lead by former Delhi High Court chief justice Rajinder Sachar, published a report reconfirming that in all measures, Muslims were by far the most disempowered and discriminated group, even beyond India's Dalits and adivasis. Indian Muslims were disproportionately unemployed, illiterate, imprisoned, and poor, leading one analyst to assert, "Muslims are India's new untouchables."1 92 While the empathy and guilt felt for SC/ST and OBCs by upper caste Hindus had facilitated the implementation of quotas to improve those marginalized groups' status, "there is no such guilt conscience about Muslims,"1 93 and quotas for Muslims were regularly dismissed as "ruinous for India's nationhood," 194 India's democratic regime did little to help matters. In 1969, well-designed survey research on political inclusion and alienation amongst various minority groups found that despite almost proportional and formal inclusion in political representation, substantively Muslims were almost totally alienated as outsiders that commanded little influence on the Indian political system.1 95 Even while the Indian state made some effort early on to ensure token political representation of Muslims, they were "kept well away from the most important and sensitive jobs," and this limited their influence as there was perceived to be a "policy of keeping Muslim ministers at a distance from the main state policies."' 96 190 Presidential address at All-India Muslim Political Convention, New Delhi, 19 Dec 1970 from Noorani, The Muslims ofIndia, 2003, p. 150. 191 192 Khalidi, 2006, p. 43. Asra Q. Nomani, "The New Untouchables," Washington Post, November 4, 2007. Asghar Ali Engineer, "Quotas for Muslims," Times ofIndia, October 22, 1994, p. 16. N. S. Saksena, "Why Muslims Do Not Fare Well," Times ofIndia, May 31, 1983, p. 8 195 Nandy, 1974. 196 Wilkinson, 2000, p. 779. 193 194 365 The most significant absence of Muslims was in the security forces, intelligence, and defense apparatus. Though well represented during the British Raj, after Indian independence, Muslims continued to hold a very small percentage of the Indian armed forces or security bureaucracy ranging between 1-3%,197 and between 3-4% in all defense and national security agencies, most of which were at very low-tier positions. 198 In 1985, just prior to the outbreak of the Kashmir insurgency, George Fernandes, an Indian MP and eventual Minister of Defence, admitted "the Muslim is not wanted in the Armed Forces because he is always suspect whether we want to admit it or not. Most Indians consider Muslims a fifth column for Pakistan."1 99 This claim was grounded in broader public opinion with the majority of Indian Hindus expressing Muslims should not be allowed in the Armed Forces.2 0 0 Furthermore, Muslims had near zero representation in intelligence agencies, specifically the Intelligence Bureau (IB) (India's domestic intelligence/FBI equivalent) and the Research and Analysis Wing (RAW) (India's external intelligence and operations/CIA equivalent) because they were distrusted and deliberately excluded, a policy that was officially revised in the 1980s only to ban the explicit prohibition of Muslim officers. 2 01 Even the domestic-oriented Intelligence Bureau did not begin inducting Muslims until 1991. As of 2006 only a "handful" of Muslims were present in the 12,000-strong IB, while there were still zero Muslims out of 10,000 employees at RAW. 2 02 The Congress Party admitted that first Muslim IB Syed Asif Director Ibrahim was recently appointed as a crude political move to garner votes from the Muslim 1% from Omar Khalidi, Khaki and the Ethnic Violence in India, New Delhi: Three Essays Collective, 2003, p. 24. Others report 2.2%. See Simon Denyer, "Muslims Excluded from India's Spy Agency," Reuters, November 6, 2006. A third account says 3% based on a CNN-IBN estimate. See Amit R. Saksena, "India's Muslim Soldiers," The Diplomat, May 20, 2014. 198 Social, Economic and EducationalStatus of the Muslim Community in India, 2006, pp. 102, 369. 199 200 Khalidi, 2003, p. 23. Khalidi, 2003, p. 23; John Anderson, "India's Muslims Fear New Physical Threat," Washington Post, March 7, 1994. 201 B. Raman, The Kaoboys ofR&A W: Down Memory Lane, New Delhi: Lancer, 2007, p. 131. 202 Saikat Datta and Bhavana Vij-Aurora, "Muslims and Sikhs Need Not Apply," Outlook India, November 13, 2006 197 Yadav says 366 community,203 and a retired senior IB officer acknowledged that he would be the "first and last" Muslim in that position.2 04 Representation in the security/defense sector might have enabled Muslims to improve their position in the country by demonstrating loyalty like Punjabi Sikhs, but they were never trusted with these opportunities. Instead, underrepresentation continued to feed a cycle of distrust and abuse. The Sachar report concluded that Muslim underrepresentation fed certain stereotypes-like "every bearded man is considered an ISI agent"-perceptions that Muslims were "foreigners," and a national discourse of Muslims as anti-national, 2 05 all which shaped the manner in which the Indian security forces dealt with the Muslim population in practice.2 06 After the riots in Aligarh, Uttar Pradesh in 1978, Prime Minister Morarji Desai even acknowledged that Muslim underrepresentation had enabled the excessive aggressiveness by security forces.2 0 7 Kashmiri Muslims While it may be defensible that Kashmiri Muslims, being a small population in the country, would have limited representation at the federal level, it is telling that they were even underrepresented in positions of authority in the civil and security sector at the Jammu and Kashmir (J&K) state level where they are the clear majority. While Muslims accounted for 65% of the J&K population in the state around the outbreak of the Kashmir insurgency, they only held 13% of government jobs in the state and less than a quarter of the senior civil service slots. Meanwhile, Kashmiri Hindus, only 3% of the population were tremendously overrepresented in Uday Mahurkar, "IB Chief Does What the Congress Didn't Want, Counters CBI on Ishrat Case," India Today, June 14, 2013. 204 Author Interview, retired Indian intelligence officer, New Delhi, India December 11, 2013 (#142). 205 Shiv Pujan Jha, "Police Finds Muslims an Easy Target" CNN-IBN, March 18, 2006. 206 Social, Economic and EducationalStatus of the Muslim Community in India, 2006, p. 14. 207 Quoted in Times ofIndia, 6 December 1978 from Noorani, The Muslims ofIndia, 2003, p. 183. 203 367 the state government and had a "monopoly on professional placements."208 Similarly, in the mid 1970s, Kashmiri Muslims held only 30% of the senior police officer positions in the state.209 The same held for the security services as Kashmiri Muslims hardly constituted any of the paramilitary forces or even the J&K Light Infantry.21 0 Kashmiri leader Sheikh Abdullah noted in his memoirs that despite his request in the 1950s for the ban to be lifted on Kashmiri Muslims in the armed forces and for there to be adequate representation, a secret circular to recruitment officers had told them not to enlist Muslims and one General justified it by suggesting their loyalty was suspect.2 Furthermore, even in the one state where Muslims held some real positions of leadership, one governor explains they were always suspected and distrusted by the center. 212 Almost two decades into the counterinsurgency, one senior Kashmiri Muslim civil servant writes, that outsiders (non-Kashmiris), "made up the bulk of the senior bureaucracy."21 3 2.2.4. "Northeasterners" While Muslims suffered from status reversal, the numerous peoples and tribal groups ofindia's Northeast (excepting Assam) were largely excluded from any representation in government or bureaucracy as they were managed as a part of the Northeast frontier. After Indian independence, there was no serious attempt to include Northeasterners in any state institutions until the 1970s.2 14 In contrast with the STs of central India tribes (where the Naxalite insurgencies would take root), the border areas (Nagaland, Manipur, Tripura, etc) were judged to Quote from Paula R. Newberg, Double Betrayal: Repression and Insurgency in Kashmir, Washington DC: Carnegie Endowment for International Peace, 1995, p. 10. Also Bose, 2003, pp. 76, 121. 209 "I.A.S. as on 1.1.1981," Muslim India, 1 (1), 1983, p. 29. 210 On paramilitary, see Bose, 2003, p. 109. The J&K Light Infantry began adding Muslims after around the 1999 Kargil episode. Author interview with Kashmiri journalist, New Delhi, India, November 28, 2012 (#96). 21 2 12 Khalidi, 2003, p. 12. Rao, 2001, p. 397. 21 Wajahat Habibullah, My Kashmir: Conflict and the ProspectsforEnduringPeace, Washington DC: U.S. Institute of Peace Press, 2008, p. 70. 21 Dhar, 2005, p. 102. 20S 368 be more difficult to integrate and assimilate, as they were less likely to have frequent interactions with India since a Hindu population belt did not surround them.215 Excluded from these interactions, North Eastern tribals-specifically Nagas, Mizos, Meiteis-are regarded by the Indian state as a people apart from the rest of mainland India, as foreigners, noble savages, and the cat's-paw of a foreign Christian conspiracy to undue India. Group Worth Northeasterners were seen as distant, backward, disloyal, and morally degenerate. While physically a part of the Indian space, the people of the Northeast were "geographically, culturally and emotionally remote." 216 One of India's founding fathers, Ambedkar, noted that while tribal areas in interior India were "more or less Hinduised, more or less assimilated with civilization and culture of the majority of the people in whose midst they live," in the case of the tribals of the Northeast "that is not the case." 217 Many acknowledged, "the identity of mainland India often excludes the Northeast." 218 Besides referring to them as Northeasterners, mainland Indians often described them as Chinese or Bhutanese "as if to suggest a distinct, peculiar, and implicitly nonIndian category." 219 One former intelligence officer explains Nagas' "total incompatibility with the rest of India-culturally, linguistically and religion-wise,"2 20 and that "the people of Manipur were ethnically and culturally very different from the plains people of India." 22 1 A Naga representative decried how mainland Indians "are not prepared to accept [Nagas] as human Schermerhorn, 1978, pp. 88-91. 216 Shiraz Sidhva "Survey of India 5: Delhi Rules with Carrot and Stick," Financial Times, 30 September 1993, p. (Accessed on Factiva, March 5, 2014) 217 Verghese, 1996, p. 35. 218 Bijoyeta Das "Racism on the Rise in India," Asia Times Online, March 26, 2014. 219 Rakhee Kalita, "Writing Terror: Men of Rebellion and Contemporary Assamese Literature," Beyond CounterInsurgency: Breaking the Impasse in NortheastIndia, Ed. Sanjib Baruah, New Delhi: Oxford University Press, 2009, p. 103. 2 Dhar, 2005, p. 156. 2 E.N. Rammohan, 2011, p. 123. 215 369 beings" 222 and one police officer serving in Nagaland in the 1950s described how fellow officers, "considered these people [Nagas] as subhuman, filthy and not worth mixing [with]" creating "a big gap... between the Nagas and the government." 223 Many Northeast tribal groups were also cast as "childlike,", 22 4 "backward and exotic" depicted in three myths as "'unchanged,' 'unrestrained' and 'uncivilized.' 22 1 Stereotypes of Nagas as "head-hunters.. .hostiles and barbarians, to whom reason does not appeal" 226 and for whom "nonviolence is foreign to their thinking," 227 permitted the state's frequent resort to "force alone."22 8 Predating the politics of counterinsurgency, Indian state leaders often described Northeasterners as "spoilt child[ren]" justifying paternalism and periodic physical rebuke to discipline them.2 2 9 Considered second tier citizens, its little surprise that the Mizos were treated with such "indifference and callousness," that resulted in a devastating famine in the 1960s,231 or that leaders from Delhi treated local Manipuri ministers and elected officials "like dirt" and "did not consider the Meiteis (of Manipur) worthy of a genuine democratic administration and rapid economic development." 232 People from Nagaland, Mizoram, and Manipur were predominantly Christians and were viewed with added suspicion by the Indian state as "agents of foreign missionaries" conspiring 222 223 Glancey,2011,p. 181. Kanwar Randip Singh, The Nagas ofNagaland:Desperadoesand Heroes ofPeace, Deep & Deep Publishing: 1987, p. 194. 224 Baruah, DurableDisorder,2005, p. 77; Kanwar Randip Singh, 1987, p. 226. 225 Duncan McDuie-Ra, "The 'North-East' Map of Delhi," Economic and PoliticalWeekly, 47 (30), July 28, 2012, .6 1. 27 Harish Chandola, "The Much Maligned Nagas," The Times ofIndia, May 13, 1956, p. 8. 22 2 Kanwar Randip Singh, 1987, p. 226. Chandola, 1956. 229 Pages 171-172 of Sanjib Baruah, "A New Politics of Race: India and its North-east," India International Centre Quarterly, 32 (2/3), Winter 2005, pp. 163-176. 0 2 Baruah, DurableDisorder,2005, p. 10. 231 Nirmal Nibedon, Mizoram: The DaggerBrigade, New Delhi: Lancer Publishers, 1980, pp. 61, 127-28. 232 Dhar, 2005, pp. 102-104. 370 with the West to weaken India. 233 Or whose political demands were guided by a "foreign hand."2 34 Finally, Northeasterners "[were] cast as immoral" and "frequently stereotyped as morally loose women in skimpy skirts who are sexually available, or good-for-nothing men who are drug addicts or insurgents. "236 Stigma, prejudice, and disdain for Northeasterners as a group continue today in India's urban metropolitan areas, most notably in Delhi, where migrants from the northeast face "pervasive racial discrimination." 237 Physical and cultural features distinguish the Northeasterners from mainstream Indians, and their cultures behaviors are deemed by mainstream Indians to be more "permissive." 238 Northeasterners are frequently derided as "chinki, momo, Bahadur [a common term for Nepalese male servants in India], Nepali, chowchow, king-kong." 239 2 4 0 An estimated 86% of the estimated 200,000 Northeasterners in the capital city of Delhi have faced some sort of discrimination, often in the form of nonpayment of salaries or assaults 24 1 as well as housing discrimination 2 4 2 confining them to residential "ethnic ghettos."243 And many admit things are much better in Delhi than in other major cities like Bangalore and Mumbai.244 Such stereotypes then manifest themselves not only in overt racism but also discrimination, harassment, and violence. The prejudicial attitudes and absence of cultural 233 234 Nirmal Nibedon, 1978, pp. 53, 55 Chandola, 1956. 235 23 6 McDuie-Ra, 2012, p. 71. Bijoyeta Das, 2014. 237 Yengkhom Jilangamba, "Let's Stop Pretending There's No Racism in India," The Hindu, May 29, 2012. 238 Nilanjana Bhowmick, "What the Death of Nido Taniam Tells Us About Racism in India," Time, February 6, 2014. 239 Bijoyeta Das, 2014. 240 Bhowmick, 2014. 241 Bijoyeta Das, 2014; Kim Arora, "Five Throw Helpline to North East Bretheren," Times ofIndia, October 17, 2013. 242 Shaswati Das and Mallica Joshi, "Delhi Won't Lend a Home to Students from the Northeast," Hindustan Times, August 9, 2013. 243 Baruah, "A New Politics of Race," 2005, p. 166. 244 Arora, 2013. 371 understanding or representation has translated into disregard for their well-being. One Indian security analyst explained "people there don't matter to our elites."245 It has also translated into a lack of police sympathy and unwillingness to register or pursue charges of sexual violence against Northeast women in Delhi.24 6 The stark difference in treatment can be found in the case of the 2010 gang rape of a Mizo woman in Delhi that aroused little to no attention 2 4 7 and almost total police neglect despite eerie similarities to the December 2012 gang rape case of an Hindu woman, whose family hailed from Uttar Pradesh, which galvanized thousands in Delhi and throughout the country to protest for days, captured national and international attention, and sparked political reform.24 8 Group Embeddedness Due to exclusion policies inherited from the British and continued by the Indian government, for three decades from independence, Northeasterners were hardly represented, if at all, in the Indian state, save for some representation from Assam. Not until the 1970s did the Nagas began to gain access to some aspects of the state bureaucracy and security institutions and even these were still scant in numbers. 24 9 For the most part, local administrations were predominantly composed of outsiders,250 and even when locals were recruited to the state administration or armed forces, the Indian government cleverly recruited predominantly plains Author Interview, Indian national security Analyst, New Delhi, December 20, 2013 (#152). Imran Ahmed Siddiqui, "Shunned, Raped Mizo Girl Quites Delhi," The Telegraph, November 1, 2011. 247 After the rape of the Mizo woman, about 48 articles in Indian print and online papers made mention of the incident over the course of a year but after the Nirbhaya rape case in 2012, over 2000 articles mentioning the subject were published within a year, with the Times of India alone accounting for over 1200 articles mentioning it in print and electronic versions. Nexis searches performed May 19, 2014. 248 "Delhi Gang-rape: Unprecedented Protests Erupt," Press Trust ofIndia, Dec 22, 2012. 249 Dhar, 2005, p. 102 25 0 Dhar, 2005, p. 100; Rammohan, 2011, p. 123; Nibedon, 1980, pp. 127-28. 245 2 46 372 Hindus from the valleys of the northeast to manage the Northeast hills peoples.25 ' At the central level, even up to 2005, most Northeast states were significantly underrepresented in the more senior levels of the Indian civil service officers, 2 though some change was developing. Northeasterners were even less represented in the security forces. Data on officer corps and armed forces shows that recruits from Northeast states were consistently underrepresented through the 1990s. 2 54 During the British colonial administration, a commission recommendation to recruit more Nagas and other hill tribes into the armed forces went unimplemented and contributed to the systemic under recruitment of these groups into the armed forces.255 The Assam Rifles, the paramilitary force that predominantly operates in the Northeast and would be the likely site of Northeastern recruits has only a small portion actually from the region. In fact, 40% of the Assam Rifles are Nepalese Gorkhas, again a historical artifact of British recruitment, while much of the rest is composed of small ethnic groups not from the Northeast but rather from Uttarakhand and Himachal Pradesh including Garhwalis, Kumaonis, and Dogras.2 5 6 Even when the Naga regiment was created in 1970, it consisted of only 1 battalion, half of which was not even composed of Nagas, 25 7 and even by 2001 with the addition of a second battalion, Naga Rammohan, 2011, p. 122. Dalal Benbabaali, "Questioning the Role of the Indian Administrative Service in National Integration," South Asia MultidisciplinaryAcademic Journal, September, 2008. 253 Ashish Sinha, "UP and Bihar India's Top Two IAS Churners," India Today, Feb 1, 2010. 2 P.K. Gautam, Composition and Regimental System ofthe Indian Army: Continuity and Change, Delhi: Shipra Publications, 2008, pp. 34, 43; Lok Sabha Unstarred Question No. 601, dated 21.11.2002 from Indiastat. 2 Stephen P. Cohen, The Indian Army: Its Contributionto the Development ofa Nation, New York: Oxford University Press, 1990, p. 70. 256 A.K. Sharma (Lt. Col.), "The Assam Rifles," IPCS brief #638, November 22, 2001; A.K. Sharma (Lt. Col.), "The Assam Rifles," Indian Defence Review, 16 (4), 2002; A.K. Sharma (Lt. Col.), "The Assam Rifles," BharatRakshak Monitor, 4 (5), March-April 2002. 2s7 Vidya Prakash Tyagi, MartialRaces of Undivided India, New Delhi: Kalpaz Publications, 2009, pp. 115-118; "The Naga Regiment" Bharat Rakshak. 2 252 373 leaders expressed concerns that ethnic Nagas did not compose a sufficient proportion of the regiment. 258 The Assamese people may be something of an exception to the rest of the people from the Northeast because of their long history of assimilation into the Indian nation, including their linguistic rootedness in Sanskrit, the Hinduization and cultural meld of the region, and intermixing between Tibeto-Burman and Aryans from the Indo-Gangetic plains.25 9 Though they still remain ethnically distinct, their identity is relatively closer to the mainland and they have a higher degree of representation in central bureaucracy and security forces, certainly within the state of Assam. 2.2.5. Distinctions from EPR Index Similar to the recent ethnic power relations index (EPR), 2 60 I try to measure the relative embedded-ness and strength of ethnic groups, based on multiple sources and types of data, and assign a value for each relevant group per decade. Though there is substantial correspondence, when these values sometimes differ from the EPR, my qualitative evidence provides extensive support for the assigned values. 2 6 1 For instance, the EPR overestimates the power of SC/ST through the 1980s when they were formally empowered rather than when they substantively began to fill government quotas and achieve a degree of proportion. It also overestimates the power of Indian Muslims and Kashmiris until the mid-1980s when in fact these groups were disempowered throughout nearly all of India's post-independence history. Similarly, it treats 258 "Recruit More Nagas in Naga Regiment: Jamir" NortheastNews Agency (Oriental Times), 3 (49-50), May 7-2 1, 2001. 259 Harekrishna Deka, "The Assamese Mind: Contours of a Landscape," India InternationalCentre Quarterly, 32 (2/3), Winter 2005, pp. 189-202. 260 261 Cederman et. al., 2010. The EPR score for a country is based on a single expert's judgment, specifically weights formal rather than substantive power, and is admittedly limited to the executive branch rather than other institutions like the civil and security bureaucracies. 374 groups like the Nagas and Manipuris as empowered after being accorded some formal regional autonomy even if it was substantively meaningless due to military occupation and political subversion. It appears the EPR-measured change in a group's power or value is largely endogenous to conflicts that emerge between them and the state but are not based on any ex ante measures. In other words, the EPR assumes Kashmiris are empowered as a group until a conflict reveals they are not, whereas a closer study of the politics shows that their marginalization since 1953 is what produced the conflict in the first place. Admittedly, the measurement of group worth and embedded-ness is crude and imprecise but it is far more grounded than the EPR in systematic historical, ethnographic, and sociological research and data. For instance, a 1974 comparative study of political inclusion and alienation involved a survey of over 1000 voting males across four states. Of the four social groups, which the EPR categorizes as powerful senior and junior partners, only Sikhs and Hindus held high levels of "insider" status but Muslims and scheduled castes were in fact "outsiders." 2 62 2.3. Predictions Based on my core-periphery theory of how territory and identity shape state counterinsurgency strategy and combining the measures of the independent variables in the Indian cases, I am able to generate a set of predictions. I expect the Indian state incumbent to employ strategies of higher effort in the regions like Punjab, Kashmir, and West Bengal, and lower effort in much of the Northeast and the Naxal belt. Furthermore, I expect the Indian state 262 The study and categorized respondents status based on his political potency, acceptance of norms and institutions, and political involvement and yielding eight categories or "types" citizenry ranging from insiders to outsiders. Sikhs had the highest level of insiders (31.9%), Hindus (19.2%) quite close to the Indian mean (20.6%), but Muslims and Scheduled castes had relatively low levels of insiders (12.6% and 7.8% respectively). And while Hindu (18.8%) and Sikh (18.1%) levels of "outsiders" fell below the Indian mean (19.4%), Muslims (27.2%) and scheduled castes (31.8%) had levels of outsiders well above the mean. See Nandy, 1974. 375 to rely on strategies of higher violence in conflicts against Indian Muslims, most Northeasterners, and SC/STs pre-1990 in the decades prior to their substantive social and political integration. The state would be likely to employ less violent and more non-kinetic strategies against Sikhs, Assamese and post-1990 SC/STs. The prediction then is for attrition (high effort, high violence) in the cases Hyderabad, Telangana, and Kashmir, population security (high effort, low violence) in the case of Punjab and Naxal I-West Bengal, enfeeblement in the cases of Nagaland, Mizoram, Manipur, and Naxal I-Andhra Pradesh, and mitigation in the cases of Naxal II (Andhra), Assam, and Naxal III. Figure 6.5: Distribution of Predicted Strategies Territory Cost-Benefit High Low Attrition Enfeeblement Hyderabad Nagaland I Telangana Mizoram Low Kashmir I Naxal I - AP Kashmir II Nagaland II Identity Worth and Manipur Embedded Mitigation Population Security Naxal I - WB Naxal II - AP High Punjab Assam Naxal III III. Coding the Dependent Variable: Strategv in Indian Counterinsurgency Cases In order to evaluate the predictions generated by the core-periphery theory of contested territory and the rebelling identity group, the next step is to carefully assess the values of the dependent variable (i.e. what strategy was implemented) for each observation of Indian counterinsurgency. Strategy is not easily ascertained with dummy or off-the-shelf variables and requires more careful scrutiny of the campaign for accurate identification. The following section achieves that by meticulously wading through the historical literature on each campaign 376 combined with extensive field research, interviews, and data collection in order to describe the strategy for each of the 14 Indian counterinsurgency campaign. Organizing the counterinsurgencies by region (Central India, Northeast, North East), I provide a narrative background and account of each campaign seeking to describe the levels of effort, violence, dominant tactics, and type of political authority wielded. I then use these accounts to specifically measure effort and violence for each strategy in qualitative terms and assign each a value for comparison. Finally, I seek to confirm each coding of strategy using observable quantifiable proxies for effort and violence collected over the course of this research. 3.1. Counterinsurgency in Central India/Naxal Belt 3.1.1. Hyderabad (1948): Enfeeblement Background Though some have treated this as an extra-territorial conflict, Hyderabad was the first rebellion in the eyes of the Indian state as "the Delhi Government regarded it, and treated it, as a subordinate of India." 263 This lesser known conflict has been described at times as an extra-territorial conflict rather than a rebellion but other civil conflicts early on India's periphery (Telangana, Nagaland, Mizoram) were also viewed by the weaker side as invasions rather than counterinsurgencies. Though the Muslim kingdom was under British colonial dominion, it was administered indirectly like many territories in the subcontinent but India rejected its claim to sovereignty. 2 64 After the Nizam (monarch) of Hyderabad refused to formally accede to the newly independent Indian state in 1948, made the "astounding assertion" of independence, 2 65 and supposedly sanctioned a "Razakar insurgency of militant Muslims" 2 66 to 263 Wilfred Cantwell Smith, "Hyderabad: Muslim 264White Paper on Hyderabad,New Delhi: 265 White Paperon Hyderabad, 1948, p. Tragedy," Middle East Journal, 4 (1), January 1950, p. 43. Government of India, 1948, p. 12. 2. 377 conduct raids and incursions into neighboring provinces, partly in collaboration with the Telangana insurgents, 2 67 the state opted to take military action. The Indian Army launched its first internal military operation with modest resources but heavy violence that amounted to an enfeeblement campaign. Nehru, consumed with events in Kashmir, deferred management of the conflict to the second most powerful Indian leader, Home Minister, Sardar Patel, an avowed Hindu nationalist, and his Hindu communalist agent, K.M. Munshi, who both harbored great disdain for the "alien" Hyderabad culture.26 8 Effort. The Indian Army launched Operation Polo in September 1948 with a number of armored brigades totaling around 30,000 troops and made quick work of the kingdom of Hyderabad's small and poorly trained armed forces of 22,000 - 40,000, another 25,000 - 30,000 police and irregulars and an additional 60,000 - 200,000 strong Muslim Razakar irregulars (but only 20% with modern small arms). 269 The conventional campaign lasted five days, with the Nizam of Hyderabad surrendering on September 23, 1948, but a six-week clean-up operation followed as well as a military occupation for almost two years.270 Indian military forces proved sufficient to defeat Hyderabad's forces in a conventional fight, and between a brigade and a division remained in place in the aftermath of Operation Polo in the region expanding over 200,000 square kilometers. While much of the surrendered Hyderabad security forces were ostensibly tasked with maintaining public order under a military 266 The Razakars were an irregular militia. Quotation from Maya Chadda, Ethnicity, Security, and Separatism in India, New York: Columbia University Press, 1997, pp. 34, 42. 267 Malik, 1972, pp. 242, 247; Sen, Slender Was the Thread, 1969, p. 13. Border incidents described in White Paper on8Hyderabad, 1948, pp. 58-75. 26 A.G. Noorani, "A.G. Noorani Blames Patel for Downfall of Hyderabad," Deccan Chronicle, November 30, 2013; Muralidharan, 2014: 129; Sukumar Muralidharan, "Alternate Histories: Hyderabad 1948 Compels a Fresh Evaluation of the Theology of India's Independence and Partition," History and Sociology ofSouth Asia, 8 (2), 2014: 119-138. 269 Figured from J.N. Chaudhuri (Maj. Gen), I ArmoredDivision in Operation "Polo", New Delhi: Indian Army Headquarters, 1949, pp. 14-15; Mohan Guruswamy, "There Once Was a Hyderabad!" India Seminar, 585, 2008; 2and "Hyderabad Police Force to be Reorganised," Times ofIndia, Nov. 8, 1948. 1 Malik, 1972, p. 254. 271 J.N. Chadhuri, 1949, pp. 64-65. 378 administration, the state police forces were quite broken opening up the vacuum for tremendous communal violence. 2 7 2 While the Hyderabad kingdom (primarily the city) was a valuable asset to secure given its wealth 2 7 3 and its location in the center of India posed a grave strategic threat so long as hostile forces remained while the state was in its infancy, 2 74 the Indian leadership regarded the rest of the kingdom as "a land of stagnant back-waters" and "the most backward amongst the Indian states" at the time.2 75 While the state helped to resettle refugees driven out by the fighting, 2 76 it did little to restore the political order it had destroyed or consolidate the state's writ in Hyderabad's rural, outer reaches for quite some time.2 7 7 Only limited Indian military forces remained for at least six weeks to target and finish off predatory bands of Razakar insurgents,2 78 but the majority of the forces were soon moved to Telangana to contend with the Communist insurgency. 279Even then, military forces were primarily focused on preventing an anti-Hindu uprising in Hyderabad and targeting Razakars, while Hindu irregulars organized in at least 17 militant camps across the border and unleashed by the Congress party (then intimately intertwined with the Indian state), to take revenge on Razakars and other Muslim civilians.28 0 Cantwell Smith, 1950, pp. 46-47. Carolyn M. Elliott, "Decline of a Patrimonial Regime: The Telengana Rebellion in India, 1946-51," Journalof Asian Studies, 34 (1), Nov. 1974, p. 38 274 White Paper on Hyderabad, 1948, P. 1, 7-8. 27s "White Paper on Hyderabad, 1948, p. 3. 276 "Hyderabad Refugees Problem, Government Plans for Resettlement," Times ofIndia, Nov 8, 1948, p. 7; "900,000 Hyderabad Refugees Return to State, Government's Plan for Speedy Rehabilitation," Times ofIndia, Nov 29, 1948. 277 Cantwell Smith, 1950, pp. 46-47; Muralidharan, 2014, p. 135. 278 "Hyderabad Relieved of Acute Shortage of Essential Goods: Second Phase of Military Government Begins," Times ofIndia, Nov 7, 1948, p. 3. 279 J.N. Chadhuri, 1949, pp. 64-65. 280 Muralidharan, 2014, pp. 133-136. 2 273 379 .- Violence. While the initial conventional operation was believed to have only killed 807 Hyderabad Army, 1373 Razakars, and 10 Indian Army, 2 8 1 mass violence motivated by anger and disdain for the Hyderabad Muslims soon followed. The state accused the Nizam's Razakar irregular forces of atrocities, when they may have been attempting to aggressively protect the Muslim regime and population from Hindu militia incursions. Furthermore, Muslims were no longer a core component of the Indian Army after partition, and Deccani Muslims (from the Hyderabad region) had purposefully been reduced in the Indian Army starting in 1932 due to British concerns about their quality and stamina.28 2 Hindu irregulars led a wave of communal and retributive violence "in the wake of the swift Indian advance" 283 for weeks after the conventional operation, resulting in the deaths of perhaps 50,000 - 200,000 Muslims (9% of the Hyderabad Muslim population2 84), including 10 20% of the adult male population, particularly concentrated in the Marathwara region of the state. 28 5 Violence certainly targeted non-combatants, potentially because even women and children were considered Razakar irregulars. 286 At the time, the state issued denials claiming, "the talk of genocide of Muslims is absolute nonsense" 287 despite the violent assault on the Muslim and Razakar population continuing for months. 2 88 However, a government taskforce dispatched by Nehru to investigate reported to the contrary though this Sunderlal commission report was never released and actually buried. Leaked accounts suggest that after being disarmed Mohan Guruswamy, 2008 says another 1647 regular Hyderabadi Army and 1911 Razakars were captured. Less is estimated by J.N. Chaudhuri, GeneralJN Chaudhuri:An Autobiography, Written (through narration) by BK Narayan, New Delhi: Vikas, 1979, p. 154. 212 J.N. Chaudhuri, 1949, p. 17. 283 Mohammed Hyder, October Coup, A Memoir ofthe StruggleforHyderabad, Roli Books, Delhi, 2012, p. 79. 284 13% of the 17 million Hyderabadi popuation were Muslims. 285 Cantwell Smith, 1950, pp. 46-47. The Marathwada region is now the Aurangabad division of the Indian state of Maharashtra. 286 White Paper on Hyderabad, 1948, p. 29. 287 "Hyderabad Settling Down Rapidly," Times ofIndia, Nov 25, 1948, p. 1 288 Claude Emerson Welch Jr., Anatomy ofRebellion, Albany: State University of New York Press, 1980, p. 289. 281 380 by the Army, much of the Razakar forces and about 27,000 - 40,000 Muslim civilians were killed throughout the region due to Hindu paramilitary groups, mob violence, and even the participation of the Indian security forces, all with the tacit acceptance of the Indian state and Congress Party. 289 This may have been motivated in part by suspicion of Muslims as a "fifth column," fears of prolonged guerilla warfare, and a lust for revenge. 2 9 0 There was evidently little effort by the Indian state to politically incorporate the Razakar base, actively police the region it had militarily occupied, stem the revenge killings, or spare thousands of civilians. 3.1.2. Telangana (1948-1951)291: Attrition Background. Immediately following the Hyderabad operations and massacre, the Indian military moved on to the Telangana region of the state of Hyderabad where a peasant rebellion had begun in 1946 against oppressive feudal landlords. After independence, the Indian state begrudgingly deferred to the Nizam to manage the rebellion lest his unseating unleash further social breakdown, but after his defeat, the Indian military began a three year campaign to stifle and defeat the Telangana peasant rebellion which had made common cause with Communist Party of India (CPI) and managed to establish a liberated zone of 2,000 - 4,000 villages primarily concentrated in 3 districts in the Telangana region. The upper castes had joined the William Dalyrmple, The Age ofKali: Indian Travels and Encounters, London: HarperCollins, 1998, p. 209-10; Mike Thomson, "Hyderabad 1948: India's Hidden Massacre," BBC Magazine, September 23, 2013.; SA Aiyar, "Declasify Report on the 1948 Hyderabad Massacre," Swaminomics Blog, Times ofIndia, Nov 25, 2012. The Telangana rebels knew the Congres party irregulars (goondas) collaborated with official state forces and later told their cadres "Do not spare the Congress goondas. What the military and police do not do the brutes (Congress goondas and spies) have agreed to do." See Mullik, 1972, p. 284. At this time, the Congress party was basically interoperable with and indistinguishable from the state itself. Sundarayya notes "Congress volunteers acted as the special intelligence branch of the military" and Congress offices were established next to every military camp and Congress volunteers closely collaborated with the military in Telangana. See P. Sundarayya, "Telangana People's Armed Struggle, 1946-195 1. Part Three: Pitted Against the Indian Army," Social Scientist, 1 (9), April 1973, p. 32. 290 Cantwell Smith, 1950, p. 42; J.N. Chaudhuri, 1949, pp. 15-16; Muralidharan, 2014, p. 135. 291 Mullik, 1972; Purshottham and Kennedy, 2012; P. Sundaryya, TelanganaPeople's Struggle and its Lessons, Calcutta: Chadha, 1972. 289 381 rebellion in response to the excesses committed by Razakar militias but quickly shifted sides by late 1948, as they became the targets of the Communist-Maoist rebellion. Both the intent and the conduct of the campaign resembled a strategy of attrition. IB Director Malik was instructed by Home Affairs Minister Rajagopalachari to develop a response "not only to contain but to completely eradicate this menace of armed struggle" from Telangana, and was given access to whatever resources he needed to muster the requisite effort.2 92 He followed through by exerting "grinding pressure" and "severer punishment" on the rebels until "they had no stomach for more," forcing the insurgents to capitulate and call off the rebellion.2 93 Though it was artificially cut short with the cessation of insurgent activity, the state had begun an administrative project in the "Telangana Special Area" to involve greater reorganization of society through development, policing, agrarian reform, and social integration with medical and educational facilities in a project of political usurpation.2 94 Effort. The Indian Army imposed a military regime on the area, launched "a massive counterinsurgency offensive" 295 with 50,000 troops to defeat the rebellion, and enlisted the help of numerous caste allies to function as auxiliary forces. As one of the first military and policing operations, this campaign required tremendous organization and leadership to coordinate so many disparate forces from different regions unaccustomed to operating together or at all.2 96 The state employed a sustained campaign of mass arrests, encirclement raids, targeting of rebel leadership as well as cadres' families, cordon and search operations, and forcible relocation of forest peoples (adivasis). This last aspect in particular forced the military to adapt out of its Mullik, 1972, pp. 277-78. 293 Mullik, 1972, pp. 60, 253, 285-86, 292-93. 294 Kennedy and Purushotham, 2012, p. 842 295 Kennedy and Purshotham 2012, p. 839. 29 6 Mullik, 1972, p. 276. 292 382 comfort zone and develop a model based on the Briggs resettlement plan used in Malaya.29 7 Some allege, "more than a million peasants were arrested and tortured." 2 98 Though likely an exaggeration, even if the actual scale of activity this was an order of magnitude less, it would still suggest significant efforts to gain control of the contested region. A new Home Minister took a direct, hands-on approach beginning in 1950 and committed to limitless resources to "not only to contain but to completely eradicate" the Telangana insurgency within a year.2 99 Violence. It is evident however, that this campaign was also quite brutal and harsh on the population. Resettlement plans involved moving peasant civilians into barebones, roadside camps where thousands died without sufficient access to food, water, or medical aid.30 0 Eschewing the use of police due to a lack of trust, the state instead purposefully chose the military, specifically because it would not be encumbered with concerns about the use of violence and could pursue a fully its offensive approach. 3 01 This was augmented by thousands of armed vigilantes or village "defense" squads who were enlisted or forcibly created to hunt and kill suspected insurgents.3 02 Still, the state lacked intelligence or language skills to selectively identify insurgents or elicit defections and had to rely on blunt force.3 03 Additionally the state emphasized killing rather than capturing insurgents, and encounter killings as well as disappearances believing it could crush the rebellion with brute force.3 04 Though the Indian state had intervened in Hyderabad ostensibly due to the heavy-handed approach of the Nizam, in roughly the same amount of time, the state appears to have killed at least twice the number of rebels and civilians as the Nizam's regime and imprisoned five times 297 298 299 300 301 302 Trinath Mishra, 2007, p. 17. Biplab Dasgupta, The Naxalite Movement, Bombay: Allied Publishers, 1974, pp. 270-71. Mullik, 1972, pp. 277-78. Sundarayya, 1973; Kennedy and Purushotham, 2012, p. 843. Sundarayya, 1973, part III, p. 34. Kennedy and Purushotham, 2012, 843. 303 Mullik, 1972, p. 253. 304 Kennedy and Purushotham, 2012, pp. 843-44. 383 the number of people. 3 05 A great deal has been made of the political accommodations during the campaign but these were piecemeal at best. The state formally disbanded the largest landholdings, rewarded mid-level landlords who collaborated with the state, but offered little to address the grievances of peasant base of the rebellion, since many of the promised developmental measures were never actually implemented.3 06 3.1.3. Naxalite I (1967-1971) Background Though the Telangana uprising was firmly dealt with early on, the state did not do enough to tackle the conditions that had given rise to the violent political mobilization, especially after the Communist party basically conceded. The same motives of the Telangana uprising remained and a new organization, would tap into those grievances to strike up a rebellion in many of the same regions less than two decades later. The Naxalite rebellion received its name after a spark was lit by a peasant uprising in Naxalbari, West Bengal in 1967. Following outrage by radical communists, the CPI-Marxist-Leninist (CPI-ML) party formed calling for armed revolution. Small pockets of class-based resistance led by Marxist ideologues but composed of rural SC/ST inhabitants emerged throughout the rural backwaters of the Red Corridor including Bihar, Orissa, Punjab, and Uttar Pradesh, but the major episodes were concentrated in northern Andhra Pradesh and West Bengal between 1969-71. In these two Cantwell Smith says the casualties of Muslims (est at 50,000) was ten times what the Nizam's regime perpetrated and since the Indian government is believed to have killed at least 10,000, the number of killings is doubled. The figure for arrests comes from P. Sundarayya book, 1972: 313 who states that 12,000 were jailed under the Nizam but 50,000 under the military regime. 306 Kennedy and Purushotham, 2012, pp. 842-44; Carolyn M. Elliott, "Decline of a Patrimonial Regime: The Telangana Rebellion in India, 1946-1951," JournalofAsian Studies, 34 (1), November 1974, p. 44; Grossman, 1992, p. 8. 305 384 regions, the central government deployed different strategies and dispatched varying levels of police, paramilitary, and even Army reinforcements as individually described below. 3 07 Figure 6.6: Map of Affected Areas in Naxalite I rebellion (1967-71) I Affected Intense Conflict 4 Naxal I - West Bengal (1967-1971): PopulationSecurity Background In West Bengal, the Indian state employed significant effort, particularly relative to the rest of the Naxal theater, and also a more restrained use of force that relied on local police and looked more like violent negotiation rather than the state unilaterally dispensing violence and public goods. Here, the state had much more to lose in Bengal, one of the most Mishra, 2007; Jennifer L. Oetken, "Counterinsurgency Against Naxalites in India," India and Counterinsurgency:Lessons Learned, Ed. Surnit Ganguly and David P. Fidler, New York: Routledge, 2009, pp. 127-151. 307 385 valuable states at the time with an urban population 25% higher than the Indian average, specifically owing to its megalopolis of Calcutta.3 08 The presence of greater social and economic mobility within Calcutta that reduced social divides, 3 09 and the participation of urban, middleclass Bengali students composing a large part of the insurgency in this region,3 10 also made the state more restrained in its use of force and reliant on both amnesties of insurgents and larger political efforts to redress grievances and devolve authority. 31' Having read Robert Thompson's treatise on counterinsurgency and circulated it amongst his men, the Calcutta Police Commissioner believed that police actions needed to be complemented with civil actions.3 12 In addition to local Bengali police forces, the Indian state deployed around 45,000 paramilitary around the city of Calcutta, which had become the target of urban terrorism,3 1 3 and dispatched three Army divisions (another 45,000 troops) to the rural regions of Bengal in Operation Steeplechase to provide an outer cordon for paramilitary and police actions to seize weapons and arrest Naxals.3 14 Insurgent targeting of Calcutta had been criticized by many other CPI-ML leaders who believed attacks on one of the most highly prized urban centers was a major mistake. 3 15 They were proved correct as the Indian central government invested substantial material support to the campaign. This included raising police force modernization spending by fourteen times the initial rate over two years,316 as well as enhancing the size and budgets of On the economy, see Bandyopadhyay, 201; On population, at the time, the Indian average was 19.9 1%, while in West Bengal it was 25%. See Registrar General of India, and Ashok Rudra, "Urban Guerrillas in Calcutta," Economic andPoliticalWeekly, 6 (28), 1971, pp. 1379-82. 3 09 Rudra, 1971, p. 1380. 310 Kennedy and Purushotham, 2012, p. 847. 311 Sumanta Banerjee, India's Simmering Revolution: The Naxalite Uprising, 1984, pp. 279-286. 3 Sumanta Banerjee, 1984, p. 279. 313 Oetken, 2009: 135 3 14 Prakash Singh, The Naxalite Movement in India, New Delhi: Rupa, 2005, p. 100; Sujan Dutta, "Clamour to Let Army Fight Maoists Rises," The Telegraph, Sept. 27, 2009. 315 Prakash Singh, 2005, p. 105; Ajay K. Mehra, "Naxalism in India: Revolution or Terror?" Terrorism andPolitical 12 (2), 2000, p. 47. 3Violence, 16 Moderinization of police force (MPF) data shows a jump from 50 lakhs in 1969 to 700 lakhs in 1971, and though this was police modernization throughout the country, given the rate of activity in West Bengal it stands to reason 308 386 paramilitary forces, upgraded vehicles, small arms, and wireless sets, intelligence coordination, the enactment of new terrorism laws, sustainment of curfews, and house to house cordon and search operations. 317 Furthermore, the more left-leaning state government in Bengal specifically harbored much greater sympathy for the Naxal movement, especially as urban middle class students played a much larger role, thus restraining the state to deploy more selective violence and nonkinetic tactics. The West Bengal police often refused to conduct operations against people with whom they had connections or to provide intelligence to Army troops to target guerillas.3 1 8 Even when the military deployed, they did not fire a bullet in the area domination exercise and instead deferred to more discriminate, local forces.3 1 9 Instead, the state sought to demobilize insurgents through non-violent means by offering formal jobs and protection to students, as well as informal deals and salaried positions as home guards to criminal opportunists who had joined the insurgency, enabling a more selective kinetic focus on the hardcore Naxals.320 It also sought to protect the Naxals' class enemies, though this may have redirected attacks on political rivals,32 1 and armed self-defense squads in rural areas to help protect populations weary of a Naxal annihilation tactics. 322 that the lions share was allocated here. See Rajya Sabha Unstarred Question No. 810 (11/27/2002) & Rajya Sabha Unstarred Question No. 1771 (8/17/2011). Accessed on Indiastat.com 7/9/14. 317 Kennedy and Purushotham, 2012, pp. 849-50; "More Intensified Anti-Naxalite Drive in Bengal," Times ofIndia, October 5, 1970, p. 9; Banerjee 1984, pp. 275-76. 318 Oetken, 2009, p. 132. 319 Shanthie Mariet D'Souza, "Naxalites: Is there a need to 'bring in' the Army?" JournalofDefence Studies 3 (3), July 2009, pp. 125-32. 320 Kennedy and Purushotham, 2012, pp. 849-51; Banerjee, 1984, pp. 284-285. 321 Kennedy and Purushotham, 2012, p. 847. 322 Oetken, 2009, p. 135; Mehra, 2000, p. 50. 387 Broader "sincere political initiative[s]" like bank nationalizations, development efforts, and debt cancellations were attempted 3 2 3 but land and tenancy reforms played a significant role and West Bengal was one of the few states to successfully institute these reforms to redress the root grievances of the uprising.3 24 Finally, instead of offensive violence, the state focused on denial tactics like imprisonment. An estimated 32,000-50,000 Naxals were imprisoned, and an estimated 18,000 in West Bengal, 3 2 5 but more than 90% had been released by 1974.326 Many accounts tend to overstate the reliance on repressive tactics, in part because of the way in which the insurgent movement seemed to rapidly shatter, but insurgent collapse primarily stemmed from rebel strategic mistakes rather than significant state repression. 3 2 7 Naxal I - Andhra Pradesh(1967-1971): Enfeeblement Background. Though the Naxal conflict stretched across numerous states, beyond West Bengal, it only seriously escalated past the "civil war" threshold in Andhra Pradesh. Elsewhere, the Naxal movement either did not escalate or Indian state did not seriously contest it, mostly relying on uncoordinated and disparate responses by local government or private actors to mitigate the uprisings or "disrupt" the Naxal organization for short periods. 32 8 In Andhra Pradesh, the conflict escalated due to its prior experience and organization of class-based mobilization against the state in the Telangana region, and the state took slightly Banerjee, 1984, pp. 281-83; Mehra, 2000, p. 46; Prakash Louis, People Power: The Naxalite Movement in CentralBihar, New Delhi: Wordsmiths, 2002, p. 57. 324 Kennedy and Purushotham, 2012, p. 851. 325 Oetken, 2009 says 32,000. The Ministry of Home Affairs Annual Report (1971-72) says 35,000. Gough says 45,000; Boyini says 50,000 (2010, p. 61). See Kathleen Gough, Indian Peasant Uprisings, Economic andPolitical Weekly, 9 (32/34), August 1974, pp. 1391-1412. Clodfelter, 2008, p. 1095 estimates 2,000 killed and 20,000 imprisoned. 326 The Govt. stated this in 1974. See "Report in Detention of Extremists Refuted," Delhi Domestic Service, September 17, 1974, Daily Report, Middle East & North Africa, FBIS-MEA-74-183 on 1974-09-19, p. U2. Accessed via Newsbank, February 19, 2014. Others report most were released by the end of the Emergency in 1977. See Kennedy and Purushotham, 2012, p. 850; Mehra, 2000, p. 47. 327 Kennedy and Purushotham, 2012, p. 850. 328 Prakash Singh, 2005, p. 100 323 388 more notice and offensive action in the form of an enfeeblement strategy. In still mostly rural backwater of Andhra, the Indian government did considerably less than in West Bengal, only sending a relatively small 10,000 paramilitary troops for a short period of a few months. Again the state relied on similar offensive tactics as those used in Telangana from 1948-51 including the razing of villages, brutal regrouping of the population, encounter killings, and the targeting of leadership, with little meaningful political-economic reform introduced to correct the imbalance between low-caste agricultural workers and high caste landowners. 329 Nearby in the Gond region of Madhya Pradesh, the state deployed the police to kill scores of adivasis.330 Ultimately the state managed to declare victory and withdraw after the Naxals collapsed with the capture and death of the CPI-ML leader Charu Mazumdar in 1972. The insurgent collapse was rapid and inevitable given the anarchical and visionless movement that had prematurely taken on too many class enemies.33 1 However, state forces left without dismantling the insurgent groups or changing the political conditions that gave rise to them. Scholars judge that the counterinsurgency campaign "did damage the organizational capacity of the insurgents in northern Telangana but.. .it failed to defeat them" allowing dalams (guerilla squads) to remain and soon regroup into a new insurgent organization, the People's War Group (PWG), a few years later due to continued exploitation of tribals with tacit government participation.3 32 329 Kennedy and Purushotham, 2012, pp. 848, 851. 330Mohan Guruswamy, "Armed Might Won't Defeat the Naxals," Rediff com, January 3, 2008. 331Oetken, 2009; Mehra, 2000; Gough, 1974, p. 1404; Kennedy and Purushotham, 2012, pp. 849-50. Kennedy and Purushotham, 2012, pp. 848-849; Prakash Singh, 2005, pp. 130-13 1. 389 3.1.4. Naxal H-Andhra Pradesh (1987-2003): Mitigation Background. After the first Naxal movement fractured by 1972, federal government policy also fractured "veering from one response to the other." 3 3 3 The Indian government punted the problem to individual states and "continued to treat the Maoist violence as a minor problem, to be handled by police with the traditional methods."3 34 While land reforms were implemented in West Bengal under a communist-leftist political front in 1977, others like Andhra and Bihar never embraced these reforms and became centers for Maoist regrouping and organizational development. In the late 1970s-80s, the insurgent movement was reorganized under the People's War Group (PWG) in Andhra Pradesh and the Maoist Communist Center (MCC) in Bihar. Significant violence did not intensify until 1987 33 prompting serious action by some state governments but only modest federal paramilitary support. The central government continued a policy towards these peripheral zones of defacto nonincorporation into mainstream administration. 336 Effort. Once again, the central government failed to deploy meaningful resources and deferred to the states to manage the insurgency. With the exception of Andhra Pradesh, most states neglected the problem, invested next to nothing to combat it, and relied on private forms of violence33 7 while the center was hesitant to provide meaningful resources. 338 Andhra's state 33 Arvind Verma, "The Police and India's Maoist Insurgency," PolicingInsurgencies: Cops as Counterinsurgents, Ed. C. Christine Fair and Sumit Ganguly, New Delhi: Oxford University Press, 2014. 33 4 Verma, 2014 33 1987 is identified as a significant spike in PWG activity. It also corresponds with the first year the conflict crossed the "100 battle dead" threshold that is the industry standard amongst civil conflict scholars as the onset of a civil war. 336 "Center Warns States About Naxalite Activities," Delhi Domestic Service, March 2, 1988, Daily Report. Near East & South Asia, FBIS-NES-88-041 on 1988-03-02 p. 54, Accessed via Newsbank Feb 19, 2014. 37 For instance, in Bihar, the class based militias like the Ranvir Sena were mobilized by landowners and elites to terrorize and punish pockets of Naxal/MCC support. See Louis, 1992. 338 "Andhra Chief Differs with Delhi on Naxalites," Delhi Domestic Service, September 3, 1989, Daily Report. Near East & South Asia, FBIS-NES-89-170 on 1989-09-05 p. 50, Accessed via Newsbank Feb 19, 2014. 390 response was sporadic and inconsistent, "mired in stark vacillation" 339 between appeasement and allowing the Naxals "utter freedom" 340 and occasional punitive measures, and eventually displacement of insurgents. Moreover, the resources at its disposal were limited. Many districts were left without police chief or promised economic programs, 3 4 1and the state remained "severely under-equipped and undermanned." 342 Neither of these tracks was ever afforded the opportunity for a sustained campaign to consolidate gains. The state government were neither able to sufficiently deal a knockout blow to the Naxals militarily, nor to invest in sufficient policing forces to restore law and order in areas where "the state [did] not exist." Andhra leaders were also unable to implement the requisite land reforms to undermine the Naxal political support and transform its political authority in rural areas from the type of colonial governance that involved heavy "corruption and criminalization of political leadership." 3 4 3 Even when the elite commando force of greyhounds was able to effectively target Naxals in Andhra Pradesh, because of an uncoordinated central government policy, the most they could do is displace them into neighboring states for periods of time, resembling a game of whack-amole.344 Even after a number of greyhound successes, one well-regarded analyst lamented this a model of "containment" offering temporary rather than permanent fixes without considerable, sustained and coordinated investments of administrative capital, policing capability and manpower, and development which the state had yet to initiate.34 5 339 PV Ramana, "Vacillating Andhra Naxal Policy," IPCS brief # 1728, May 4, 2005. 340 Ajai Sahni, "'Naxalism': The Retreat of Civil Governance," Faultlines, 5, 2000. 341 PV Ramana, "Vacillating Andhra Naxal Policy," IPCS # 1728, May 4, 2005 342 Account from Oetken, 2009, pp. 137-38; Sahni, 2000. 341 Sahni, 2000. 344 Trinath Mishra, 2007, pp. 112-116; Also see Ajai Sahni, "Andhra Pradesh: The State Advances, the Maoists Retreat, South Asia Intelligence Review, 6 (10), September 18, 2007. 341 Sahni, 2000. 391 In Bihar, the MCC had also gathered strength in the 1980s but Bihar's policy was mostly characterized by inaction and a lack of effort. 34 6 Instead of developing or requesting regular forces and sustained or comprehensive actions, the state government relied on senas (private militias) that periodically massacred members of the MCC civilian base, or the state used cheap, poorly trained home guards, deployable in small numbers, to conduct limited operations and serve as easy targets for the Maoists.3 4 7 Violence. Initial actions in Andhra appeared as restraint if not outright appeasement between 1987-1991, offering development palliatives, addressing social and economic justice issues, and even releasing Naxal prisoners as a good-will gesture.34 8 This liberal and "soft attitude" 3 4 9 gave way to some more aggressive actions, but they involved state police targeting Maoist leaders with limited civilian casualties,350 and relied heavily on arrests, active surrender, and rehabilitation policies rather than pure attrition. The ratio of arrests to insurgents killed was 15:1 and ratio of surrendered to killed, 34:1.351 And these were pursued alongside efforts at development investments, poverty alleviation, employment generation, and improvement of tribal welfare, with leaders regularly restating that the roots of the problem were "social, economic, and political."3 52 What is evident is that during this period, there was a considerable absence of interest or serious effort by the center in these conflicts. In this period, the Naxal conflict theater produced hundreds and sometimes almost a thousand annual fatalities but triggered no alarm by the center, 346 Oetken 2009, p. 141. Verma, 2014. 348 Prakash Singh, 2005, p. 139; "Minister Urges Plan to Solve 'Naxalite Problems,"' Delhi All India Radio Network, August 3, 1991, Daily Report. Near East & South Asia, FBIS-NES-91-150 on 1991-08-05, p. 60. Accessed via Newsbank on Feb 19, 2014. 349 3 5 0Mishra, 2007, pp. 110-112. Verma, 2014, p. 294. 351 Prakash Singh, 2005, p. 142 says the offensive involved 248 killed, 3,434 arrests, 8,500 surrenders. 352 Prakash Singh, 2005, pp. 139, 145; Sahni, 2000; Sahni, 2007. 347 392 while far less intense violence in Punjab, Kashmir, and even some Northeast states had elicited major responses by the central military and paramilitary. Throughout the 1980s and 1990s, Ministry of Home Affairs' annual reports continued to dismiss mounting Naxal violence as the work of inchoate left-wing extremists and offer little strategy for how to contend with them. 3.1.5. Naxalite III - Central India (2004-Present): Mitigation Background Over the past decade, following the merger of the PWG with the MCC, Naxal influence and violence has poured out of Andhra and Bihar to neighboring districts and states including Chhattisgarh, Jharkhand, Orissa, and some parts of Maharashtra, Madhya Pradesh, and West Bengal. It has spread from 53 districts in 2001 to 252 districts in 2009313 and escalated to a peak in 2010 with 1,200 deaths and 2,200 incidents, rivaling the intensity of the Kashmir insurgency in its early years. The Prime Minister has called it "biggest internal security challenge ever faced by our country," 35 4 yet the Indian state response has been criticized for being "timid and weak" lacking a national policy, a cohesive strategy, central coordination, and punting responsibility to the states. 355 Despite an estimate 10,000 - 20,000-strong Maoist cadre,3 56 50,000 militia support, 357 and perhaps 100,000 over-ground activists, one former IPS officer and Indian governor writes, "Vacillation and ad hocism has been the hallmark of the government's counter-terrorism policies." 35 8 Manish Tewari, MP Lok Sabha debate, July 8. 2009 "Need to formulate an effective strategy to combat the menace of naxalism in the country", 15'h Lok Sabha 354 Manmohan Singh, "PM's Valedictory Address at the Seminar on the Occasion of Golden Jubilee of National Defence College," October 22, 2010. 35 Marwah, 2009, pp. 130, 147. 356 Shaikh Azizur Rahman, "Amid Violence, Hamlet Joins Electoral Politics in Bengal," India Ink blog, New York Times, July 24, 2013. 31 Chakravarti, "Introduction," 2008. 35 Marwah, 2009, pp. 140, 147. 353 393 e 6.7: Naxal Affected Re (Source: Institute for Conflict Management) Effort. Other noted analysts derided government efforts as "timid and weak," 35 9 "fabricated," 360 "pathetic," 361 and indifferent. 3 62 Meanwhile, many states have pursued individual "ostrich-like polic[ies]" that "let sleeping dogs lie" so as not to invite retaliation.3 63 The Indian government has deployed something between 60,000 - 100,000 paramilitary forces to make it look like they're doing something, even those forces lack the capabilities to fight the Naxals Ved Marwah, "Naxal Gambit State Response," Tehelka, 45 (6), Nov. 14, 2009. Ajai Sahni, "Maoists: Enduring Strengths," South Asia IntelligenceReview 10 (39), April 2, 2012. Author Interview, retired senior Indian Army officer, Gurgaon, India, November 9, 2012 (#67). 362 Saiful Haque and Sujit Nath, "Operation Green Hunt Fizzles Out in Red Zone," India Today, Feb 22, 2010. 363 Mishra, 2007, p. 116. 359 360 394 effectively 3 64 and are sparse in an area four times the size and containing eight times the population of Kashmir.3 65 The Indian state's exertion of effort, whether for security or development purposes, has been quite limited in terms of force presence, resources, skill, sustainment, and political focus or oversight.3 6 6 Since Delhi has opted not to deploy the more seasoned and professional forces of the military as it did in past insurgencies, it has been forced to rely on dilapidated federal paramilitary organizations that lack quality leaders due to what one former senior paramilitary official alleges is a promotion policy based on "bootlicking." 367 Meanwhile state police forces are the main line of defense, despite a 20% under strength force, often misused resources,3 68 and active avoidance of engagement with the Naxals by holing up in their stations.3 69 Police can prove effective counterinsurgents with major investments and overhaul as demonstrated in Punjab, and eventually Kashmir. But neither state nor federal police forces are adequately trained, equipped, supported, or reinforced enough to deal with the Naxal threat. 7 Even when federal paramilitary battalions are sprinkled throughout states, they are used sporadically3 7 ' while offensive operations are "fitful," "transient" and short-lived. 2 Government interest in the insurgency only flares up after a major Maoist strike but "except for a few 364 Author Interview, Indian Army Officer, New Delhi, India, Nov 6, 2012 (#65); Author Interview with Indian national security journalist, New Delhi, Dec 17, 2013 (#147). 361 Marwah, 2009, p. 93. 366 Bhavana Vij-Aurora, "Armed and Dangerous," India Today, April 20, 2011; "Is the Strategy Against the Naxals Working?" India Today, Nov. 6, 2009. 367 Shoma Chaudhury, "Bringing on the Army Against the Naxals Will Be a Disaster," Tehelka, 7 (23), June 12, 2010. 368 Vishwa Mohan, "Over One-fifth of the Country's Police Force Just on Paper," Times ofIndia, Dec. 24, 2012; Rukmini Srinivasan, "Massive Misuses of Funds Bleeds War on Naxals," The Economic Times (India), April 10, 2010. 369 V.K. Shahikumar, "Police Cannot Take on Maoists," Indian Defence Review, 25 (4), Oct-Dec 2010. 370 Ajai Sahni, "Maoists: Surviving Adversity," South Asia Intelligence Review, 12 (37), March 18, 2014. 371 "India's 372 Naxalite Insurgency: Not a Dinner Party," Economist, Feb 25, 2010. Sahni, "Maoists: Surviving Adversity," March 18, 2014. 395 ' bureaucratic responses, soon all the sacrifices of these Martyrs fades into oblivion." 37 3 Data collected from 2005-08 shows that Naxal operations outpaced police operations by a ratio of 4:3 and this ratio had worsened through 2013."4 One journalist travelling through this region noted that the central government had basically abdicated the Dandakaranaya region of Chhattisgarh as well as "vast chunks of Jharkhand and Orissa" to the Maoists. 375 Even in its developmental approach, the government had done little more than offer platitudes on development in the tribal regions, 376 and has still failed to offer any political and economic reforms, 37 7 or improve human development. 378 When taking into account the affected population size, the annual spending on security related expenditures or the Integrated Action Plan for Naxal-affected districts is a drop in the bucket compared to annual spending in Kashmir and Punjab. General development schemes have had some positive effects 379 but these are non- targeted programs with inconsistent spending, while targeted programs suffered from poor implementation, weak monitoring, and numerous leakages and corruption. 380 After new schemes for training, intelligence, equipment and elite federal police forces were announced, promised resources never arrived.3 8 Violence. For the most part, the state has eschewed the deployment of battle-tested counterinsurgency forces in favor of an often-criticized "holistic" or "developmental" 373 Purushottam Singh, 2011, p. 29. 2005-08 data from Khanna and Zimmerman data, 2014; 2013 data from "India Maoist Assessment: 2014" SATP. Chakravarti, 2008, p. 185. 376 Development Challenges in Extremist Affected Areas: Report of an Expert Group to PlanningCommission, New Delhi: Government of India, April 2008. 377 Marwah, "Naxal Gambit State Response," 2009. 378 Kennedy and Purushotham, 2012. 379 Khanna and Zimmerman data, 2014. 380 "How Large if the NREGS?" Research Unit for Political Economy, March 23, 2012; Chandrani Banerjee, "A Sieve of a Scheme?" Outlook India, April 19, 2010; "Is the Strategy Against the Naxals Working?" India Today, Nov. 6, 2009. P.V. Ramana, "A Critical Evaluation of the Union Government's Response to the Maoist Challenge," StrategicAnalysis, 33 (5), 2009, pp. 749-750. 381 Nishit Dholabhi, "Schools Shut, CRPF Suffers," The Telegraph, March 14, 2014. 374 37 396 approach, 3 82 dismissing the military option and relying primarily on local and federal police. These lightly armed forces are employed reactively or for defensive actions that ensures low collateral damage relative to other counterinsurgency campaigns, but also much lower exchange rates that favor the rebel forces.3 3 Commanders have defended their "selective approach as the appropriate one give that Naxals are simply "disgruntled" 385 or "misguided." 3 8 6 Even when their proxies have committed certain excesses, such as with the vigilante Salwa Judum movement, state institutions like the Ministry of Home Affairs have sought to monitor and police them 387 while the Supreme Court has quickly reigned them in, 388 in stark contrast to freedom of action in Kashmir and the Northeast. Though the state appears to have gone more proactive and kinetic after Maoist violence escalated in recent years with Operation Green Hunt in 2009-10 and the creation of Commando Battalions for Resolute Action (CoBRA),3 89 these remain heavily balanced by political reforms and economic development initiatives to gradually win over the public.390 The strategy has been guided by serious efforts at restraint to "alleviate" rather than "defeat" the rebellion because Naxal grievances are perceived by a large part of the state to be legitimate.39 1 One officer writes, "'Alleviate' is the key word not 'defeat' because attempts to defeat a peoples movement ultimately end up defeating the nation itself, and tearing it apart in the process." 392 But alongside Ajai Sahni, "India's Maoists and the Dreamscape of 'Solutions'," South Asia Terrorism Portal, February 2010. Bhavana Vij-Aurora, "Armed and Dangerous," India Today, April 20, 2011. 384 "Naxal Response has to be More Humane: CRPF DG," The Hindu, September 24, 2013. 385 "They're Disgruntled Tribals, Not Naxals: KJ George," DNA India, June 19, 2013. 386 Sujay Mehdudia, "Sonia Extends Olive Branch to Maoists," The Hindu, April 22, 2009. 38 D.M. Mitra, Genesis and Spreadof Maoist Violence andAppropriateState Strategy to Handle It, Sponsored by Bureau of Police Research and Development, New Delhi: Institute of Social Sciences, 2011, pp. 19-20. 388 J. Venkatesan, "Salwa Judum is illegal, Says Supreme Court," The Hindu, August 3, 2011. 389 Its still telling that CoBRA battalions have still had a 4:1 exchange ratio. See "Introduction," Central Reserve Police Force website, last accessed April 25, 2014. 390 Purushottam Singh, 2011, p. 26. 391 Author interview, Advisor to the Indian National Security Council, New Delhi, India, November 20, 2012 (#78). 382 383 392 Purushottam Singh, 2011, p. 29. 397 this low violence, low effort strategy, the state has done little to alter its patterns of authority over the Naxal-affected regions. 3.2. Counterinsurgency in the Northeast Though the insurgencies in the Northeast have arisen out of varying circumstances, the Indian state has generally applied a strategy of low effort and high coercive violence, applying intermittent indiscriminate military force with a focus on large unit operations and kill-ratios, 393 a "divide and rule" policy 3 94 often encouraging fragmentation and fratricidal violence, 3 95 abetting corruption,396 accepting militia governance, 3 97 and offering paltry "development"398 and accommodations that often perpetuate conflicts. Admittedly, Assam has involved less violence, and Mizoram is the one Northeast conflict that has been firmly and formally resolved, but they all share many of the same properties and are governed by a "garrison mentality,"3 99 similar to an administrative occupation with a "parallel political structure" autonomous from local input and direct controlled by Delhi.4 00 Marwah talks about the poor intelligence, 2009, p. 235; Cline, 2006, p. 142 on large unit ops; Baruah, Durable Disorder, 2005, pp. 54, 66-67 on military governors and kill ratios. 394 Marwah, 2009, pp. 233-235. 395 Strategies of "coercion, co-option and corruption of political processes" have "encourage[d] the politics of exclusion and retribution." See Sanjay Barbora, "Rethinking India's Counter-Insurgency Campaign in North-East," Economic and PoliticalWeekly, 41 (35), September 2-8, 2006, p. 3807. 396 An estimated 80-90% of development aid returns back to Delhi through contractors, bureaucrats, and corrupt politicians. Rammohan, 2011; Patricia Mukhim, 2005, p. 179; Baruah writes that corruption and extortion are considered "part of the natural landscape" and normal codes of morality or accountability "do not apply to India's modem but wild Northeast Frontier." See Baruah, "A New Politics of Race," 2005, p. 174. 397 One former governor points out that "in a large part of the region, the writ of the government simply does not run." See Marwah, 2009, p. 249. 398 Though the Indian state claims to offer the highest per capita assistance to these states, "there is no evidence of that to be seen in these states." Anand Teltumbde, "Race or Caste, Discrimination is a Universal Concern," Economic and PoliticalWeekly, 44 (34), August 22-28, 2009, p. 17. 399 400 Barbora, 2006, p. 3811. Baruah, DurableDisorder, 2005, p. 44 393 398 Continued insurgent activity and predation by dozens of militant groups is tolerated by the state with a "do nothing" approach 40 ' so long as it remains below a certain threshold-what one scholar of the region terms "durable disorder." 402 Incompetent and poor quality officers dumped in the region provide little human security and when they do, it is narrowly focused on settlers from the heartland.4 0 3 Rather than gradualist approach to suffocating the insurgency,4 04 policy is guided by "short- term reaction to political exigencies." 405 One former intelligence officer warns this cynical approach has "resulted in the retreat of the state from the crucial areas of governance and subversion of democratic politics" and breaths new life into this "dystopia" with militias having "a free military run" in parts of the Northeast.40 6 As a result, the government's expectation is that there "low levels of disorder will continue for the next few decades."40 7 3.2.1. Nagaland I (1955-19 75) and H (1976-Present): Enfeeblement Background. The ethnic Nagas, located in India's furthest regions of its Easternmost province of Assam, initiated the first of what would be many ethnic tribal groups to secessionist rebellion against the state led by the Naga National Council (NNC). Ethnic Nagas, seeking an independent state, rebelled against the Indian government in 1956 and though the intensity of violence dipped in 1964 after the creation of a Naga state, it continued as a robust insurgency until the signing of the Shillong Accord with a faction of Naga leaders in 1975, marking the end of the first insurgency. However a large section, the NSCN, broke away from the accord in 1980 401 Author 402 Interview, retired Indian Army officer, New Delhi, India, November 10, 2012, (#68). Baruah, DurableDisorder,2005 403 Barbora, 2006, p. 3810; Marwah, 2009, pp. 249, 234. 404 Anit Mukherjee, 2009, p. 142. 405 Cline, 2006, p. 143. 406 RN Ravi, "Nagaland: Descent into Chaos," The Hindu, January 23, 2014. 407 Author Interview, retired senior Indian civil servant, New Delhi, India, November 23, 2012, (#83). 399 to continue fighting, and then again split in 1988 into two main factions (NSCN-K and NSCNIM) based on tribal cleavages. Further ceasefire deals were struck but without demobilization, fratricidal violence and violence associated with local politics continued and even rose. In other words, the Indian government routinely settled for short-term fixes rather than investing the requisite efforts to defeat the insurgency, whether through a massive disarming campaign or more targeted methods combined with economic and political measures. For the most part, state strategy in Nagaland has looked much like enfeeblement with considerable underinvestment resulting in cheap tactics like periodic strikes and sweeps and encouragement of inter-ethnic divisions and active fostering of fratricidal violence by intelligence agencies in order to weaken the insurgency.40 8 But there was neither sufficient resources and pressure to crush the insurgency, nor political offers to strike a meaningful deal. Effort. The Indian Army did concentrate some material assets on Nagaland in the early years but never in sufficient numbers to control the unforgiving terrain or seal the border and prevent ex/infiltration.4 09 While area domination was the declared strategy, this was never actually accomplished by the Army.410 In the wake of partition and communal violence, the response to Naga agitations was "subdued and on a low-key" since "there were more pressing problems, mammoth in scale and intensity. ... Who cared for an uprising staged by a few hundred tribals!" 41 1 Intelligence sources and estimates were also underdeveloped for Nagaland. 412 Some have decried the "total lack of resources" employed in this fight,4 13 though there is a recognition that the utilization of these resources results from the central government's indulgence of fiscal Rammohan, 2011: 107; Ravi, 2014. 41 Nibedon, 1978, pp. 75, 101. 0 Rajesh Rajagopalan, "'Restoring Normalcy': The Evolution of the Indian Army' Counterinsurgency Doctrine," Small Wars & Insurgencies, 11 (1), Spring 2000, p. 52. 41" Nibedon, 1978, pp. 54-55. 41 Dhar, 2005, pp. 186-87. 413 Quotation from page 8 of Brig M.D. Phadke, "Countering Insurgency: Thrust in a Broader Dimension," Pratividrohi,4 (2), September 1992, pp. 5-12. 408 409 400 indiscipline (or cynical promotion of corruption),414 as well as the tolerance of insurgent taxation of business and government employees. 415 Despite development spending on paper, Nagaland remains "a dead place with very little opportunity for self-employment and other income generating activities." 4 16 While it has expanded garrisons in the state, it has deferred most authority to militias opening up a power vacuum and signaling a "retreat of the state" that most resembles administrative occupation of an enfeeblement strategy. 417 Even today, "the army can live with the NSCN as long as things don't get past a certain threshold." 41 1 Violence. In Nagaland, the military "applied themselves with full bloodied vigour to the business of 'softening up' the recalcitrant Nagas," 4 19 deploying massive force in the 1950s, the use of heavy weapons, indiscriminate targeting, and concentration camps that resulted in the deaths and displacement of thousands of Naga civilians. 42 0 There was a standing retaliatory policy that whenever a unit was ambushed near a village, that village would be burned down. 421 Despite some political accommodations such as the granting of statehood in 1963, the state has never been highly motivated to negotiate a resolution,422 and encouraged the degeneration of the insurgency through factionalism, splinter groups, fratricidal and predatory violence, and criminalization of the movement in Nagaland.42 3 This has encouraged fragmentation and 414 Glancey, 2011, p. 243; Rammohan, 2011, p. 116; Nari Rustomji, ImperilledFrontier:India's North-Eastern Borderlands,New Delhi: Oxford University Press, 1983, pp. 85-86. 415 Page 214 of Ashok Krishna (Maj. Gen. (Retd)), "Counter Insurgency and Internal Security," The IndianArmy: A BrieffHistory, Ed. Ian Cardozo, New Delhi: United Service Institution, 2007, pp. 192-227. 416 Mukhim, 2005, p. 185. 4 17 RN Ravi, 2014. 418 Author Interview, Indian national security journalist, New Delhi, India, December 17, 2013 (#147). 419 Rustomji, 1983, p. 31. 42 Nibedon, 1978, p. 74, 80; Lt Gen Narahari acknowledges the indiscriminate use of force. N.S. Narahari (Lt. Gen.), Security threats to North-EastIndia: the Socio-ethnic Tensions, New Delhi: Manas, 2002, pp. 110-111. Author Interview with an Indian scholar of the Northeast, New Delhi, India, November 15, 2012 (#74). 422 Sanjib Baruah, "Confronting Constructionism: Ending India's Naga War," JournalofPeaceResearch, 40 (3), 2003, p. 335. 423 R.N. Ravi, 2014; D.B. Shekatkar, "India's Counterinsurgency Campaign in Nagaland," India and Counterinsurgency:Lessons Learned, Ed. Sumit Ganguly and David P. Fidler, New York: Routledge, 2009, pp. 13- 401 fratricidal clashes, abetted corruption, tacitly accepted violence committed against people so long as it doesn't threaten the state, failed to disarm insurgent organizations, and neglected to consolidate control over the region after the insurgency was defeated. 3.2.2. Mizoram (1966-1986): Enfeeblement Bdckground. After the Indian government's apathetic response to a major famine in Mizoram in Assam that left thousands dead, its neglect of economic reforms and development for the region, and growing nationalism and assertiveness of their identity, the Mizo people's discontent was mobilized by the Mizo National Front (MNF) towards separatist ambitions. After a few years of political mobilization and covert military training and preparation the insurgency began with the MNF launching a major surprise operation to physically takeover political and security targets and assume the reigns of government as an independent political entity. Violence. The sudden mobilization of the MNF caught an alarmed Indian government (GOI) by surprise. Before it could reinforce its limited Assam Rifles force normally based in the region with an additional brigade, the GOI had to heavily rely on the use of airpower to fend off the Mizo offensive. The state continued to rely on heavy weapons and extensive retaliatory violence against Mizo civilians, 4 2 4 believing "fear was the key to the operations." 42 5 A journalist described this as, "a policy of psychological warfare combined with armed attrition to wear down the rebel forces,"A 26 and one officer admits that "anything and everything" was morally sanctioned in a campaign regarded as a "dirty job," 4 2 7 such as the tactical use of terrorist-like 14. Vivek Chadha, "India's Counterinsurgency Campaign in Mizoram," India and Counterinsurgency:Lessons Learned, Ed. Sumit Ganguly and David P. Fidler, New York: Routledge, 2009. 424 Jafa, SATP? 1999; (Nibedon, 1980, pp. 112-113. 42 'Nibedon, 1980, p. 95. 426 Nibedon, 1980, p. 147. 427 Nibedon, 1980, p. 140. 402 "pseudo-gangs". 4 2 8 A senior general admitted the definition of insurgents was broadly defined to cast a wide net,4 2 9 and a retired Brigadier clarified the brutality was not just one of an undisciplined force but a deliberate "colonial" approach.4 3 0 Perhaps the greatest violence committed against civilians was the forcible and coercive relocation of 82% of the population into "protected and progressive villages" (PPVs), resembling concentration camps. Villagers did not move willingly for their safety but were relocated by the Army through coercion, forged consent papers, and brute force. One officer confessed to have "done the job of an executioner." 431 Plans for development were totally neglected and limited provisions and relief efforts deliberately sustained a "near-famine situation" in all the grouping centers. 43243 Here, the food supply could be regulated4 33 and "halted at will when collective punishment was found necessary,",434 resulting in the estimated deaths of thousands of civilians.435 Effort. Despite the harsh measures, the GOI never seemed to deploy sufficient numbers of well-trained forces in the region, nor sufficient fiscal or political resources to consolidate their hold, allowing the insurgency to drag on two decades. This was the problem early on with only one brigade posted to the area.43 6 Even after the conflict escalated, no more than an army division equivalent was ever deployed to control Mizoram and the surrounding areas, and 42 Maj L Menezes, "Intelligence - Weapon of a Counter Insurgent," Pratividrohi,4 (2), September 1992, pp. 15-18 - this quote from page 16. 429 Nibedon, 1980: 91 4 30 Nibedon, 1980: 192 431 Vijendra Singh Jafa, "Counterinsurgency Warfare: The Use & Abuse of Military Force," Faultlines 3, 1999; On forged consent papers, also see Abheek Barman, "Air Attacks in Mizoram, 1966: Our Dirty, Little Secret," Economic Times, February 19, 2013. 432 C. Nunthara, Impact ofIntroduction of Grouping of Villages in Mizoram, New Delhi: Omsons Publications, 1989, p. 4 433 Ali Ahmed, "Mizo Hills: Revisiting the Early Phase," Scholar Warrior,Autumn 2011, p. 17; Jafa, 1999; Barman, 2013; Nibedon, 1 9 80, p. 118. 434 Nibedon, 1980, p. 111. 435 Based on demographic irregularities in the 1966-71 period suggesting as many as a few thousand died from these actions. 436 Ali Ahmed, "Mizo Hills," 2011, p. 15. 403 oftentimes far less, a force presence acknowledged to be at least half of what was required. 3 After 82% of the population was regrouped into 248 villages, only slightly more than a brigade's worth of troops were deployed, insufficient to "protect" them and prevent infiltration.438 Largeunit cordon-and-search operations were derided by one officer as "a total waste of time."43 9 Military degradation of the MNF was more an accidental byproduct of the 1971 war with Pakistan since the MNF had heavily relied on East Pakistan as a safe haven. Furthermore, the PPVs were also heavily under resourced and security was never sustained. From a governance perspective, the Assam government, which was ostensibly responsible for its immediate management, viewed the Mizo territory with "indifference and callousness" 440 both prior to and after the famine. Once the intensity of insurgent activity declined in the 1970s, "New Delhi adopted a rather ambivalent attitude," 441 and all indicators pointed to a political management system of administrative occupation. In 1972 when the MNF was in its "death throes", 44 2 the government maintained an "ambiguous" policy that exhibited real "contradictions."4 43 The MNF leader-after being displaced from safe havens in East Pakistan, his guerilla forces sharply reduced, and his people mostly in concentration camps-sought a peace agreement as early as 1977. However, the vicissitudes of Indian politics and government leaders' unwillingness to expend political capital prolonged the conflict another decade until a peace agreement was finally signed in 1986.444 437 Jafa 1999 says at least another infantry division was required. 248 villages were each manned by 20-man pickets totaling 4,960 troops. Rajagopalan, 2000, p. 56; Sajal Nag, Contesting Marginality:Ethnicity, Insurgency and Subnationalism in North-EastIndia, New Delhi: Manohar, 2002, p. 247-48. Brig Dal Singh, quoted in Rajagopalan, 2000, p. 57. 440 Nibedon, 1980, p. 61. 441 Nibedon, 1980, p. 193. 442 Nibedon, 1980, p. 172. 443 "Mizoram: False Hopes," Economic and Political Weekly, 17 (27) July 3, 1982, p. 1083. 444 Chadha, 2009, pp. 41-42. 438 404 3.2.3. Manipur (1982-Present) 445 : Enfeeblement Background. While an organized push for self-determination began in earnest in Manipur in the 1960s, the secessionist rebellion did not commit major acts of violence until the late 1970s due to the dissatisfaction of the ethnic Meiteis (the dominant ethno-linguistic group in the state), resentment over Naga gains, and spillover effects of neighboring insurgencies. Though similar to Nagas and Kukis in physical appearance and members of the Tibeto-Burman language family, Meiteis in the valley converted to Hinduism somewhere between the seventeenth and eighteenth century, but separatists began reviving their pre-Hindu faith of Sanamahi.4 46 The insurgents suffered some initial reverses in the early 1980s due to a leadership targeting effort led by the Army.44 7 Believing these strikes to have "decimat[ed]" insurgent groups,44 8 the Army withdrew around 1983, but their efforts only proved a temporary setback as the insurgency revived around 1988 and has continued ever since as the bloodiest of insurgencies in the Northeast. Despite a population of roughly 2 million over this period, Manipur has suffered some of the most intense insurgent violence, in part because the Indian state has relied on a limited but violent containment strategy of enfeeblement intended to deter the escalation of violence rather than defeat the insurgency, while tacitly accepting insurgent predation, control of some regions, and violence under certain thresholds. Effort. The state has expended little effort in all aspects of the counterinsurgency campaign. Manpower resources to protect the borders were in short supply or of poor quality, 449 445 Accounts from Marwah 2009; E.N. Rammohan 2011; Krishna, 2007; Prakash 446 Subir Bhaumik, "Ethnicity, Ideology and Religion: Separatist Movements in Singh, 2008; Vivek Chadha, 2005. India's Northeast" Religious Radicalism & Security in South Asia, Ed. Satu P. Limaye, Honolulu, HI: Asia Pacific Center for Security Studies, 2004, p. 237. 447 E.N. Rammohan, 2011, p. 157. 448 Krishna, 2007, p. 203. 449 E.N. Rammohan, "Blue Print for Counter Insurgency in Manipur," USIJournal, 548, April-June 2002. 449 Mawarwah, 2009, pp. 208-09. 405 due in part to the sidelining and neglect of local policing forces 45 while the vast majority of fiscal resources made available by the center have leaked through corruption, extortion, and repatriated by contractors. Political and administrative leadership was also absent, exemplified through bad and unreformed governance, corrupt development schemes, misappropriation of funds, and tolerance and/or collusion with rebel taxation. While battling insurgents, units from different security organizations deployed to the region could never overcome friction and implement joint coordination. 45 ' But instead of being a byproduct of challenging environment, one analyst explains that the corruption and undergovernance were deliberate choices: "Its not a failure of the state. It's a bloody strategy." 45 2 Military operations were "intermittent" 45 3 and "sporadic,", 454 abiding by a policy of "watch and wait." 455 They were followed by little administrative follow through and rapid troop withdrawals after the situation achieved fragile stability. Security gains frequently unraveled soon after displacing insurgents who would soon return to their established liberated zones. 456 Finally the state issued no serious corrective measures to turn around the strategy despite two decades of failure and repeated audits confirming the same deficiencies in capabilities and morale. 457 Instead it has turned a blind eye to insurgent "liberated zones" where militants operate with impunity,45 8 while using and pitting militants against each other.459 "Manipur remains the second most conflict ridden state in the NE" Access Northeast(NENA), 3 (41), Jan 1-15, 2008. 451 Mukhim, 2005, p. 186; Chadha 2005, p. 317; Routray, 2008. 452 Author Interview, Indian think-tank analyst, New Delhi, India, November 23, 2012 (#82). 41 Routray, 2008. 45 4 Prakash Singh, 2008, p. 41. 455 N. Somorendro Singh, 2006, p. 337. 456 Prakash Singh, 2008; Narahari, 2002, pp. 159-160; E.N. Rammohan, "Manipur: A Degenerate Insurgency," USI Journal, 547, Jan-March 2002. 457 Routray, 2008; Cline, 2006, p. 143; "Security mutiny sparks fears of rebel attacks in northeast India," Agence France-Presse, December 12, 2000. 458 Bibhu Prasad Routray, "Manipur: The Nexus again," South Asia IntelligenceReview, 6 (14), October 15, 2007. 45 9 Rahul Bedi, "Need a Cop? Just Call your Local Guerilla," Courier-Mail,July 7, 1993; Routray, 2007. 450 406 Violence. The state made no attempts to restrain violence by doing little to protect the population from militants, fostering inter-ethnic violence that resulted in tremendous fratricidal and collateral violence of a "degenerated insurgency," 460 and directly conducting thousands of extra-judicial executions (over 1,500 of which have been documented) along with rape, torture, enforced disappearance, and arbitrary detention as well as victimizing civilians indirectly (displacement, disease, starvation).46 1 Since the initial strategy of leadership decapitation, state strategy has had an offensive orientation due to Army dominance of the theater, sidestepping the civilian administration4 6 2 and sidelining the local police to nonexistent roles in operations.4 63 They make no secret of their preference for killing insurgents over arrests or rehabilitation, a fact substantiated by casualty figures. Poor intelligence due to personnel lacking an understanding of local language and customs exacerbates this.464 Nasty tactics that foster ethnic divides to counterbalance and exhaust militants have produced heavy inter-ethnic violence and collateral damage (that in one year killed over 1,000 people) 465 and resulted in over 30 well-armed and capable insurgent organizations.46 6 Meanwhile no serious attempts to redress grievances 4 67 with dialogue, 468 or substantively offer surrender and rehabilitation to Meitei insurgents have been undertaken. 469 Finally, despite 460 Praveen Swami, "Despair and Democracy," Frontline, 21 (12), June 5-18, 2004. 461 E.N. Rammohan, 2011, p. 127; Manipur:A Memorandum on Extrajudicial, Summary or Arbitrary Executions, Guwahati: Civil Society Coalition on Human Rights in Manipur and the UN, March 28, 2012; Yumnam and Konsam, "Militarisation & Impunity in Manipur," Article 2, 5 (6), Dec 2006, pp. 2-3; Laishramcha Janine "The Militarisation of Manipur," Article 2, 5 (6), Dec 2006, p. 10; Soutik Biswas, "Living on a Razor's Edge," India Today, October 15, 1993. 462 Chadha, 2005, p. 320. 463 Routray 2008; Wasbir Hussain, "Insurgency in India's Northeast: Cross-border Links and Strategic Alliances," Faultlines, 17, 2004. 464 Dhar, 2005, pp. 101-104. 465 John Rettie, "India Sends in Troops to Stop Ethnic Slaughter," The Guardian, September 24, 1993, p. 16; Chadha, 2005, p. 313. 466 Wasbir Hussain, "Manipur: Impending Collapse of Governance," 2004. 467 Dhar, 2005, p. 105. 468Anoop Sarbahi, "Insurgent-Population Ties and the Variation in Trajectory of Peripheral Civil Wars, ComparativePoliticalStudies, 47 (10), September 2014, pp. 1470-1500. 407 large appropriation of fiscal resources, with the appointment of unscrupulous, incompetent (and often absentee) bureaucrats to the region, the state has failed to translate or invest this into public goods to yield actual development gains. 4 70 They have allowed government salaries, corruption, and contractors to siphon nearly all of these resources. 47 ' At the same time, the state accepts and sometimes colludes with insurgent predation,4 72 revealing a blatant disinterest in the Manipuri public, whether in the valley or the hills. Ultimately, the state has employed an aloof approach, treating Manipur as "enemy territory, which needs to be subdued by force" 473 but also as a dependency, "remote-controlled by political agents but not administered on a day-to-day basis." 474 The political authority of administrative occupation is also evident to outside observers. In a confidential cable, an American Consul General to India noted the degree of military occupation made Manipur "less a state and more a colony of India." 475 The state has employed high levels of routine violence, corruption, and collusion with extortion networks "with hardly any visible impact on the rebellions." It has been content with this low-effort, high violence containment strategy of enfeeblement functioning as "a deterrent to a more vigorous armed campaign or acts of violence by the rebels." 476 4 69 Phanjoubam, (2009: 149) points out that in 16 years of a nominal surrender policy, only 377 have taken it (in contrast to over 10,000 in Assam) despite similar estimated insurgent levels, though even these deals have not been honored by the state. See Pradip Phanjoubam, "Northeast Problems as a Subject and Object," Beyond CounterInsurgency: Breaking the Impasse in NortheastIndia, Ed. Sanjib Baruah, New Delhi: Oxford University Press, 2009, pp. 147-169. 470 Marwah, 2009, pp. 208-09. 471 Rammohan, "Blue Print," 2002; Dhar, 2005, p. 102; Rammohan, "Manipur," 2002; Marwah, 2009, p. 209. 472 Rammohan, "Manipur," 2002; Rammohan, "Blue Print," 2002; Routray, 2007. 47 Patricia Mukhim, "Life Under Martial Law," The Telegraph, September 21, 2004. 474 Subhir Bhaumik, 2004, pp. 219-220. 475 Suresh Nambath, "Manipur More a Colony of India," The Hindu, March 21, 2011; Sanjib Kr Baruah, "Manipur CM is Corrupt: Wikileaks," Hindustan Times, March 21, 2011 476 "Manipur remains the second most conflict ridden state in the NE," 2008. 408 3.2.4. Assam (1990-Present) 7 : Mitigation Background. The "sons of the soil" political mobilization by the Assamese ethnolinguistic group, sparked by increasing numbers of illegal Bangladeshi immigrants in the 1970s, produced periodic spats of political violence in the 1980s and escalated to an ethnic separatist insurgency by 1990 led by the United Liberation Front of Assam (ULFA). ULFA was formed in 1979 and ran something of a parallel government in the state from 1988-90 until Delhi finally decided to crack down on it with Army operations.47 8 The Indian central government declared Presidents Rule in November 1990 and launched two major military operations in quick succession-Operation Bajarang from November 1990 to April 1991, followed by Operation Rhino from September 1991 to January 1992-though both were abruptly stopped prematurely due to political inconvenience. 479 Nevertheless, the ULFA was able to form links with other insurgent organizations in the Northeast, continue its violent agenda, penetrate the state government and develop extortion networks as the state pursued a minimalist strategy of mitigation. Effort. The state did not invest serious material or political resources nor was it able to sustain operations. After circumscribed operations, the state maintained an ostensible presence with roughly two Army divisions along with a range of police forces, but poor coordination between the different organizations limited their effectiveness. 480 While these deployments offered some force presence, this was hardly the level of manpower deployed to Punjab or Kashmir, despite Assam having roughly the same size militant cadre and being larger in area and 477 Account drawn from Marwah, 2009; Bhat, 2009. 478 "Why is this Region Asia's Insurgency Haven?" Tehelka, August 7, 2004. 479 Gurmeet Kanwal (Brig.), "Strategic Stalemate in Assam," FinancialExpress, April 12, 2009; Bhat, 2009, p. 66; Rammohan, 2011, p. 69. 480 "Anaemic Assam," The Hindu, November 1, 1998, (p. Factiva, accessed 3/5/14); Wasbir Hussain, "Multi-force Operations in Counter Terrorism: A View from the Assam Theatre," Faultlines, 9, July 2001. 409 population. Intelligence collection was quite limited 481 and political actors, constrained by electoral incentives, refused to invest in substantive actions. 482 Like the two major operations in Assam in 1990-91, the campaign regularly involved operations that were not sustained as the state sought to counter ULFA militancy "on the cheap" 4 83 with "divide and rule tactics" that "[gave] away long term gains." 484 The state also relied on substantial numbers of flipped militants-surrendered ULFA (SULFA)-as the lead counterinsurgent force, a tactical signature of a mitigation strategy.4 85 One former intelligence officer described Assam as "a portrait of national mismanagement" 486 and a former army officer lamenting the limited efforts of the state, suggested "the army was literally fighting both bound and blindfolded"4 87 The result of these limited efforts was a focus on containing militancy below a certain threshold rather than addressing the root sources. One police officer remarked, "[The Army's] successes against terrorists have been exemplary and manifold. Their successes against terrorism,however, have been few." 488 Violence. Initially from 1990-96, operations were heavily restrained in terms of violence with the government adopting a softer line expressing deep concern about both collateral damage ,489 and "our boys.,6 Politicians were able to prematurely curtail the military operation, limiting both civilian victimization and an offensive orientation. This was followed by a four-year period of brutality where SULFA cadres were unleashed to conduct secret killings and "dirty war" Rammohan, 2011, p. 69. Rammohan, InsurgentFrontiers:Essaysfrom the Troubled Northeast. New Delhi: India Research Press, 2005, pp. 19, 30. 483 Editorial, "Anarchy in Assam," Hindustan Times, January 16, 1998, p. Factiva, accessed 3/5/14 84 4 Phadke, 1992, pp. 10-11. 41s Ajai Sahni and Bibhu Prasad Routray, "SULFA: Terror By Another Name," Faultlines, Vol. 9, July 2001. 486 Dhar, 2005, p. 397. 4 1 Chadha, 2005, p. 247. 488 Wasbir Hussain, July 2001. 4'9 Baruah 1999, pp. 145-46; Rammohan, 2011, pp. 59-63; Subrata Nagchoudhury, "Going Back to Blood and Tears" India Today, June 15, 1996. 481 482 E.N. 410 tactics "...at least with the State's acquiescence." 490 Having gone rogue and posing a threat to the state as well, 49 1the SULFA was rapidly wound down and subjected to thorough scrutiny, investigation, and public condemnation. 4 92 For the most part, police and local militias (backstopped by the Army) led the campaign, making it far less offensive and kinetic than other campaigns in the Northeast. Simultaneously, the state sought to make accommodations with militants and offer broad public goods to sap the drivers of the insurgency, another indicator the state was pursuing a mitigation strategy. A generous program, started within the first two years of the campaign, facilitated the surrender and amnesty of thousands of militants to rehabilitate them with economic packages, skills development and training, and fonnally incorporate some in state police forces, all alongside continued dialogue, ceasefire negotiations, and attempts at peace talks with ULFA. 4 93 By 1998, "The army's strategy was voiced by Assam governor, retired LtGeneral S.K. Sinha: "The ULFA is not our enemy. They are our own misguided youth" "494 The Assamese governor also sought to complement security measures with reorienting the police towards community policing, 495 as well as economic reforms and development programs, in part enabled by his ties to the federal center to create opportunities addressing the demand side of rebellion and win public support for the state.496 But limited resources Baruah, 1999, p. 167. Rammohan, 2011, p. 81; Avirook Sen, "Assam: Jungle Justice" India Today, Nov 8, 1999, Factiva accessed 3/5/14. 492 Mrinal Talukdar, Utpal Borpujari and Kaushik Deka, Secret Killings ofAssam, Guwahati, India: Nanda Talukdar Foundation, 2009; Sushanta Talukdar, "Mahanta at the helm of secret killings in Assam: Saikia Commission," The Hindu, 16 Nov 2007. 491 Phanjoubam, 2009, p. 150; Rohan Mukherjee, "Promoting Peace Through Development: Assam State, India, 490 491 2001-2009," Princeton Innovations for Successful Societies brief, 2010, p. 11; "Over 10,000 militants surrendered since '98," Assam Tribune, March 16, 2010; "Assam Government to Recruit Surrendered Militants," The Pioneer, May 21, 2008; "Gogoi urges PM to revamp Brahmaputra Board," Press Trust ofIndia, Feb 7, 2012 ; Avirook Sen, "Assam: Disarming Action," India Today, Aug 10, 1998; "Assam Timeline, 1993" South Asia Terrorism Portal. 494 Avirook Sen, "Assam: Disarming Action," India Today, Aug 10, 1998 495 496 Mukherjee, 2010, pp. 9-10. Mukherjee, 2010, pp. 4-9. 411 constrained development of infrastructure and public welfare schemes as much as it did kinetic activity and both were "fitful" at best, 497 and largely based on the governor's personal influence with the Prime Minister rather than broad state material or strategic interests 49 8 The combination of limited efforts without substantial material and political commitments, and also some concern for the welfare of the public and civilians, has produced a mitigation strategy designed to reduce violence and co-opt militants but it is hardly extensive enough to neutralize the rebellion or political subversion. Relative to other Northeast states, Assam is more a "mainstream" state, 499 but as predicted by my theory, its overall peripherality has limited serious efforts such that the best the state could produce was "stable anarchy," according to one retired Lieutenant General. 500 A retired Army brigadier judged this situation could erupt again into full blown militancy. 501 Another local journalist concurred with the limits of the mitigation strategy: "Scales of violence have gone down, but we are nowhere near resolving the issue."5 0 2 3.3. Counterinsurgency in North India50 3 3.3.1. Punjab (1983-1993)5o4: Population Control Wasbir Hussain, 2001; "Anemic Assam," The Hindu, November 1, 1998, p. Factiva, accessed 3/5/14. Mukherjee, 2010; Gogoi Seeks Special Fund from PM to Rehab Former Ultras," Rediff com, February 7, 2012. Marwah, 2009, p. 249. 500 Sahni and Routray, July 2001. 501 Kanwal, "Strategic Stalemate in Assam," 2009. 50 2 Journalist Wasbir Hussain quoted by Mukherjee, 2010, p. 13. 503 The North Zone is the Northwest of the country defined by the Indian Ministry of Culture as Chandigarh, Haryana, Himachal Pradesh, Jammu & Kashmir, Punjab, Rajasthan, and Uttarakhand.. 104 The account is based on Chadda, 1997; Marwah, 2009; K.P.S. Gill, "Endgame in Punjab: 1988-1993," Faultlines:Writings on Conflict and Resolution, 1, 1999; C. Christine Fair, "Lessons from India's Experience in the Punjab, 1978-93," India and Counterinsurgency:Lessons Learned, Ed. Sumit Ganguly and David P. Fidler, New York: Routledge, 2009, pp. 107-126; Jugdep S. Chima, The Sikh SeparatistInsurgency in India: Political Leadershipand EthonationalistMovements, Los Angeles: SAGE, 2010; Paul R. Brass, Ethnicity and Nationalism: Theory and Comparison,New Delhi: Sage Publications, 1991; Mark Tully and Satish Jacob, Amritsar: Mrs. Gandhi's Last Battle, London: Cape, 1985; K. S. Dhillon, Identity and Survival: Sikh Militancy in India, 1978-1993. New Delhi: Penguin Books, 2006; Mark Tully, No Full Stops in India, New Delhi, India: Viking, 1991. 4978 49 499 412 Background. After decades of counterinsurgency containment strategies in the Northeast, the Khalistan insurgency, that was of India's own making, threatened India's heartland, and prompted it to field a much more comprehensive strategy. Political maneuvering by Indira Gandhi and her Congress party led to sponsoring radical Sikh militants to counterbalance and fracture the power of the ethnically based Akali Dal Party. The radical Sikh groups led by Jamail Singh Bhindarwale soon went rogue, joined forces with the Akalis, and mobilized young followers towards political violence for the cleansing and establishment of a separate state of Khalistan. As the center sought to crackdown on the Sikh militant movement, the militants took control of and fortified the Sikh's holiest site, the Golden Temple in Amritsar. In 1984, the Army conducted Operation Blue Star, a poorly planned, bloody assault on the temple resulting in the deaths of hundreds of militants, civilians, and security forces. Combined with the sweep operations of Operation Woodrose in rural Punjab, these actions sparked Sikh public outrage, the assassination of Indira Gandhi by her Sikh bodyguards, and a subsequent wave of communal violence and anti-Sikh pogroms throughout the country, claiming the lives of roughly 3,000 Sikhs in Delhi alone. As a result, a horrified and alienated Sikh population concentrated in Punjab was mobilized in support of the separatist insurgency that lasted for more than a decade. The Indian state responded with a population control strategy involving sustained inputs of manpower, fiscal, and organizational resources to heavily degrade insurgent combat power but also win some public support. They also made tactical and strategic deals to defeat the insurgency involving a healthy level of power-sharing with political groups once aligned with separatists. Effort. The Indian state, alarmed at the degeneration and possible separation of Punjab, marshaled significant effort in the campaign including material resources expended to: deploy 413 multiple army and paramilitary divisions; expand and upgrade the local police force size and capabilities (through training, equipment, and pay); and develop dense intelligence collection and analytic capacities. Costly political and administrative investments were made to empower the police director generals as virtual theater commanders and enable coordination with multiple forces. Despite some early obstructions, continuous force presence, patrolling, and strikes against leaders was maintained at a high operational tempo in the latter half of the campaign to pressure insurgent organizations resulting in significant fragmentation as well as some escalation of desperate insurgent violence. Moreover, the state continued to study and evaluate operational effectiveness to continuously improve and experiment with new tactics for intelligence collection and infiltration. This intelligence gathering supported efforts to provide village security, acquire popular support, and offer exit-ramps to demobilize insurgents without killing them. Violence. After an initial ham-handed Army operation proved disastrous, the state imposed significant restraints on the use of force, even cancelling some planned but provocative operations, as it deployed the Punjab police as the primary force to directly confront insurgents, backstopped by army and paramilitary forces. It also leaned heavily on negotiations with separatist political leaders to resolve the disputes, which faltered early on, but eventually paid off. Conventional accounts depict the campaign as one that brutally crushed the Sikh insurgency but the campaign's success specifically derived from its proactive reliance on defensive and nonkinetic substitutes. No doubt, serious levels of state-directed violence escalated (including a number of extra-judicial killings) beginning in 1988 as the state began to more directly confront the myriad of insurgent groups that had splintered after 1984 and degenerated after 1988. But alongside this, the state also invested significant effort in: selective, intelligence-driven targeting (to kill or capture high level rather than rank-and-file insurgents and minimize civilian 414 casualties); rural patrols and deployments to protect the population from militant violence and predation; public goods, infrastructure provision, and employment schemes especially in rural regions; an extensive surrender policy; both overt political efforts and covert intelligence efforts to bolster (even armed) moderates; and finally openings for some insurgent-affiliated parties to re-enter "normal" electoral politics and eventually take power and the restoration of autonomy.505 All of these activities-particularly extensive reliance on amnesties and macropolitical deals-suggested revealed the state's appreciation of insurgents' and their ethnic base's agency requiring consensual obedience through violent negotiation rather than pure force. 3.3.2. Kashmir I (1990-2003) and H (2004-Present) 506: Attrition Background The history of Kashmir was long ripe for conflict. A range of conditions provided the spark for a broad populist uprising that exploded in violence beginning in 1989. This conditions included: historic Hindu dominance over Muslim subjects in the princely kingdom of Kashmir; the seemingly underhanded process by which Kashmir acceded to the state of India; the Indian center's dismissal of autonomy agreements; political interference, manipulation, and imposition perceived as centralized colonial control since 1953; periodic spats of terrorism since the 1960s; and finally, a blatantly rigged election in 1987. Kashmiri youth streamed over the border to Pakistan for arms and training and the nationalist Jammu and Kashmir Liberation Front (JKLF) led a separatist insurgency against the Indian state. Eventually the JKLF would be eclipsed by Islamist insurgent organizations like Hizbul Mujahideen (HM) 505 Tillin, 2007, p. 60. Cites Manor, 1996 and Brass 2001. 506 Chadha, 2005; Telford, 2001; Wallace, 2007; Bose, 2003; Newberg, 1995; Wirsing, 1994; Adrian Levy and Cathy Scott-Clark, The Meadow: Kashmir 1995 - Where the TerrorBegan, London: Harper Press, 2012; Manoj Joshi, The Lost Rebellion, New Delhi: Penguin, 1999; Paul Staniland, "Kashmir Since 2003: Counterinsurgency and the Paradox of "Normalcy," Asian Survey, 51 (5), 2013, pp. 931-957; Ahmed, Ali, "Countering Insurgency in Jammu and Kashmir: Debates in the Indian Army," Confronting Terrorism,Ed. Maroof Raza, New Delhi: Penguin, 2009, 58-79. 415 around 1993 as the vanguard of the insurgency. By the end of the 1990s, HM would be overtaken by Pakistani-based militant groups infiltrating in to Kashmir like Lashkar-e-Taiba (LeT) and Jaish-e-Mohammad (JeM), though it would still have a popular base of support and continue to attract indigenous recruits who composed half the militants even by 2003. Believing the Kashmir insurgency to be a proxy war with Pakistan, 507 the Indian state relied on a strategy of attrition that involved significant military manpower and indiscriminate force to degrade insurgent combat power by killing insurgents faster than their replacements. They continuously adapted to the challenges of a dynamic insurgent landscape, but never seriously considered redressing political grievances. This was not the first time the state had attempted to rapidly crush a nascent secessionist revolt in Kashmir with overwhelming force, but this time the insurgency was able to get off the ground. 508 Between 2003-2005 observers noted a change as the Indian state announced in strategy focused on winning hearts and minds, 509 but while this period was less violent in terms of fatalities, the tactics bore strong resemblance to the pre-2003 period. Effort. The attrition strategy involved saturating the region with massive amounts of Army and paramilitary forces (estimates over time range between 400,000-700,000), primarily in the Jammu and Kashmir divisions, because local police were deemed incapable and untrustworthy, resulting in "security forces everywhere-at least one man every fifty feet or so in a city." 51 0 The military invested considerable organizational resources to develop a new counterinsurgency force, the Rashtriya Rifles, to be permanently deployed in the region and 507 Author Interview with retired senior intelligence officer and government advisor on national security, New Delhi, India, Dec 5, 2012 (#101). 508 "Police Crush Revolt of Kashmiri Secessionists," APP, August 7, 1980, Daily report, South Asia, FBIS-SAS-80154 on 1980-08-07, p. E3. Accessed via Newsbank on February 19, 2014. 509 Mona Bhan, 2008, p. 12; Staniland, 2013; Anant notes the Army Doctrine and Sub-conventional Doctrine were released around this time in 2004 and 2006. See Arpita Anant, Counterinsurgencyand 'Op Sadbhavana'in Jammu and Kashmir, New Delhi: IDSA Occasional Paper, 2011. 510 Newberg, 1995, p. 27. 416 retain accumulated knowledge of best practices and terrain. Though local governments changed hands, the central government also managed to maintain a tight grip on the region through physical force and presence, direct political rule for seven years (under Presidents rule) and more indirect and subtle forms of political subversion in order to continuously apply pressure on any and all political challengers, regardless of their levels of participation in violence. The Indian response was "not simply repression, but a concerted effort to reshape life in Kashmir."51 1 Violence. In combination with the overwhelming physical domination of space, the Indian state employed tremendous amounts of brute force and violence to "wear down the insurgency" through offensives by kinetically winning the "military battle." Under the longest stretch of Presidents rule (with no elected government), imposed on Kashmir for almost seven years from January 1990 through October 1996, the military operated with impunity against militants and civilians. With a focus on "attrition,"5 13 success was measured in kills of militants regardless of their commitment levels and units were rewarded based on kill counts. 514 This was accompanied by the levels of state-perpetrated civilian fatalities including acknowledged collateral damage, recently discovered figures of encounter killings, disappearances, and mass graves, retaliatory destruction of property,51 5 and individual accounts of torture. The state later sought to keep its hands clean by utilizing renegades to break the back of the militancy between 1995-99 but also commit horrific acts of violence against civilians that completely alienated the 511 51 2 Newberg, 1995, p. 10. Newberg, 1995, p. 74. 513 G.D. Bakshi, Kishtwar Cauldron: The Struggle Against ISI's Ethnic Cleansing, New Delhi: Pentagon Press, 2013, pp. 45, 48. 514 Interviews with retired Indian Army officer and Indian national security journalist, New Delhi, India, November 23, 2012 (#84) and December 17, 2013 (#147); Interviews with Kashmiri journalists, December 2012. Measures of progress discussed in Ministry of Home Affairs Annual Reports also indicate this. Recent studies by Lyall in Afghanistan seems to suggest this may be even more devastating and incite even more anger than the death of a family member as it destroys the entire family's livelihood and source of sustenance. 417 Kashmiri public.5 16 Though the state achieved population control, it was purely by force and hardly to leverage support from the population trapped within "a vast prison camp." 5 17 After 2003, a ceasefire with Pakistan was credited with a drop in violence and state strategy was said to have shifted, but the Indian state approach in Kashmir did not fundamentally change.5 18 The approach over the past decade still does not look anything like violent negotiation with real devolution of authority as in West Bengal or Punjab. Instead, it appears much more akin to what Colin Jackson describes as Model 2 counterinsurgency or "political war" with some carrots to accompany overwhelming force, impose and compel obedience, and subject the civilian population to "non-lethal attack."5 19 After roughly fifteen years, coercive violence levels eventually began to drop, largely as a result of their success in degrading the insurgency along with some exogenous factors, and the state advanced some nominal public welfare programs such as OperationSadbhavana,cloaked in the rhetoric of Winning Hearts and Minds (WHAM). However, these have also been described as top-down photo-op projects with little substance,52 0 window dressing and avenues for corruption, 521 narrow club goods, 522 or a cynical and deliberate cultivation of a dysfunctional dependent political order,5 23 much closer to a political project of usurpation with little meaningful powersharing. Meanwhile, civilians continue to suffer daily humiliations and One former official (PM), Nov 20, 2012 acknolwegdes the mistake of using such a brutal force; One former officer describes it as a tactical success but strategic failure. Interview with a retired Indian Army officer, New Delhi, India, November 23, 2012 (#84). 516 517 51S Bose, 2003, p. 115. 519 Jackson, 2008, pp. 111-116. 520 521 Interview with an Indian think-tank scholar, New Delhi, India, November 23, 2012 (#83). A charge made by two retired Army captains who served in Kashmir. Author interviews, New Staniland says this is after 2003; The ceasefire was signed in November 2003; Anant says the strategy shifts sometime between 2004-06. Delhi, India, November 20, 2012 (#79) and November 23, 2012 (#84). 522 What concessions have been made have been targeted at the small number of pro-India Kashmiris but still 8085% Kashmiris who covet independence have received no concessions. Author interview with longtime Kashmiri journalist, New Delhi, India, Nov 28, 2012 (#96). 523 Staniland, 2013. 418 treatment as militants or their collaborators. 52 4 For all the money being poured into the region, there has been insufficient investment in the state police to reduce the military footprint. 25 There is also shockingly little evaluation research on efficacy,5 2 6 and substantial evidence denying this including still high levels of popular grievance, demands for separation,5 and recruitment. 528 27 and militant sympathy Former intelligence and army officers readily admit the state sees the status quo as stable and has made no serious efforts at political accommodations.529 3.4. Measuring Strategy Based on the accounts provided in this chapter and qualitative measures of the dependent variable, a distribution of Indian strategies for each campaign can be generated in graphic form (see figures 6.8 and 6.9). Based on qualitative assessments in each campaign of the effort (in terms of resources, organizational investments, sustainment, and adaptive learning) and violence (in terms of collateral damage, enemy-centricity, accommodations, and public welfare) the Indian state employed, I am able to assign a value and plot it on and XY graph. Based on these qualitative codings, the Indian state deployed the following strategies: attrition in the cases of Telangana and Kashmir (I and II); population security in Naxal I (West Bengal) and Punjab; enfeeblement in Hyderabad, Nagaland (I and II), Manipur, and Mizoram; and mitigation in the cases of Naxal II, Assam, and Naxal III. Naxal I (Andhra Pradesh) occupies the border between Author Interview with retired Indian Army officer, New Delhi, India, Nov 20, 2012 (#79). Anit Mukherjee, 2009, p. 153. 526 Author interview with longtime Kashmiri journalist, New Delhi, India, Nov 28, 2012 (#96). 527 These range from 66% to as high as 95% in the Valley. See Adam Plowright, "Two Thirds in Indian Kashmir Want Independence: Poll," Agence France-Presse,September 12, 2010; Alastair Lawson, "'First' Kashmir Survey Produces 'Startling' Results," BBC News, May 27, 2010. 528 Author interview with longtime Kashmiri journalist, New Delhi, India, Nov 28, 2012 (#96). 529 Author Interviews, retired Indian Army officer, New Delhi, India, November 23, 2012 (#84) and with a retired senior intelligence officer and government advisor on national security, New Delhi, India, December 5, 2012 (#101). 524 525 419 enfeeblement and mitigation but after further examination of the case, it seems to fall closer to the strategy of enfeeblement Figure 6.8: Qualitative Distribution of the Dependent Variable Hyderabad Kashmiri Nagal Naga2 Manipur *Mizoram * 4 Telangana North East * Kashmir2 4 Punjab Naxalsi-AP K4 * Assam Punjab -A * 4 Naxals2 Naxalsl-WB Punjab-B SNaxalsl 0 Naxals3 0 0 4 8 Effort Figure 6.9: Quantitative Distribution of the Dependent Variable -2 SMizoram Hyderabad Nagaland Telangana Kashmiri 4 Manipur Naxal I - Andhra# -4.5 Punjab 4 Assam Kashmir 2 Naxals II -Andhra Naxal I - W. Bengal Naxals III -70 Log(Manpower/Area) 420 Quantitative measures of deployed manpower to measure effort and state-inflicted civilian casualties to measure violence (while controlling for area, population, and duration)1 3 0 help confirm values assigned by the qualitative measures with two slight departures: in Manipur, India appears to have exerted more effort and Kashmir II, it appears to have used less violence. However, a closer examination helps to explain why the quantitative measures differ but still confirm the initial qualitative measures. In Manipur, military manpower ostensibly assigned to the counterinsurgency campaign was inflated since some were tasked with managing the border against external incursions from Myanmar.53 ' Actual counterinsurgency operations seemed to occupy considerably fewer forces, about half of total deployed armed forces. 3 2 In Kashmir, some suggested that the strategy had changed around 2004 since the violence appeared to decline along with the insurgency due to a substantial reduction in Pakistani support and the construction of a border fence.5 33 However, this did not substantially change Indian strategy, its WHAM rhetoric and policies notwithstanding. Only observable fatalities declined while the state continued its focus on killing insurgents, disregard for civilian dislocation, 53 4 and unwillingness to countenance redress of Kashmiri grievances. 35 The distribution of a strategy's effort is further substantiated by alternative quantitative measures such as state spending on a campaign, regardless of whether utilized for more or less violent tactics. Though data is only available for some of the more recent campaigns, figure 6.10 530 Note the actual measure of effort is the Log[Manpower/Area of operation] and the measure of violence is Log[(state-inflicted civilian casualties/yr)/Area population]. 531 "Assam Rifles Likely to Replace Army, BSF to Guard Manipur's border," Times ofIndia, March 5, 2013. 532 Operations were estimated to consume 16,000-17,000 Army, Assam Rrifles, BSF, and CRPF troops with only about 5,000 on active patrols at any point in time. 5 Stephen Tankel, Domestic Barriersto Dismantlingthe MilitantInfrastructure in Pakistan, Washington DC: US Institute of Peace, September 2013; Zulfiqar Ali, "Kashmir Militants Give Up Fight and Head Home," BBC Urdu, May 31, 2012; Author interview with a retired senior intelligence officer, New Delhi, India, December 5, 2012 (#101); Data trends from South Asia Terrorism Portal confirm this. 534 The best evidence of this is the lethal force employed and number of unarmed civilians killed by Indian security forces in the 2008 and 2010 street protests. 535 Author interviews with retired Indian Army officers, New Delhi, India, November 20 and 23, 2012 (#79, #84). 421 measuring security related expenditures confirms what we expect to find. The Indian state spent substantially more in high effort campaigns in Punjab and Kashmir while spending far less in low effort campaigns in the Northeast and Central India/Naxal belt. An alternative quantitative confirmation of a strategy's level of violence in a campaign was constructed by measuring the "dirty-ness" of a campaign-that is the relative percentage of noncombatants killed by the state. 536 Figure 6.11 shows that the "dirty" campaigns mostly match up with strategies of enfeeblement and attrition, confirming previous measures of violence.53 7 Despite claims of normalcy, India's strategy in Kashmir continues to be a dirty one, but which generates fewer fatalities.53 8 The secrecy, violence, and limited access surrounding Manipur makes data collection for this case especially difficult and civilians killed in encounters are often staged to look like they are insurgents. Nevertheless, qualitative evidence confirms that India's approach to insurgency in Manipur is indeed an extremely dirty campaign. 53 9 The one result of this measure that is truly puzzling is the Naxal II campaign in Andhra, as there is nothing to suggest this was at a high level of civilian targeting during this campaign. This involves constructing a percentage by (non-combatants killed by state/total killed by the state) I draw this formula from Madelyn Hsiao-Rei Hicks and Michael Spagat, "The Dirty War Index: A Public Health and Human Rights Tool for Examining and Monitoring Armed Conflict Outcomes," PLOS Medicine, 5 (12), December 2008, pp. 1658-1664. There are admittedly problems with this formulation as pointed out by Hoover et al, 2009 but this is counterbalanced by a range of other qualitative and quantitative measures. See Amelia Hoover, Romesh Silva, Tamy Guberek, and Daniel Guzman, "The 'Dirty War Index' and the Real World of Armed Conflict," Unpublished Paper, Human Rights Data Analysis Group, May 23, 2009. 57 The cases that differ are Northeast, Manipur, and the Naxals I. 538 Staniland, 2013. 539 Teresa Rehman, "Murder in Plain Sight," Tehelka, August 8, 2009; Shoma Chaudhury, "Life in a Shadow Land," Tehelka, 32 (6), August 15, 2009. 536 422 Table 6.5: Modernization of Police Force Allocations from Central Government (Measuring Effort) (Crores/Million people) State/Region 1980-2000 1997-2013 2007-2013 Andhra Pradesh 0.68 14 6 Bihar 0.58 24 20 J&K 1.92 107 64 0.74 16 7 Punjab NE States 2.23 77 59 Naxal States 0.54 14 8 Indian Avg 0.58 16 9 Source : Rajya Sabha Unstarred Question No. 1552, dated 31.07.2002 & No. 810, dated 27.11.2002; Lok Sabha Unstarred Question No. 966, dated on 29.11.2005 & No. 3055, dated on 13.12.2005.; MHA; Rajya Sabha Unstarred Question No.1199, dated 7.03.2001; No. 1552, dated 31.07.2002; No. 810, dated 27.11.2002; No. 3183, dated on 08.12.2010; No. 1620, dated on 28.03.2012; Lok Sabha Unstarred Question No. 966, dated on 29.11.2005; No. 3055, dated on 13.12.2005; No 558, dated 23.05.2006; No. 4117, dated on 22.04.2008; No. 446, dated on 21.10.2008; No. 48, dated on 07.07.2009; No. 2247, dated on 23.11.2010; No. 4968, dated 29.04.2008;No. 5194, dated on 08.05.2012. Accessed on Indiastat.com July 9, 2014 Figure 6.10: Security Related Expenditures by Theater (Measuring Effort) 45 I40 13-38 S35 a 30 17-23 25 High 20 >15 -- ILow 4-37- 2 10 0 Naxal Belt North East Kashmir Punjab (Data Source: MHA Reports, IndiaStat, Economic Times, India Spend, Census data (http://www.indiaspend.com/investigations/link-between-government-spend-naxal-surrender); (http://articles.economictimes.indiatimes.com/2013-04-25/news/38817341_1_additional-central-assistance-schemecrore-cent-grant-basis); "Timely Waiver," India Today, August 4, 1997, p. 17 (http://indiatoday.intoday.in/story/i.k.gujral-writes-off-outstanding-rs-8500-crore-debt-punjab-accumulated-since-1984/1/276375.html), Sidhu and Bains, The Indian Police Journal,55 (1), 2008, p. 57) 423 - Table 6.6: Use of President's Rule (Measuring Effort) State High Effort Punjab West Bengal Kashmir (I) Telangana* Days % of Campaign 2,442 1,017 2,456 496 61% 56% 45% 44% Low Effort 11% 1,481 Nagaland (II) 8% 567 Mizoram 4% 440 Manipur 2% 214 Assam 0.40% 26 Bihar 0% 0 Andhra Pradesh *Though the campaign was prior to President's Rule and Article 356 of the Indian Constitution, this measures the period under military occupation rather than civilian administration Figure 6.11: Dirty-ness of Campaign (Non-combatants killed by State / Total killed by State) (Measuring Violence) 0.8 0.6 0.4 0.2 0 b,4 Table 6.7: Ratios of Kills, Arrests, Surrenders, Rehabilitation (Measuring Violence) High Violence Ratios Kill: Arrest Kill: Surrenders Kill: Rehab Rehab: Surrenders Kill: Arrests & Surrenders Total Renegades (Source: Numerous) Manipur 1.02 36.62 27.58 1.33 0.99 Mod-Low Violence J&K 0.9 8.05 53 0.15 0.81 Punjab 0.42 4.53 2.78 1.63 0.38 Assam 0.52 0.49 1.06 0.46 0.25 5000 300 4843 424 Very Low Violence Naxal II Naxal III Andhra 0.09 0.05 0.28 0.11 .27 .33 1 0.33 0.07 0.03 Table 6.8: Detention and Complaints (Measuring Violence) (all per million people) Complaints Against Armed & Paramilitary Forces High J&K Manipur Nagaland Moderate Assam Punjab Low 15.99 9.15 Persons Detained Under Natl. Security Act -- 0.5 205.8 107.5 1.84 13 -- -- Persons Detained Under TADA Persons Detained Under TADA/Yr 1423 871 197 136 -- -- 567 754 57 76 --1.3 0.22 23.5 -112 Andhra Pradesh 0.11 India 0.6 3.4 89 15 (Data estimated from following sources: Complaint data (1998-2001): Lok Sabha Unstarred Question No. 2353, dated 04.12.2001. via Indiastat.com. Detention data (2003-06): Lok Sabha Unstarred Question No. 2474, dated 14.03.2006 via Indiastat.com. TADA data (1985-June 30, 1994): Home Ministry, via PUDR, via "Black Law and White Lies: A Report on TADA, 1985-1995," Economic and Political Weekly, 30 (18/19) May 6-13, 1995, p. 977981. (Table 1 and Table 3). TADA/Yr data: "TADA-A Monster Handcuffing Human Rights," Lawyers for Human Rights International, Jan 2, 2002 (based on Union Home Ministry Status Paper in April 1995); PUDR Lawless Roads: A Report on TADA: 1985-1993, Delhi: People's Union for Democratic Rights, September 1993, p. 24 (based on parliament replies); S Ramakrishna, "TADA Repealed But Accused Still Suffer," India Express, August 16, 1998.) Naxal States After confirming values of counterinsurgency strategy in each of the 14 Indian campaigns, we can now begin to evaluate how the core-periphery dynamics of territory and identity fare in explaining this distribution of strategies. 425 IV. Evaluating the Theory 4.1. Correct Predictions The thick qualitative descriptions along with quantitative measures that identify each counterinsurgency campaign's strategy provide a baseline against which to evaluate the predictions of my theory (table 6.9). The Medium-N analysis allows one to maintain concept validity while examining the entire set of Indian counterinsurgency campaign observations to identify how variation of the dependent variables of effort and violence correlates with the hypothesized independent variables of territory and identity. Not only does this analysis allow us to estimate how many of the cases can be explained with the theory, the process of data collection and coding the variables for the medium-N analysis also revealed the cases most ripe for further exploration through intensive and focused comparisons to be conducted in chapter 7. Table 6.10 shows that in fact the theory does successfully explain patterns of sub-national variation in India. Based on the coding procedures for the dependent variable, the theory explains 13 of the 14 Indian counterinsurgency cases. Even if two borderline cases that could be disputed were re-coded, the theory would still predict 11 of 14 (and predict at least one dimension of strategy in each campaign-12 of 14 on the dimension of effort, 13 of 14 on the dimension of violence). Two cases require further examination: the Hyderabad campaign is not well-predicted by the theory; and the strategy of the Naxal I AP and Naxal II AP campaigns proved challenging to code, bordering between two potential values. 426 I --- I NJ" Table 6.9: Detailed Predictions of Strategy IV2: Identity IVI: Territory Costs Moderate Projection on Value Moderate Worth Low Embedded Low Predicted Strategy Attrition High High Low Moderate Low Attrition Incentiv es High 19481951 Telangana region of Hyderabad, now Andhra Nagalim (includes parts of Assam, Arunachal Pradesh, Manipur, and Nagaland) Mizoram CPI Identity Group Deccani Muslims SC/ST NNC Nagas 19551975 Low High Low Low Low Enfeeblement MNF Mizos Low High Low Low Low Enfeeblement CPI-ML SC/ST Moderat e Moderate Low Moderate Low Enfeeblement CPI-ML SC/ST Moderate High Moderate High Nagas Low High Low Low Low Population Security Enfeeblement 8 Manipur Manipur Low High Low Low Low Enfeeblement 9 Punjab Punjab Sikhs 19831993 High Low High High High Population Security 10 Naxal II -AP Andhra Pradesh NSCN - IM, NSCN-K PLA, UNLF, KCP, KNF, PREPAK Sikh Militant/ Panthic Committee PWG 19671971 1976ongoing 1982ongoing High Nagaland II Srikakulam and Telangana regions, Andhra Pradesh West Bengal (& Calcutta) Nagalim 19661986 19671971 SC/ST Low Moderate High Mitigation Assam Assam ULFA, NDFB Moderate Low Moderate Moderate Mitigation 12 Kashmir I Kashmiri militant groups High Moderate High Low Low Attrition 13 Naxal III CPI-M SC/ST High Low High High Mitigation Kashmir II Kashmiri militant groups Kashmiri Muslims 2004ongoing 20042013 Low 14 Jammu and Kashmir Divisions of J&K State Central India/ Naxal Belt Jammu and Kashmir Divisions Assamese, Ahom Kashmiri Muslims Moderat e Low Moderate 11 19872003 1990ongoing 19892003 High Moderate High Low Moderate Attrition No Observation 1 Hyderabad 2 Telangana 3 Nagaland I 4 Mizoram 5 Naxal I - AP 6 Naxal I - WB 7 State/Region Hyderabad state, Insurgent Group(s) Razkars Meiteis Date 1948 Table 6.10: Evaluating Predictions No. Predicted Strategy Attrition Effort Low Correct Prediction? X Dependent Variable Correct Violence Prediction? High 4 1 Observation Hyderabad 2 Telangana Attrition High High Attrition Enfeeblement Low High Enfeeblement 3 Northeast (All Combined) Nagaland I Enfeeblement Low High Enfeeblement 4 Mizoram Enfeeblement Low High Enfeeblement 5 Naxal I - AP Enfeeblement Low Moderate High Low -Enfeeblement (borderline Mitigation) Population Control 4/? 6 Naxal I - WB Population Actual Strategy Enfeeblement Security 7 Nagaland II Enfeeblement Low High Enfeeblement 8 Manipur Enfeeblement Low High Enfeeblement 9 Punjab High Low Population Control 10 Naxal II -AP Population Security Mitigation Low Moderate Mitigation Low Low -Mitigation (borderline Pop. Control) Mitigation Kashmir I Attrition High High Attrition 13 Naxal III Mitigation Low Low Mitigation 14 Kashmir II Attrition High /? 11 Assam 12 High I__ -Political War) Attrition (with 00 4.2. Under-Predicted Cases While the theory correctly predicts high levels of violence in Hyderabad, it does not correctly anticipate the limited effort the state employed. This may partly stem from Hyderabad simply being a bizarre case whose inclusion amongst India's rebellions may be disputed. While there is evidence to suggest there was an insurgency, this may also be an extra-territorial conflict or a war of annexation since accession to the Indian union was in theory voluntary and Hyderabad certainly tried to remain independent. Nevertheless, some scholars point out how the Indian state (primarily the key leaders like Nehru and Sardar Patel) perceived the Kingdom of Hyderabad as already being a part of India and meant to be integrated through peaceful accession. 540 They also saw Operation Polo not as an invasion of Hyderabad but as a "police action" to crack down on the Hyderabad ruler's Razakar irregulars who supposedly "let loose an orgy of violence" and ensure the region's peaceful transition into the Indian union.' Once the Nizam's forces were defeated in a conventional operation, he surrendered, willingly acceded to the Union, and the conflict was largely over from the state's perspective. This may explain the state's limited effort after this clean victory despite continued Razakar activity and the wave of communal violence unleashed upon the Muslim population of the region as retribution. It is also conceivable that the low manpower may have been based on an expectation that state forces would not need to be deployed everywhere since the insurgent base was Muslim and concentrated around the city of Hyderabad but the vast majority (87%) of the region was Hindu, thus would not necessitate significant force deployments. Low sustainment may have been a function of the unexpected rapidity of their success over the Razakar forces. Based on Phillips Talbot, "Kashmir and Hyderabad", World Politics, 1 (3), 1999, pp. 324-25. L.N. Subramanian, "Operation Polo: The Liberation of Hyderabad" BharatRakshak, 2(3), November/December 1999. 542 Talbot, 1999, p. 323. 540 541 429 planning concepts described by the officers who led the campaign,5 4 3 the Indian state appeared prepared and willing to deploy more forces, stay longer, and adapt politically and operationally as needed. However, they were able to smash Hyderabad conventional forces within five days and following their surrender, the military turned their attention to the Telangana insurgency located within the Hyderabad region. Irregular Razakar forces were certainly a point of concern, but it appears that Hindu irregulars and militia were able to liquidate them within a few weeks. If the state knew about and anticipated this, as some evidence suggests, then its strategy may have been much closer to attrition as the theory predicts. Furthermore, had the resistance proved enduring and widespread, in all likelihood, the Indian state would have rapidly escalated its force commitments and sustainment of the campaign. In the case of Naxal I-AP, the state's use of violence borders between high and low in both qualitative and quantitative measures. Nevertheless, the theory's prediction of an enfeeblement strategy seems validated by the evidence since many of the tactics and force employment in this campaign described earlier were similar to the Telangana campaign, but conducted with fewer resources for a shorter period of time. There appears to be little evidence of the state attempting to address the demand-side of rebellion or devolve authority to those composing the rebel base the way it did in West Bengal. Moreover, violence is likely to have been less observable and underestimated in rural Andhra relative to urban West Bengal where casualties were meticulously documented, 544 and more accurate figures would likely confirm this strategy of enfeeblement. Finally in the Naxal II case, the Indian state periodically deployed high effort and some signature tactics of population control, particularly in the latter part of the campaign. Some have 1 J.N. 544 Chaudhuri, 1949. Ashoke Kumar Mukhopadhyay, The Naxalites Through the Eyes of the Police: Select Notificationsfrom the CalcuttaPolice Gazett, 1967-1975, Kolkata: Deys Publishing, 2006. 430 likened this campaign to the one deployed in Punjab based on the efforts to develop the specialized Greyhound force, integrate within an upgraded state police force, and use intelligence-driven police-initiated targeting of insurgents. 54 5 However, this effort was not operationally sustained until the early 2000s, which is why it took roughly eighteen years for the Greyhound force to produce meaningful results.54 6 Even in this instance, the state government forces escalated effort with a limited degree of central government help, because by the late 1990s, there was much more under threat in the region. The Telugu elite had begun to see the potential for the transformation of this rural backwater into a center of capitalist growth in the Hyderabad region and was no longer tolerant of Maoist insurgents mucking about.5 4 7 Rather than taking the requisite measures to preserve these gains, whether by fully degrading insurgent combat power or integrating them into the political fold of the state, Andhra's strategy seemed to end after the displacement of insurgency to neighboring states where the insurgency continued to thrive just across the border into Chhattisgarh and Orissa.548 While the state of Andhra Pradesh achieved control, perhaps with a bit of good luck from some insurgent mistakes, 549 the Indian central government persisted with its strategy of containment that simply mowed the grass for a few years, only to see it crop up elsewhere soon after. 4.3. Evaluating Competing Explanations For the purposes of my larger study, I have controlled for state-level structural variables like strategic/organizational culture, civil-military relations, doctrine, as well as state strength and regime type in order to explore the sub-national variation within a country (India and 545 Sahni; Verma, 2014 546 Mishra, 2007; Sahni 2008. 547 Author Interview, Indian national security journalist, New Delhi, India, December 17, 2013, (#147). 548 Vij-Aurora, 2011. 549 Janyala, Sreenivas. "When They Talked And Failed," IndianExpress, March 14, 2010. 431 Pakistan). Strategic culture and civil-military relations generally do not temporally vary much at all but some structural conditions can change in theory. India is relatively constant in its macroconditions as a weak, democratic state. Its political institutions have been consistently rated as democratic and free based on polity and freedom house scores but there are some minor fluctuations. And while it has exhibited substantial economic growth since the 1990s, this has not fundamentally altered its state strength the way it did for the East Asian NICs as it still remains in the bottom third of development in terms of GDP/capita as well as measures of political extraction. 550 Therefore, in the Indian cases these structural conditions are relatively constant, but they can be relaxed slightly to explore whether some more fine-grained variation can explain patterns of the dependent variable. 4.3.1. State Strength If examining a finer variation over time, one could argue that the Indian state leviathan fundamentally changed in capacity and strength after its economic takeoff by roughly the mid1980s to early 1990s, 551 and that this accounts for high effort strategies in cases like Kashmir. This argument cannot be dismissed out of hand, as even a selection of strategy is bounded by the available resources of the state. But state strength is not sufficient or complete enough to account for high effort strategies prior to this takeoff (notably in Telangana, West Bengal, and Punjab). Additionally, the state capacity argument cannot explain the continued anemic efforts in Nagaland and Manipur as well as new low-effort campaigns of Assam and the Naxal III in central India. All these campaigns have continued to be fought with low-effort amidst a dramatic rise in state capacity due to decades of robust economic growth since 1990. Furthermore, there 55 0 Ashok Kotwal, Bharat Ramaswami, and Wilima Wadhwa, "Economic Liberalization and Indian Economic Growth: What's the Evidence?" JournalofEconomic Literature,49 (4), 2011, pp. 1152-99. 551 Kotwal, 2011. 432 / are clear divergences in effort between campaigns, despite direct overlap in the same period of time, for which macro-structural factors simply cannot account (for instance the Naxal I campaigns in West Bengal and Andhra from 1969-71, the Manipur and Punjab campaigns beginning in the early 1980s, or the Kashmir and Assam cases beginning in 1989/90). The Indian Figure 6.12: Indian State Strength/Capacity, 1947-2008 (CINC Scores and GDP/Capita) I U.7 $3500U 0.06 0.06 $3,000 0.05 0.05 $2,500 $2,500 0.04 0.04 $2,000 0.03 $1,500 Cu 0 0.03 0. 1 0.02 Cu U C.' 0 $1,000 4. 1~ $500 0.01 A $0 $0 1947 1957 1967 1977 1987 1997 2007 (Sources: Correlates of War and Maddison GDP data) Figure 6.13: Indian State Strength/Capacity, 1960-2007 (Relative Political Extraction (RPE) and Reach (RPR)) 1 0.8 0.4 0.2 0 w rnrrrrrrrl 1960 1966 1972 1978 RPE (Agr) -- 1984 w 1990 RPE (GDP) rrrrr1 1996 RPR (Source: Relative Political Capacity data, Kugler et al) 433 2002 cases prove that just because a state is stronger and richer does not mean it will necessarily choose to muster all of its money, manpower, materiel, and political will into defeating rebellions if they do not threaten what it most values. 4.3.2. Regime Type By commonly used quantitative measures, India has remained a free democracy through nearly all of its history save for a short period between 1975-77, but one can argue that real competitive electoral politics only emerged after the defeat of the Congress party's stranglehold on national politics in 1977. Again there is no denying a secular trend that has on average constrained the state's freedom of action (particularly with respect to tactics of forcible resettlement of populations), but this true throughout the world and has more to do with broader normative and informational constrains. Furthermore, even under the Congress-dominated regime, the state still exhibited constraints in West Bengal, while in the more democratic age, Indian has managed to regularly employ quite brutal tactics and deliberate civilian targeting (often selectively but sometimes even indiscriminately) in campaigns in Kashmir and Manipur.55 2 Similar to the British, the Indian leaders have routinely suspended or subvert normal democratic procedures in a region whenever it has suited them. 5 3 In Hyderabad and Telangana, military regimes were put in place to combat rebellion. In Nagaland and Mizoram, the village relocation procedures made a mockery of purported Indian democracy and civil liberties, and 552 Levy and Scott-Clark, 2012. For evidence of subversion of Kashmir democracy, see Staniland, 2013. On the Northeast, see M.K. Dhar, "The Lake of Fire: Manipur Bums, Delhi Bungles," Maloy Krishna Dhar Blog, May 20, 2010. In some cases, the Indian government has suspended state legislatures and instituted President's Rule via an appointed governor whenever it believed it a state is not able to function. It has implemented this at various points in twenty-five of its twenty-eight states regardless of the region. However, there appears to be no correlation between implementation of this of President's Rule in the event of a rebellion and strategy -- it did this during rebellions in Assam, Punjab, J&K and Manipur despite the fact that different strategies were employed in each of them. 553 434 even today, detailed accounts of the politics of J&K and the Northeast reveal, "politics...are not conducted according to democratic norms" due to the absence or deliberate weakening of democratic institutions. 54 In J&K and the Northeast, the state has been protected by the Armed Forces Special Powers Act, a highly undemocratic law that that gives blanket immunity to security services stationed for killing civilians and thus a wide latitude to define who insurgents are.5 55 The fact that the Indian government has been able to regularly find routes around normal democratic processes suggest that these institutions are not particularly restraining as the regime type literature seems to suggest. The core-periphery theory performs better than the structural variables of state strength and regime type predicting at least 12/14 cases of effort and at least 13/14 cases of violence, but more importantly, the theory predicts and explains cases that state strength and regime type cannot. Figure 6.14: Indian Regime Type, 1950-2012 (Polity and Freedom House) 6 10 Thesol DemocracyTheol 0 "Freedom" ThresholdE -10 + 1950 rTT4 1960 1970 1980 1990 2000 0 2010 (Source: Polity IV, Freedom House) 554 555 Marwah, 2009, p. 184; Staniland, 2013. Author Interviews with Indian human rights activists, New Delhi, India, November 14 and 30, 2012 ((#73, #99). 435 4.3.3. Reputation Theories of reputation and internal deterrence have been offered to explain why states fight, but after some analysis, it cannot effectively explain how the Indian state fights. Though the reputation theory conflates effort and violence in fighting rebellion, ultimately it is making predictions about the level of violence a state is willing to use. A cursory glance at Indian strategies over time would appear to validate the reputation theory that predicts a gradual shift from high violence to low violence strategies. However, it cannot explain significant departures, such as the intensity of violence the Indian state is willing to use in Kashmir, after countless campaigns where it fought and signaled its tenacity or early campaigns where it employed violent negotiations in West Bengal and even accommodative gestures in Telangana and the Northeast. Furthermore, since reputation expects states to eventually make accommodations after it has sufficiently signaled its reputation for toughness, the theory accounts for some of the longstanding conflicts where the state has refused to make the necessary concessions to end conflicts that have endured (e.g. Nagaland, Manipur, Kashmir) or recurred (Naxal conflicts) for multiple decades. More importantly, a close look at the Indian cases reveals a number of departures from the reputation theory's observable predictions and mechanisms. Since the reputation theory is chiefly concerned with the tough signals the state emits in its early campaigns rather than the substance of its policy, it cannot account for significant state efforts to conceal the level of violence it committed in early campaigns (e.g. Hyderabad, Telangana, Nagaland, Mizoram), which belies claims of reputation theorists who argue states will seek to publicize the costliness of rebellion to deter future insurgents. 436 Additionally, the theory cannot account for the accommodative gestures the Indian state made in early campaigns of Telangana and Nagaland (respectively offering land reforms and statehood), despite never meaningfully following through on them. While this was not a substantial part of the strategy nor did it do much to mitigate the conflict, these gestures nevertheless undermined the state's ability to signal toughness and deter future challengers. In fact, many scholars have noted that these half-hearted, half-baked accommodative gestures of statehood in the Northeast have not only perpetuated conflicts and increased violence but "official policy has merely encouraged separatist tendencies." 56 Empty concessions to the Nagas in 1963 emboldened the Mizo revolt in 1966 and the Manipur mobilization in the 1970s, and specific Indian policies to encourage ethnic militias to counterbalance insurgencies have resulted in new separatist claimants like the Bodos of Assam. Even when the stakes are highest and you would expect concerns for reputation to dominate the minds of Indian decision makers, they apparently do not. For instance, the reputation theory has a hard time grappling with why the Indian state would have endorsed sweeping political deals like the 1986 Longowal Accord or award broad amnesties and political concessions (between 1991-93) to Sikh separatists in Punjab, when Kashmiri separatism was boiling in the mid-1980s and insurgency erupted in 1990. Clearly reputation theory has its limits. While it performs admirably in explaining the initial Indian state decision to routinely fight violent challengers, whether large or small, regardless of their location, it offers considerably less purchase in explaining how the Indian state has fought and specifically what type of military strategy it adopts or has adopted in the past. 156 K.C. Khanna, "The Troubled North-East: Impermanent Frontiers," Times ofIndia, March 14, 1974. 437 4.3.4. Learning/Organizational Culture Another of argument contends strategy may be a product of Indian security organizations learning over time, but there is substantial evidence that specifically contradicts the learning hypothesis. While organizational culture may be largely stagnant, one could argue that some militaries are "learning organizations" that allow for adaptation over time, and this may correspond with the observable secular trend of fewer high-violence violence strategies. However, this theory is particularly challenged by the persistence of high-violence strategies in Nagaland and Manipur, and more importantly in Kashmir, where despite a near limitless supply of material and political resources to learn and adapt, the state has persisted with the same strategy. A core gap in this argument is the problem of missing strategic adaptation, which should have been visible soon after the setup of the first Counterinsurgency and Jungle Warfare School in the late 1960s,557 or at least after doctrinal changes in the 1980s. 5 While some tactical changes were instituted-for instance, lighter forces, fewer population relocations, and a prohibition on area weapons-it is noteworthy that the Indian military generally persisted with large-scale, kinetic-centric strategies in old campaigns (Nagaland and Mizoram) as well as newer ones (Manipur and Kashmir). 559 Additionally, there is strong evidence to suggest that the secular trend of less violent campaigns has occurred in spite of "the operational culture of the Indian milita-y, which inhibits organizational learning." 60 This is due to the fact that in campaigns with strategies of low violence, local police forces rather than the military have been tasked as the lead. See page 25 of Rajesh Rajagopalan, "Innovations in Counterinsurgency: The Indian Army's Rashtriya Rifles," ContemporarySouth Asia, 13 (1), March 2004, pp. 25-37. " 8 Rajagopalan, 2000, pp. 58-64. "9 56 0 Cline, 2006, p. 143. Mukherjee, 2009, p. 140. 5" 438 4.3.5. External Support Recent scholarship has pointed to factors in the international environment that might motivate states to fight harder in counterinsurgencies, specifically links between insurgents and an external rival that render expectations of a future external threat environment all the more threatening.561 However, with the exception of the Naxal I and Telangana rebellions, all Indian insurgencies have received material support from major strategic rivals of Pakistan or China,562 a product of tit-for-tat exploitation of vulnerabilities common throughout South Asia. Pakistan's ISI is believed to have supported every insurgent group from the 1980s onwards including the Nagas, ULFA, Kashmiris, Sikhs, and even Tamil insurgents fighting the IPKF. "The Indian authorities soon realized that the ISI was anywhere and everywhere, where there was a chance of sparking a revolt."5 63 Indian security officials have certainly expressed a belief that even the present Naxal insurgency is receiving external support from Pakistan and China. 564 External support cannot account for why the state has employed more violence and more effort against past Naxal insurgencies (and their antecedents like Telangana) with no external support, than against the present Naxal conflict with at least some evidence of external links. An argument could be made that Pakistani material support to Kashmiri insurgents was uniquely larger and more involved but once again, close examination of new empirical work and interviews reveals similar efforts in Punjab, where a less violent strategy could not be accounted for by external support. The Pakistani state not only provided arms and safe haven to Sikh militants, but also generated an estimated 5,000 insurgents in training camps, many alongside 561 Butt, 2012. 562 List sources Joshi, 1999, pp. 168-69. This notion is reiterated in an editorial in a Hindi newspaper alleging Pakistani and ISI involvement in a number of insurgencies. "The Enfeebled Internal Security," DainikJagran(Delhi,), November 28, 1996, p. 6. Daily Report. Near East & South Asia, FBIS-NES-94-232 on 1994-12-02, p. 34. Accessed via Newsbank, February 19, 2014. 564 See table on external support; Author interview with a retired senior Indian Army officer, Gurgaon, India, December 11, 2013 (#141). 563 439 Table 6.11: External Support for Indian Insurgents No. Observation Support? Actor Type Source(s) Hyderabad Yes Pakistan Arms; Potential conventional intervention/air support feared JN Chaudhuri, 1949: 17; Guruswamy, 2008 2 Telangana No China, USSR NA - Ideological 3 Nagaland I Yes China, Pakistan Arms, Sanctuary, Training Cline, 2006: 130; Rammohan, 2011: 57-58, 103-104, 107 4 Mizoram Yes China, Pakistan Arms, Sanctuary, Training 5 Naxal I - AP No China NA - Ideological Bhat, CLAWS, 2009: 4; Ali Ahmed, Scholar Warrior, 2011: 17; Nibedon, 1980: 97, 125, 140, 164-165; K Roy, 2010, p. 209 MK Dhar, 2005: 88; Mishra, 2007: 26 1 6 Naxal I - WB No China NA - Ideological MK Dhar, 2005: 88; Mishra, 2007: 26 7 8 Nagaland II Manipur Yes Yes Arms, Sanctuary, Training Arms, Sanctuary, Training Bhat, CLAWS, 2009: 4; Bhat, CLAWS, 2009: 4; ENR articles 9 Punjab Yes China, Pakistan China, Pakistan, Myanmar (tacit) Pakistan, Sikh diasporas 10 Naxal II -AP Yes Pakistan (some via Northeast insurgent organizations) Arms, Financing, Intelligence, Operational Guidance, Sanctuary, Training Arms MK Dhar, 2005: 321, 324, 375; Kumar, Faultlines, 2007; Sharma, 2006: 270, 285; Phadke, 1992: 11; P Swami, 2007: 147-148; Laurent Gayer, 2009: 242-43 T Sunil Reddy, "PWG Links with Kashmir Ultras Bared," Indian Express, September 29, 1998; "Home Secretary: ISI aiding Naxalites," Tribune News Service (Chandigarh), March 7, 2000; NS Jamwal, "Counter Terrorism Strategy," Strategic Analysis (IDSA), 27 (1), JanMar, 2003, p. 13; Soumen Bhattacharjee, "Intelligence alert on ISI operations - South Bengal new haven for Pak agents, says IB" The Telegraph (Calcutta), Sept. 21, 2004; Kanchan Gupta, "Naxals, India's Enemy Within" Rediff.com, November 25, 2004 11 Assam Yes Kashmir I Yes 13 Naxal III Yes China, Pakistan Arms, Financing, Logistics, Operational Guidance, Sanctuary, Training Arms, Financing, Intelligence, Operational Guidance, Sanctuary, Training Arms, Financing, Training Bhat, 2009: 8-10; 22-23; 32-46; EN Rammohan, 2011: 54, 85;"Assam Militants are Spying for Pak - Army," Times of India, June 26, 1999 12 Pakistan, Bangladesh, Myanmar, Bhutan Pakistan Praveen Swami, 2007; Staniland, 2014; Oberoi, 2007; Chadha, 2005; Levy and Scott-Clark, 2012; Sinha, USI lecture, 2005 Author Interviews; Uday Kumar, Maoists Terrorist and ISI: A collaboration of Naxalite and ISI against India, New Delhi: Ankit Publications, 2011; Pervez Iqbal Siddiqui, "Lethal Combo, Maoist-Naxal Nexus," Times of India, November 20, 2004; Deepak Kumar Nayak, "Naxals' Foreign Links," Geopolitics (India), September 2013, pp. 6062; Ben West, "Pakistan and the Naxalite Movement in India," Stratfor, November 18, 2010; Dwaipayan Ghosh & Neeraj Chauhan, "Cops Nail China Link With Naxals," Times of India, October 8, 2011 14 Kashmir II Yes Pakistan Arms, Financing, Intelligence, Operational Guidance, Sanctuary, Training Praveen Swami, 2007; Staniland, 2014; Oberoi, 2007; Chadha, 2005; Levy and Scott-Clark, 2012; Sinha, USI lecture, 2005 Afghans during the 1980s.5 65 Additionally, at least 45,000 suspected Sikh insurgents were captured along the Punjab/Pakistan border and potential received tactic support from Pakistan.56 6 Additionally, external support cannot explain different patterns of violence within the periphery like the Northeast. Specifically the state relied on higher levels of violence in Nagaland and Mizoram in the past, and Manipur today, than it has in Assam. While my theory does not explicitly rule out this external support argument, it more likely interacts with the variable of insurgent group identity. In the case of Kashmir, it was not simply the realistic prospect of secession and alignment of a future Kashmiri state with Pakistan that drove high levels of Indian effort and violence, but it was specifically the proximity in identity of the Kashmiri and Pakistani people-and perhaps even conflation-that permitted and motivated an attrition strategy. Violence was permitted because of Kashmiri "otherness." As one prominent Indian human rights activist said, "we see them as Pakistanis and our hatred for Pakistan runs deep" and with their Muslim identity they are "definitely dispensable enemy aliens." Such an otherness "was not so pronounced with the Sikhs." 5 67 4.3.6. "Sons of the Soil" Similar to rebel aims variable, there is a notion that a certain type of rebellion, a "sons of the soil" rebellion where the rebelling group fears a loss of local/regional power through demographic encroachment of mainland migrants, can explain the dynamics of a conflict, particularly its duration.568 In the rebellions of India, there is evidence from most conflicts that D.P. Sharma, The PunjabStory: Decade of Turmoil, New Delhi: APH Publishing Corporation, 1996, p. 270; Laurent Gayer, "The Khalistan Militias: servants and Users of the State," Armed Militias ofSouth Asia: Fundamentalists,Maoists and Separatists,Ed. Laurent Gayer and Christophe Jaffrelot, New York: Columbia University Press, 2009, pp. 242-43. 566 Swami, 2007, p. 148. 567 Author Interview, Indian human rights activist, New Delhi, India, November 30, 2012 (#99). 568 Fearon, 2004; Fearon and Laitin, 2010; Mukherjee, 2014. 565 441 the rebellion was in part motivated by real or perceived demographic encroachment and a sense of entitlement to the territory they held. 569 The duration may in fact be endogenous to strategy of the state and whether it takes meaningful actions to quell the rebellion, and potentially allay the fears, but my evidence suggests this has more to do with the geography of the insurgency rather than the particular source of insurgent grievances. Distinct state strategies towards "sons of the soil" rebellions are evident in the cases of Assam, Manipur, and Punjab. Separatists in Assam and Manipur have continued to rebel on the basis of Hindu/mainland settlers and encroachment, and this too was one of the grievances listed by Sikh militants who wanted to carve out an independent Khalistan. The state's different strategies, motivated by the independent variables hypothesized in this project, were able to allay those concerns in Punjab and thus quell the rebellion whereas the state's moderate to low efforts have allowed the Manipur and Assam rebellions to persist indefinitely in a chronic, steady-state of low-level violence. At the same time, conflicts without sons of the soil have elicited limited effort and still endured for decades such as the Naga and Mizo conflicts. 4.3.7. Insurgent Threat Theories of strategic interaction often highlight the attributes of the rebel to explain threat perceptions and the dynamics of conflict. Indian officials and practitioners will readily acknowledge that higher levels of effort in Punjab and Kashmir were driven by a greater sense of insurgency and threat perception, but this does not necessarily correspond with how the civil 569 Fearon and Laitin classify India's Northeast as part of this phenomenon, but the Nagas and Mizos rebellions had nothing to do with in-migration. There is also evidence this sons of the soil dynamic occurred in Punjab (which they acknowledge though neglect the fact that this triggered a rebellion in the early 1980s) and the Naxal conflicts. See Kennedy and Purushotham, 2012, p. 853. 442 conflict literature measures insurgent threat. A handful of the most common variables fail to offer any traction on the dependent variable. Rebel Size. Large rebellions in India have not necessarily produced state strategies of high effort or violence. Estimates of strength from large datasets that are generally removed from the specific details of the literature and since these figures are highly problematic, only a clear difference in the order of magnitude would persuasively distinguish categories of rebel size. Yet in all of India's insurgencies, there have roughly been between 1,000 - 10,000 insurgents. 570 By numbers, the strongest insurgencies have been the Mizo and recent Naxal insurgencies57 1 yet the Indian state has responded with weak effort while putting forth much greater effort against smaller rebellions in Telangana and West Bengal. While some theories expect higher levels of violence to be explained by this rebel manpower figure,5 72 it is difficult to explain how either the Assam or Punjab rebellion, with roughly the same (or by some estimates greater) rebel strength, such as the Kashmir rebellion, were met with a much less violent state counterinsurgency strategies.57 3 Centralized Control. Nearly all of India's 28 rebel groups identified in the NSA dataset are coded to have centralized control except the Sikh and Kashmiri insurgencies. To begin with, these are highly suspect codings, likely brought on by an observation bias of big dataset collection with abundant information on the infighting Kashmir and Punjab insurgent groups but a lack of information on other cases where this same infighting also prevailed. A deeper knowledge of the cases and literature would reveal that both Manipur and Assam had far less 570 Non-State Actor Dataset from Cunningham et. al., 2009. For instance, current estimates of Naxal manpower are around 10-20,000 hardcore plus up to 50,000 people's militia. See "We Must Root Out Naxal Meance, Build On Our Successes: Manmohan Singh," DNA India, November 23, 2013; Chakravarti, 2008, p. 3-4. 572 Valentino et al, 2004. 573 My figures find them to be closely matched at around 5,000-6,000 but the NSA data suggests Punjab had twice the insurgent size as Kashmir (9,000 to 5,000) 5 443 centralized insurgent organizations, which resulted in significant fragmentation during the insurgencies. Furthermore, it is highly atheoretical and counterintuitive to contend that low centralization (and therefore presumably a lower threat) triggers higher levels of state effort in Kashmir and Punjab, and even if true, it could not explain high effort against centralized insurgencies in West Bengal and Telangana. Rebel Strength and Fighting Capacity. The invariance of relative rebel strength in India cannot explain the different strategies the state has deployed. Again, nearly all 28 rebel groups categorized as "much weaker" than the state (except the Maoists in Telangana and the Naxal rebellions which are simply "weaker") and the groups are uniformly coded as having a lower fighting capacity relative to the state with highly professional military and paramilitary forces inherited from the British Raj. 574 At a more precise qualitative level, practitioner accounts contend the Nagas fought with the greatest skill, the Kashmiri militants (like LeT) with the greatest moxie, Punjab militants established the most fortified defensive position, and today the Naxals have exercised significant discipline in company-sized operations, but none of these can account for the variation of the dependent variable.57 5 Arms Procurement.Arms procurement might be an indicator of insurgent strength but again it is not accurately measured by existing datasets and when it is, it reveals little variation between cases to explain different strategies. The NSA dataset only codes the Kashmir insurgency as having arms procurement capacity but a close examination of the cases shows this is simply not true. However, the Sikh insurgents along with nearly every Northeast insurgent group were able to procure arms via porous borders through external supporters like Pakistan and China and other militant groups and arms bazaars frequently found in the ungoverned spaces of 574 Non-State Actor Dataset from Cunningham et. al., 2009. 575 Practitioner accounts based on author interviews in 2011-2013 444 Bangladesh and Myanmar. 576 Even the People's War Group in Andhra Pradesh was able to procure weapons by forging links with the LTTE in the early 1990s, 577 and today the Naxals have both sophisticated internal arms production as well as numerous external sources of arms procurement through porous borders and smuggling networks that have serviced many of India's Northeast insurgencies as well (see figure 6.15).578 Figure 6.15: Naxalite Weapons Supply Network (Source: Stratfor Intelligence) Gayer, 2009. Trinath Mishra, 2007, p. 110. 57 This argument has also been made in an author interview with a retired senior Army officer, Gurgaon, India, December 11, 2013 (#141). 576 57 445 TerritorialControl. While territorial control is considered important for rebel survival, the only group believed to have it is the ULFA of Assam. First, it is debatable whether the ULFA have had any more territorial control than Naxal insurgents recently held 579 or than Naga and Mizo insurgents had in their first decade of their insurgencies. Second, it is then difficult to explain why, if this poses a threat, the state would adopt a low effort, low violence mitigation strategy in Assam. Goals/Aims. Separatism is usually used to explain violence or intensity but in the case of India, separatist insurgencies have elicited every type of counterinsurgency strategic response. Except for the Naxal and Telangana rebellions, insurgent aims have generally been non-variant as they have all been separatist movements seeking secession. Even if the commitment to these goals varied within the insurgent organization(s), they were generally believed and consistently feared by the Indian state. Likewise, all these insurgencies except the batch in central India have operated on India's borders and benefitted from foreign sanctuaries as well as active foreign support. Ultimately, my theory does not discount a state's threat perceptions but contends that they are heavily interactive if not dependent on the specific territory that is threatened. If insurgents are able to contest valuable territory or project violence out to areas that are valuable, then the state will have a greater incentive to respond with a high-effort strategy. The Sikh and Kashmiri insurgents as well as Naxal insurgents in Bengal, while already contesting territories valued by the Indian government, were also able to project violence throughout parts of the country including the capital or key urban centers. Some Kashmiri militants have been involved with high-profile terrorist attacks in Delhi, while Sikh militants committed the most terrorist 579 For instance, the Naxals have at least enough territorial control to have been running 84 training camps within their areas of operation in 2011. Lok Sabha Starred Question No. 493, 4/30/2013, accessed on Indiastat.com, July, 9, 2014. 446 attacks outside Punjab into Haryana, Chandigarh, Delhi, Kashmir, and Maharashtra. By contrast the Northeast insurgent groups and most Naxal/Maoist groups have not projected violence outside their regions on to more valued territories (sometimes due to limited capacity, sometimes deliberately), thus not providing the incentive for the state to take decisive measures with a high effort counterinsurgency strategy. In a final analysis, the theory performs better and accounts for more variation than the litany of insurgent threat arguments. V. Conclusion: India and the Core-Periphery Theory The medium-N analysis of the Indian counterinsurgency case validates the theory's key predictions of how core and peripheral territories and identity groups shape different strategies to combat insurgency. The Indian state exerted significant counterinsurgency effort to regain control of contested core territory in its Northwest, but it allotted far less effort to contain insurgencies in its peripheral zones in the Northeast and rural central corridor. Rebels from core identity groups who were valued by and embedded within Indian institutions restrained state force employment from unleashing the kind of violent offensives states and militaries prefer, but rebels from peripheral groups did not activate the same restraints and in come cases, motivated greater state violence. Identifying meaningful patterns of variation in Indian counterinsurgency strategies is a contribution in its own right, especially when the theoretical and empirical literature glosses over this feature. The types of strategy are difficult to deny, particularly when the campaigns are subjected to rigorous and consistent measures. First there is demonstrable variation in the level of effort the state deployed in certain campaigns. India opted for limited-effort containment strategies when state capacity was high, allowing it to conduct high-effort campaigns elsewhere 447 and even when there were no other simultaneous campaigns to siphon resources. Furthermore, the Indian state chose more offensive strategies targeting the supply-side of insurgency in particular cases while in others it choose more defensive strategies that also addressed the demand-side of insurgency and displayed significant restraint and self-correction. A review of each of the cases suggests the state pursued an attrition strategy in Kashmir, population control in Punjab, enfeeblement in most of the Northeast campaigns, and mitigation in most of the Naxal/Central India campaigns. Territory and identity both played significant roles in shaping Indian counterinsurgency strategies. The Indian state spared no expense, manpower, or political capital when Punjab and Kashmir were threatened with insurgency because they were not only central to the economy and national defense but also to the "idea" of India. The same state leaders were simply not motivated to dispatch the same sustained flow of resources to counterinsurgency campaigns in far-flung mountainous regions of the Northeast or the Maoist-infested jungles of central India, which ran contrary to their material and strategic incentives and remained disconnected from the national imagination. Apart from the resources, Indian leaders found it difficult to unleash and sustain a nasty, brute force counterinsurgency campaign against Punjabi Sikhs highly embedded in Indian security institutions or against the highly sympathetic and legitimately aggrieved SC/STs. These inhibitions compelled the state to restrain kinetic force while substituting selective targeting, non-kinetic tactics, and politico-economic inducements. The same restraints were simply not present when the state faced down numerous Northeast separatist rebels or Kashmiri Muslim rebels. Locating these ethnic groups outside the fold of state institutions or moral worth lowered the costs of brute force, evoked distrust and disdain, expanded the potential 448 set of insurgent participants and irreconcilables, and permitted the state to apply greater violence and indiscriminate force. Along the way this chapter has made additional contributions. The India cases reveal the limits to the utility of macro-level theories to explain and predict counterinsurgency strategy. High state strength pushes the upper boundary of a state's potential capacity but it does not guarantee state leaders exhaust this potential. A democratic regime's norms and audience costs on average may restrain a state, but not necessarily against a group that has no moral standing or political strength and is perceived as an external enemy. Mountainous terrain alone cannot prevent a state from expanding its reach and control if it has sufficient strategic motivations. Organizational learning in professional organizations may not occur in every theater when the returns to costly adaptation are discounted. Even a military's organizational or strategic culture may not have as much influence if the state leadership chooses to work around this with intelligence agencies, irregular auxiliary forces, or cheaper and less professional paramilitary forces. Inclusion of multiple, lesser-known cases like the Telangana campaign, which is often glossed over by most histories of the Indian counterinsurgency, 580 offers empirical contributions as well as theoretical insights. By examining cases under varying structural conditions, it reveals how a fairly weak and newly independent India was still able to crush the Telangana insurgency in an attrition campaign. It also reveals the limits of organizational learning since it shows the state committed and persisted with a number of counterinsurgency "mistakes" much earlier than commonly recognized. This chapter has studiously avoided evaluation of counterinsurgency outcomes, though it should be noted India's different strategies have not necessarily corresponded with success. 5 Mukherjee, 2009, p. 141. 449 Punjab is often held up as a model case for its use of police and political concessions central to population control, but brute force seems to have successfully crushed insurgencies in Kashmir and Hyderabad. Low-effort enfeeblement in Mizoram and mitigation in Andhra Pradesh appears to have produced qualified successes, while high effort strategies in Telangana and West Bengal did not prevent the recurrence of insurgencies in the same region. This insight might provide the basis for future theoretical contributions. Though the relationships between territory, identity, and strategy revealed by this medium-N analysis are compelling, it is worth asking whether there is any evidence directly connecting contested territory to state calculations of strategic effort or rebel identity to state intent on the use of violence. The next chapter will seek to observe direct evidence of the hypothesized mechanisms linking strategy affective responses and strategic calculations of territorial and group worth (or spatial and social distance) through tactical choices and strategic decisions. 450 Chapter 7 Core-Periphery Mechanisms in Puniab, Kashmir, and the Naxal Belt I. Introduction The medium-N analysis of all Indian counterinsurgency campaigns reveals a relationship between territory, identity, and strategy that forms a distinctive pattern, but it is worth asking whether there is any evidence directly linking contested territory to state calculations of strategic effort or rebel identity to state intent on the use of violence. Through a close analysis of the mechanisms described in chapter 3, this chapter will present observable evidence that in some cases, the Indian state strategy was directly constrained or motivated by its value of the contested territory and/or the identity of the rebel social base. This chapter closely examines three casesPunjab, Kashmir, and the ongoing Naxalite (III) rebellion-to trace the mechanisms that connect my theory of territory and identity to strategies of effort and violence, specifically mechanisms structuring incentives, constraints, and opportunities. These Indian counterinsurgency cases are chosen for their data-richness having been well trod by historians, practitioners, and journalists. Because the India adopted different strategies in each of these cases-populationcontrol in Punjab, attritionin Kashmir, mitigation in the Naxal III campaign-all the theorized mechanisms structuring incentives and constraints on effort and violence can be fully explored. First, I expect to find evidence that Indian state actively weighed the costs of a controlling a particular contested territory relative to the incentives it posed, and that this led to observable choices of counterinsurgency effort in material, political, organizational, and operational terms. Second, I expect to find evidence of Indian state empathy, trust, vulnerability, and informational access with rebelling groups of high moral worth and embeddedness, (and evidence of their absence with the opposite groups) and that this had an observable effect on the 451 counterinsurgency campaign's tactics and use of violence. Finally, in each case I evaluate alternative explanations. I intend to present evidence linking territory and identity to observable properties of counterinsurgency campaigns-which my theory associates with distinct strategies-based on journalistic and expert historical accounts of choices, decision processes, and critical junctures as well as from state officials' memoirs, public statements, and interviews I conducted during fieldwork. The Punjab and Kashmir counterinsurgency cases both offer opportunities to examine when and how the state judged the incentives for high-effort strategy much greater than the costs, but they also illustrate how rebel identity produced major divergences in strategies of violence. In these campaigns, the position of Punjabi Sikh rebels triggered all four hypothesized mechanisms restraining state violence but the contrasting position of Kashmiri Muslim rebels triggered none, resulting in high observable violence. The Naxal III case is one where the state judged the incentives did not substantially outweigh the costs and therefore opted for a low-effort strategy of containment. The case also reveals how affective mechanisms of empathy and trust of the rebel adivasi and dalit (SC/ST) identity base have constrained the state's use of violence, and to a lesser extent the strategic mechanisms of vulnerability and trust have as well. This restraint is a marked contrast to the state's strategy against previous rebellions composed of the same identity base (Telangana, Andhra Pradesh) when such restraining mechanisms were absent. The Naxal III case reveals how the rise in the identity group's status and embeddedness over the course of three to five decades eventually yielded corresponding changes in state restraint. The Punjab, Kashmir, and Naxal III counterinsurgency cases selected for process tracing offer a number of distinct advantages. They occur in a period when many structural conditions 452 are constant, not just for India but also the international environment (norms, post-cold war, media access, integrated global economy). These relatively recent cases offer more external validity and relevance for understanding contemporary conflict, which has changed substantially from the days of Boer concentration camps or even Malayan village relocation. Recent campaigns also improve internal validity with a larger array of better sources, richer data, and key decision makers, participants, or observers with first hand access to events still alive and available to be interviewed. My argument is highly contingent on the quality of accessible information , which admittedly is often quite limited, particularly on the guarded affairs of a highly secretive state's national security decision making. Nevertheless, I attempt to compensate for this with the extensive data collection, fieldwork, and methods described in chapter 3 or to make a contribution on the motives and processes that undergird observable strategic choices often taken as given by most scholars of these conflicts. This chapter contains five sections including this introduction to examine the Punjab, Kashmir I, and the Naxal III campaigns in order to evaluate the causal argument of the theory. In the second section, I evaluate the Punjab case with a focus on observing both the territory and identity mechanisms. The third section evaluates the Kashmir case identifying similar territory mechanisms but detailing the noticeable absence of identity mechanisms constraining violence. The fourth section evaluates the Naxalite III case contrasting the level of effort to the Punjab and Kashmir cases and contrasting levels of violence to previous counterinsurgency campaigns against the same social base in Telangana and Naxal I. The fifth section concludes with implications for the India cases and for a broader understanding of counterinsurgency strategy. 453 II. The Puniab Insurgency: 1983-1993 Sikh fundamentalists launched an ethno-religious separatist insurgency in Punjab in the early 1980s. The insurgency erupted in 1984 due to a serious of high-profile state blunders and communal clashes, which deeply antagonized India's Sikh minority and galvanized their support for the insurgency. Violence continued to escalate into 1988 and again through the early 1990s, but a combination of Indian state security forces blanketing the region, intelligence-driven targeting, and political outreach brought insurgent violence down to a trickle by the end of 1993 as the state declared victory after roughly 25,000 people had died in the conflict. In Punjab, the GOI adopted a population control strategy because the strong incentives and low operational costs motivated high effort while the group worth and embeddedness of the Sikh rebel base compelled the state to employ a relatively low violence strategy, limiting collateral damage, targeting selectively, and deploying non-kinetic substitutes and political deals to soundly defeat the insurgency. 2.1. Territory and Effort 2.1.1. Territory Costs and Incentives In contrast to neglect of the Northeast and Naxal belt, the Indian central government relied on high effort strategy in Punjab since its calculation of incentives far outweighed the nominal costs incurred for state power projection. Based on the description of the value of the Punjab territory in chapter 6, it is apparent that the incentives would outweigh any costs, though in Punjab, the costs were minimal. Punjab was almost adjacent to the capital and the plains of Punjab were highly accessible with few hills or forest cover.' Roads that had been built 1Medha Chaturvedi, "Early Warnings: Is Naxalism Expanding in Punjab?" IPCS brief #3398, June 9, 2011 (http://www.ipcs.org/article/india/early-warnings-is-naxalism-expanding-in-punjab-3398.html). 454 throughout Punjab to facilitate counteroffensives and logistical supply lines in the event of a war with Pakistan also made the state easily penetrable by Indian state forces. This would enable mobile police forces to actively patrol and quick reaction teams to respond to insurgent incidents. Human terrain costs were also minimal, as the population had long grown accustomed to centralized state authority and presence after being a critical node of the British Empire and before that the Mughal and Sikh empires. The region had prospered in terms of fiscal and human capital because of its relationship to centralized state authority, particularly the Army, and was therefore intertwined with the productive aspects of state disciplinary power. As a result, state provision of public goods efforts during the counterinsurgency campaign to upgrade infrastructure and development, restore the security of daily life, and resurrect a peaceful, stable political process would all be viewed as meaningful and legitimate rather than regarded with distrust and suspicion as it would in some other regions like the Northeast. The incentives for high effort counterinsurgency were even more evident than the permissive conditions. Economically, at the time of the conflict, Punjab was known as India's breadbasket, one of its richest and most industrialized states.2 Safeguarding it was therefore the difference between food exports and food insecurity for the nation. In ideological terms, Punjab was also central to the idea of India, and disorder there hit much closer to home. One analyst noted, "Half of Delhi is Punjabi - when a person dies in Punjab, Delhi screams." 3 With more expectations of the state in this region, containment would not be sufficient to provide security for the Punjab public. 2 William K Stevens, "Mrs Gandhi Says Terrorism Will End," New York Times, October 16, 1983, p. 3 Interview, Indian national security analyst, New Delhi, India December 20, 2013 (#152). 3 Author 455 Regaining full control of the region was also a strategic imperative because events in Punjab "had to do with the total situation of the security of the country." 4 Long term deployment of the Army and paramilitary as well as early declaration of President's Rule indicated the state took this seriously. Particularly after Pakistan upgraded its conventional capabilities with American military assistance in the 1980s and the 1987 Brasstacks crisis made conflict between the to countries increasingly likely, the Indian state was deeply concerned about how the Punjab insurgency might exacerbate external threats and undermine rear-area security.5 Interviews with retired military officers, civil servants, and scholars confirm that Punjab was a core interest and priority for the nation due to its value and history of state reach.6 These incentives remain today. One retired intelligence director explained, "Ifthe Punjab insurgency were to flare up [today] for argument's sake, we'd have to divert and add a lot of resources there." 7 2.1.2. Effort Given the stakes of Punjab's territory, the state invested an unprecedented level of effort in material, political, organizational, operational, and intelligence resources that were sustained, expansive, and adaptive over time. MaterialInvestments The clearest evidence of effort in Punjab was the deployment of manpower. At its peak, the state had concentrated 280,000 security forces to the Punjab, an area 1.5% of the country. This included 120,000 regular Army troops from three Corps including a strike Corps (I Corps) 4 Stevens, 1983, p. 3. 5 Rajagopalan, 2004, 6 Author pp. 28-29. interviews with a number of retired Army and civil servants, New Delhi, India, November 2012 (#65, 68, 71, 83). 7 Author Interview, retired senior Indian intelligence official, New Delhi, India, December 18, 2013 (#150). 456 with another prepared for deployment (II Corps at Ambala). Such a large number of Army troops had likely never been concentrated in such a small area other than in the midst of conventional warfare. The deployment also included 70,000 Central paramilitary forces, between 53-60,000 Punjab police (more than doubled from a decade prior8 ) including 10,000 trained in special weapons and tactics and 4,000 as part of the elite Punjab commando force, 15-28,000 home guards, and 10,000 special police officers (SPOs) as part of a village defence scheme.9 Even before the intensification of efforts between 1988-93, the government consistently maintained over 100,000 troops in the state between police, paramilitary, and army deployments." The increase in manpower was directly responsible for increased arrests, weapons seizure, and denial of militant movement, activities, and footholds." Fencing the Punjab border from 1989-1993 inhibited infiltration and retasked forces from static to active duties. In 1989 alone, redeployed security forces ensured 43% of all villages in Punjab were combed by patrols' 2 and 45,000 persons were arrested attempting to cross the border illegally between 1986-92.13 When army units deployed to the border, they maintained close surveillance on 900 border villages and conducted 600-700 patrols per day.' 4 8 Chandon, 2013, p. 145; Fair, 2009; Kang argues the swing was even larger, from 32,855 officers in 1984 to 70,228 in 1994. Charanjit Singh Kang, "Counterterrorism: Punjab a Case Study," Masters Thesis, Simon Fraser University, Spring 2005, pp. 106-107. 9Telford 2001; Fair, 2009; Gill 1999; Wallace 2007; Manoj Joshi, Combating Terrorism in Punjab: Indian Democracy in Crisis, Conflict Studies 261, London: Research Institute for the Study of Terrorism, May 1993, pp. 12-13; Prem Mahadevan, "The Gill Doctrine: A Model for 21st Century Counter-Terrorism?" Faultlines, 19, April 2008; Jugdip S. Chima, "Controlling the Sunni Insurgency in Iraq: 'Political' and 'Military' Strategies from Successful Counterinsurgency in Punjab-India," Small Wars & Insurgencies, 18 (4), December 2007, pp. 626, 630; Manoj Joshi, "Learn from Past Before Sending in the Army," Mail Today, June 3, 2010, Factiva accessed 3/5/14. 10 D.P. Sharma, 1996, p. 84 on 1985 deployments; Victoria Graham, "6,000 Additional Troopers Sent to ViolenceTorn Punjab," AssociatedPress, March 30, 1986; Dilip Ganguly, "75,000 Security Forces on Alert in Punjab With AM-India-Gandhi Bjt," Associated Press, May 12, 1987. " Kang, 2005, p. 107. 1 5,280 out of 12,188 villages. See Chandan 2013, pp. 153, 165. 13 Swami, 2007, p. 148. 1 Joshi, 1993, p. 13. 457 Beginning in 1985,15 the Indian government began a very costly effort for the Punjab police's "reorientation, modernization, and reorganization" that involved "across-the-board reforms in its working modes, reorientation of training systems and syllabi, motivational techniques and a large-scale expansion in its armed battalions and upgrading of their weaponry."' 6 The state invested in upgrading the capability of police force in terms of skill and materiel. Pay was increased, training was provided by the military itself, and capital investments were made in better equipment, firepower, communications, and mobility/transportation. 17 Central government finances underwrote thee material investments. The state security budget increased fifteen to twenty-fold between 1985-1992 "by a bureaucratic system that understood its priorities remarkably well"' 8 but the center absorbed the massive accumulated debt, nearly all of which would be forgiven. 19 The full financial resourcing for this campaign is still not yet known due to reports of secret funds, but in contrast to the Northeast, counterinsurgents in Punjab applied these resources directly to the campaign rather than lining their pockets. 15 The spike in funding to the police begins in 1985, plateaus for a couple years, and then begins to rise dramatically in 1988. See H.S. Sidhu and S.S. Bains, "Public Expenditure on Police Services in India: With Special Reference to Punjab," Indian PoliceJournal, 55 (1), 2008, p. 57. 16 K.S. Dhillon, "A Decade of Violence, 1983-92," Punjab in Prosperityand Violence: Administration, Politicsand Social Change, 1947-1997, Ed. J.S. Grewal and Indu Banga, New Delhi: K.K. Publishers, 1998, p. 114 " Gill, 1999; Mahadevan, 2008; Sidhu & Bains, 2008. 18 The irony of this statement is that the Indian bureaucracy, especially at this time, was not known for its rapid adjustments. However, the massive injection of funds, particularly towards intelligence came after police officers were bribed by an insurgent with a sum of money larger than the entire state's annual intelligence budget. Following a complaint and lobbying by KPS Gill, funds were quickly made available. See Prem Mahadevan, "Counter Terrorism in the Indian Punjab: Assessing the 'Cat' System," Faultlines,18, January 2007. 19 Pettigrew says 15 times. See Joyce Pettigrew, Sikhs of Punjab: Unheard Voices of State and Guerilla Violence, Atlantic Highlands, NJ: Zed Books, 1995, p. 61. Suneel Kumar (2007) says from 150 million Rs to 3 billion Rs. 20 Shekhar Gupta and Kanwar Sandhu, "True Grit," India Today, April 15, 1993 (http://indiatoday.intoday.in/story/Police-chief-k.p.s.-gill-tums-the-tide-in-punjab-with-controversial-and-ruthlessmethods/ I/302060.html). 458 OrganizationalInvestments and Adaptation Investments in police personnel not only put more boots on the ground but changed the nature and activities of the organization. The state purged and overhauled the police force of risk averse personnel, incentivized initiative, and created a much more vertical hierarchical command for better internal communication and information sharing.2 1 The appointed Punjab Police Director General (DG), KPS Gill, who served as the de facto theater commander, bucked the entrenched system of Indian bureaucracy by promoting officers based on demonstrated skill and merit rather than seniority or political status. The Punjab police adopted new tactics focused on protection, psychological operations, and intelligence gathering and a shift from the safety of static checkpoint positioning to active patrolling and response to protect villages. Mobile response time of police was reduced to 3-5 minutes in urban areas and 15-20 minutes in rural areas. 2 The police also developed their own internal technology development cell for flexible response to demands that arose during the campaign like bulletproofing, infrared, explosives management, and detonation timer detectors.2 4 The police overhaul also beefed op the intelligence wing, once a "skeletal structure." 25 Gill lured in large numbers of younger, smarter officers to the operational and administrative command center "brain" to conduct more extensive and sophisticated intelligence collection and analysis, effectively discern patterns of the Sikh insurgency, and devise means to target, infiltrate or defeat it. At the federal level, the Intelligence Bureau created its first strategic intelligence estimate of the situation in Punjab, something that had been notably absent in counterinsurgency 2 22 23 24 25 Kang, 2005, pp. 100-101, 104. Chandan, 2013, pp. 119, 122, 138; Gupta and Sandhu, 1993.; Mahadevan, 2008. Chandon, 2013, p. 145. Chandan, 2013, pp. 146-48. Chima, 2007, p. 628. 459 campaigns in the Northeast, 2 6 and the central government also tasked federal intelligence officers from the IB, CBI, and R&AW to supplement the Punjab endeavor and coordinate with state police. 2 Intelligence not only made the Punjab police more lethal but also more discriminate, conducting operations in a "more surgical and effective way at the grassroots level without significantly antagonizing or alienating the general population." 29 Though covert intelligence operations at times seemed to be working at cross-purposes with the policing strategy by continually opening negotiations with insurgents, it proved to be the perfect complement to eventually bring back moderate separatists into the fold.30 The Indian state invested an unprecedented level of political authority into the campaign, particularly in DG Gill, especially by 1992. The "free hand" given to Gill is sometimes confused with impunity for excess force, but it had more to do with Gill's campaign remaining "untrammeled by politically-imposed handicaps." 3 1 This helped to overcame the innumerable coordination problems, bureaucratic obstacles, and interference by other politicians that normally hampered the sustainment or tempo of operations and ultimately the effectiveness of Indian counterinsurgency campaigns (including the early years in the Punjab campaign3 2 ). Gill often possessed a direct line to the Prime Minister in formulation of strategy. As a result, Gill was able to deflect criticisms and overrule objections by other powerful ministers in Delhi.34 Additionally, the government empowered the once heavily-restricted security forces to 26 27 28 30 31 32 Dhar, 2005, p. 186. Kang, 2005, p. 129. Kanwar Sandhu, "A Timely Reprieve," India Today, March 15, 1991, p. 135; Dhar, 2005, p. 411; Chandon, 2013, 127. Chima, 2007, p. 628. Dhar, 2005. Mahadevan, 2008. Kang, 2005, p. 103. Author interview with senior Punjab police official, New Delhi, India, December 20, 2013 (#153); Chandan, 2013, p. 137. 3 3 Tully, 1991, pp. 176-77. 460 effectively conduct the counterinsurgency campaign by passing a series of national legislative measures regarding the use of force, arrest and detention.3 5 Integration and coordination of forces also extended to the military deployments. Army units deployed in the area to backstop the Punjab police did not formally take orders from Gill but in practice exhibited operational deference to Gill's policing strategy "on an exceptional basis, due to overriding strategic considerations." 36 Such a "radical change" in the Army's usage and deployment patterns was critical for the local counterinsurgent strategy to play out and not be subverted, though this was normally out of the question for an Army generally unwilling to take a back seat to local security forces as a matter of prestige. 37 Police ensured coordination by tasking a police inspector general to each of the Army Corps deployed in the region. 38 At first, the Army resisted this deployment without freedom of action, but civilian authorities in Delhi, signaling political commitment, overruled them.39 The fact that the "prickly Army" deferred to the Punjab police is remarkable given that it was "very conscious of its preserve and distrustful of the 'civilians"' as it displayed in Kashmir. 40 Expansiveness Because Sikhs were not geographically confined but in fact were integrated throughout the social fabric of the country,41 Sikh militants possessed networks to perpetrate a number of attacks in valuable neighboring regions like Haryana, Maharashtra, and Delhi, which added Fair, 2009, pp. 110, 119-20. Mahadevan, 2008. 37 Author interview with retired senior Army officer who served in Punjab during operations, New Delhi, India, November 14, 2012 (#71); Mahadevan, 2008; Chandan, 2103, pp. 151, 184, 191; Gupta and Sandhu, 1993. 38 Gupta and Sandhu, 1993. 39 Joshi, 1993, p. 20. 40 Joshi, 1999, p. 227. 41 Dhar, 2005, 275. 3 36 461 further to the state's alarm.42 As the insurgents pushed their violence into other regions beyond Punjab (including as far as Assam and West Bengal), the counterinsurgency forces were able to coordinate and conduct criss-border operations in Chandigarh, Haryana, Delhi, and the Terai region of Uttar Pradesh.4 3 Sustainment The Indian state sustained its security operations during this campaign seeking completion rather than backing away after a modicum of stability had been restored as was the case elsewhere. Its "long-fuse strategy" sought to counter insurgent influence and activity until it degenerated and eroded its own popular base of support allowing the state to more effectively compete for the people's allegiance.44 The declaration President's rule, where the federal center took charge from 1983-85 and again from 1987 for almost another five years, enabled the longterm approach. By 1992, "the militant movement appeared to be effectively 'contained,' but not necessarily 'crushed.' 45 However, the state persisted by maintaining force presence, developing local capabilities, and combining pressure with political maneuvers to whittle away moderate support for the militants. By the middle of 1993 it had soundly defeated the insurgency.4 6 Unlike other regions where India has eliminated the threat to the state but left residual insurgent elements in tact allowing protracted, low-level rebellion and dysfunctional politics (particularly in the Northeast and even the Naxal belt), here the campaign did not stop with the fragmentation of insurgents. Instead state forces remained in place to consolidate their hold and 42 B Raman, 2007, p. 93; Dhar 2005, pp. 303, 310; Tully, 1991, p. 170. Gill, 1999; "Decision of the Authority Delivered by RPG Haines," on Refugee Appeal 523/92, Refugee Status Appeals Authority, March 17, 1995 (http://www.refugee.org.nz/rsaa/text/docs/523-92.htm). 44 Chima, 2007, p. 619. 4 4 46 Chima, 2010, p. 230. Chima, 2010, p. 235. 462 prevent any re-escalation of violence while political and intelligence leaders focused on tying up loose ends for a stable, non-violent political order by allowing aggrieved political elites "to unite, moderate, and re-enter the 'normal' political process.' 47 2.2. Identity In chapter 6's medium N analysis, I found that in broad terms, India employed a relatively low-violence population security strategy to combat the Punjab insurgency due to the value placed on the rebels' identity base, Punjabi Sikhs. There is little doubt that Sikhs played were highly esteemed and embedded within the Indian state since its earliest days and even stretching back further to the British Raj. In the event of a Sikh rebellion, my theory then expects that the positional status of Sikhs should constrain state-directed violence through mechanisms of empathy, trust, vulnerability and information. This section reveals strong evidence of all these mechanism at work in the Punjab case. First, Sikh's embeddedness in the state security forces made Indian state leaders (including many Sikhs in positions of leadership) acutely aware of their country's strategic vulnerabilities. A high violence counterinsurgency campaign could anger the Sikh communities in these institutions, triggering backlash and defection, which could prove strategically costly by compromising India's ability to defend itself. A major backlash and mutiny following the botched 1984 Operation Blue Star generated widespread awareness of this vulnerability and motivated the state to reduce its risk exposure and quickly adopt a strategy of lower and more discriminate violence. Second, genuine empathy for Sikh civilians, especially amongst both Sikhs and non-Sikh leaders in positions of power, led the state to avoid to high-violence strategies and favor tactics 47 Chima, 2010, p. 265. 463 like policing, arrests, rehabilitation, and addressing grievances above brute force. Third, Sikh coethnics in the security forces afforded the state strategic advantages such as personnel with language and cultural skills for increased collaboration with the population, lower friction, and better information for more discriminate targeting and minimal collateral damage. Finally, the prevailing co-ethnic ties between many state officials and the rebel base maintained a modicum of trust, allowing the state to address the demand side of rebellion to deplete the insurgency with side deals, public goods, and political power in as substitutes for pure force. Reciprocal trust by Sikh rebels, moderate separatists, and the Sikh community writ large made them more inclined to accept these gestures, negotiate in good faith, and re-enter peaceful politics, which could produce dynamic returns on these state investments and expand them to constitute a larger part of the overall campaign strategy. All four of these restraining mechanisms, stemming from the state's relationship to the rebel identity base, played important roles in shaping the counterinsurgency strategy to one of population control, which balance if not privileged minimizing civilian victimization, non-kinetic tactics, and engagement of moderates with killing insurgents to. Below, I describe each of these four restraining mechanisms in detail. 2.2.1. Vulnerability Early on in the Punjab counterinsurgency campaign, the state learned it was acutely vulnerable to significant strategic costs if it chose to rely on high levels of violence. If the state was considering a strategy of sustained high levels of violence to attrite the Sikh insurgency, this notion was quickly beaten back during the failure and aftermath of Operation Blue Star launched in June 1984. The aftermath of Blue Star exposed the state to great strategic vulnerability as it 464 faced internal upheaval from the military in the form of retired officer criticism, internal dissent amongst senior officers, and a mutiny from younger soldiers. This initial backlash convinced the state that it needed to quickly adjust and adopt a more careful approach to the Sikh insurgency that invested substantial resources to counter political subversion but emphasized precise targeting and violent negotiation above pure force and violence. Background ofBlue Star Since the late 1970s, the Sikh nationalist party, the Akali Dal, was increasingly discontent with ongoing economic and political shifts and pushing for greater autonomy. Delhi, fearing a separatist political movement on its hand and a threat to its political power in the state of Punjab, propped up a radical fringe group led by Jarnail Singh Bhindarwale to counterbalance and discredit the Akalis. This eventually backfired as Bhindarwale broke free from Delhi's leash, began his own political agenda of purifying Punjab through terrorism and violence against Hindus, civilians, and rivals, and directly challenged the center calling for a Sikh territorial homeland of Khalistan. Bhindarwale's more powerful political movement soon attracted the alignment of the Akalis and initiated separatist violence in the early 1980s including the murder of Punjab's Deputy Inspector General of police in 1983. By 1984, Delhi judged the Sikh-based insurgency had escalated beyond the control of state police and opted for a more proactive course of action. Bhindarwale and his armed cadres had stockpiled arms and holed themselves up within the complex of the Golden Temple, the holiest of shrines for the Sikh religion, located in Amritsar. On June 3, 1984, Army and paramilitary units were sent in to the Golden Temple and throughout the rural countryside's 37 suspected Sikh temples to deal a major blow to militants and defeat the nascent insurgency. 465 Operation Blue Star, and its partner Operation Woodrose in the rural countryside, resulted in a blundering, bloody disaster that killed hundreds of militants as well as Sikh civilians, defaced a sacred space, and infuriated much of the Sikh population. 48 What followed these operations may have been much worse. After Prime Minister Indira Gandhi's Sikh bodyguards assassinated her in retribution for Blue Star on October 31, 1984, communal riots broke out, primarily targeting Sikh civilians, and an estimated 3,000 may have been killed in Delhi and 8,000 around the country. 49 All these events of 1984 combined to dramatically escalate the Punjab insurgency. OutrageandAnxiety While the Indian Sikh community was reeling from all these events, the blunders of the Indian military's Blue Star served as a lightning rod for criticism and a focal point for a collective "sense of outrage." 5 0 A number of prominent senior Sikh officials including four senior retired officers (three Lt. Generals and one former Chief of Air Staff) who still held sway within the military publicly voiced their anger with the government. President Zail Singh, the first Sikh to hold that position, threatened to resign, having been left out of the operation's planning .5 Sikhs had composed a large number of the commanding officers in Blue Star but nevertheless, Sikh officers were deeply upset by the strategic choice. 88% of Sikh senior officers surveyed soon after believed Blue Star was unnecessary and 26% of non-Sikh officers agreed. This divide was a concern for the state as was the across-the-board frustration at the central 48 Pages 52-53 of Apurba Kundu, "The Indian Armed Forces' Sikh and Non-Sikh Officers' Opinions of Operation Blue Star," Pacific Affairs, 67 (1), Spring 1994, pp. 46-69. 49 Sandeep Unnithan, "The Untold Story Before Operation Blue Star," India Today, January 31, 2014. 50 Shekhar Gupta and Kanwar Sandhu, "Pakistan Has Lost; Operation Bluestar has no Champions Today: KPS Gill," India Today, April 15, 1993, p. 43 (http://indiatoday.intoday.in/story/you-can-say-i-have-vested-interest-inrestoring-normalcy-in-punjab-k-p-s-gill/l/302041.html); Author interview with senior Punjab police official, New Delhi, India, December 20, 2013 (#153). 51 Tully, 1991, pp. 192-94, 198. 52 Kundu, 1994, p. 57. 466 government for overall mismanagement of the conflict. Gen. KV Krishna Rao who had just retired as Army Chief of Staff in 1983 and serving DGMO VK Nayar both expressed serious concern and anxiety over how Army operations criminalized the Sikh community and risked 54 losing the loyalty of Sikh officers and soldiers. 55 Anxiety over the strategic consequences of high-violence counterinsurgency strategy in Punjab was shared by a number of senior officers in the intelligence community and the army who had anticipated and warned of this risk before the commencement of Blue Star. R&AW and other intelligence agencies feared a major backlash and indiscipline within the Army should the military be deployed in the frontline against Sikh militants.5 6 Additionally, then Maj. Gen. V.K. Nayar serving as the Additional Director General of Military Operations warned his boss, Army Chief of Staff Vaidya, and leading state decision makers that using heavy military force in Punjab would "blow up" with a "very adverse impact within the services," especially the Army. 57 In late 1983 and early 1984, Nayar successfully quashed a plan to deploy Army forces for internal security in Punjab.58 Eventually, a decision would be made to send in the Army during a meeting that excluded Nayar and Vaidya and was instead guided by the advice of the hawkish Western Forces Commander, Lt. Gen Sundarji. 59 Considering the counterfactual, if Vaidya, Nayar, or a Sikh voice of moderation like Zail Singh or KPS Gill had been in the room, perhaps they might have been able to successfully push back against the decision and averted the disastrous operation. 6 06 1 Kundu, 1994, p. 54 Rao, 2001, p. 217. 5 V.K. Nayar (Lt. Gen), From Fatigues to Civies: Memories of a Paratrooper, New Delhi: Manohar Publishers, 2013, p. 202. Rao (2001, p. 285) also voices this in his memoir. 56 Raman, 2007, p. 96. 57 Nayar, 2013, p. 193, 188. 58 Nayar, 2013, p. 193. 59 Nayar, 2013, pp. 206-07. 60 Tully and Jacob, 1985, p. 144. 5 5 467 " Strategic Costs The direct costs of such a counterinsurgency misstep in Punjab manifested in the form of unprecedented indiscipline and a mutiny within the Indian army. Immediately after Operation Blue Star a number of Sikh Army units-humiliated, emotionally, traumatized, and enraged at the perceived injustices visited upon their community-mutinied or deserted their cantonments. Between 2,000-2,800 Sikh troops, mostly young recruits, mutinied throughout northern India states of Bihar, Jammu, Rajasthan, and Maharashtra. Angry troops of the Sikh Regiment's 9 battalion stormed the armory of their headquarters, drove through the streets of a neighboring town firing indiscriminately, and both Sikh officers and enlisted personnel made for Delhi or the Pakistan border. Many formed convoy, the largest one around 1400 troops, and set out to defend the Golden Temple in Amritsar. Most of these troops were rounded up by other Army units mobilized to contain them, though at least 35 soldiers were killed in firefights.6 2 The anger of the young Sikh troops and was palpable even amongst retired Sikh military personnel and could not be overlooked. At the time there were at least 500,000 retired Sikh soldiers based in the Punjab, which means roughly 6.8% of the adult male population had serious military experience. If not dealt with carefully, this community could potentially set Northern India ablaze from within.63 South Asia scholars agree that Indian leaders recognized the Sikh backlash and mutiny posed a major threat to the Indian military and state. Tully writes the mutinies were "the most serious crisis of discipline the Indian Army had faced since independence,"64 and Barua confirms it "threaten[ed] the very cohesiveness and fabric of the largest professional army in the Chandan, 2013, pp. 116-117. Barua, 1992, p. 130; Telford, 1991; Tully, 1991, p. 194; Kundu, 1994, p. 65; Cohen, 1988, pp. 134-35; Mary Anne Weaver, "Sikh Mutiny Spreads in Indian Army," ChristianScience Monitor, June 12, 1984; Birbal Nath, The UndisclosedPunjab: IndiaBesieged by Terror, New Delhi: Manas Publications, 2008, pp. 201-202. 63 Cohen, p. 134. I estimate this by looking at population estimates of adult males in Punjab (age 15 and over) in from census data which roughly amounts to 7.3 million. 64 Tully, 1991, p. 197. 61 62 468 world.. .hitherto immune to the ethnic conflict that frequently ravaged the country."65 Not surprisingly, this was a "shock to both the army and the government." 66 Even though most Sikh soldiers had not participated in these activities, the mutinies strained relations between Sikhs and other ethnic groups in the military, further threatening the cohesion and morale of mixed units. 67 In the aftermath, one retired Brigadier at the time wrote, "after Operation Blue Star, Sikhs [sic] officers may well be undependable... It has shaken my confidence in Sikh troops, maybe unjustifiably."68 Another non-Sikh Major General suggested this may lead some to question the loyalty of Sikh troops who composed 25% of the force and warned "the army's fighting effectiveness will be endangered." 6 9 The strategic consequences of the mutiny and cohesion problems were particularly farreaching for an Indian military locked in a regional rivalry. In what was described as a "hysterical climate" after Blue Star, "many in India's security and intelligence establishment believed there was a real threat to the country's existence... "70 One of the foremost experts on the Indian military argues the mutiny's effect on cohesion even altered the regional military balance: The overall integrity of the Indian armed forces, especially the army, may have been badly, if temporarily, weakened. If Sikhs comprise about 12 percent of the army, then the effective fighting strength of the army was probably reduced by at least that figure (more, if non-Sikh units must be deployed so as to contain another mutiny). This decline in the efficiency and reliability of the Indian armed forces in 1984-85 was probably greater than the increase in the capabilities of Pakistan's armed forces since 1982; the two combined do not add up to a shift in the strategic balance between India and Pakistan, but it must reduce Indian dominance to mere superiority. .A worsening of the Punjab crisis could lead to.. .the breakdown in the integrity of the Indian armed forces if their critical Sikh component should be removed. 65 Barua, 1992, p. 131. 66 Barua, 67 Tully, 1992, p. 131. 1991, p. 197. 68 Kundu 1994, p. 63. 69 Kundu, 1994, p. 65. Swami, 2007, pp. 157-58. Cohen, 1988, pp. 137-138. 70 71 469 At this point, the state and the military were extremely vulnerable and needed to attend to both the Sikh soldiers who had mutinied as well as the Sikh insurgency without triggering further backlash. The magnitude of Bluestar's fallout provided the impetus and urgency to shift gears very early in the Punjab counterinsurgency campaign and shaped the strategy for the next decade. Course CorrectionsandNew Directions After 1984, the Indian state set a new course of action that prioritized broad negotiations, bolstering moderates, amnesty provisions, development, and intensive intelligence-driven policing power. Following Indira Gandhi's assassination, her son and political heir, Rajiv Gandhi, became prime minister, cleaned house of his mother's political advisors, and initiated serious dialogue with Sikh political leaders to settle some of the outstanding grievances. His turn to negotiations with leaders of the more moderate faction of the Akali Dal (AD) suggested he and his government "realized that the heavy-handed counterinsurgency had if anything worsened the insurgency" 72 and that security operations alone "could not provide the answer to communal strife." 7 3 These negotiations produced the Rajiv-Longowal Accord, with eleven points seeking to redress grievances and devolve power to the state of Punjab, and set up an election that returned the AD to the helm of the state government.7 4 Though some planks of the accord were never implemented and it fell apart two years later, these attempts alongside the counterinsurgency campaign signaled a different approach to engender good will and address political demands and Shale Horowitz and Deepti Sharma, "Democracies Fighting Ethnic Insurgencies: Evidence from India," Studies in Conflict and Terrorism, 31 (8), 2008, p. 759. 7 Tully and Jacob, 1985, p. 224. 72 74 Nayar, 2013, p. 224. 470 grievances. In 1985, the Bains committee under the new AD Barnala government released an estimated 2,000 people from prison, some extremists and known terrorists, in order to win further support.75 While criticized as appeasement, the state was attempting to end combat the insurgency using minimal violence and non-kinetic political tactics (something noticeably absent in Kashmir).76 Most importantly, the Indian government adopted a major course correction by placing state police as the lead counterinsurgents with the Army as a backstopping force. Even as it was being built up and overhauled, the Punjab police emphasized "minimum use of force" even against militants. 77 Meanwhile, the Army took the unprecedented step of deferring to the Punjab local police command. 8 Following the chaos of 1984, senior military officers and officials recognized the army was too blunt an instrument for much of the counterinsurgency and would cause internal problems as it did after Blue Star. 79 Such a deployment threatened "the army's reputation for fairness and integrity" and could "turn [the population of Punjab] against them with potentially disastrous consequences," compromising internal and external security operations.8 0 Learning from the strategic costs of Blue Star sunk in quickly as demonstrated by an episode the following year. When now Lt. Gen. Nayar was posted to run the Corps Command in Jalandhar, Punjab, he was able to modify a plan that might have further antagonized the Sikh community and troops with another Army operation. After receiving a command from Army Headquarters to deploy to Punjab and engage in similar sweeping operations to clear the Mand Chandan, 2013, p. 124. Chandan, 2013, pp. 129-131, 166, 173. 77 Chandan, 2013, pp. 129-131, 166, 173. 78 Mahadevan, 2008. 79 Author Interview, retired senior Indian Army officer, New Delhi, India, November 10, 2012 (#68). 80 Tully, 1991, p. 204. 75 76 471 region (as the Army did in Woodrose), he objected to the order and convinced the Army Commander to put it on hold. This time Nayar 's warnings were heeded and he then managed to limit Army deployment and condition it on the local police leading as the direct face of the operations, administering a curfew, establish surveillance, and dealing directly with any terrorists they encountered. The police engaged in serious encounters while the army moved in at night, backstopped the police, swept the area of illegal activity, and returned to the barracks by morning without causing public alarm.8' India's civil and military leaders had begun a course correction that would open the door for the Punjab police to lead its only decisive counterinsurgency victory. 22.2. Empathy Leaders at the highest levels of the Indian state expressed empathy for the aggrieved Sikh population and an unwillingness to use the military against them. State officials described insurgents and mutineers alike as "misguided youth."8 2 83 Soon after the 1984 debacle, Arun Singh, a close advisor to Prime Minister Rajiv Gandhi, who himself was a clean-shaven Sikh, "was convinced it was an 'identity problem' which could not be solved except at the emotional level" and would go on to advise dialogue with militant leaders and representatives as a central part of the counterinsurgency campaign. Enlisted personnel in the Army, paramilitary, and police shared sympathy for the Sikh militants with whom they shared had kinship ties and sometimes tipped them off. 85 The external intelligence agency, R&AW, expressed concern for 8i Nayar, 2013, pp. 220-24. 82 Anant Mathur, "Secrets of COIN Success: Lessons from the Punjab Camapign," Faultlines,20, January 2011 (http://www.satp.org/satporgtp/publication/faultlines/olume20/Article2.htm). 83 Raman, 2007, p. 112. 84 Tempest, 1985; Janardan Thakur, "The Return of Arun Singh," Rediffcom, July 10, 1999 (http://www.rediff.com/news/1999/jul/10arun.htm). 85 Suneel Kumar, 2007. 472 and faith in their large number of Sikh officers and sought to protect them, particularly during the communal riots in 1984. The surprising number of Hindus (and even dominant caste Brahmins) who participated in and occupied top rungs of the Punjab insurgency was further evidence of both the strange closeness of identity and community ties as well as another constraint on the Indian state's freedom of action. 86 The state's expressions of empathy for the Sikh population ultimately translated into choices for "softer" counterinsurgency tactics including the rapid rehabilitation of mutineers, active police-led efforts to protect the population, new steps at civic engagement as non-kinetic substitutes to force, and concerted efforts to avoid or minimize collateral damage. Amnestying Mutineers, RehabilitatingMilitants Guided by the empathy felt for misguided militants at both the leadership and ground levels, the Punjab campaign involved significant use of surrenders, reintegration, and rehabilitation, a clear indicator of what Jackson terms "violent negotiation" rather than pure attrition. The state's rehabilitation of many mutineers was the first revelation of this approach as it chose not to punish many of the soldiers and in fact sought to downplay their transgressions. Despite even the killing of some commanding officers, the Ministry of Defence had no choice but to take a "generous view" describing the mutineers as "misguided." Retired Sikh officers lobbied the government for a more sympathetic treatment of them, again with the concern of military cohesion. 87 Sympathy for their emotional state eventually allowed the state to rapidly rehabilitate at least 900 of the mutineers back into the military by 1985.8 Out of concern for 86 Ramesh Vinayak, "Hindu Youths Join Hands With sikh Militants in Punjab, Leave Police Baffled," India Today, March 31, 1992. 87 Tully, 1991, pp. 198-99. 88 Kundu, 1994, p. 69. 473 ' divided loyalties, some policymakers wanted all Sikh battalions dating back to the British to be watered down as mixed units. However, at the advice of former officers like former Army chief of Staff K.V. Krishna Rao who expressed confidence in the Sikh troops, this effort was rejected. 89 Furthermore, the loyalty of Sikhs more broadly was affirmed. A government White Paper published to explain the issues around Blue Star took great pains to explicitly affirm the status of the Sikh community's worth and embeddedness. 90 Emotional appeals to mitigate the repressiveness of the counterinsurgency campaign were also voiced. An editorial in the Indian Express called for "a salve to be applied to the hurt collective psyche of this proud and valiant community before any lasting damage is done." 9 These amnesty policies were not just afforded for soldiers; the state used surrenders and rehabilitation as a common tool to deal with insurgents of all stripes, particularly in the final years of the counterinsurgency campaign. Punjab Chief Minister Beant Singh embraced militant surrenders, stating, "Boys, you are like my sons, you have resorted to guns against us, I am a Sikh like your fathers are. I suggest you with authority to shun violence. We will embrace you with warmth."92 Even after thousands of prisoners released under the Barnala government were recognized as premature and a tactical mistake, the state took the risk of rehabilitating more than 2100 militants who surrendered over the next decade. 93 Those who were not rehabilitated may have remained imprisoned, some turned as renegades (an estimated 300 - an order of magnitude less than those used in Kashmir), and many others were unofficially released after serving as 89 Rao, 2001, p. 285. 90 White Paperon PunjabAgitation, New Delhi: Government of India, 1984, p. 58. 91 Tully, 1991, p. 206. 92 Chandan, 2013, p. 198. 93 Sarbjit Dhaliwal, "Amarinder's No to CBI Probe into Sukhi Case," Tribune News Service (Chandigarh), February 25, 2006 (http://www.tribuneindia.com/2006/20060225/punjabl.htm). 474 informants by "disappearing" with anonymous identities and exfiltrating out of the region to avoid retribution by their former colleagues. 94 PolicingandProtectingthe Population The Indian state made a crucial decision shortly after Blue Star to principally rely on the Punjab state police, which was 65% Sikh,95 to conduct operations with greater restraint and sensitivity to the local population. This approach was deliberately chosen over heavy reliance on the army or on more nefarious methods used in Kashmir. Even after Longowal Accord's breakdown by 1986 and the resumption of high insurgent violence, Prime Minister Gandhi contemplated the use of "terrorizing tools to destroy the terrorist" but ultimately decided against it in favor of doubling down on the policing strategy advocated by KPS Gill.9697 The substitution of the police for the Army as the lead counterinsurgent was a critical move because the Army was inherently "trained in a manner different from the police" and prone to rely on force rather than non-kinetic substitutes. 98 The predominantly Sikh police would be more likely to possess compassion, work with the population, gather information from them for more precise targeting, and elicit the trust of them and insurgent cadres when it came time to cut deals. Out of concern for the population's well-being, the Punjab police actively sought to protect them from coercion and violence to deny insurgent influence rather than just pursue militants offensively. Dismounted patrols conducted by local police attuned to the cultural and religious sensitivities of Punjabi Sikhs improved ties with the community. Night domination 94 Both interviews with counterinsurgents as well as external audits more than a decade later confirm the renegades who remained off the books were only 200-300. See Mahadevan, 2007; "Dal Khalsa Seeks 'Police Cats' List," Tribune News Service (Chandigarh), February 25, 2006 (http://www.tribuneindia.com/2006/20060225/punjabl.htm). 95 Mahadevan, 2008. 96 Dhar, 2005, p. 275. 97 Author interview with senior Punjab police official, New Delhi, India, December 20, 2013 (#153). 98 Raman, 2007, p. 96. 475 ' exercises were one such example. Though the term conjured up images of police brutality, they were developed based on intelligence analysis of where most militant violence took place at night in order to deter it. After their regular duties, the police would gather at night in sensitive areas typically dominated by the insurgents in order to patrol and prevent militant activity, especially in rural areas. 99 "Focal point patrolling" where police units were concentrated in one sector at night was used to create the impression of a much larger security presence again to reassure villagers and deter insurgents.1 00 The police also organized and equipped locals as special police officers (often army veterans) for the first line of defense against militants while backing them with professional security forces and mobile police units that could rapidly respond.' 0 Even as terrorism began to decline, the police remained committed to population protection to allay their fears and help them resume normal life through. The police pursued civic outreach, sought continuous liaison with community leaders, set up public meetings with residents, and even dispatched 20-30 men to periodically camp in rural areas to protect and reassure villagers. Co-ethnic ties helped inform and enable these interactions with the population. 0 2 Police hosted and protected markets, cultural events, celebrations, and even musical shows late at night to signal that it was safe for the public to travel and be outdoors at all hours. As a symbolic defiance of the militants' authority, even "drinking was made a counterterrorist measure. 10 3 As the police assumed a more visible community protection role with their own civic outreach programs, people who were suffering at the hands of the insurgency were 99 Chandon, 2013, p. 195; Kang, 2005: 112. Also described in interviews with senior Army and police officers who served in Punjab, New Delhi, India, November 2012 (#71) and December 2013 (#153). 100 Chandan, 2013, pp. 150-51. 101 Chandan, 2013, pp. 152-53. 102 Mahadevan, 2008. 103 Chandan, 2013, pp. 201-202; Author interview with senior Punjab police official, New Delhi, India, December 20, 2013 (#153). 476 more apt to turn over intelligence to the state. 1041 05 Believing media relations were central to countering the Sikh community's fears and suspicions, the police also sought to protect newspapers in order to ensure transparency and distribute unbiased news.106 Protection extended to safeguarding rival political parties (even those known to be sympathetic to the separatists) to freely participate in election campaigns. During the February 1992 elections, the police deployed 220 companies to protect each candidate with a platoon.10 7 Confidence in the protective authority and legitimacy of the state's writ rapidly increased over the course of the final year of the campaign as voter turnout jumped from 24% in the February 1992 state elections to 75% in the September 1992 municipal elections to 82% in the January 1993 panchayat(village council) elections.10 8 Military Adaptation The army had learned its lessons from the grave mistakes of Blue Star, which largely kept it in a backseat role, but when it was dispatched to backstop the Punjab police in the border provinces, it emulated policing activities and balanced targeted operations with tactics to defend and support Sikh villagers. First, the Army accepted playing a backstop role to the police (whereas in other campaigns it refused) mostly along the border areas to interdict infiltration that threatened the border villages. Among the forces deployed were troops from the region including 104 Kanwar Sandhu, "Punjab: A Golden Opportunity," India Today, November 30, 1992, p. 38. Author Interview, Indian human rights activist, New Delhi, India, November 14, 2012 (#73). 106 Gill was quoted as stating, "if we can't protect your newspaper, we cannot protect the people of Punjab." Chandan, 2013, pp. 153-54, 170-72. Ironically, it is the media's "unprecedented access to Police operations" that may have produced far more stories of police repression relative to other conflict theaters where access to information was stymied by the military. See Mahadevan, 2008. In social science terms, the uneven distribution of access may have led to journalists oversampling of the kinetic practices of the Punjab police creating an impression that the Punjab counterinsurgency was relatively more brutal than other campaigns. 107 Chandan, 2013, p. 185. 108 Chandan, 2013 pp. 195-96. 105 477 battalions from the Sikh Light Infantry.1 09 "The secret of the Army's success was in maintaining a low profile" as well as providing public goods and services, which were welcomed and won the confidence of the population. 10 Additionally, the Army allowed the predominantly Sikh police to remain the primary face of the campaign and even take credit for Army successes.'11 Second, the Army changed tactics to correct its past errors and not only held territory but also abided by restrictive rules of engagement to "allay the fears of the local population and, if possible, win them over." The Army's unwritten objective was "to instill confidence in the local people with the aim of creating an environment in which the civil administration, [at the time] emasculated, [could] function once again."" 2 With company and battalion headquarters deployed just outside villages, the Army and paramilitary forces coordinated multiple cordons during police dragnets without antagonizing villagers. 1 3 Troops and units deployed were particularly selected based on their sensitization to the socio-religious conditions of the region. They were forced to abide by a number of restrictive rules of engagement and operation including restrictions on entering civilian households, encamping in vacant areas, the damaging of crops, and the use of their weapons.' 1 4 Sikh villagers expressed relief at the security provided by rural deployments of the Army where they patrolled day and night." 109 5 Within a three and a Sandhu, 1991. Mukesh Sabharwal (Col.), "Role of Security Forces in Counter Insurgency," Pratividrohi,9 (2), Autumn 1997, p. 27-32. Gupta and Sandhu, 1993. 1 2 Sandhu, 1991, p. 133. 113 Sandhu, 1991, p. 135. 114 Sandhu, 1991, p. 137; Patrol commanders ordered their men not to shoot unless fired upon. Harinder Baweja, "A Double Edged Sword: Operation Rakshak II" India Today, February 15, 1992 (http://indiatoday.intoday.in/story/operation-rakshak-ii-our-aim-is-to-bring-militancy-under-manageablelimits/1/306145.html). 115 Author interview with a retired senior Army officer, New Delhi, India, November 14, 2012 (#71). 110 478 half month span, the Army had conducted almost 11,000 patrols in the rural areas and laid nearly 5,000 ambushes that arrested 1600 militants. 116 Finally, the Indian army units also proactively engaged in a number of unorthodox civic activities and missions to earn goodwill amongst the population where they were operating and serve as non-kinetic substitutes to counter insurgent influence. They set up medical camps, sometimes attracting more than a thousand cases in a day, and brought in specialist medical practitioners like gynecologists, offered transportation to army base hospitals for the seriously ill, and subsidized the costs of medicine. When schools had closed because teachers had left in fear, enlisted soldiers substituted as teachers for some time, but more importantly, provided security to entice teachers to return to the classroom. Villages were "adopted" by local units that carried out civic projects like the construction of roads, bridges, drainage systems, and the stringing of power lines. One villager in extolled the army's civic activities as "a godsend" and another in a rural village around Amritsar enumerated the feeling of public safety stating, "now we can sleep in peace and the boys no longer need to stalk the area." 17 Excesses Punished In their commitment to minimize collateral damage and brute force, the government also made a more concerted effort to punish transgressions committed by security personnel in Punjab. This was certainly not a perfect record but there was a much higher rate of investigation and prosecution in Punjab than in the equally intense Kashmir, despite lower levels of violence and abuse against civilians. Around 270 were being punished through jail time or prosecution 116 Baweja, 1992. 117 Sandhu, 1991, p. 137; Baweja 1992; India West, February 28, 1992; Times ofIndia, February 28, 1992. 479 along with another 85 investigations." 8 One analyst pointed out the difference in the two campaigns was that police and officials from Punjab counterinsurgency continued to be investigated and brought up on charges for their role in these operations even after the campaign while there was near total silence on Kashmir."1 9 The knowledgeable local police also kept turned militants tasked to infiltrate the Sikh militant groups on a much tighter leash by police whereas in Kashmir, the Army lacked the capacity or will to hold their renegades accountable. 20 Turned militants were primarily utilized to collect information with only a few of them used for direct action (see Table 6.7 on arrests surrenders), but as an institutional check, they were subject to inquiries in case excess force was suspected. 2 1 2.2.3. Information Owing to a "home-team advantage" with a high number of Sikhs in the security forces and local police (including leadership positions), the Indian state benefitted from access to better local intelligence, information networks, and forces with situational awareness, and a robust understanding of the human terrain. Consequently state forces conducted more discriminate violence, minimized friction with the public even during operations, and made more informed strategic judgments to manage public perceptions and accurately measure success. "1 Telford, 2001. This stands in stark contrast to the over 1000 claims, 44 investigated, and only 11 pending. Author Interview, Indian human rights activist, New Delhi, India, November 30, 2012 (#99). 119 120 Author Interview, Indian human rights activist, New Delhi, India, November 14, 2012 (#73). Harinder Baweja and Ramesh Vinayek, "A Dangerous Liaison," India Today, March 15, 1996. 121 Mahadevan, 2007. Only 300 Renegades were used "Dal Khalsa Seeks 'Police Cats' List," Tribune News Service (Chandigarh), February 25, 2006 (http://www.tribuneindia.com/2006/20060225/punjabl.htm). 480 Intelligence A predominantly Sikh police force cultivated population-generated intelligence for more precise targeting 22 and "discriminate operations." 123 Punjab police were predominantly Jat Sikhs like the caste identity of militants. These "sons of the soil were familiar with the way in which rural Punjab worked, and often had families in the same villages as the militants." 24 Commanders and constables possessed language skills and knowledge of "local geographic, demographic, and cultural factors that simply lay beyond the reach of their federal counterparts." This made the police adept at penetrating insurgent organizations, recruiting local spies, collecting information from civilians, and managing violence.12 5 DG Gill initiated major intelligence gathering and analysis for a more systematic picture of the insurgent organization and violence. By 1986, Punjab police, injected with a boost in morale, began regularly patrolling and developing local intelligence networks.1 2 6 Intelligence was collected by documenting and mapping incidents at the thana (local police station) level for closer analysis, as well as collecting human intelligence through policing, financial inducements, civic action, and investments in public goods. The result was more discriminate use of force against hardcore militants while minimizing collateral damage. 27 The police utilized intelligence to sort insurgents by commitment level into A, B, and C categories and focused their efforts on targeting leadership (A category) while minimizing targeting of B and C rank-and-file cadres to minimize total casualties. Only 10% of insurgents were considered committed hardcore separatist insurgents and the other 90%--termed "misguided youth" joined for reasons of 122 Dhar, 2005, pp. 323, 336; Kang, 2005, p. 144. Horowitz and Sharma, 2008, pp. 759-760. Chandan, 2013, p. 169; Gupta and Sandhu, 1993. 125 Mahadevan 2007; Mahadevan, 2008; Kang, 2005, pp. 130, 137, 143-44; "Case Study - Long Rope (Punjab)," Pratividrohi,4 (1), March 1992, pp. 47-54. 126 "Punjab Cops' Morale Rising," Times ofIndia, July 15, 1986, p. 17. 2 1 Chandan 2013, p. 122; Mahadevan, 2008. 123 24 1 481 livelihood or adventure. This intelligence essentially attempted to separate the few who needed to be dealt with through pure force from the vast majority of insurgents who could be dealt with through addressing the demand-side of rebellion. B and C militants were usually afforded opportunities for surrender, amnesty or rehabilitation, sometimes in exchange for information. 2 8 "As the police acted on specific intelligence and analysis, it reduced public harassment and increased public confidence in the police force.. .The swift and clinical targeting of the terrorists due to intelligence analysis enabled the police to win over the 'conditional support' of the public."1 29 OperationalInformation andSituationalAwareness Co-ethnicity was central to all targeted operations based on gathered intelligence. By 1987, a premium was placed on arresting rather than killing militants. One officer reported, "We no longer harass people if they were forced to give food or shelter to terrorists. Instead, we encourage them to talk to us and we try to win them over to our side. This change in our policy has helped us a lot. Now people are beginning to inform us about the whereabouts of these terrorists.'"1 3 0 Small businesses affected by the violence were particularly forthcoming with information.131 Even well-armed commando forces conducting larger operations against insurgents clusters or fortified targets, did so "selectively and with minimum force," producing few collateral casualties and minimal property damage. 28 32 Punjab commandos carried out most Chandan, 2013, pp. 149-50; V.K. Singh, India'sExternalIntelligence: Secrets of Research & Analysis Wing (RA W), New Delhi: Manas, 2007: 179; KPS Gill, Punjab: The Knights ofFalsehood, New Delhi: Har-Anand, 1997; Author interview, Indian national security journalist, New Delhi, India, December 17, 2013 (#147). 129 Chandan 2013, p. 143. The same was true with BSF units. See Chaman Lal, "Terrorism and Insurgency," India Seminar, #483, November 1999. 13 0 Dipankar Gupta, Sumanta Banerjee, Dinesh Mohan, and Gautum Navlakha, "Punjab: Communalised Beyond Politics," Economic and Political Weekly, Aug 13, 1988, pp. 1678-79. 131 Author Interview, Human Rights activist, Delhi, India, November 14, 2012 (#73). 132 Horowitz and Sharma, 2008, pp. 759-760. 482 - of the kinetic operations. Their "neatly-groomed beards and headgear similar to many militants, helped to ameliorate Sikh perceptions that the central government's Hindu dominated security forces were killing their co-religionists." 133 Consequently, civilians were more inclined to help the police in their efforts rather than feel like the victims of state repression. Even amongst paramilitary units like the BSF-whose members still had significant ties to the Sikh rebels' 34 a certain amount of emphasis was placed on persuading the population, coordinating operations with local leaders based on good intelligence, and trying to rehabilitate rather than create permanent enemies of militants.135 Informed Strategicjudgment Aside from ground-level information and interaction, the Indian state also benefitted from many Sikhs in positions of leadership in the armed forces, paramilitary and police, who could make strategic-level judgments based on their more intimate knowledge of co-ethnic insurgents and population. A hard charging Sikh officer, KPS Gill had been appointed to run the Punjab Armed Police in September 1984, left for Delhi in early 1986 but returned by June 1986 to run the paramilitary police in the region. Though he began running things by late 1986, he was formally installed as the DG Punjab police in 1988 led the campaign as the area commander for the next six years. Better local knowledge by the area commander was not only useful for morale, and targeting of kinetic and non-kinetic activities, but Gill's comprehension of Sikh public attitudes and psychology also enabled him to influence and ensure better strategic decisions made by the 133 134 135 Chima, 2007, p. 630. Gayer, 2009, p. 250. Chaman Lal, 1999. 483 federal center to prevent inflammation or escalation. 13 6 When Sikh militants again took refuge in the Golden Temple in 1988, Gill's advice to the PM prevailed and this time, given "the army's lack of knowledge of local circumstances and customs," the Punjab police led the raid, OperationBlack Thunder.13 7 The operation was much more sensitive to civilian casualties and Sikh public perceptions, employed far less violence and more discriminate force to clear the area, and incurred higher risks to police forces to secure the compound.13 8 Gill's strategy appeared to be a siege that wore down the militants until they surrendered rather than directly engaged them. 139 The police were keenly aware of the public relations challenge and focused on being transparent to the media and using the minimal amount of force both against civilians as well as militants. One journalist who was present during both Blue Star and Black Thunder observed that in the second operation, security forces were very respectful to the civilians when requesting to search houses or enforce curfew and even provided food to townspeople during the siege.1 4 0 In the aftermath of the operation, another journalist noticed civilians felt much more secure in speaking out against the militants.141 Gill made it clear that more than dislodging the militants, half the purpose of the operation was "psychological, not to upset the Sikh community."1 42 Better information collection was not simply used to attrite insurgents but to broadly measure the changing nature of the insurgency and its effect on the population in order counteract political subversion. Success was not measured by terrorist kills alone but by other indicators of public support and security. The police read the declining number of successful 136 137 Tully, 1991, pp. 176-77. 99 Tully, 1 13 8 Kundu, 1,p. 164. 1994, p. 69; Tully 1991, p. 163; Chandan, 2013, pp. 134-35. 139 Gupta et al, "Punjab: Communalised Beyond Politics," 1988, p. 1679; B Raman, 2007, pp. 164-65. 140 Tully, 1991, pp. 162, 165. 141 Tully, 1991, p. 178. 142 Tully, 1991, p. 170. 484 bandhs (general strikes), bhog ceremonies (memorial service for martyrs), family distress migrations, and high turnout at public events as evidence of a public disaffected with, and no longer beholden to, the militancy.14 3 2.2.4. Trust The mechanism of trust based on rebel identity also played a significant role in limiting state violence. Within a strategy of violent, two-way negotiation, trust proved instrumental to both tactical and strategic substitutes for violence such as surrenders, broad-based public goods, and negotiations with separatists and even some militants to empower moderates and devolve authority. The state leaders and forces maintained implicit trust that given Sikhs past record, the demands of rebels' social base could be satisfied through other means, separatists could be rehabilitated, and negotiated tactical and strategic bargains would hold. Sikh separatists and some insurgents held some trust that state public goods would be meaningful, devolution would continue after conflict, and co-ethnics within the state would ensure their survival and interests. Throughout the campaign, the state remained acutely aware of the demand side of the insurgency. Quoting Clausewitz, one officer examining the Army role in Punjab writes, "'the defeat of enemy armed forces does not automatically lead to the attainment of the political objective.' Instead, the key is eradication of conditions conducive to violence and instability. In all cases, political, psychological, and economic methods must be fully integrated with military force."14 4 The offer of public goods and political deals to address grievances is frequently invoked in counterinsurgency campaigns but they often may not take, endure, or offer any traction if there is deep mistrust between the parties. Efforts in this direction without dynamic 143 144 Chandan, 2013, pp. 195-96. Sabharwal, Pratividrohi,1997, p. 31 485 returns can be seen by either or both sides as non-serious and quickly wither or foster distrust before playing no meaningful role in the state strategy. In Punjab, this was not the case. SurrenderPolicies Co-ethnic empathy allowed the state to trust that insurgents could be demobilized through non-kinetic means. Very early on, an open policy was offered to uncommitted youth militants to surrender and be given jobs. 145 This proved effective because the police discovered through extensive interviews and interrogations that many Sikh youth were joining the insurgency not out of religious fervor or ideology but out of a lure for money and power.1 46 Consequently, "the police.. .made money power the central plank of their strategy to terminate the movement" and "buy off captured terrorists" sometimes hiring them as special police officers, particularly from the lower caste Mazhabi Sikhs, or even the regular police force.1 47 The recruitment Punjab Home Guard were both designed for their own protection in more rural areas, but also as a means of "a means of empowerment" for the poorer rural Punjab youth to soak up a potential recruitment pool.1 48 Security forces in Punjab also deputized a number of low-caste Sikhs as special police officers to neutralize their incentive to join the insurgency as much as to protect the community. 49 Specifically, coethnicity between state forces and the rebel groups in Punjab allowed trust to overcome credible commitment problems that often plague such interactions, resulting in higher levels of surrenders and rehabilitation. On an individual basis, the state was able to 145 "Stalling Security Belt Major Victory: Barnala," Times ofIndia, Sept. 29, 1986. Gill, 1997. 1 Mahadevan, 2007; Gupta and Sandhu, 1993. 146 47 Chima, 2007, p. 630; Sunil Dasgupta, "Paramilitary Groups: Local Alliances in Counterinsurgency Operations," Brookings Counterinsurgency and Pakistan Paper Series, No. 6, June 2009. 149 Gupta and Sandhu, 1993. 141 486 negotiate the surrender of individual insurgents by communicating with their families through retired police officers. Surrenders were taken in, used as informants for a while, but later incorporated into the police or quietly and anonymously rehabilitated into the mainstream. 150 Empathy and trust opened up opportunities to address the demand side of rebellion but also brought with it certain risks. Many of the Punjab police were known to have sympathies with or be compromised by the militants, which is how insurgents were able to perpetrate some major attacks including one against the DG Mangat in 1991.5 In early negotiations, some deals burned the state such as the release nearly 280 hardcore mi.litants in the end of Oct 1985, many of whom returned to the insurgency.152 Nevertheless, the Indian state stood by the local police, while improving and professionalizing them and assumed the risks of devolution, bargaining, and rehabilitation as non-kinetic tactical substitutes that would eventually payoff. Public Goods The Indian state also utilized its pre-existing identity links and networks of trust with the Sikh community to demobilize insurgents with tactical substitutes. Quite the opposite of measuring success by killing ones way out of an insurgency as the state had tried in other campaigns, Indian intelligence pursued "prudent political, administrative and economic measures" for "less counting of body bags." In addition to targeting information, intelligence was gathered in order to "[devise] strategic approaches that could yield lasting solutions of the human problem and healing of the fault line." 5 3 Ultimately, the combination of enhanced Mahadevan, 2007. 151 Kanwar Sandhu, "New Upsurge: Police Morale at Low Ebb in Punjab," India Today, February 28, 1991. (http://indiatoday.intoday.in/story/police-morale-at-low-ebb-in-punjab/1/317930.html). 152 "Playing with Fire," Times ofIndia, Oct 26, 1985. "'Dhar, 2005, p. 320. 150 487 security operations and fresh political initiatives through security, economic, and other civic provisions made by security forces improved the security situation of Punjab. 54 Bringing Moderates Back Into the PoliticalFold The Indian state had sufficient trust that even some violent insurgent factions as well as separatist political leaders could be rehabilitated, and the state pursued concessions and political deals with them to return them to the fold of normal politics, even as it intensified kinetic activities . The purpose of bargaining with insurgent factions was not to induce insurgent defection that would escalate violence but to use them as bridges to consolidate control, dial back violence, and open space for a political deal. Covert Efforts. Prior to the more open moves that brought the Akali Dal back in to normal politics, the state used its Intelligence Bureau to draw militant factions into the election process, even without preconditions. One intelligence operative was tasked to bring "terrorismtainted leaders" to contest or engage in elections in order to "expose the militant sponsored politicians to the mainstream democratic process." 55 The presence of Sikhs in government and the security forces enabled the state to use them as emissaries to credibly reach out to militants for potential deals - including an attempt with former minister and political leader Talochan Singh Riyasti.1 56 Co-ethnic links also allowed for covert alteration of the insurgency. Interviews with a number of former Khalistani militants finds that links or familial ties between low-level members of the security force and the militants provided Indian intelligence agencies 154 11556 Dhar, 2005, p. 414. Dhar, 2005, pp. 374-75. Dhar, 2005, p. 329. 488 opportunities to "turn" members of the movement and use them to sideline more radical elements.157 In February 1988, one covert peace initiative was launched with the release of Bhindarwale's nephew and three high priests along with the provision of some arms to help them consolidate control over the insurgent movement and to set up a peace track. 158 While some of this infiltration was certainly designed to flip key leaders, drive wedges with others, and gather information for targeting, the primary goal was to move these groups towards a negotiations before they violently fractured like in India's Northeast. 159 Part of this involved ad hoc arrangements and "personal compensation packages" to woo militants, political leaders, affiliates.1 60 Another major covert attempt in 1991 authorized at the highest levels and led by senior intelligence officer M.K. Dhar again sought to bolster certain militant leaders who could facilitate a deal and rehabilitation.!61 Ultimately these larger attempts failed, vetoed both by militant leaders based in Pakistan as well as their ISI handlers who feared losing leverage. Despite this, the very attempt suggested a major strategic difference from Kashmir that would yield benefits down the road. Overt Efforts. Even as kinetic and non-kinetic innovations yielded some tactical success, the Punjab campaign sought more than the defeat of the insurgents. Accounts published in 1990s noted the dramatic success of the state in bringing down levels of violence but victory was not a foregone conclusion. The future of Sikh militancy, Khalistan separatism, and political subversion Laurent Gayer, 2009, p. 250. Dhar, 2005, pp. 334-37. "96 0 Dhar, 2005, pp. 383-87. 1 Dhar, 2005, p. 409. 161 Praveen Swami, "Fading Memories of a Promised Land," Frontline,24 (20), Oct 6-19, 2007 (http://www.hindu.com/thehindu/thscrip/print.pl?file=20071019503404200.htm&date=fl2420/&prd=fline). 157 15 489 still very much hung in the balance and depended on combining these pure military approaches with major political moves to consolidate control and prevent the resumption of insurgency.1 62 Throughout the Punjab counterinsurgency campaign even before informal negotiations, the Indian state had remained in communication with the insurgency through various individuals and backchannels for escalation management. Initially, the state had been restrained to a fault because President Zail Singh and Home Minister Buta Singh managed to stay its hand due to sympathy for and links to Bhindarwale, having been instrumental in propping him up in the late 1970s.163164 Even prior to open deals, Sikhs like Jiwan Singh Umranangal with close ties to both the Akali separatist movements and the state functioned as a bridge to curb the state's predilection for violent tactics while also reaching out to disenfranchised Sikh families to dissuade them from supporting the insurgency.1 65 In 1989, militants conveyed to DG Gill their desire for an off-ramp and though he thought nothing came of it after he reported this up the chain of command,1 66 its possible this served as the prelude to covert and overt attempts at accommodation that soon followed. The release of the Jodhpur detainees (held on sedition charges after Operation Blue Star) served as a goodwill gesture, but also freed these separatists to contest and even win seats the November 1989 elections.1 67 Alongside the kinetic operations, the state counterinsurgency campaign continued to create the space and provide institutional off-ramps to moderates and separatist-leaning political groups rather than repressing them all.1 68169 This period was marked by repeated serious efforts See for example, Sandhu, "Punjab: A Golden Opportunity," November 30, 1992; Gurharpal Singh, "Punjab Since 1984: Disorder, Order, and Legitimacy," Asian Survey, 36 (4), April 1996, pp. 410-421. 163 Tully & Jacob, 1985, p. 105. 164 Gayer, 2009, p. 238; Dhar, 2005. 165 See Nayar, 2013, p. 246; "Akali Leader Umranangal Passes Away," The Tribune (India), Nov 8, 1998 (http://www.tribuneindia.com/1998/98nov08/head2.htm). 66 1 Chandan, 2013, p. 164. 167 Paul Brass, The Politics ofIndia Since Independence, New York: Cambridge University Press, 1994, p. 196. 168 This is the crux of Chima's argument, both his 2010 book and his 2007 article. 162 490 at negotiations and political concessions, even though they may have at times inadvertently contributed to sustaining the insurgency.17 0 Throughout the campaign, the trust in the political reintegration of Sikh separatists was actively encouraged by a number of unbiased Sikh officials advising the Prime Minister including Arun Singh and KPS Gill. Dhar writes: Gandhi's decision to encourage a section of the Punjab militants to take part in the democratic process was a rare instance of his long-term strategic understanding of the problems of the people of Punjab... But Rajiv enjoyed an advantage in Punjab that he missed out in Kashmir. The new governor was not biased by prejudices.. .The Director General of Police, KPS Gill, contrary to the general perception believed in the intrinsic strength of the democratic process.17 1 By mid-1991, the state observed that certain militant groups had been "softened" and the government "exhibited boldness in extending the hands of conciliation to certain exhausted and disillusioned groups." 172 The political route to redress grievances was reopened and by 1992 the state began to resume state and local elections to motivate Sikh political elites, both extremist and more moderate Akali political leaders, to re-enter the normal democratic process.173 Though the Akalis did not officially participate in the September 1992 elections out of fear of insurgent targeting, some parties fielded independent candidates who did remarkably well, incenting them to re-enter the normal political process and climb down from their radical platforms.1 74175 The more moderate Akalis, still intimately tied to insurgent organizations and commanding influence with more extremist political groups, also served brokers. 176 A slightly Sikh majority in the state legislature also imbued security operations with greater legitimacy. 177 By 1993, full Akali Horowitz and Sharma, 2008, p. 761. Chima, 2010, pp. 170-71; 175-76. 171 Dhar, 2005, p. 376. 172 Dhar, 2005, p. 412. 173 Chima, 2010, p. 234. 74 169 170 1 Chima, 2010, p. 227. Telford, 2001 Chowdhury and Krebs, 2009, pp. 388-392; For instance, The Akali Dal head repotedly maintained close ties to insurgents groups. See "Prakash Singh Badal was Getting Protection From Khalistan Commando Force During Militancy: Capt Amarinder Singh," Times ofIndia, February 17, 2014. 177 Chima, 2007, p. 622. 175 176 491 participation in local elections and over 80% turnout signaled high confidence in the legitimacy of the state. Meanwhile, internal competition between Akali factions propelled a race back to moderate positions. 178179 Separatist groups were not only being drawn back into normal politics by 1992 with devolution, concessions, and accommodation, but the Akali Dal led by Badal went on to compete and win elections in 1997, effectively sealing the deal on political (re)integration and institute even more rehabilitation of former militant leaders to close the door on conflict recurrence.' 8 0 As on intelligence officer explained, the separatists had been won over and quieted by "the loaves and fishes of political office."' 8 ' Today, Sikhs have continued to dominate state politics whether through Akali alliance's with the BJP, or Sikh leaders who "dominate the Congress party" ensuring there are always Sikh state leaders to prevent the renewal of Khalistani violence. 182 2.3. Alternative Explanations of Punjab Strategy Two alternative explanations of the Punjab case warrant some attention. The first argument proposes an alternative interpretation of the strategy contending Punjab was not as balanced as the analysis above argues and was in fact incredibly brutal. The second argument recognizes the Punjab campaign was less violent but implies this was the "easier" insurgency (especially relative to Kashmir) due to lower conflict intensity, insurgent size, and external support. Both these arguments are found wanting. 17 Chima 2010, p. 231. Chima, 2010, pp. 229-30, 238-39. Gayer, 2009, pp. 250-51; Wallace, 2007, p. 461. Manraj Grewal details this in Dreams After Darkness: A Search for a Life Ordinary Under the Shadow of 1984, New Delhi: Rupa, 2004. 181 Author Interview, retired senior Indian intelligence official, New Delhi, India, December 11, 2013, (#142). 182 Wallace, 2007, pp. 426, 429, 434. 179 180 492 2.3.1. Alternative Explanation 1: Brutality A number of scholars have argued that the Punjab campaign was far more brutal pointing to episodes like Blue Star, the use of renegades, and the intensification of insurgent targeting between 1991-93 to supposedly finish off the insurgency. "state terror,"' 83 but new research reveals these accounts both "ahistorical and intellectually lazy."' 8 4 During this time, there was a belief that the Punjab police had adopted a "bullet for a bullet" policy as articulated by then police chief Julio Ribiero. In fact, this was merely a policy to correct extreme restraint and allow police to return fire if fired upon by militants, something that previously had been prohibited from doing. After this pacifist policy resulted in nine armed constables being killed without firing a shot in self-defense as a militant was freed from police custody, Ribiero decided to change it but that did not make it offensive. 85 Likewise, the claim that Punjab counterinsurgency was successful due to a "free hand" does not imply reckless violence but rather the lack of political and bureaucratic interference that in previous counterinsurgency campaigns generally hampered Indian pursuit of a sustained strategy.1 86 Contrary to these claims, one insider confirms that the Prime Minister "was keen on simultaneously pushing of the peace and war process" and Ribiero executed this by "finely blend[ing] the ingredients of tough policing and political approach."' 8 7 Nevertheless, after the blowback from Blue Star, the police still implemented a predominantly restrained and "defensive Paul Wallace 2007; Patricia Grossman, "India's Secret Armies," Death Squads in GlobalPerspective: Murder with Deniability,Ed. Bruce B. Campbell and Arthur D. Brenner, New York: Palgrave Macmillan, 2000, pp. 261286; Cynthia Keppley Mahmood, "Trials by Fire: Dynamics of Terror in Punjab and Kashmir," Death Squad: Anthropology of State Terror, Ed. Jeffrey A Sluka, Philadelphia: University of Pennsylvania Press, 2000. 184 Mahadevan, 2008; One Army Colonel makes a distinct case for why Kashmir involved much more force than 183 Punjab. See Ali Ahmed, 2009. 18 Tully, 1991, p. 161; G.V. Gupta, "Bullet-for-bullet is not My Baby," Tribune India, March 7, 1999 (http://www.tribuneindia.com/1999/99marO7/book.htm#1). 1 86 Chandan, 2013, p. 199. 117 Dhar, 2005, pp. 323-25. 493 strategy" for much of the campaign until the intelligence yields finally enabled more targeted offensives. 188 There was undoubtedly a reliance on force alongside all these non-kinetic tactics, and even some nasty tactics as nearly all counterinsurgencies are bound to involve,18 9 (e.g. bountykillings, renegades, torture, and fake encounters).1 90 The fact is, even if all these are accepted at face value, their prevalence is overstated. For instance, some turned insurgents were unleashed as renegades for kinetic activity targeting other insurgents but most of them served as informants. Hyperbolic claims of brutality were quite common soon after the campaign because few understood how the Punjab insurgency rapidly and suddenly collapsed.191 Most importantly, the state's tactical and strategic deals reveal there was no attempt to "finish off' the insurgency in the final. By the end of 1991, the insurgents were estimated at 10,000 strong and between 1992-93, only 2911 were killed, 2371 were arrested.1 92 The insurgency certainly had the numbers to continue (as numerous insurgencies in Nagaland and Kashmir have persisted for decades at lower levels). It therefore seems more plausible that what ushered the abrupt end to the insurgency were the state's use of negotiations and accommodations alongside violence with numerous individual deals in the surrender and rehab policy as well as broader political accommodations. The quantitative differences in civilians Kang, 2005, pp. 98-100. This is important to recognize that almost 90% of counterinsurgencies have involved targeted violence against civilians, particularly because it is found to be effective. See Johnston, 2009. The notion of clean, "hearts and minds" tactics as the pivotal element of a counterinsurgency campaign is simply a fantasy. 190 Romesh Silva, Jasmine Marwaha, and Jeff Klingner, "Violent Deaths and Enforced Disappearances During Counterinsurgency in Punjab, India: A Preliminary Quantitative Analysis," Human Rights Data analysis Group, Palo Alto: Benetech, January 26, 2009; Ami Laws and Vincent lacopini, "Police Torture in Punjab, India: An Extended Survey," Health and Human Rights, 6 (1), 2002, pp. 195-210. 191 Mahadevan 2007 explains this of Grossman, 2000. 192 Chandon, 2013, p. 200 and arrest data from Pramod Kumar "Violence in Retrospect," Punjab in Prosperityand Violence: Administration, Politics, and Social Change, 1947-1997, Ed. JS Grewal and Indu Banga, Chandigarh: Institute for Punjab Studies, 1998, p. 13. 188 189 494 killed as well as numerous other measures (see for instance, Table 6.7) reveal Punjab was quite a different campaign from Kashmir or many cases in the Northeast.' 93 Some have made the argument that Operation Blue Star itself contradicts the theory as it reveals the state willing to employ brute force against an ethnic identity group valued and embedded in the state, but a closer look at the operation does not bear this out. First, it could be argued that Blue Star in a way was not really part of the counterinsurgency campaign but in many ways triggered the insurgency's onset. Second, while the actual casualties are not known, experts seem to concur the total killed was around 1,000 with as many as half being insurgents and between 80-140 security force casualties, but a far cry from exaggerated claims of tens of thousands.1 94 It is possible the operation was even restrained to a degree with five of the seven commanding army officers being Sikhs.' 95 Simultaneous Army operations in the countryside labeled Operation Woodrose were alleged to have been equally as bad, but independent journalists who critique Blue Star contend that allegations are overblown. Isolated incidents were exaggerated but "there was no widespread repression, no organized brutality."19 6 After a little Silva et al, 2009 estimate something between 1,700-2,000 potential civilian casualties at the hands of the state. However this is significantly lower than what is estimated for Kashmir or earlier Northeast campaigns. Furthermore, Valentino and Uldfelder estimate and order of magnitude difference between civilians killed by the state in Kashmir versus Punjab. 194 Tully estimates 1,000 in a recent interview "I doubt Margaret Thatcher Played a Role in Operation Bluestar, Says Former BBC Bureau Chief Mark Tully," DNA India, January 14, 2014; Tully and Jacob, 1985, p. 185 say roughly 1,600 unaccounted for. While some say 492 included militants and civilians, others say the Army now claims it was just militants and report Army fatalities at 142. See Harpeet Bajwa, "The Operation Blue Star Papers," Sunday Standard(New Indian Express), February 9, 2014. Others put the figure at 1,000-1,200. See Puneet Singh Lamba, "The New York Times on Operation Blue Star," Sikh Times, June 2, 2004. This seems in line with a news report estimating 863 bodies of Sikhs being cremated soon after, which would mean at least 371 civilian casualties. See "Death Toll at Sikh Temple May Be as High as 2,000," Globe and Mail, June 12, 1984. Part of this high violence had to do with the intensity of the operation an estimated one-third of the roughly 1,000 security forces deployed for the operation suffering casualties. An estimated 83 were killed and 249 injured. Jacob and Tully, 1985, p. 184. 195 Tully and Jacob, 1985, pp. 147, 152; Weaver, 1984. 196 Tuly and Jacob, 1985, p. 205. 193 495 more than a month into Woodrose, roughly two-thirds of these apprehended in the sweep had been released.1 97 Most scholars acknowledged the high violence in Blue Star was not a result of deliberation but rather its absence, poor planning, and a lack of intelligence - more a Branch Dividian compound botched raid rather than a My Lai massacre. The operation raised alarm in Delhi due to sheer difficulty and unpreparedness of Indian Army forces against a ragtag bunch of militants, insurgents' preparations, and the degree of resistance. 198 Blue Star was indeed a colossal blunder by the Indian military and if the state had persisted with this approach, it would have proceeded down a path of attrition. But it is noteworthy how quickly this forced the Indian state to alter its approach to the Punjab counterinsurgency for the next decade. Even if one accepts the Indian state similarly altered its strategic approach in Kashmir, it took the state nearly fifteen years to do so while in Punjab it occurred in a matter of months. Finally, there are also three reasons to be confident in the qualitative assessment of the campaign. First, the above analysis is supported by numerous interviews with former practitioners, journalists, and analysts who served or covered the campaign at the time as well as work by other academics. 99 Second, even at face value, the levels of alleged violence are still much lower in Punjab relative to most conflict theaters in India after accounting for the scale of violence and the population size. Third, there is an inherent overestimation bias in Punjab since information on this campaign is relatively more accessible than on Kashmir or the Northeast since the conflict has been resolved, more transparent civilian rather than military sources 3,000 of the 4,508 persons apprehended, were released. "'Disturbed' Status for Punjab Stays," Times ofIndia, July 8, 1984, p. 1. 197 198 199 VK Singh, 2007, p. 170. Mahadevan, 2008; Chima, 2010; Chowdhury and Krebs, 2009; Horowitz and Sharma, 2008. 496 collected the information, and opposition parties that came to power (e.g. the Akali Dal) were able to closely scrutinize and publicly reveal it. 2.3.2. Alternative Explanation 2: The "Easier"Insurgency? An argument could be made the scale of the Punjab insurgency was simply not on par with that of Kashmir in terms of violence, insurgent size, or external support from India's archrival Pakistan, therefore explaining why the Indian state took a "softer" approach, but again, a close look at the evidence rules this out. 200 Conflict intensity was actually on par when accounting for conflict duration. The Punjab insurgency lasted 11 years in that time, insurgents killed an estimated 13,410 civilians and security forces. Meanwhile, in the first 11 years if Kashmir's counterinsurgency campaign, insurgents killed an estimated 11,744-13,708 civilians and security personnel. The overall difference in intensity during this time was the far greater number of Kashmiri insurgents and civilians killed by the state. 2 0 1Additionally, though the size of the insurgency is dynamic and endogenous to the conflict, estimates track closely. Various state sources estimate insurgent size for both at 5-6,000 at the low end and 9-10,000 at the high end.202 This is frequently the response offered when asking Indian security officials why Punjab looked different from Kashmir. Author interviews, 2011-13. 201 SATP data provides the Punjab and higher Kashmir estimate; Swami (2007) data provides the lower. At the same time, during this period the Indian state killed 80% more insurgents (based on Swami data) and acknowledges killing over 2,800 civilians (J&K govt data). 202 Sikh insurgents were estimated at 6,000 at their peak (Telford, 2001); Kashmiri insurgents were estimated at 5,000-6,000 (LTC Singha, Prathividrohi, 1995, p. 34); At the high end, insurgents estimated at 9,000-10,000 by Punjab police sources (Chandan, 2013; SIPRI 1987-88, William K. Stevens, "800 Sikh Militants are Said to Have Died in Raid," New York Times, June 11, 1984) and in Kashmir by Army-intelligence sources (Rao, 2001, pp. 446, 473; UCD database; Wirsing, 1994, p. 129-30). This 10,000 figure was echoed by Gov. GC Saxena in 1991. See Edward A. Gargan, "Behind Its Mountain Walls, Kashmir Wages Vicious War," New York Times, October 28, 1991). Though some have estimated as high as 45,000 potential insurgents (Victoria Schofield, Kashmir in Conflict: India, Pakistan and the Unfinished War, London: IB Tauris, 2000, p. 157), this again tracks closely with the number of potential insurgents caught along the Punjab-Pakistan border (Swami, 2007, p. 148). 200 497 Finally, Pakistan's involvement with the Punjab insurgency was also robust, and in a way, a training ground for Kashmir. The Pakistani state not only maintained links to Sikh militants but after Blue Star, Sikhs flooded across the border for material support, training, and safe haven, and an estimated 5,000 were trained in Pakistani insurgent training camps, many alongside Afghans during the 1980s. 2 03 Additionally, at least 45,000 suspected Sikh insurgents were captured along the Punjab/Pakistan border along with 26% of all Kalashnikov rifles capture in Punjab.20 4 New research based on reviews of classified documents and interviews with former insurgents reveals Pakistan was deeply involved in all aspects of the insurgency including coordinating different insurgent groups within the Panthic Committee organization and even directing tactics, targets, and operations. 2 05 Between 1986-89, 60% (375) of all Pakistani spies captured in India were in Punjab.206 Pakistani intelligence had established extensive ties with Khalistani separatists almost a decade earlier and been involved with a number of operations including terrorist plane hijackers, 2 07 and even today Indian intelligence officers judge the key reason for the onset, escalation and endurance of the Punjab insurgency was Pakistan.2 08 Strategically, Punjab and Kashmir served a similar purpose for Pakistan, which is why Operation K2 (Khalistan and Kashmir) intended to align and link both theaters, though it was ultimately foiled by the rapid progress of the Punjab police.2 09 It is hard to defend that Pakistan's involvement in Punjab was equal to that of Kashmir. However, the difference in degree of support cannot explain the order of magnitude difference in the state's strategy of violence. Moreover, Pakistan's deeper involvement in Kashmir, and 203 D.P. Sharma, 2006, p. 270; Gayer, 2009, pp. 242-43. Swami, 2007, p. 148. 205 Gayer, 2009, pp. 242-247; Swami, 2007, pp. 145-150; Suneel Kumar, 2007. 206 D.P. Sharma,1996, p. 285. 207 Swami, 2007, p. 147. 204 Author Interview, senior Indian intelligence official, New Delhi, India, December 12, 2013 (#143). Dhar (2005) also talks about Pakistan vetoes of deals. 208 209 Swami, 2007, pp. 154, 193. 498 India's perception of this, may have much to do with the closer proximity of Pakistani Muslim and Kashmiri Muslim identity, making the insurgency more dangerous and the conflict more charged. But if this is the case, it confirms the salience of identity, and locates it prior to or constitutive of material support. By combining multiple sources, personal accounts, and data, this section also reveals new insights about the campaign itself. For instance, the Punjab has been routinely studied as an example of an effective counterinsurgency campaign for the tactics employed, but few studies sought to explain why this campaign was allowed to adapt to a model of successful practices and stable politics. Counterinsurgency strategy is neither just a product of states living up to their material, institutional, or organizational capacity nor is it just about the right individual agency at the right time. The accounts from military, intelligence, and politics round out the picture to reveal a state informed by costs and consequences and working to adjust and experiment with new tactics, including less brutal ones.2 10 The theater commander was able to guide a strategy of population because of the background roles of a lot of influential military, civic, and political leaders, all motivated by an meso-level incentive-cost structure. Punjab Police Director General Gill lead the campaign through the right doors but these other actors worked to close certain doors (like the use of the military for intense violence) and open others (such as the tremendous supply of resources and organizational capital, and attempts to bring moderate separatists back into the political fold). III. The Kashmir Insurgency: 1990-Present After decades of repression and manipulative rule by the center, a rigged 1987 election in the Muslim-dominated Kashmir valley sparked a crisis and provided the impetus for a rebellion. 210 I'm referring here to Nayar, 2013; Dhar, 2005; and Chima, 2010. 499 Kashmiri youth, mobilized by the countless insurgent organizations funded by Pakistan, took up arms against the state amidst a groundswell of popular support and disaffection. After a slow start in 1989, insurgent violence erupted in 1990 in response to the harsh federal crackdown, control by President's Rule, imposed curfews, and harsh tactics of federal military and paramilitary forces. Violence would escalate to a peak in 1996, dip slightly and then rise again through 2001 with the increase of foreign militants. All the while the Indian state concentrated efforts on wiping out insurgent organizations and creating vacuums for new groups to succeed them. Attrition focused first on the nationalist JKLF, followed by the Islamist HM and HuA, and subsequently moved on the foreign tanzims with local support base like LeT and JeM. Though the Indian government claims to have changed strategy after 2003 when violence began to fall dramatically, much of the state posture remains the same and the conflict is ongoing with over an estimated 70,000 killed. When insurgency broke out in Kashmir, the state calculated, like in Punjab, the incentives of a high-effort campaign for total control of the region far exceeded the costs. However, instead of population control, the GOI opted for an attrition strategy. The Kashmiri Muslim identity of the rebel base simply did not guilt, motivate, or compel state leaders to deploy a restrained, less violent strategy of selective targeting, non-kinetic substitutes, or violent negotiation. Instead, state disdain and distrust of the rebel identity base encouraged the state to adopt a high-violence strategy. 3.1. Territory As far as territory was concerned, Kashmir was essential to the Indian state. Kashmir held all the same strategic imperatives as Punjab as a focal point of contestation with Pakistan and 500 strategic terrain for its territorial defense,2 1 ' provided critical water resources from its rivers as well as foreign exchange from a robust tourism industry,2 12 but most importantly possessed the emotional and ideational trump card. Indian government leaders considered Kashmir "the real jewel in India's crown" 213 and "the cornerstone of its identity as an inclusive, secular state"214 which motivated the state to deploy everything in its arsenal to regain control of the region. The insurgents' ability to project force beyond Kashmir into the capital and even kidnap Western tourists from a Delhi guesthouse required the Indian state to respond seriously.2 15 Unlike in Punjab, the costs of state reach and control were not cheap, particularly due to difficult physical terrain, but the incentives exceeded these costs to justify a high-effort counterinsurgency campaign. The physical terrain costs of Kashmir posed significant challenges, especially the innumerable mountainous passes that enabled insurgent infiltration from Pakistan, but India needed to control this terrain regardless in order to protect against a potentially revisionist adversary. Other campaign costs were more manageable. Kashmir was in close proximity to the majority of Indian Army forces as well as paramilitary forces making logistics and supply much easier, and also making the insurgent threat that much closer to the political center. Despite the difficult physical terrain, the state had a long history of asserting disciplinary power over the human terrain. For centuries dating back to the Sikh and Dogra empires, outsiders had ruled over the Muslim subjects of the valley with an iron fist, so that even while Swami, 2007, p. 46. Chindu Sreedharan, "Reviving the Economy," Blood In the Snow: Ten Years of Conflict in Kahsmir, Rediffcom, Dec 17, 1999 (http://www.rediff.com/news/1999/dec/17blood.htm); Surinder Singh Oberoi, "Kashmir is Bleeding," Bulletin ofAtomic Scientists, 53 (2), March-April, 1997. 213 Andrew Hogg, "Gandhi's 'Jewel' Ricked by Bombs," The Sunday Times, Feb 26, 1989 211 2 12 2 215 Bose, 2003, p. 111. Joshi, 1999, p. 429; Levy and Scott Clark, 2012, p. 141; Praveen Swami, 2007: 198-99. 501 the Muslim population was resentful and resistant, they were accustomed to centralized state authority, therefore reducing expected human terrain costs. 2 16 With this incentive structure, the state was motivated to take decisive counterinsurgency actions. Learning from the Soviet Union's mistake of only committing "relatively modest" resources in Afghanistan, the Indian state "would demonstrate its willingness and ability to commit enormous military resources" to safeguard its prize of Kashmir.2 17 In late 1998, a former RA&W chief appointed as Governor for a second stint stated publicly in an interview, "You will never see a loosening of India's grip on Kashmir." 218 The belief that Kashmir is a "core interest" has been repeated ad infinitum by a military, which believes that if "we let Kashmir go, we're unraveled." 2 19 To this end, the Indian state wasted no time or effort launching an attrition campaign as soon as Kashmir began to boil to an insurgency. As anti-government sentiment and pockets of unrest percolated after the 1987 election, the state government was plagued with paralysis and hesitated to take serious actions (much like in Assam and Andhra Pradesh), but once violence escalated, the central government took decisive action and seized control by declaring federal rule in January 1990.220 Inchoate insurgent violence sought to "strip [the state] of its authority" rather than target it directly with only 6 of 390 violent incidents in 1988 and 49 of 2154 incidents in 1989 aimed against security forces.2 2 ' Days before the crackdown began, police estimated 2 16 Mridu Rai, Hindu Rulers, Muslim Subjects: Islam, Rights, and the History of Kashmir, Princeton, NJ: Princeton 2University Press, 2004. 1 Praveen Swami, 2007: 170 218 "India'11 Never Loosen its Grip on Kashmir," Interview of Girish Chandra Saxena, IndianExpress, October 17, 1998. 219 22 0 221 Author Interview, retired Indian Army officer, December 2013 (#145). Marwah, 2009, p. 48; Bose, 2003, p. 108. Praveen Swami, 2007, pp. 164-65. 502 only 50 hardcore militants behind the instability. Nevertheless, the Indian state still chose to crack down hard on the nascent insurgency by deploying 80,000 BSF paramilitary troops and declaring President's rule and a region-wide curfew.2 23 The Indian Army, typically averse to heavily populated environments, took the extraordinary measure of patrolling urban curfewed areas of the Kashmir valley.22 4 With these aggressive measures of rapid security force deployments, emboldened by the immunity of AFSPA, the Indian state demonstrated its resolve to retaining control of Kashmir, defeating any militant movements, and staying until the job was done Violence, but these heavy-handed measures quickly escalated and intensified the conflict. 225 The Kashmir campaign demanded massive manpower and the government saturated the region with 300-400,000 security forces by 1993 and up to 500,000 by 1995.226 Troop deployments would hover around this level for the next two decades. Though some have reported deployments in the 500-700,00 range,2 27 even if forces levels reached no higher than 400,000 troops including police, 22 8 this was still the largest and most dense deployment in India's counterinsurgency history. In a region where the population was 7.5 million at the time Sanjoy Hazarika, "A Vale of Tears and Anger: Anti-Indian Militancy is on the Rise in Kashmir," New York Times, January 14, 1990, p. 3 223 The authors say one-third of BSF's 240,000 strong force was deployed. Levy and Scott Clark, 2012: 43. 224 Sanjoy Hazarika, "Violent Protests Growing in Kashmir," New York Times, January 10, 1990, p. 3; Bose, 2003: 222 146. Initially the violence began small. 80 people had been killed in bombings in 1989 by mid Dec 1989. And from mid 1988 to mid 1989, about 40 Muslim youths were killed by police (so estimating half in 1989). Within 7 days of curfew, more than 70 people were killed and even journalists were told that they could be shot on sight if they broke the strict curfew in place. See "Delhi Offers to Free Kashmiri Separatists in Kidnap Deal," Sydney Morning Herald, Dec 12, 1989 (org form Reuters, Guardian), p. 10; SATP estimates 79 civilians killed in 1989. Tony Allen-Mills, "India and Pakistan Play Kashmir Charades," The Independent, August 10, 1989, p. 9; The same Moulvi Farooq said that 20 had been killed by Feb 1989. See Andrew Hogg, "Gandhi's 'Jewel' Ricked by Bombs," The Sunday Times, Feb 26, 1989; Dilip Ganguly, "Troops Search House-to-House for Separatist Moslems," AssociatedPress, January 26, 1990. 226 Wirsing, 1994: 146; Newberg, 1995: 27; Barry Bearak, "Kashmir a Crushed Jewel Caught in a Vise of Hatred," New York Times, August 12, 1999 227 Angana P. Chatterji says 671,000 in Sept. 2011 including counter-terrorism and special forces. See Angana P. Chatterji, "The Militarized Zone," Kashmir: The Casefor Freedom, London: Verso, 2011: p. 106. 228 Author Interview, Indian journalist, Srinagar, India, Dec 8, 2012 (#108). 225 503 Figure 7.1: Disposition of Indian Counterinsurgency Forces in Kashmir, 1993 C IS C H INA K A S H M I R J */ z - h GAN UND_ -' _ _ __ _ _ _ _ __ _ _ _ _ _ UOASPU Harmrruo 4c4Ax IARMOUR I RZEULAA TROW" 2 PARA MWUTARY I.Skw J SMBMi IS~ INDIAN SECURITY FORCES IN INDIAN HELD KASHMIR SAtNGDUTE Source: Robert G. Wirsing, India, Pakistan, and the Kashmir Dispute, New York: St. Martin's Press, 1994, p. 271 (Appendix II). Note - the author disclaims the map was supplied by Pakistan military intelligence and was not independently verified. 504 and the Kashmir valley between 3-4 million, these force levels amounted to "security forces everywhere-at least one man every fifty feet or so in a city." The purpose of this saturation was to "deter insurgents, intimidate civilians, and even more, make clear that the Indian government dominates the state [of Jammu & Kashmir]." 2 2 9 Today, the overwhelming coercive footprint of the state is still evident in places like Gurez Valley where a contingent of 26,000 troops from an infantry division and artillery brigade surrounds a population of 31,800. Similar to Punjab but in contrast to most other Indian campaigns, state security forces in Kashmir aimed for transformative control rather than containment. The Indian strategy was "not simply repression, but a concerted effort to reshape life in Kashmir." 2 3 After briefly losing control, troop deployments fanned out across the region beyond population centers to retake insurgent-established no-go zones by 1991-92 and "brutally battened down" the countryside. 3 Bunkers and checkpoints across Kashmir resembled the trench warfare of World War I where "land was retained painfully and at a high human cost."234 India's organizational innovations reveal another indicator of its commitment and effort during this campaign. Early in the campaign, the GOI experimented with the creation of an unorthodox unit, the Rashtriya Rifles (RR), as a specialized and permanent counterinsurgency force. Initiated in 1990, the RR required significant organizational and tactical innovation creating a new organization with different mission sets within the Army. It also required raising new troop strength within these units to create a permanent battalion deployment without a rotation of troops and extensive coordination across the ministries of defense and home affairs. 229 Newberg, 1995: 27 Iftikhar Gilani, "Exfiltrating Cattle Are the New Border Town Blues for Kashmiris," Tehelka, July 5, 2011 (http://archive.tehelka.com/story main50.asp?filename=WsO5O7 11 Exfiltrating.asp). 23 Newberg, 1995: 10. 232 Levy and Scott Clark, 2012: 398. 233 Praveen Swami, 2007: 179. 2 34 Joshi, 1999: 136. 230 505 Gradually, the RR assumed a central role in India's counterinsurgency strategy in Kashmir. 235 The importance of regaining firm control allowed this force to thrive despite traditionally powerful parochial interests. The RR expanded from a Rs 5.2 million, six battalion experiment in 1990 to 30 battalions by the mid-1990s, to a 66 battalion (or six division) force by 2005 that cost the government more than Rs. 7280 million annually and served as the lead counterinsurgent force in Kashmir. 23 6 It is noteworthy that when similar experiments like dedicated Army counterinsurgency "I (insurgency)-battalions" were attempted in the Northeast campaigns, they were terminated within two years due to the perceived tradeoffs, but in Kashmir, such organizational innovations were nurtured and prevailed.2 3 7 The rapidity and scale of expansion in Kashmir also demonstrated organizational flexibility uncharacteristic of the Indian military. Before 1989, there was only 1 battalion posted in Srinagar and after the outbreak of insurgency, but this would later increase to increased to 150-200 military battalions at the peak of the conflict.2 38 Three corps commands within the state were established in the Kashmir Valley (XV Corps), in Jammu (XVI Corps), and for the Ladakh region (XIV Corps). 239 Additionally, a unified command overcome the objections of those seeking to defend their bureaucratic prerogatives and facilitated coordination by placing the paramilitary units, police, and civil bureaucracy underneath a single chain of command, effectively controlled by a military general.24 23 5 Joshi, 236 1999: 404-05. Rajagopalan, 2004, pp. 25-37; Author Interview, retired Indian Army officer, New Delhi, India Nov 2012 (#84). 237 See Rajagopalan, 2004: 25-26. These were disbanded despite far less external threat pressures in the period after India's victory over Pakistan in the 1965 war and when the conventional military balance shifted even further in its favor. 238 Author Interview, retired Indian Army officer, New Delhi, India Nov 2012 (#84); Rajagopalan, 2004: 33 239 Author Interview, retired Indian Army officer, New Delhi, India Nov 2012 (#84). 240 Rao, 2001: 448-49; Joshi, 1999: 228. 506 3.2. Identity Kashmiri Muslims' low moral worth in the eyes of the Indian state and public, and their near total exclusion from national government institutions and security forces at allowed the Indian state to run roughshod over Kashmiri rebels and their civilian base with a high-violence strategy. Chapter 6 makes clear Indian Muslims positional status had dramatically declined since India's independence, but Kashmiri Muslims of the valley had been continuously repressed and subjugated long before 1947 since the days of the Sikh and Hindu Dogra empires as a lower class that was dis-embedded from state institutions. My theory then expects that when the Kashmiri Muslims rebelled against the state, their lack of worth or embeddedness would evoke little sympathy and impose few constraints on state violence. Scrutiny of the Indian attrition strategy in Kashmir reveals the total absence of those identity-induced restraining mechanisms at work in Punjab. With such social distance and non-embeddedness of Kashmir Muslims, the Indian government and security forces were unrestrained by mechanisms of empathy or vulnerability, and lacked mechanisms of information or trust to develop substitute tactics to pure force. First, the state regarded Muslims and particularly Kashmiri Muslims as conspicuous, of low moral worth, disloyal, and in league with India's archrival Pakistan, evoking little empathy and limiting the state's acknowledgement of grievances, concern for civilian casualties, or interest in insurgent rehabilitation. Second, the very limited number of Kashmiris or Muslims in positions of institutional influence or power (such as the civil bureaucracy or security forces) at the state or national level limited the state's exposure to internal dissent, political costs or defection, or any other 507 consequences of excessive violence or brute force that would otherwise prompt a course correction. Third, the scarce number of Muslims in the administrative, security, or intelligence apparatus during the campaign limited access to the rebel identity base, constrained strategic information on rebel dynamics and tactical information on insurgent organizations or for selective targeting, and a prompted a resort to crude, indiscriminate tactics like the use of brutal, unaccountable renegade militias. Fourth, state perceptions of Muslim disloyalty and mutual distrust with no interlocutors to bridge this gap prevented the state's credible outreach to Kashmiri Muslims whether through strategic negotiations, tactical surrenders, or public goods. This limited the scope of non-kinetic tactics or non-lethal demobilization of insurgents, compelling the state to rely heavily on violence with a strategy of attrition. All four of these absent mechanisms are explored in further detail. 3.2.1. Empathy Both the Indian state and the Indian public, partially influenced by rising Hindu nationalism, perceived Kashmiri Muslims as outsiders, foreign, and disloyal evoking little sympathy for their grievances or suffering during the counterinsurgency campaign. Consequently, the state deployed security forces and leaders that would have no qualms prosecuting an attrition campaign that brooked no restraints and measured success in kill rates. The result was tremendous indiscriminate violence and civilian victimization. 508 Disdain/Distance From the outset, Kashmiris were perceived as different, distant, and potentially disloyal. Kashmiris themselves were considered of low moral worth -- "not their own people but someone to rule"2 4 ' and not necessarily "true Indians."2 4 2 Former chief minister Farooq Abdullah once remarked on the exclusion of the Kashmiri people, "Kashmir is geographically part of India but it is certainly not part of Indian democracy" and retired Kashmiri officials claim this remains constant.243 Historically J&K had "overwhelmingly never identified with the Indian state, with the idea of India," 244 and retained closer ties to Western Punjab (in Pakistan) and Central Asia due to historical trade routes. 245 Violent political mobilization supported by Pakistan provided early evidence of disloyalty to the Indian state. 24 6 Perceptions of Kashmiri disloyalty may have been intensified by the rise of Hindu nationalism in the 1980s.2 47 At the time of the Kashmir insurgency's outbreak, there was an intensified national "mood" shaped by Hindu nationalist feelings, that manifested itself in: (1) Strong resentment against what many felt was the government's "pampering" of Muslims; (2) the conviction that Muslims were polygamous, bred faster, and in general were culturally inferior to Hindus; (3) the strong suspicion that Muslims were basically loyal to Pakistan; and (4) the equally strong suspicion that foreigners (and not just Pakistanis) were meddling in India on behalf of the Muslim minority.24 8 Author Author 243 Author 244 Author 245 Author 241 242 246 247 248 Interview, Interview, Interview, Interview, Interview, Kashmiri journlist, New Delhi, India, December 11, 2012 (#114). retired Kashmiri police officer, Srinagar, India, December 9, 2012 (#111). retired Kashmiri police officer, Srinagar, India, December 9, 2012 (#111). Kashmiri journalist, Srinagar, India, December 7, 2012 (#107). Kashmiri journlist, New Delhi, India, December 11, 2012 (#114); Bose, 2003, p. 31. Bose, 2003, pp. 32, 98, 148. Author Interview, Indian journalist, Srinagar, India, December 6, 2012 (#105). Wirsing, 1994: 165. 509 ' Leaders in Delhi rode this wave of Hindu nationalism and escalated Kashmiri resistance "into a nasty religious war" in the early 1990s seeking to firm up "the Hindu government's writ on Muslims in the valley." 249 Indian perceptions of Kashmiri distance and foreignness, allowed "collective punishment on a disloyal population"250 rather than little restraint or concern for collateral damage. One scholar traveling through the region attributed the intensity of violence and human rights abuses partly to "the ethno-religious differences that divide the security forces from most of the residents in the valley and that intensify the distrust between them." 25 While empathy played an important restraining role in the Punjab campaign, such restraint was out of the question for even uncommitted Kashmiri militants or accidental guerillas. With the insurgents classed as anti-national elements, "the militant had gone out of the pale of humanity for the ordinary officer or jawan [enlisted soldier]. This was not a matter of nomenclature, but of the brutal battle being conducted."25 2 Kashmiris as a group did not deserve human rights or rule of law. As one lawyer put it, "They don't consider us human. So where does the question of rights come in?...The law here is the security personnel. If they want to kill you, they will kill you..."2 5 3 Counterinsurgencies against Sikh and Naxalite insurgents involved distinct efforts to rehabilitate militants but "humbug efforts to bring 'misguided' people into the national mainstream" were pilloried in Kashmir arguing, "The morality bug and human considerations are opium for inaction." 25 4 249 Levy 250 and Scott Clark, 2012: 128. Navnita Behera, Demystifying Kashmir, Washington DC: Brookings, 2006: 48-49. Wirsing, 1994: 162. Joshi, 1999: 155. 253 Chindu Sreedharan, "Wronged Rights," 1999. 25 251 252 4 Ali Ahmed, 2009: 63 quoting another author's article. "Joint civil-Military Responsibility for Terrorism: Critique, Part II," Combat Journal,December 1992: 83; "Defusing Violence and Militancy: An Increasing Role for Armed Forces," CombatJournal, December 1991. 510 Indian brutally in Kashmir was abetted by a total incomprehension of the demand-side of rebellion or what propelled mass participation and support of the insurgency by average, frustrated people: Most did not consist of battle-hardened mujahids, but of students, mechanics, tailors, carpenters, bakers, schoolteachers and the unemployed, an army of amateurs supported by hundreds of thousands of poor country people whose resentment of India kept their lips sealed. All too often the price of their loyalty was a terrible retribution from the Indian security forces: ancient wooden villages incinerated and angry garrisons parked outside the ruins, a corrosive cycle of military operations - crack-down, cordon-andsearch and catch-and-kill - that saw thousands of residents detained without warrant, tortured and killed without trial, their bodies dumped in unmarked graves scattered throughout the district's woods. Despite a new strategy launched around 2004 to purportedly win hearts and minds was, the federal security forces deployed to Kashmir remain incapable of communicating with Muslim civilians and routine prejudice, suspicion, and excessive violence persist despite a decline in insurgent violence. 256 The continued aggressive posture and approach to counterinsurgency has been justified and likened to drowning an adversary until the final breath bubbles to the surface, because prematurely releasing him could quickly undo the progress. 25 7 Kashmiri scholars allege the Indian state has still never been "sensitized to the emotions and psyche of Kashmiris." 258 Even democratically elected Kashmiri leaders who expressed a degree of empathy and willingness to address insurgent grievances continue to be labeled "antinational" by some in the security forces.2 59 Levy and Scott Clark, 2012: 386-87. Jafa, 2005: 156 257 Author interview, retired senior Indian Army officer, Gurgaon, India, Dec 14, 2013 (#145). 258 Author Interview, Kashmiri journalist, Nee Delhi, India, December 2012 (#114). 259 A.G. Noorani, The Kashmir Dispute: 1947 -2012, New Delhi: Tulika Books, 2013, p. 468 identifies Lt Gen Sinha referring to Mufti Mohammed Syed. 255 256 511 Proxy War The Indian state made a point of distancing itself from the violence it perpetrated by alleging insurgency was reducible to the work of a foreign power and therefore conventional conflict rules applied. Even in the early years, many Indian state elites "often exaggerated the foreign component" 260 and describe the insurgency as a "proxy war" led by Pakistan or a fight of "Islamic Jihad." 2 61 One former state official told me, "When [the Indian Army] fights in Kashmir, they think they're fighting Pakistanis" 262 an opinion echoed by numerous Kashmiri journalists.2 63 Gov. Garish Saxena, the former RA&W director dispatched as governor during the campaign, viewed the Kashmir insurgency not as an insurgency but "a 'low intensity conflict' launched by Pakistan and the Indian Army could hardly refuse to give battle." 264265 While convenient to label the violence foreign terrorism in order to justify force and discredit the movement, 266 in fact, the data suggests the militancy did not become heavily populated by foreign militants until the late 1990s and even then only estimated around one-third of all militants. 2 67 Perception of a proxy war stemmed from leaders and troops conflating the identities and goals of Kashmiris with Pakistanis 268 and justified "a free hand" 269 in the use of violence in order 2 60 261 Praveen Swami, 2007: 194 Ved Marwah, 2009: p. 66, 71. Author Interview, retired Kashmiri state official, Srinagar, India, Dec 9, 2012, (#112). 263 Author Interviews, Kashmiri Journalists, Srinagar and Delhi, India, Dec 2012 (#107, 114). 262 2 64 265 266 Joshi, 1999: 138. Joshi, 1999: 77. Newberg, 1995: 72. Praveen Swami, 2007: 194 makes this point. Government data shoe local militants killed outstrips foreign militants until at least 2000 (SATP data). Furthermore, locals composed an average of 40% of infiltration. See MHA 2006-07 data. Data also shows the percentage of foreign terrorists killed out of total terrorists to average 18% between 1990-2006 and to never cross the 36%. See MHA data 11361 & 12402 "Militants Activities and Persons Killed in Jammu and Kashmir (1990 to 2010)" Indiastat. Also see Swami (2007) data. 268 In one incident earlier in the conflict, soldiers were overheard referring to all Kashmiri Muslims as Pakistanis. 267 Bose, 2003: 113 269 Author Interviews, Kashmiri Journalist, Srinagar, India, Dec 7, 2012 (#107). 512 to control the territory with little regard for the people. 2 70 The attitude of the state towards even peaceful political protests was "if you are a Kashmiri, you're a Muslim, you're pro-Pakistan and you have to be dealt with accordingly."271 A former Army major general acknowledges that any Kashmiris who ever received training and material support across the border in Pakistan were treated like foreign enemy combatants rather than insurgents who could potentially be reconciled. 2 Today, military officers admit there is far less political and social pressure from politicians and NGOs that challenges operational freedom when the Army can claim its counterinsurgency is really fighting a proxy war with Pakistan. Forcesand Leaders Deployed In quite the opposite of Punjab, the Indian state maintained and fostered a continued insensitivity to the grievances of Kashmiris or potential collateral damage. The GOI consistently deployed more offensive military forces to lead the counterinsurgency campaign, sidelining the local police, and it appointed Governors from the intelligence or military services to lead the state, sidelining most civilian bureaucrats, leave alone Kashmiri political leaders. Initially central paramilitary forces of the CRPF and BSF that "consisted almost entirely of non-Muslim Indians from outside IJK"2 74 were dispatched to the state when violence broke out in 1990. Soon the army, which lad even less Muslim or Kashmiri representation, would arrive. The utterly disarrayed police, about half Kashmiri Muslims, were sidelined for the first few years but even as they were integrated into operational responsibilities, top officials at the Author Interview, Kashmiri academic, Srinagar, India, Dec 8, 2012 (#110) Riaz Punjabi, "Drift to Disaster (Roundtable Discussion)," India Today, August 31, 1991. (http://indiatoday.intoday.in/story/Political-expediency-brings-simmering-problem-of-kashmiri-secessionism-to270 271 boil/1/3 1873 8.html) Author Interview, retired senior Indian Army officer, New Delhi, Dec 11, 2012 (#113). Author Interview, retired Indian Army officer, New Delhi, India, Nov 20, 2012 (#79). 274 Bose, 2003: 109; The same claim is made by Wirsing, 1994: 145; Iffat Malik, Kashmir: Ethnic Conflict, InternationalDispute, Karachi: Oxford University Press, 2002, p. 303. 272 273 513 state level and in each of the state's fourteen districts were "almost invariably Hindu or Sikh." 2 71 No Kashmiri Muslims served as heads of police (DG) or IGs during the time of the insurgency. There had been two in the early and mid 1980s but they were either forced into early retirement due to poor treatment or transferred out due to fears about their loyalty.276 The Governors of Kashmir who commanded extraordinary power were frequently from the military or paramilitary,27 with little connection to the Muslim community and at best oblivious to Muslim grievances and at worst, Hindu nationalists or "stinking communalists." 27 8 Rather than acknowledging a need to reach out to Kashmiri Muslims with Governors who could calm tensions as it often tried to do in Punjab, the Indian state displayed an uncanny knack for provocation. For instance, as tensions with Kashmiri public mounted in the aftermath of a rigged 1987 election and rising acts of violence, the Indian state made a deliberate choice to appoint Jagmohan Malhotra-who leading Kashmiri politician Farooq Abdullah described as "a man who hates the guts of Muslims" 27 9 -to serve as governor in early 1990. Though not steeped in the security forces like the governors who would follow him for the next eighteen years, his reputation of being tough on Muslims was considered an asset by the Indian state (shaped by the BJP's influence on the coalition government) rather than a liability that might antagonize the population or instigate violence. 2 75 Wirsing, 1994: 144 Alee Andrabi, "J&K Police: Estd. 1914, Kashmiri Muslim Chiefs: 2" Greater Kashmir, Aug 22, 2008 (http://www.greaterkashmir.com/news/2008/Aug/22/-k-police-estd-1914-kashmiri-muslim-chiefs-2-45.asp); Author Interview, Kashmiri journalist, New Delhi, India, December 11, 2012 (#114). 277 e.g. Gen. KV Krishna Rao, 1989-90, 1993-98; IPS officer and Head of external intelligence G.C. Saxena, 199093, 1998-2003), and Lt. Gen. Sinha (2003-08). 278 Jagmohan was referred to as a Hindu Nationalist; Sinha as a communliast. Author Interview, retired senior Kashmiri police officer, Srinagar, India, December, 9, 2012 (#111). 279 Schofield, 2000: 147. 276 514 Brute Force From the very beginning of the campaign, the state believed victory could come from attrition and brute military force, not politics, surrenders or rehabilitation. Governor Jagmohan who presided over Kashmir during the initial explosion of insurgent violence in 1990 stated in an interview that year, "The bullet.. .is the only solution for Kashmiris. Unless the militants are fully wiped out, normalcy cannot return to the Valley." 280 Such a strategy involved non-targeted house to-house searches with arbitrary arrests, strip-searches, and open fire on civilians breaking curfew. 28 1One scholar wrote, "Jagmohan saw the insurgency as a movement, abetted by Pakistan, which had to be brutally crushed, even if it meant targeting virtually the entire population." 282 The center doubled-down on this strategy even after recognizing that Jagmohan had thrown oil on the fire in 1990 and committed huge strategic blunder. Instead of replacing him with a more conciliatory choice, the GOI appointed an equally hard-nosed former RA&W Director Girish Saxena to manage the Kashmir campaign and "the bloodletting increased." 2 83 Five years in, the state was still trying to "wear down the insurgency" by winning militarily. 284 While unable to curb violence in Kashmir in the early years, state forces were happy "to impose a steady attrition on the militant groups and their leaders" 285 with special forces conducting offensive, deep interdiction missions. 2 86 After recognizing that mass imprisonment was 280 281 282 283 2 84 285 286 Behera, 2006: 48. Levy and Scott Clark, 2012: 42. Schofield, 2000: 150. Levy and Scott Clark, 2012: 43. Newberg, 1995: 74. Joshi, 1999, p. 248. Joshi, 1999, p. 406. 515 generating new waves of insurgents, the Army adopted a "catch and kill" policy where they reported civilian deaths as insurgent fatalities, increasing custodial deaths to dozens a month. 2 87 The Indian state routinely measured its progress in the counterinsurgency campaign in terms of exchange ratios and attrition rates to make sure it was regularly killing sufficient numbers of militants.2 88 Cash rewards were offered to regular and irregular forces to kill more militants creating perverse competitive incentives. 2 89 In early 1993, the state police chief declared 91% of the top militants had been arrested or killed and expected normalcy to return within six months. 29 0 Even after Hizbul Mujahideen admitted it had lost 10,000 cadres by the middle of 1994, violence would soon escalate again.2 9 1 The mentality that victory was always only just a few more kills away persisted for more than a decade. In 1998, Home Minister LK Advani reaffirmed this approach stating, "We are satisfied with the progress we are making. Daily eight to ten militants are being eliminated. The process of attrition is on ... There is no other solution but just to eliminate the terrorists." After successfully killing off militants for almost a decade and a half, around 2004 the strategy of kinetic attrition through countermilitancy shifted to one closer to attrition through exhaustion and indefinite military control. The strategy continued a political authority of usurpation involving a number of more subtle, sinister methods but involved little devolution of authority and few positive inducements. 2 92 2 87 Newberg, 1995: 48; AG Noorani, 2013, p. 517; Chindu Sreedharan, 1999. Swami, 2007: 180, 193 describes the Indian security forces' concern for this; MHA report; "More Troops Can Be Moved Out of Kashmir," PTI/IndianExpress, February 6, 2006. 289 Author Interview, Indian national security journalist, Delhi, India, Dec 17, 2013 (#147); Chatterji, 2011: 112. 2 90 Joshi, 1999: 225. 291 Joshi, 1999: 291. 292 Bose, 2003: 136; Staniland, 2013. 2S8 516 No Restraints The scale and extent of permissible violence in Kashmir was qualitatively different than that of Punjab because there were no restraints on state action or serious attempts to limit collateral damage or deliberate civilian victimization. From 1990-2002, the state security forces could pick up anyone and kill him with impunity.2 9 3 One Indian journalist writes, "The security forces were without a moral compass and on autopilot. ... their job was to defeat the militants and they were doing it the only way they knew how." 294 In Punjab, protection ofjoumalism was prioritized to ensure transparency and information flows. In Kashmir, during periods of curfew government officials told journalists "Ifyou go out of the hotel, you can be shot, and if we find out you have gone out, you can be imprisoned for six months." 29 While in Punjab, militants were sometimes the target of extra-judicial killings (usually in response to the killing of police officers' families), in Kashmir the government exercised "'official' vigilantism" even against journalists, lawyers, activists, and other civilians.29 6 Security forces "frequently did not discriminate between combatant and non-combatant," 297 sometimes as collateral damage but also to punish civilians with a retaliatory assault or send a message, "not dissimilar to those used by the Germans in Europe during World War II".2 9 8 In a rare bout of honesty, the BSF commander acknowledged to a researcher the government's inexorable resort to high violence was not a few soldiers gone rogue but the strategy directed by Delhi.29 9 293 29 4 Author Interview, Indian journalist, Srinagar, India, Dec 8, 2012 (#108). Joshi, 1999: 155. 295 Dilip Ganguly, "Troops Search House-to-House for Separatist Moslems," Associated Press, January 26, 1990; Confirmed with Author interview with former American journalist, Washington, DC, December 18, 2012 (#116). 29 6 Joshi, 1999: 414. Quote on p. 418-422. Joshi, 1999, p. 232. Joshi, 1999, pp. 229, 235. For instance a doctor was killed for providing information to an NGO on civilian casualties. 299 Newberg, field research, 1995: 30. 297 298 517 The state also never corrected its heavy reliance on violence or excess force in Kashmir. Despite decades of COIN experience, the Army treated this primarily as a low-intensity conventional war and did not offer a COIN capsule in the officer-training regimen until roughly eight years into the conflict. 30 0 While the RR proved to be an organizational innovation, this did not change its doctrine or the way in which Indian state employed force in Kashmir, relying on large scale rather than small unit operations. 3 0 1"Brutality, at times in its most egregiously repugnant forms, has been institutionalized by the security forces in Kashmir" and that "curbing these violations would take far more massive and sustained effort than the government.. .seemed willing or able to make." 302 For the first few years, the government had done "virtually nothing.. .to act against cases of 'excessive use of force. ' 3 03 Eventually the Army began to crackdown on human rights violations and improved discipline, but this stemmed from being less involved in day-to-day COIN operations or in confrontations with the Kashmiri militants and civilians. The BSF and CRPF did not improve and remained immune from any such formal corrections, 304 and some argued when the Army was forced to confront these groups, it too resorted to the same tactics. 3 05 The Indian Supreme Court has come to the defense of civilians affected by counterinsurgency campaigns against the Naxals and in Punjab, but it has never allotted the same oversight in Kashmir.3 06 Indian officers and policymakers proudly remind everyone the state never used airpower on the Kashmiri insurgency 3 07 but the chief reason had less to do with restraint and more to due Author Interview, retired Indian Army officer, Delhi, India, November 23, 2012 (#84). Rajagopalan, 2004, pp. 34-36. 30 2 Wirsing, 1994: 158. 303 Joshi 1999, p. 434. 304 Joshi, 1999, p. 449. 3 05 Wirsing, 1994: 159 306 Author Interviews, Kashmiri journalist, New Delhi, India, December 2012 (#114). 307 Author Interviews. Also, see former military intelligence officer RSN Singh, "Indian Army New Challenge," Canary Trap, November 10, 2011 (http://www.canarytrap.in/201 1/1 1/10/indian-army-the-new-challenge/). 300 301 518 with strategic prudence. The Indians feared that if they resorted to airpower for counter infiltration missions in the mountains of Kashmir, they might risk escalation in the tenuous Line of Control. More importantly, they feared external backers of the insurgency (i.e. Pakistan) would counter airpower against insurgents with anti-air capabilities (like the Mujahideen received in Afghanistan) that could deplete Indian helicopter gunships critical for potential conventional missions. 308 IndiscriminateViolence and Civilian Victimization The identity distance and disdain for Kashmiri Muslims that factored in to an attrition strategy resulted in incredibly high levels of civilian victimization. The high levels of civilian casualties most likely took place in the first five years of the conflict. Scholars estimated through police and hospital casualty records that by 1991, 12,000 civilians had been killed, and by 1995, around 20,000 civilians had been killed (while human rights organizations estimated this to be closer to 30-40,000).309 This is four to ten times higher than the government figures, which only counted civilians killed by militants as well as some caught in the crossfire.3 10 By 1994, the state admitted that at least 1500 civilians had been killed "caught in the crossfire," though this was often a euphemism for deliberate targeting of civilians. 31 ' Excessive force was the nomenclature used for arbitrary violence, sometimes out of panic in reaction to an ambush causing security Joshi, 1999: 398, 403 Newberg, 1995: 48; Jalil Andrabi, a Kashmiri Human rights lawyer, stated that as many as 40,000 had been killed by 1996. See Shubh Mathur, "Impunity in India," Guernica, February 1, 2013. Schofield, 2000: 183 also reflects these figures. 3Data from Swami (2007: 175) based on internal MHA reports shows 1564 civilians killed through 1991 and 5619 killed through 1995. Data on Civilians killed in crossfire from J&K govt shows 1063 killed through 1991 and 2228 killed through 1995 311 Schofield, 2000: 183; Levy and Scott Clark, 2012: 44. 308 309 519 forces to spray bullets in the vicinity, but often-times retaliatory violence against civilians, their homes, and their livelihoods.3 12 External outlets like the BBC and AP as well as independent analysts and human rights groups routinely reported civilian casualty figures dramatically higher than government figures.3 13 This may in part be explained by the ways in which the government was able to suppress information on violence through censoring data, re-categorizing identities, and dispersing violence. First, data on casualties was censored because orders were issued to the Kashmir police in 1992 not to register complaints against the armed forces or enter them into their logs (and without these first information reports, there was no way to track or account for those acts of violence). Even when reports filed by lower level officials called for investigations into military malpractice, these were often dismissed or whitewashed.31 5 Second, the military could routinely re-categorize identities and claim civilian actors were armed insurgents after-the-fact. In the early years of the campaign, one analyst noted data on insurgent casualties was routinely vague and contradictory with civilians included in their figures. 316 Twenty years later, more probing investigations reveal Kashmiri civilians were routinely abducted by the RR, "summarily executed" but then dumped at local police posts where they would be labeled "foreign militants." 317 Third, the state continued to kill civilians at high rates but make bodies "disappear" by dispersing smaller numbers of victims in rural areas where they Joshi, 1999: 113, 157; The CRPF burned down a marketplace in Doda that was 80% Muslim. Bose, 2003: 118 For instance, the government estimates were 45,000 while AP would regularly report 68,000. See "Troops in Fatal Avalanche," BBC, April 14, 2009 (http://news.bbc.co.uk/2/hi/southasia/7997989.stm); "India Protests Widen After Kashmir Deaths," Associated Press, August 13, 2008. 34 Author Interview, Kashmiri journalist, Srinagar, India, December 2012 (#107). 315 Habibullah, 2008: 79. 3 16 Newberg, 1995: 48 317 Levy and Scott Clark, 2012: 87. 312 313 520 ' would not be recorded.3 18 Villagers in rural areas report that security forces would often bring by corpses with instructions to bury them. 319 In addition to the 3,222 civilians killed in the crossfire by the Indian state and the thousands unaccounted for in fatality estimated by government data, an estimated that 10,000 or more Kashmiris were believed to be missing in two decades. In one unmarked grave in north Kashmir, 2730 bodies were unearthed, at least 574 who were civilians unconnected with the insurgency. In the Jammu districts of Rajouri and Poonch, another 4,000 unmarked graves including one with at least 2,500 corpses were dug up. The GOI also made extensive use of non-lethal violence. In 1993, the number of Kashmiris under preventative detention was roughly 10,000 or .2% of the valley's population.32 And another 35,000 or .8% of the population had been apprehended in the first decade. 322 In a two-stage cluster sample survey conducted in 2005 of-500 Kashmiris in two districts, high numbers of people reported being subject to various forms of violence including mistreatment (44%), forced labor (33%), arrest (17%), torture (13%) and sexual violence (12%). Meanwhile, 83% had been exposed to roundup raids, 67% witnessing torture, and 13% witnessing a rape among a host of other traumatic experiences resulting in 33% diagnosed with mental health problems. 323 Though torture was common to policing throughout India, one analyst conducting field research and interviews in the region in 1993 described it as "absolutely universal" in Kashmir Sreedharan, "Wronged Rights," Dec 16, 1999; Habibullah, 2008: 111. Levy and Scott Clark, 2012: 479-48 1. 320 Levy and Scott Clark, 2012: 479-481. 321 Wirsing, 1994: 157. Wirsing gives a range of figures but I take the one offered by a Kashmiri government official and take it as a percentage of the Valley's 4.5 million population. 322 Surinder Oberoi, "Fear and Loathing in Kahsmir," The Washington Quarterly,24 (2), Spring 2001, pp. 195-99. 323 Kaz de Jong et. al, "Conflict in the Indian Kashmir Valley II: Exposure to Violence," Conflict andHealth, 2 (10), 2008 (http://www.conflictandhealth.com/content/2/1/10) Kaz de Jong et. al, "Conflict in the Indian Kashmir Valley II: Psychosocial Impact," Conflict and Health, 2 (11), 2008 (http://www.conflictandhealth.com/content/2/1 /11). 318 319 521 and "more systematic and extreme." An estimated 60,000 have been tortured in Kashmir interrogation centers, 324 which one retired Indian officer described as "vast, 'beyond the imagination of the Indian public.'"3 2 5 Some accounts reveal the type of torture by Indian security forces' in Kashmir often focused less on collecting intelligence and more on humiliating and punishing individuals for participating in the rebel movement.3 26 Because of this, victims sometimes died in custody due to excessive torture and a fake encounter had to be concocted or the body disposed of, often by dumping it in the river. By 1997 almost 500 bodies had been found floating down the Jhelum river.3 27 After the purported strategy change in the second Kashmir campaign around 2004, the state typically relied on more subtle forms of coercion and violence but still regularly deployed overt violence against civilians. In 2008 and 2010, the Indian state returned to visible lethal force against unarmed civilians who had mobilized in large protests through the Kashmir Valley. More than forty civilians were killed in 2008 and over 100 civilians, mostly teenagers, in 2010. Unsurprisingly, counter-protests by Hindus in Jammu elicited tremendous restraint by the state, even when it came to making preventative arrests.3 28 3.2.2. Vulnerability When the state brought the hammer down early on the Kashmir insurgency, unlike in Punjab, the GOI faced no major backlash from within the state or its critical security institutions and no strategic vulnerabilities because of it. Kashmiri Muslims and other Indian Muslims who Chatterji, 2011: 109 Wirsing, 1994: 160 326 Basharat Peer, Curfewed Night, New York: Scribner, 2010, p. 138. Chapter 11, "Papa-2" describes this in depth. Punjab, by contrast, did involve torture by primarily for extraction of information. 327 Shubh Mathur, 2013; Joshi, 1999: 156 328 Amitabh Sinha, "PM to Tell All Parties Today: Jammu Versus Kashmir will Help Our Enemy," Indian Express, August 5, 2008; "Top Kashmir Separatists Detained," BBC News, September 5, 2008; Sheikh Mushtaq, "India to Free Protesters in Kashmir Peace Move," Reuters, September 30, 2010. 324 325 522 might have felt solidarity with the civilian victims in fact had little to no presence or voice in central institutions like the security forces, intelligence, or civil bureaucracy and therefore had no capacity to voice dissent or leverage the threat of exit to influence state strategy. When the state declared a generalized curfew across the valley and acknowledged killing more than 500 civilians in the first year (and likely many more), it paid no substantive price that compelled a strategic adjustment and continued down a path of offensive attrition. The only resistance came from the Muslim rank and file of the J&K police, who were deeply frustrated with and opposed to the state's harsh counterinsurgency approach, and at times cooperated with the militancy. The mostly Hindu military and paramilitary were in charge throughout the 1990s while the J&K police were hardly utilized, overruled, and even treated as suspects. In April of 1993, 3,000 state police staged a walkout and marched in a demonstration to protest federal troops killing a Muslim police officer but the government was unmoved and the police were only further marginalized in the campaign. 329Because the state did not rely on nor care for the Muslim contingent of the J&K police and Kashmiri Muslims played a virtually inconsequential within state institutions leave alone within the total security forces, the GOI could afford to ignore this in a way that it could not disregard the 1984 Sikh mutiny. Rather than feeling a strategic incentive for restraint, the state believed it had a strategic and tactical incentive to go on the offense to compel Kashmiri submission. One senior officer stated: "The Kashmiri tends to gravitate towards the stronger force. If a battalion operates aggressively, the Kashmiri invariably starts giving out information. But if a battalion is perceived as 'soft', the support for militants begins to rise. This attitude tends to attract tough conduct." 330 "Three Policemen, Two Militants Killed in Gun Battle," Associated Press, April 23, 1993. (http://www.apnewsarchive.com/1993/Three-Policemen-Two-Militants-Killed-in-Gun-Battle/id329 fc708896e8b77c377c4e89ab2d05f93b). 330 Sreedharan, "Wronged Rights," Dec. 16, 1999. 523 Figure 7.2: Growth of Police in Punjab and Kashmir 60,000 50,000 6 40,000 -I 30,000 I 20,000 1 2 3 4 Punjab (1983-1992) 5 6 7 " 'Kashmir 8 9 10 (1990-1999) (Sources: NCRB data) Table 7.1: Police and Local Forces in Puniab and Kashmir J&K 41,807 Punjab 57,771 Home Guard Special Police Officers & Village Defense Councils 0 26,544 28,000 10,000 Central Forces (Army & 400,000 190,000 Paramilitary) TotalForces % Police % Local Police 10 year Growth 468,351 8.9% 14.6% 46% 285,771 20.2% 33.5% 101% Police (Sources: NCRB data; SATP data based on a document submitted to the Jammu and Kashmir State Assembly as reported in Kashmir Times (Srinagar), March 14, 2001; Wirsing, 1994; Barry Bearak, "Kashmir a Crushed Jewel Caught in a Vise of Hatred," New York Times, August 12, 1999; Chima, 2007: 626, 630; Telford 2001; Joshi, 1993: 12-13; Mahadevan, 2008; Fair, 2009; Gill 1999.) Eventually the police would be carefully reconstituted with non-Muslim leadership and with the prevailing dominance of the Army and central paramilitary forces, they posed no risk of defection or strategic costs. Even when the police took on a more active role, the core component of this was the SOG, a predominantly non-Muslim unit within the police used for kinetic actions. 524 In essence, by excluding Muslims from critical institutions, the Indian state had completely insulated itself from Kashmiri or Muslim reactions to its high violence strategy. The threat of exit had no resonance when Kashmiri loyalty was already doubted and it had no consequence when Kashmiris were so dis-embedded from the state. 3.2.3. Information Without Kashmiris embedded in state institutions, the Indian state faced a scarcity of information, particularly early in the campaign due to a lack of personnel familiar with the Kashmiri people, language, behavior, and social organization. This shortage of intelligence and inability to identify, target, infiltrate, and manipulate the sinews of society as it had done in Punjab forced the state to rely on indiscriminate tactics and forces like unaccountable militia proxies that terrorized the population and escalated violence. Indistinguishable Stereotypes certainly stoked Indian perceptions of Kashmiris collective disloyalty that militated against selective targeting, but so too did simply bad information. This started with state leaders' cognitive biases and priors about Kashmiri Muslims and Muslims in general. One senior IB official admits his training to understand the subcontinent's Muslim political movements and their links to pan-Islamic forces was "shallow and abstract" and based on civilizational narratives of persistent conflict that bred hatred for Muslims. 33 1 For the Indian state, "a 'surgical' response was not feasible" because it could not ably distinguish actual insurgents from supporters given the popularity of the uprising. 332 331 332 MK Dhar, 2005: 87, 91. Bose, 2003: 112. 525 Consequently, all Kashmiris were treated as collectively guilty. One Indian paper wrote, "The face of the Kashmiri has dissolved into a blurred, featureless mask. He has become a secessionist-terrorist-fundamentalist traitor." 333 For the first seven years of the campaign and even beyond that under a period of de facto martial law, "becoming as suspect was as unavoidable as catching a cold"3 34 and substantial numbers of noncombatants including women and children were killed with little attempt at distinguishing them from militants. As described earlier, the GOI had a motivated bias to view Kashmiris as Pakistanis. 335 Furthermore, allies and enemies were identified in stark dichotomous terms, 33 6 rather than the multiple tiers of commitment identified in Punjab where B and C category militants could often be demobilized through non-lethal means like arrests, surrenders, and rehabilitation. Had deployed security forces been equipped with a degree of cultural sensitivity, language skills, and kinship ties to the regional population, one former police official speculates the response might have been more cautious and targeted.3 3 7 Limited Policing, Limited Intelligence At the onset of the insurgency, intelligence was quite poor as the central Army and paramilitary forces were comprised of "ethnic outsiders ill-equipped to develop channels to the local population." 338 India's decisions sidelining and collapsing the police cut off the state's eyes and ears and led to "a cessation of information flow." 339 This stemmed from distrust of the mostly Muslim local police force, deemed suspect and compromised, and "an undifferentiated 333 Bose, 2003: 113. Levy and Scott Clark, 2012, p. 133. 33 Author Interview, Kashmiri journalist, Srinagar, India, December, 2012 (#107). 336 Author Interview, Indian journalist, Srinagar, India, December 2012 (#105). 337 Author Interview, retired senior Kashmiri police officer, Srinagar, India, Dec 2012 (#111). 33 8 Wirsing, 1994: 144. 339 Joshi, 1999: 413. 34 526 collection of traitors."340 Hundreds of police were dismissed and thousands were disenfranchised. 34 1 Despite the apparent success of the Punjab police efforts in 1991-92, the GOI was not applying these lessons to Kashmir." 342 A few years in to the campaign, one analyst traveling through the region observed "little communication between the locals and security forces" and a population fearful and averse to engaging state forces. 343 The lack of information manifested itself in massive cordon-and-search operations conducted by the Army conducted operations with units in excess of a brigade. Described as a "sledgehammer to kill fly," the operations involved little intelligence, leading the state down a "bloodier, messier [route]." In 1992, the state launched an assassination program similar to the Phoenix program in Vietnam intending to wipe out militant leadership, but again, it lacked the requisite informant networks and frequently mistook its targets. 34 6 Even if security forces could use signals intelligence to identify an insurgent's name, they still embarked on a "wild goose chase" 347 without any information on his appearance, background, network, or residence. Tracking down militants became a complicated by "looking for a needle in the haystack without knowing what the needle looked like." 348 By 1993, the state claimed to introduce a more targeted intelligence-driven strategy but remained "plans on a piece of paper." 349 The signature tactic of an information-scarce Indian security force was the cordon-andsearch of a village in which all the inhabitants were forced to vacate their homes, assemble, and squat in a designated area at gunpoint for between four to six hours, sometimes through 340 341 Praveen Swami, "Tackling Terror," Frontline, 15 (1), May 9-22, 1998; Wirsing, 1994: 144. Joshi, 1999: 41-42, 116; 132, 224-227. 342 Praveen Swami, "Tackling Terror," Frontline, 15 (1), May 9-22, 1998. 343 Newberg, 1995: 28. 344 341 346 34 7 348 349 Joshi, 1999, pp. 132-133. Joshi 1999: 151. Joshi, 1999: 153. KV Krishna Rao, 2001: 442. Joshi, 1999, p. 133. Joshi, 1999, p. 229. 527 inclement weather or all night while the entire village was searched. The harassment, disruptiveness, fear, collective blame, and humiliation of these operations was evident to observers and even to Indian officers who conducted or presided over them." An IG of the J&K police had raised concerns about the alienating effects of this approach just days before a curfew was installed but these concerns were sidelined along with the police.351 Even when intelligence improved, it did not result in very selective violence as state force still used collective punishment and reprisals as a tool against whole villages suspected of harboring militants.35 2 When the eventually began to employ them, intelligence-driven operations did not have the same domino effect as in Punjab because groups the Indian state did not comprehend how the groups were more cellular and easily replaceable by a legion of willing Kashmir volunteers disaffected by the GOI's initial responses. 35 3 The state certainly worked to improve this information gap, primarily through mass arrests, interrogation, and torture, but developing credible and reliable human intelligence networks through revitalization of the police proved elusive for years if not a full decade to rebuild and modestly integrate the police into the security plan.354 As the lessons of Punjab began to sink in, the state slowly began to improve its police forces but its access to and use of information as well as its subsequent operations looked very different from Punjab. Proposals to rebuild the police in 1994 began to take shape by 1995 but primarily with non-Kashmiri Muslim personnel in new units like the special taskforce (STF) and 350 Joshi, 1999: 133; Levy and Scott Clark, 2012: 417; Wirsing, 1994: 155; Interview with Captain' KV Krishna Rao, 2001:442. 351 Sanjoy Hazarika, "A Vale of Tears and Anger: Anti-Indian Militancy is on the Rise in Kashmir," New York Times, January 14, 1990, p. 3. 352 Schofield, 2000: 158. 353 JoShi, 1999: 413. 35 Author Interview, Indian national security journalist, Delhi, India, December 2013 (#147). 528 special operations group (SOG). Real investment in the general police did not begin until 1997 or materialize until 1999 and they still remained bit players as one-tenth of the total counterinsurgency forces in the area of operation. 356 Even by 2001, a former governor admitted the police force still required much more reconstruction work.35 7 Both the STF and the SOG were actually composed primarily of non-Kashmiri, nonMuslim personnel (including Sikhs, Dogras, Gujjars) "to create the impression that the counterinsurgency effort had local support." 358 Even when Muslims were hired, most were brought in from other regions, not from the valley that was the heart of the insurgency. 359 Consequently, they still had far less access to information than the Punjab police who grew up in the same villages as their insurgent counterparts. Because the police were never in the lead during the campaign, information they generated was not necessarily trusted or acted upon by decision makers, and wide gaps existed between the main operating forces and the population.3 60 Furthermore, even when the state began to turn to the use of local forces, the reliance on surrendered militants turned into armed thugs "detracted from efforts to revive the J &K police." 361 Instead of a low violence strategy with more discriminate operations, the SOG partnership with renegades between 1996-2002 initiated a ruthless attrition campaign with any manner of atrocities.362 KV Krishna Rao, 2001: 480; Joshi 1999: 132, 428-33. Praveen Swami, "Tackling Terror," Frontline,15 (1), May 9-22, 1998; Total police forces remained 41,807 through 1999. This jumped in 2000 by about 50%. See NCRB data. 357 KV Krishna Rao, 2001: 546. 358 "Behind the Kashmir Conflict," Human Rights Watch Report, 1999 (http://www.hrw.org/reports/1999/kashmir/back.htm). 359 Joshi, 1999, pp. 429, 431. 360 Author Interview, retired senior Kashmiri police officer, Srinagar, India, Dec 2012 (#111). 361 Joshi, 1999, p. 428. 362 Author Interview, Indian journalist, Srinagar, India, Dec 8, 2012 (#108). 355 356 529 Local IrregularForcesandRenegades Even when the Indian state did managed to access better information through local sources and irregulars, it did not lead to lower levels of violence. The Indian state certainly brought in some Kashmiri Muslims as irregular forces, particularly those who were disaffected by Pakistan funding being funneled primarily to certain insurgent groups, but this was done to "outsource harsh measures" and ruthlessness while claiming "plausible deniability." 363 About five years into the campaign, the India state did manage to cultivate a useful local information source through surrendered insurgents enlisted to serve as special police officers (SPOs). These ikhwans (named after one of the Kashmiri militant organizations to defect en masse), more generally termed "renegades" coordinated closely with the SOG.36 4 Instead of using these SPOs to improve selective targeting, the state unleashed them to terrorize insurgents and civilians alike. Sometime around 1995 the renegades were empowered with off the books funds to serve as "India's premier attack dogs," 365 and were tasked with controlling and "sanitis[ing]" villages in remote regions. 3 66 Renegades also used the opportunity to establish unfettered power, settle scores, and terrorize locals. Those who surrendered but did not serve in irregular militias received little rehabilitation support.3 67 The renegades' actions cannot be dismissed as rogue actors because their actions were known to their government handlers, who encouraged indiscriminate violence with a "cash for corpses incentive scheme," allowing renegades unfettered opportunity to "shoot whosoever [they] choose." 368 Irregulars even targeted journalists and human rights activists. 3 69 These 363 Author Interview, Kashmiri journalist, Srinagar, India, Dec 7, 2012 (#107); Levy and Scott-Clark, 2012: 396; Swami, 2003. 364 Bose 2003: 133-134. 365 366 367 368 Levy and Scott Clark, 2012: 396. Baweja and Vinayek, 1996 Baweja and Vinayek, 1996 Levy and Scott Clark, 2012: 389-90. 530 irregulars answered only to the military and intelligence and were not locally accountable to their villages or local police, who were powerless to investigate them for their extortion, revenge killings, and sadistic crimes against civilians. 3 70 One notable renegade leader Basir Ahmad Wagay reportedly dismissed police attempts to rein him in saying, "I'm the fucking law now." 3 71 Today, members of the national security establishment concede the renegades were a tactical success but a strategic blunder, becoming "the most hated symbol of Indian rule" in Kashmir. 3 72 While they provided real value in helping destroy the armed capacity of some insurgent organizations between 1995-99 (principally the Hizbul Mujahideen, the strongest Islamist insurgent organization), their value diminished over time. As the renegades became liabilities with their routine violence and abuse of the population engendering tremendous alienation and resentment amongst Kashmiri Muslims, 373 the state had little use for them. Promised jobs and rehabilitation often never materialized, which may have motivated the turn to extortion and illicit economies. The GOI, perhaps eager to slough off the liability and bury secrets, eventually stood by and allowed many to be liquidated.3 74 A related tactic was the use of 18,000 strong village defense committees (VDCs) again to use local information for security but they predominantly operated in the Hindu-dominated Jammu region and were nearly all (96%) composed of Hindus and Sikhs, along with militant Hindu nationalist Bajarang Dal volunteers from the Indian mainland.3 75 These VDCs could 36 9 Noorani, 2013, P. 593. Levy and Scott Clark, 2012: 394-95; Baweja and Vinayek, "A Dangerous Liaison," 1996 371 Levy and Scott Clark, 2012: 387-89. 37 2 Ahmed, 2009: 77 quoting Pankaj Mishra. 373 Author Interview, Army Captain (Retired), New Delhi Nov, 23, 2012, (#84). Also echoed in Author Interview Advisor to Natl. Security Council, Nov 20, 2012 (#78). 374 Praveen Swami, "A Beleaguered Force," Frontline, 16 (3), Jan 30-Feb 12, 1999; Praveen Swami, "India's Forgotten Army," The Hindu, September 14, 2003 (http://www.hindu.com/thehindu/2003/09/14/stories/2003091406170800.htm). 375 Thomas A. Marks, "India: State Response to Insurgency in Jammu & Kashmir - The Jammu Case," Low Intensity Conflict & Law Enforcement, 12 (3), Autumn 2004, pp. 122-143; Barry Bearak, "Kashmir a Crushed Jewel 370 531 provide little information to either protect the majority Muslim civilians or help the state employ more discriminate violence. Fifteen years into the campaign, the Indian military finally began inducting local Kashmiri youth into Home and Hearth Battalions to provide some local security but this was quite limited in scale (beginning with only 3 battalions totaling 2700 recruits).37 6 The difference from turned insurgents and irregulars in Punjab is while they too were critical sources of information, they were used for more discriminate violence and for the protection of locals, while their utility increased rather than decreased over time with many of them protected, rehabilitated, or integrated into the regular police forces. 3.2.4. Trust The Indian state's contempt and distrust of the Kashmiri rebels' social base and Kashmiri distrust of a state that viewed them in such terms and lacked any credible representatives for outreach both ensured the scarcity of negotiations and other non-kinetic substitutes for counterinsurgent violence. While the Punjab campaign involved violence alongside negotiation and inducements, in Kashmir the Indian state neglected or failed to provide meaningful public goods, embrace surrenders and rehabilitation, or entice moderates back into the political fold. Distrust Any reciprocal interactions between the Indian state and Kashmiris were fraught with distrust inherent to the adversary's identity. At the tactical level, "In the eyes of the several Caught in a Vise of Hatred," New York Times, August 12, 1999. The VDC forces eventually reached 25,000. Riyaz Wani, "The Powder Keg of Jammu," Tehelka, 36 (10), September 7, 2013, (http://www.tehelka.com/the-powderkeg-of-jammu/); Joshi confirms this though estimates that in Doda in 1996, Muslims comprised 17% of the VDCs. See Joshi, 1999: 201; Jaffrelot, 2009: 213; Bose, 2003: 152. 376 "First Batch of Kashmiri Youth Joins Armed Forces," Rediff com (PTI), May 19, 2004. (http://www.rediff.com/election/2004/may/19kash.htm). 532 hundred thousand soldiers and paramilitary troops flooding the Valley, the whole population was suspect-not just disloyal to India but, much worse, in league with the enemy state across the LOC." 3 77 The feeling was mutual as forces deployed to the region were also broadly seen by the people as foreign occupying forces. 378 Even by 2001 after India had gained a firm grasp of the region militarily, negotiations were impossible. One retired Lt. Gen. wrote, "The Government of India does not trust Kashmiris and Kashmiris do not trust the Government of India.. .the bane of political or any dialogue or negotiations" and concluded that the prejudicial views of intelligence and bureaucrats will never bridge this gap. 3 79 It is difficult to expect different treatment when the primary counterinsurgency force of the paramilitary still harborss an "irrational dislike and distrust of the Muslims of Kashmir." 38 0 Distrust of Kashmiri leadership dated back prior to insurgency onset during Indira Gandhi's tenure as Prime Minister and continued to reproduce itself. Even when trouble was brewing in Kashmir during the late 1980s and the first ever Muslim was appointed Home Minister, distrust of Muslim leadership ran so deep that a governor like Jagmohan, known for his Hindu nationalist prejudice, was appointed as counterbalance the GOI's confrontational rather than conciliatory approach. 81 Jagmohan then sidelined every local institution and the police, whose loyalties he suspected, and instead relied on biased information from intelligence services predisposed to suspect all interactions with Muslims and reinforce distrust. His successor, former R&AW chief Saxena, also relied on intelligence and security forces rather than political Bose 2003: 113; This claim was reiterated by a Kashmiri professor. Author Interview, Kashmiri academic, Srinagar, India, Dec 8, 2012 (#110). 378 This point is made by both the military and Kashmiri side. Author interviews, Kashmiri academic, Srinagar, India, Dec 8, 2010 (#110); Author interview, retired senior Indian Army officer, Nov 10, 2012 (#68). 3 79 Nayar, 2013: 372-73; Confirmed by longtime Kashmiri Journalist Surinder Oberoi, 2001, p. 196. 377 380 381 Jafa, 2005: 157. Habibullah, 2008: 69-70; Schofield, 2000, p. 147. 533 initiatives as the only means to deal with the insurgency as did leaders for two decades in Kashmir. 3 82 Political dialogue could not gain any traction because Kashmiris did not trust the Indian state had any goal in mind other than to white wash the stain of their counterinsurgency campaign in the eyes of the international community. India's attempt at state elections in 1995 were dismissed by Kashmiri civilians as "political prattle as opposed to political initiative" and one JKLF leader, Shabir Shah, who had been recently released from prison refused to participate stating "We have no trust in Delhi" since they had continued to renege on agreements since 1953. Other leaders saw India's goal as merely to "create a government, a chief minister, an administration and then stop." 3 83 Limited Outreachto Muslims The Indian government was conscious of their Muslim problem in their counterinsurgency campaign, but even when there was interest to include them, Muslims were scarce in the Indian state's bureaucracy and security forces. Wajahit Habibullah, one of the few Kashmiri Muslims in the Indian civil service, was brought in to help as a divisional commissioner for short stints in 1990 and 1993 but his influence was limited by the dominance of the security forces. Gov Rao sought out one Muslim advisor S.M. Murshed, a retired IAS civil servant from West Bengal,3 84 but in his memoirs, Rao admits he had a very difficult time convincing the central government to fill his request for suitable Muslims officers. 38 ' Lt. Gen. Zaki, who initially commanded the 15 Corps in Kashmir through 1991 and then served as a Habibullah, 2008: 74-75. Schofield, 2000: 167 384 Joshi, 1999, p. 279. 385 KV Krishna Rao, 2001: 481. 382 383 534 military advisor to the governor occasionally provided a bridge to the Kashmiri Muslims but ultimately this was insufficient as "Muslims in senior command positions... [were] few and far between."386 Moreover, one professional consultant to the Indian army argued that this would have little influence because unlike Gill in Punjab, Zaki "had not been given real authority over anything." 3 87 His tussles with the center, threats to replace him, and the discounting of his advice led him to resign abruptly in 1995.388 Only in 2010 would a Muslim return to a prominent military position in Kashmir.3 89 Much was made of Rajesh Pilot, a Hindu cabinet minister of the Gujjar caste/ethnicity, who managed to flip some of the minority Gujjar Muslims to abandon the Valley-based rebellion and enroll in the BSF and central police forces around 1993.390 But this may have been less significant because the Gujjars, nomadic herdsmen living in the remote mountains, were nonKashmiri Muslims inactive with the insurgency except serving as paid mountain guides and only later as mercenaries. 39 1 Even if they had played a more serious role, this was nothing like Punjab recruiting Sikhs into the police. As a distinct minority group, Gujjars "nothing in common" with Kashmiris,3 92 their partial "defection" proved inconsequential, and deployment of Gujjar paramilitary recruits to Kashmir was deemed "extremely improbable." 393 Wirsing, 1994: 146. Wirsing, 1994: 178. 388 386 387 3 89 KV Krishna Rao, 2001: 514. Manu Pubby, "After Two Decades, Muslim Officer to Command Srinagar-based 15 Corps," Indian Express, October 20, 2010 (http://archive.indianexpress.com/news/after-two-decades-muslim-officer-to-commandsrinagarbased-1 5-corps/699847/). 39 0 Joshi, 1999, p. 278. Joshi, 1999, p. 427; Levy and Scott Clark, 2012: 48; Bose, 2003, p. 118; Sumantra Bose, "Session 3: Geography, Politics and the Fighters of Kashmir: An Eyewitness Account" The Conflict in Kashmir," LSE Seminar Notes (http://fathom.lse.ac.uk/seminars/10701013/index.html), Accessed July 21, 2014. 391 392 Schofield, 2000: 185. 393 Wirsing, 1994: 145. 535 No Public Goods Though stark poverty was never a major problem in Kashmir as it has been in other parts of the country, the state paid little attention to delivering public goods. Claiming the root of discontent was economic deprivation rather than political resentment at the history of "cynical authoritarianism," 394 the state proposed development initiatives that never actually materialized3 95 even as their irregular forces destroyed public goods like Kashmiri schools.3 96 Following the 1996 elections to restore normalcy, the state neglected crucial needs like job provision as well as economic revitalization, reconstruction, compensation for victims, or rehabilitation of surrendered militants. 3 97 Essential amenities were still not being provided by the state like water, electricity, schools and hospitals, which continued to elicit public anger.398 Money spent on social services has not actually yielded resulted any results in education, health care, infrastructure, or encouraging integration into the national economy, to say nothing of the more salient political demands of autonomy. 39 9 In the second phase of the campaign (Kashmir II) starting around 2004, the Indian state rhetorically embraced public goods through a campaign of winning hearts and minds, offering some goods like school construction accompanied by a much bigger public relations campaign. Ironically, the Army's acronym for this is WHAM. Thus far, security efforts have concentrated on Hindu families rather than the broader population and the new strategy has still been criticized by Army officers for being "on paper" and cynical "window dressing," Bose 2003: 98-99. Newberg, 1995: 5. Schofield, 2000: 182. 397 Oberoi, 1997; KV Krishna Rao, 2001: 546. 398 394 395 396 Nayar, 2013: 373. 399 Nayar 2013: 384-385. 400 Author Interview, retired Indian Army officers, Nov 20 and 23, 2012, (#79, 84). 536 400 while described by Kashmiris as "generosity at gunpoint" 40 ' and a "substitute for a well thought out political strategy." 402 Even proponents who describe its merits admit the approach makes no efforts to account for local input or measure the utility or impact on public opinion. 403 The WHAM efforts are still not central to the campaign and though money has been pumped into the state from the center, corruption and misgovernance in Kashmir is tolerated or even encouraged by Delhi, while little has been done to stem rising unemployment, especially amongst educated youth.4 4 One U.S. ambassador cabled the US state department alleging "Kashmir politics is as filthy as Dal Lake." He argue that India (as well as Pakistan) had managed to soak the state in corruption and made all Kashmiri political leaders "dependent on handouts." Despite "rivers of money" given to the state in the period India alleged to have focused on winning hearts and minds, the ambassador observed a near total absence of meaningful development.405 Cynical Surrender/RehabilitationPolicy As detailed in a previous section, surrender policies in Kashmir were extremely cynical at best and limited at worst, producing more mayhem and destruction but little rehabilitation. Instead of using surrenders to reduce violence, the Indian intelligence agencies willingly participated in "encouraging dogfights and jealousies... and fomenting betrayals" believing fragmentation of the insurgency would serve its strategic purpose en route to wiping out all insurgents, even if violence escalated.406 Only 4,000 militants in Kashmir were officially Author Author 403 Author 404 Author Interviews, Delhi and Srinagar, India, Nov-Dec, 2012 (#94, 105, 107, 112). Interview, Kashmiri academic, Srinagar, India December 7, 2012 (#106). interview, Indian think-tank scholar, Delhi, India, November 27, 2012 (#94). Interviews, Indian journalist, Srinagar, India, Dec 8, 2012 (#108); retired senior Kashmiri police officer, Srinagar, India, Dec 2012 (#111); retired Indian Army officer, Nov 23, 2012 (#84). 405 "Kashmir is a Money Game: David Mulford," Economic Times, September 5, 2011. 406 Levy and Scott Clark, 2012: 132, 413. 401 402 537 surrendered,407 most who were used as mercenaries or renegades (~3,000 to 5,000 between 19942001)408 to ruthlessly target their former colleagues and settle scores with civilians. Rather than serving as a useful bridge between the state and the insurgency, eventually the renegades were wiped out which served the Indian state just fine. Even if not all interactions with militants ended up like the renegades, the Indian state maintained a bad track record with surrenders in Kashmir. Occasionally state security forces were known to opt for the direct route of deliberately eliminate militants attempting to surrender by claiming it was in a firefight.409 When amnesty was granted in the release of a JKLF leader in 1992, he was soon reported killed in a cordon and search operation. 410 Few of the 20,000-30,000 former militants were offered meaningful rehabilitation and were frequently harassed, detained, and forced to work as informers for police. 41 1 Two decades after the insurgency began, "instances of complete rehabilitation are hard to find,"4 12 and unlike Punjab, in Kashmir they "remained on records only."A1 3 "Over 4,000 militants surrendered in J&K since 1990: Omar," Zee News, March 26, 2013 (http://zeenews.india.com/news/jammu-and-kashmir/over-4-000-militants-surrendered-in-jandk-since- 1990omar_838002.html#) 401 M L Kak, "Govt Withdraws Support to Renegade Militants" Tribune News Service (Chandigarh), February 14, 2006 (http://www.tribuneindia.com/2006/20060214/j&k.htm); "India: Country Reports on Human Rights Practices 2001" Bureau of Democracy, Human Rights, and Labour, US State Department, March 4, 2002 (http://www.kashmirlibrary.org/kashmirtimeline/kashmir-files/dos_2002.htm); Ritu Sarin, "J&K's Friendly Ultras Say Pay More, Or Else..." Indian Express, May 5, 1999 (http://www.kashmirlibrary.org/kashmirtimeline/kashmirfiles/renegades.htm); Amnesty International, "Recent Political Developments...," India: "Ifthey are Dead, Tell Us" - "Disappearances"in Jammu and Kashmir, London: Amnesty International, February, 1999, pp. 5-9 (ASA 20/002/1999) (http://www.amnesty.org/en/library/asset/ASA20/002/1999/en/1 53ef5a8-e34c-1 1 dd-a06d 790733721318/asa200021999en.html). 409 Oberoi, 1997; Author Interview, retired senior Indian police officer, Nov 27, 2012 (#92). 41 1 Joshi, 1999: 159. 411 Peerzada Ashiq, "Yearning for a Normal Life, Sordid Tale of Kashmir's Former Militants," Hindustan Times, July 16, 2013 (http://www.hindustantimes.com/india-news/yearning-for-normal-life-sordid-tale-of-kashmir-sformer-militants/article1 -1093286.aspx). 412 Binoo Joshi, "Surrenders Militants are Kashmir's Unwanted Lot," Indo-Asian News Service, April 27, 2010 (http://www.thaindian.com/newsportal/india-news/surrendered-militants-are-kashmirs-unwantedlot_100354421.html) 4 "Former Kashmir Militants," Kashmir Global, March 27, 2013 (http://www.kashmirglobal.com/2013/03/27/foriner-kashmir-militants-some-narrate-their-ordeal-others-cry-for407 538 Absent Two- Way Politicsand Missed Opportunities Though the Indian state has taken the threat in Kashmir very seriously and addressed it through military force, Delhi never seriously engaged in any concessions on autonomy, redress of grievances, or negotiations instead always postponing them until after elections while striving to keep the Hindu nationalist right at bay. Accommodations have been resisted or failed "because no Indian wants to see softness towards [Kashmiris]."414 As a result, there are a litany of decision points that reveal missed opportunities. Since 1990, political efforts have always been trumped by military considerations dictated by governors of military backgrounds and the Army, particularly after the one senior Muslim officer, Lt Gen Zaki, had retired.4 " Even its political efforts to restore democracy were "miniscule in comparison with its military efforts," limiting their efficacy so long as the military component of intensified resentment.4 16 Democratic principles were abrogated in Kashmir and "the military was used as a weapon to close politics, not to guarantee its opening."41 Rather than a violent negotiation, the state believed that after crushing the insurgency it could dictate the outcome and "craft its relationship to Srinagar on its own terms." 418 Multiple critical decision points emerged where opportunities for politics presented themselves but the state and Army opted to exclusively rely on kinetic operations. Early on, the state was responsible for the killing of moderate voices of non-violent separatism. 4 19 As the nationalist JKLF was moderating and beginning to embrace politics and renounce violence, the help-dumped-in-the-name-of-rehab-policy.html); Mehboob Jeelani, "The Departed: The Return Home of Kashmir's Disillusioned Militants," Caravan, September 1, 2012 (http://www.caravanmagazine.in/reportage/departed) 414 Author Interview, retired senior Kashmiri police officer, Srinagar, India, Dec 2012 (#111). 415 4 16 Joshi, 1999, pp. 286-87. Wirsing, 1994: 184-85. Newberg, 1995: 10. 418 Newberg, 1995: 3. 419 Habibullah, 2008, pp. 81-82. This Indian civil servant argues the state killed Dr. Abdul Ahad Guru, and then covered it up by killing the militant they had made a deal with. 4 17 539 Army relied on HM informants to wipe out the JKLF.4 20 In 1994, the JKLF and HM had been badly battered by BSF and Army forces offering the state an opportunity to propose a political solution. The ministry of external affairs offered recommendations to build good will and trust with Kashmiris including the loosening restrictions on the ground but the military Governor vetoed these. 4 21 Furthermore, the state not only doubled down on permanent military forces with the deployment of the Rashtriya Rifles but also chose to escalate to crush the separatists by unleashing the Ikhwans and other renegade groups, eventually opening the door to foreign fighters. 42 2 In 1996, the central government brought violence down to a level where it believed it could pursue democratic elections. However, it refused to entertain any autonomy concessions to the National Conference and the Home Ministry appeared to be working to set up another puppet government and sideline the National Conference entirely. The state again refused to curtail the activity of renegade militants feigning a lack of control. 2 Even after the 1996 elections were declared a success, the "elasticity of decision" was "sharply restricted" by a bureaucracy and military bereft of Kashmiri or Muslim representation, "which [held] the key to any plan," according to one parliamentarian, "[had] a stranglehold on policy," and was deeply opposed to serious negotiations and concessions on autonomy.4 2 4 Around 2000, the Indian state leadership convened intense deliberations after the end of the 1999 Kargil War and considered the next opportunity that would go missed. 420 Bose 2003: 130. 421 KV Krishna Rao, 2001: 476. 422 Author Interview, Indian journalist, Srinagar, Kashmir, December 2012 (#105); A.G. Noorani, "Jammu and Kashmir: Contours of Militancy," Frontline, 17 (20), September 30-October 13, 2000. 423 KV Krishna Rao, 2001: 516. 424 Wirsing, 1994: 171-72. 540 The state contemplated significant political gestures by reducing forces and reducing the degree of harassment of the population, the degree of which far outstripped the actual levels of violence, but again this opportunity was never seized.42 In 2000, HM commander Abdul Majid Dar declared a ceasefire and sought to find a negotiated settlement with the government and organized support from HM field commanders and the political leadership of the Hurriyat Conference. However, spoiler violence and India's unwillingness to countenance concessions on territory, inclusion of Pakistan in discussions, or reduced military activity demonstrated they were not seen as sincere in their efforts.426 In subsequent negotiation attempts in 2001 the state remained unwilling to engage the Islamist insurgents. 427 Eventually Dar became a target due to his outreach to the Indian state and after repeated failures, he was assassinated. Some suspect this blow to HM leadership was the objective of bad-faith negotiations all along.42 8 Since 2004 when the Pakistani state began to reduce support for Kashmiri insurgents, foreign infiltration, and cross-border artillery fire allowing the Indian state to build a robust fence across the LOC, the Indian state has continued to foreclose on political reforms that would meet some local aspirations.4 2 9 Some former military acknowledge there has never been a genuine effort at redressing Kashmiri alienation or arriving at some meaningful accommodation, such as refraining from political manipulation. 3 While the central government continuously experimented with devolution and autonomy in Punjab, it opposed even a "quantum of autonomy" as good faith to win popular support for negotiations kowtowing to an increasingly Author Interview, Indian journalist, Srinagar, Kashmir, December 2012 (#105). Praveen Swami, 2007: 188-93. 427 "The Rediff Interview Abdul Majid Dar, Chief Commander of the Hizbul Mujahideen," Rediffcom, April 7, 2001 (http://www.rediff.com/news/2001/apr/07inter.htm). 428 Author Interview, Kashmiri journalist, Srinagar, India, Dec 8, 20102 (#109) 429 Staniland, 2013: 938; Cohen and Dasgupta, Shootingfor a Century, 2013, pp. 50-52. 430 Author Interviews, Srinagar and Delhi, Nov-Dec, 2012 (#84, 106, 114). 425 426 541 powerful Hindu chauvinist political force. 4 3 1 Empowering local political control like the state did with the Akali Dal in Punjab has never been allowed in Kashmir.43 2 The state has granted Buddhists in Ladakh autonomy measures but not Muslims of the valley, and the state has never truly been willing to devolve power to local government.4 3 3 The government has continued to repeat the "farcical exercise" of soliciting political recommendations by civilian committees, sometimes government-appointed, that remain ignored, buried, or unimplemented. These are designed to avoid tackling issues ofjustice and the drivers of conflict including the rigging of elections, oppressive force presence, and political high-handedness in New Delhi.43 4 Even today when militancy no longer challenges the governing structure, the degree of militarization remains and continues to manipulate the political system. It seems to suggest the Indian state "does not believe that its possible to retain these people without this scale of military force," either due to fragility of their victory or the profound disloyalty.4 35 3.3. Alternative Explanations Few serious scholars or journalists dispute the above characterization of the Kashmir counterinsurgency campaign, but an alternative explanation in the South Asia literature alleges Kashmir was more a foreign-backed proxy war. This companion to Punjab as the "easier" insurgency casts Kashmir as the "hard" insurgency or even sponsored terrorism with far more Joshi, 1999: 160. Author Interview, Indian journalist, Delhi, India, December 11, 2012 (#114). 43 Author Interview, Kashmiri academics, Srinagar, India, Dec, 2012 (#106, 110); Author Interview, retired Indian Army officer, New Delhi, India, Nov 20, 2012 (#79). 434 Amitabh Mattoo and Happymon Jacob, "Farce in Three Parts," The Hindu, June 4, 2012 (http://www.thehindu.com/opinion/lead/farce-in-three-parts/article3487152.ece). 43 Author Interview, Kashmiri journalist, Srinagar, India, Dec 8, 2010 (#109). 41 432 542 external backing and led primarily by foreign fighters. 4 36 This explanation simply does not comport with the evidence and data on the insurgency and to the extent that it only reflects the Indian state's and military's perception of the insurgency as foreign serves to validate the coreperiphery theory's hypotheses on identity. While foreign dominance did not occur at least until more than a decade into the insurgency if not later, the military believed this to be the case by about 1993 if not sooner, 43 7 revealing more about perceptions of Kashmiris and motivated bias rather than of the actual material effect of foreign fighters. As one scholar who skillfully dismisses this argument has noted, this explanation manages to lay responsibility of the insurgency at the doorstep of Pakistan while exonerating the Indian state of decades of mismanagement and neglect as well as violent overreactions at the start of the insurgency. 4 3 8 The Indian state certainly has had a motivated bias in identifying the insurgency as foreign to acquire international legitimacy for its actions while discrediting its rival of Pakistan. This is why the state has even gone to such lengths as secretly indulging if not directly supporting the abduction and killing of foreign hostages 1995.439 A closer look at the data reveals foreign dominance of the insurgency is both exaggerated and inaccurate, certainly through 2002.440 Additionally, from 1990 through mid-1993, exfiltration figures closely matched infiltration numbers revealing that local Kashmiri Muslims 4 36 An alternative variant of the "Kashmir was harder" argument contends there was far more popular support for insurgency, but this neglects how highly endogenous this support is to the prior identity relations with the state as well how endogenous collective action is to state repressive measures. See Kalyvas and Kocher, 2007. 437 Bose, 2003: 149, 156-58, 161 says until 1999, but military believed this was the case by 1993. Author interview with retired Indian Army officer, New Delhi, Nov 6, 2012 (#65). 438 Wirsing, 1994: 114-15. 439 Levy and Scott Clark, 2012: 422, 428-3 1. 44 0 Praveen Swami, 2007: 194. 543 on the Indian side were travelling to Pakistan for arms and training.44 ' MHA reports also show foreign fighter fatalities do not surpass local insurgent fatalities until 1999.442 Inferring foreign dominance from casualties is a problematic barometer because foreign militants were more ideologically driven and daring, and therefore more likely to get themselves into dangerous firefights or be targeted by the Indian state resulting in a higher number of deaths. Moreover, security forces claiming killed civilians were foreign militants would routinely inflate these numbers.4 4 3 To the extent there has been an actual shift to foreign fighters, it is more a result of the Indian state's attrition strategy than a cause of it as a sequential analysis shows Indian forces degraded local Kashmiri militant groups so quickly that a vacuum emerged to be filled by foreign recruits.444 Nevertheless, local militant organizations like HM that are 95% Kashmiri still compose a large section of the insurgency. 445 Though they may supply the majority of high profile violence today, militant organizations based out of Pakistan like LeT and JeM (and its earlier instantiation HuA) in fact rely on a local logistical and popular support-as revealed by the massive Kashmiri protests against the state in 2008 and 2010-and increasingly on local Kashmiri recruits in recent years.446 Furthermore, even the Islamic identity of Kashmiris was not something imposed by Pakistani militants during the insurgency as budding communalism was Wirsing, 1994: 153 reports that Exfiltration amounted to 18,750 and 16,900 infiltrated while 1,699 were killed (836) or arrested (863). 441 442 Ministry of Home Affairs data. Levy and Scott Clark, 2012: 87. BSF Director General EN Rammohan is quoted on this. See A.G. Noorani, "Jammu and Kashmir: Contours of Militancy," Frontline, 17 (20), September 30-October 13, 2000. The point is also made by Lt. Gen. Nayar, 2013: 372. 445 HM is estimated at 1500 cadres which is at least half the insurgent size estimated at about 3,000-3,500. See C. Christine Fair, "Insights from a Database of Lashkar-e-Taiba and Hizb-ul-Mujahideen Militants," Journalof StrategicStudies, 2013, p. 7. 446 Levy and Scott-Clark, 2012; BSF Director General EN Rammohan is quoted on this. See A.G. Noorani, "Jammu and Kashmir: Contours of Militancy," Frontline, 17 (20), September 30-October 13, 2000; Bose, 2003: 149, 156-58, 161; Author Interview, Kashmiri academic, Srinagar, Dec 8, 2012 (#110) and a Kashmiri journalist, Delhi, Nov 28, 2012, (#96). 443 444 544 growing within the state since the 1970s, best evidenced by the electoral politics. 447 To the extent that the Indian state believed that there was a higher foreign presence or that Kashmiris loyalties were equivalent to Pakistani militants, this confirms my characterization of Kashmiris as socially distant from the state, which confirms with my theory and mechanisms. IV. The Naxalite (III) Insurgency: 2003 - Present 4.1. Background Ideologically-based movements mobilizing lower and marginalized social groups--tribals and lower castes (SC/ST), have repeatedly clashed with the Indian state for decades. Generally labeled "Naxals" and presently led by a Maoist organization, the CPI-M, this resistance movement has been punctuated by periods of intense rebellion in 1948-51, 1967-71, 1987-2003, and 2004-present. The present Naxal rebellion led by the CPI-M (what I have termed Naxal III) is a descendent of a number of failed political movements-cum-rebellions such as the Communist Party of India's (CPI) rebellion in Telangana (1948), the CPI-Marxist-Leninist rebellion from 1967-71, and their regional remnants that regrouped in the 1980s-90s such as the People's War Group (PWG) of Andhra Pradesh and Maoist Communist Centre (MCC) of Bihar. The Indian state keenly observed Naxal growth and violence but largely chose to ignore it for at least a decade. India's Home Ministry had been collecting and publishing detailed data on Naxal activities since 1997 (and chronicling the activities of "Left Wing Extremists" long before) and the Intelligence Bureau began sounding the alarm on Naxal links to the ISI as early as 1998.448 Meanwhile, the Inidan state tracked the rapid expansion of Naxal recruitment almost 441 Swami, 2007: 122-27, 156. 448 "Home Secretary: ISI Aiding Naxalites," Tribune News Service (Chandigarh), March 7, 2000; T. Sunil Reddy, "PWG Links with Kashmir Ultras Bared," IndianExpress, September 29, 1998. 545 doubling in size between 2000-2004441 and violent activity rising by 66%.450 The GOI took whene two largest Naxal organizations, the PWG and the MCC, merged to form the Communist Party of India-Maoist (CPI-M) in 2004. Finally in 2006, Prime Minister Manmohan Singh identified Naxalism as India's "single biggest internal-security challenge ever faced by our country" threatening one-third of India's districts, and began mobilizing resources to address it.45 1 Despite the state's awareness and promises of action, many analysts and practitioners derided government efforts, describing it as "vacillation and ad hocism," "timid and weak," inchoate, fitful, and indifferent. 5 Unlike nearly every other campaign, the Indian state chose not to send in the military or even a coordinated force to take over in affected regions. Instead, it dispatched a limited number of federal paramilitary forces in discrete, company or battalionsized units to different districts and states for them to be utilized (or not) as each state government saw fit. As a result, state police forces provided the main line of defense despite being on average 20% under strength, far less capable, and often misused.4 5 3 Without the organizational overhauls and military backing the GOI provided in Punjab and Kashmir, both state and federal forces remained inadequately trained, equipped, or reinforced enough to contend with the Naxals. 4 54 Ajit Doval "Growth of Maoists Fuelled by Politicians," Ajit Doval's Perspective (blog), September 10, 2008 (http://ajitdoval.blogspot.com/2008/09/growth-of-maoists-fuelled-by.html); Anuj Chopra, "India's Failing Counterinsurgency Campaign," ForeignPolicy, May 14, 2010 (http://www.foreignpolicy.com/articles/2010/05/14/india_s_failingcounterinsurgencycampaign). 450 My estimate of non-Andhra incidents shows they rise from 754 in 2000 to 1223 in 2004. 451 "Ending the Red Terror," Economist, Feb 25, 2010; "Chidambaram Inaugurates DGPs/IGPs Conference," Ministry of Home Affairs, Sept 14, 2009 (www.pib.nic.in/release/release.asp?relid=52610). 452 Haque and Nath, 2010; Sahni, 2012; Sahni, 2014; Marwah, "Naxal Gambit State Response," 2009. 45 Mohan, "Over One-fifth of the Country's Police Force...," 2012 454 Bibhu Prasad Routray, "India's Response to Maoist Growth,: Tactics and Strategies," Naxal Violence: The Threat Within, Ed. Gurmeet Kanwal and Dhruv C. Katoch, New Delhi: CLAWS, 2010: 180; G.D. Bakshi, "Envisaged Roles for the Army," Naxal Violence: The Threat Within, Ed. Gurmeet Kanwal and Dhruv C. Katoch, 449 New Delhi: CLAWS, 2010, pp. 219-221; Ajai Sahni, "Maoists: Surviving Adversity," 2014. 546 The central government's hands-off approach provided little coordination between these disparate federal forces and local police forces in the same locale, let alone across state lines. Naxal violence escalated despite of federal force presence because, like a balloon, they could relocate across the border to another state while continuing to recruit, tax, and conduct ambushes and raids across state lines. Years after the Naxals began to pose a serious threat, the government finally attempted to coordinate federal security forces and operations synchronized with the states under the name Operation Green Hunt in 2009. Even as the central government proudly announced a unified command in 2010, experts quickly dismissed it as a waste of time45 5 and the effort soon died out after a year while violence continued to climb dramatically to at least 2,200 incidents and 1,200 fatalities in 2010 alone. The Indian central government initiated a laundry list of measures to support the states in their efforts but slacked on implementation. Funds for security-related expenditures, infrastructure, road construction, police station fortification, and development proved to be tiny compared to past campaigns, inconsistently funded, significantly depleted by corruption and leakage to insurgents, and lacking a human development component. Well-funded, nationwide poverty alleviation schemes were also instituted, but they were not specifically targeted for Naxal-affected districts and also suffered from implementation and leakage problems. 4.2. Explaining Effort in Anti-Naxal Campaigns as a Function of Territory The Naxal III campaign had all the evidence of low-effort. Despite rhetoric of a holistic, developmental or two-pronged approach, the state reluctantly expended effort in the Naxal III campaign, merely trying to contain insurgency below a certain threshold rather than defeat it to 4 11 Vibhuti Agarwal, "Experts Slam 'Unified Command' Naxal Plan," Wall Street Journal:IndiaReal Time, July 15, 2010 (http://blogs.wsj.com/indiarealtime/2010/07/15/experts-slam-%E2%80%98unifiedcommand%E2%80%99-naxal-plan/). 547 regain control of the contested zones. Substantial evidence suggests Indian state's mitigation strategy against the ongoing Naxal (III) rebellion derived from active assessments and deliberate considerations of the incentive-cost structure. Difficult physical and human terrain raised the costs of any campaign for control, while such control and outright victory.yielded few incentives. 4.2.1. Costs The physical terrain of the Naxal-affected regions has imposed major costs on Indian state power projection as insurgents haven chosen to operate specifically in the mountainous, forested, and hilly areas of the Central Indian states. 456 One Indian government supported study detailed the increased difficulty of penetrating the current Naxal strongholds in regions almost three times more densely forested than those contested in previous Naxal rebellions.457 The population in many villages of the Naxal belt reside in inaccessible areas, located far off the road, sometimes separated by natural barriers like deep ravines. One scholar traveling through the region discovered that in many of the Naxal areas, travelling on horseback was the fastest mode of communication and trekking the only mode of commuting.458 Accessibility issues have often precluded the use of vehicles forcing security personnel to move on foot.459 One party official of the ruling government publicly advised against paramilitary operations in the region "as the terrain is so difficult." 460 A former intelligence director had described at length the difficulty of sustained operations in the Central Indian terrain: 45 4 6 57 4 58 Kennedy and Purushotham, 2012, p. 852. DM Mitra, 2011, p. 5. P.V. Ramana, "Naxal menace: Security Forces Challenged in Difficult Terrain" The Tribune (Chandigarh), July 10, 2006 (http://orfonline.org/cms/sites/orfonline/Modules/analysis/AnalysisDetail.html?cmaid=2562&mmacmaid=820). 459 "3 CRPF Jawans Killed in Two Naxal Attacks," Times ofIndia, April 10, 2014. 460 "Digvijay Latest: Why Use Paramilitary Against Naxalites," Indian Express, May 15, 2010. 548 Anti-insurgency work.. .involved long marches over hills and jungles, often no food and water for many hours and exposure to danger from both the flanks. They had to stay overnight under the sky in the jungles and keep a continuous vigil against guerilla attacks.4 1 The human terrain added to these access costs in tribal regions "where the state's ability to project its disciplinary power remained limited" because the populations were never politically or socio-economically dominated by centralized state-governed societies of the plains."4 62 One retired Army Lt. General cautioned that below average human development, high poverty, mortality, illiteracy rates, and poor access to medical facilities made even benign operations to upgrade the area extremely difficult.4 63 In describing the persistence of limited state reach on to these populations, the Home Minister explained, "People cannot read [for lack of schools]. People cannot communicate [for lack of telephones]. Nothing can move in or out [for want of roads] .464 The history of state exploitation combined with limited service provision bred distrust and perceieved illegitimacy of the state's authority. 4 65 A Ministry of Tribal Affairs report acknowledged that the tribals of Central India "view the State as their exploiter and enemy,"4 66 and the Home Minister stated this "trust deficit" challengeed any good intensions of the state. 467 Even after reportedly successful operations in Andhra Pradesh that pushed insurgents out to neighboring states, one major survey reported people in the newly Naxal-free did not view the state as a legitimate authority.46 8 461 Mullik, 1972, p. 276. Kennedy and Purushotham, 2012, p. 852. V.K. Ahluwalia (Lt. Gen.), "Future of the Naxalite Movement," Presentation to the Institute for Defense Studies and Analysis, New Delhi, India, June 5, 2014 (http://idsa.in/event/FutureoftheNaxaliteMovement.htnl). 464 K Balchand, "No End to Naxal Problem Without Winning People's Trust," The Hindu, May 13, 2010. 462 463 465 Ministry of Tribal Affairs, NationalTribalPolicy: A Policyfor the Scheduled Tribes of India,New Delhi: Government of India. 2006, p. 15. 467 "People Distrust Government in Naxalite-Hit Areas: Chidambaram," Times ofIndia, May 13, 2010 (http://timesofindia.indiatimes.com/india/People-distrust-govt-in-Naxalite-hit-areasChidambaram/articleshow/5922935.cms). 468 "In Naxal Land Govt Faces Trust Deficit," Times ofIndia, September 28, 2010 466 549 4.2.2. Incentives While posing high costs to physical control and extertions of state authority and disciplinary power, the Naxal belt has offered little material, strategic, or ideational incentive for such a high-effort campaign. The Mughal, Maratha, and British empires largely left the tribal regions alone due to their limited material value as a predominantly slash and bum or subsistence economy that did not produce an extractable agricultural surplus. One retired officer explained the logic of these empires: "There's no point in trying to consolidate our hold, extend infrastructural and administrative penetration into tribal areas because of the simple fact that there is nothing you can get out of this. So all three empires left it in splendid isolation." 469 Consequently, the modem Indian state still does not take the larger Naxal insurgency seriously and instead "treat[s] it like some kind of developmental joke." Additionally, the region posed no strategic value as it was not critical for self-defense nor did counterinsurgency forces deployed against a "heartland" insurgency offer dual use value to deter foreign enemies as they did in some "rimland" insurgencies. 470 The Naxal belt also offered little ideational value. A former Home Minister explained that while part of Indian civil society is sympathetic to the Naxals, the other camp, "to put it uncharitably is unconcerned." 47 1Public opinion surveys confirmed this assessment. Barely half the country (54%) even knew about the Naxal insurgency at its peak around 2009-2010, and less than half of those surveyed specifically in Naxal-affected districts expressed concm. 4 72 In the same year, another survey question assessing perceptions of India's greatest threat found three Author Interview, retired Senior Indian Army officer, Gurgaon, India, November 9, 2012 (#67) Author Interview, retired Senior Indian Army officer, Gurgaon, India, December 14, 2013 (#145) 471 "People Distrust Government in Naxalite-Hit Areas: Chidambaram," Times ofIndia, May 13, 2010 (http://timesofindia.indiatimes.com/india/People-distrust-govt-in-Naxalite-hit-areasChidambaram/articleshow/5922935.cms) 472 "State of the Nation Survey," Programme for Comparative Democracy, 2010 (http://www.lokniti.org/stateofthenation.php). 469 470 550 times as many people ranked the LeT (42%) and twice as many ranked Pakistan (33%) as compared to Naxalism (16%). Presently, the Indian public expresses far more interest in the economy, corruption, and regional threats over the Naxal threat.474 Despite Naxal expansion and escalation in the past decade, they are still not regarded as a real threat because their power projection remains confined to areas that largely do not matter. One former police officer explains the disposition of many in leadership positions was "why invite retaliation! Let sleeping dogs lie." 475 This low threat perception had much to do with their targeting. Though Naxals had begun to organize in urban areas, they had not begun targeting them with violence, a somewhat strategic calculation, and until and unless they did so, analysts judged the state was unlikely to deploy the same kind of escalated effort it had in Punjab and Kashmir.476 One advisor to the Indian government on national security explained that insurgent attacks in a remote jungle and attacks in Delhi were simply "not on the same plane." 477 This costbenefit calculation could not motivate the Indian government to seriously tackle the insurgency, whether with kinetic actions or political steps such as full implementation of the Forest Rights Act, a ban on mining, and the revival of local self-government. 478 Low effort containment strategies in the Naxal belt are based on a similar cost-benefit calculation as those in India's Northeast. The state tolerates the disorder these insurgencies produce because, as one senior intelligence officer explains, "they're not the kind of problems 473 Global Attitudes Project, "Indians See Threat From Pakistan, Extremist Groups," Washington, DC: Pew Research Center, October 20, 2010, p. 1. 474 "Modi is Middle Class Messiah," India Today, Feb 15, 2014 (http://indiatoday.intoday.in/story/modi-is-middleclass-messiah-kejriwal-tops-non-upa-non-nda-pm-probables-india-today-cvoter-survey/1/343742.html) 47 Mishra, 2007: 116. 476 Author Interview, Indian think-tank analyst, New Delhi, India, November 12, 2012 (#70). 47 Author Interview, advisor to Indian National Security Council, New Delhi, India, November 20, 2012 (#78). 478 Ramachandra Guha, "The Continuing Tragedy of the Adivasis," The Hindu, May 28, 2013 551 ' that will threaten the state." 479 An interview with a senior national security journalist with significant insight into the establishment's threat perceptions explicitly summarized this calculus: "The Indian state has been quite comfortable living with under-governance in a number of places.. .The Northeast insurgencies and the Maoist insurgency, in fact for all the talk about it being an existential threat, I don't think have ever been perceived-- If you speak to politicians in Chhattisgarh, at the end of the day they'll turn around and say that this is a bunch of desperately poor tribals fucking each other in some corner of the state, which no one gives a shit about, so what is the problem? ... I'm always struck how none of the legislators actually give a shit. [Bastar] is one district-sure it might be large-but there's nothing to be made off of it. There's nothing there, so if a bunch of people feel like roaming around the countryside shooting at each other, why should anyone actually give a shit?" 480 Some analysts have argued India's national patrimony is at stake and the mineral wealth contained within the tribal areas should augment the state's material incentives for decisive action. While aware of the potential resources, the state has not actually been moved by this. Much of this mineral wealth constitutes amorphous future opportunity costs rather than more disruptive present costs. Second, even such material incentives offer no corresponding strategic and ideational incentives to balance out the high costs. Finally, while minerals are lucrative, this type of material incentive is purely an extractive rather than the dynamic, value-added kind dependent on state control, and mineral wealth has been exploitable even amidst the Naxal insurgency, as it has for many decades. 4 4.3. Identity and Violence The Naxal III case also contains evidence that identity mechanisms hypothesized by the core-periphery theory restrained the counterinsurgency campaign's level of violence against the Naxal insurgents and their social base. Over the past six decades, tribals and lower castes rose in 47 Author Interviews, retired senior Indian intelligence officer, New Delhi, India, Dec 18, 2013 (#150) and Indian academic, New Delhi, India, November 15, 2012 (#74). 480 Author Interview, Indian national security journalist, New Delhi, India, Dec 17, 2013 (#147). 481 Author Interview, Indian national security journalist, New Delhi, India, Dec 17, 2013 (#147). 552 group worth but particularly in group embeddedness within state institutions and this changed the way the state managed their violent challenges. Because the campaign against the Naxal III insurgency had not escalated to a high-effort strategy, not all the restraining mechanisms were as visible as in the Punjab case, particularly the strategic mechanisms of vulnerability and information, but the affective mechanisms of empathy and trust played a significant role guiding a low-violence mitigation strategy. 4.3.1. Empathy The moral worth of the Naxal social base, SC/STs, and perceived legitimacy of their grievances engendered empathy for the Naxal movement amongst state leaders, politicians, and the broader public. This empathy has limited permissible levels of collateral damage, narrowed the definition of an insurgent, and expanded the potential pool of those eligible for amnesty and rehabilitation. Furthermore, there is some evidence that the Naxal insurgency has attracted the participation of other backwards classes (OBC)4 82 in Andhra, Bihar, Jharkhand, and Chhattisgarh.4 8 3 Since OBCs have risen even higher in stature, power, and worth over the past 30 years, their participation would cause the state to further balk at the heavy use of violence. Indian national security leaders embraced a strategy of low violence against the Naxals specifically because they are deemed both worthy and central to the Indian identity. Former Home Minister Shivraj Patil defended the minimal use of force against Naxal insurgents, arguing, "They are our children. They are angry and we have to show them the right path with OBCs are something like lower peasants but above the more marginalized "untouchables" or tribals. 48 For instance, 38% of Naxals interviewed by one sociologist were OBCs. See Chitralekha, "'Committed, Opportunists and Drifters': Revisiting the Naxalite narrative in Jharkhand and Bihar," Contributions to Indian Sociology, 44 (3), 2010, pp. 299-329; "Inclusion in OBC List: Govt Sops to Naxal-affected States," Indian Express, July 22, 2010 482 553 affection. We have the forces to deal with violence but that is not the only approach."484 Recently, the head of the CRPF, the main federal paramilitary force tasked to combat Naxals, reiterated this position explaining, "Naxals are our own people. They are not from the outside" and therefore this requires security forces to be "more humane" and "selective in our action." 485 Another former paramilitary director had issued a directive to all company commanders not to treat Naxal insurgents as real enemies, "but rather call them adversaries who are tactically opposing you in your approach." 4 86 Sonia Gandhi, leader of the Congress Party that ruled from 2004-14, expressed confidence that such "misguided youth" could be rehabilitated into the mainstream with minimal violence.487 Other senior members of her party and government diagnosed Naxals as "misguided friends and misadventurous youths" who have "lost faith in the system."488 This empathy expressed by both civilian and security officials restricts how broadly state institutions define the category of insurgent, the threshold of irreconcilability, and the levels of permissible violence. Commensurate with these statements and strategic intentions, surrenders and rehabilitation appear to be the primary methods for dealing with insurgents and civilian casualties are far lower than in previous campaigns, even other ongoing ones like in Manipur (Tables/figures from chapter 6).489 In fact, the state has accepted over 18,000 surrenders of Naxals by 2009.490 Exchange rates between the government forces and Naxal insurgents were some of the lowest in India's counterinsurgency history, in part due to poor quality forces, but 84 Ajai Sahni, "Bad Medicine for a Red Epidemic," South Asia Intelligence Review, 3 (12), October 4, 2004 (http://www.satp.org/satporgtp/sair/Archives/3_12.htm). 485 "Naxal Response Has to Be More Humane: CRPF DG," The Hindu, September 24, 2013. 486 Brijesh Pandey and Prakhar Jain, "Q&A Vijay Kumar, DG, CRPF: 'We Are Not a Ragtag Army or Irresponsible Militia'," Tehelka, 28 (9), July 14, 2012. 487 Sujay Mehdudia, "Sonia Extends Olive Branch to Maoists," The Hindu, April 22, 2009. 4 4 ' Routray, 2010, p. 184. 489 Therefore it cannot be said there is a secular trend toward greater surrenders as their rate is far lower in Manipur and Kashmir. 9 0 PV Ramana 2009. 4 554 also due to the deliberate restraints placed on the use of force and the types of tactics permitted by the GOI. Empathy for Naxals is shared by a vast majority of surveyed Indians who held a sympathetic or neutral image of Naxalites, with 47% expressing explicit sympathy. 491 This has steered the Indian state toward a "holistic" or developmental approach with a concerted emphasis on "address[ing] the genuine grievances of people." 492 Concern for the rebel social base has motivated kinetic restraint as well as the laundry list of development initiatives, even if they are poorly implemented or pilfered by corruption endemic to governance in India. What made a case like Punjab stand out as high effort campaign was that the stakes pressured careful and thorough implementation of all initiatives because failure was too unnerving. Nevertheless, in both Punjab and Naxal III, the state made a concerted effort to treat armed rebellion as a symptom of a larger political grievance rather than the main challenge to be crushed. Indian leaders have routinely rejected Army deployments to bolster the quality of counterinsurgency forces because of what some proponents of professional military deployment allege is a false differentiation of Naxal violence and intentions. One scholar assesses: This 'anti-Army' belief is largely based on a premise that differentiates left-wing extremism from militancy in Jammu & Kashmir and the insurgencies of the Northeast. The 'Naxalites' in spite of their indulgence in wanton violence, are still considered "our people" who "do not demand secession". An analysis of the pattern of left-wing extremist violence, however, indicates little difference between these different categories of violent extremism.493 Critical junctures during the campaign offered the state opportunities to justify and launch a more aggressive military offensive but it deliberately refrained from doing so. After a State of the Nation Survey, 2010, Table 3.1 (sympathetic or neutral is 38%, negative 7%, other 6%) and Table 3.5 (47% sympathy, 53% hostility) and Table 3.6 shows fairly even distribution of this sympathy across class groups. Table 5.3 shows those who favor negotiation or development amonts to 44% whole those favoring coercion is 19%. and this is in most affected areas In Table 5.3, 67% supported development and in Table 5.4, 56% supported unconditional ceasefires. 492 Col Purushottam Singh, Pratividrohi,2011, p. 26 491 493 D'Souza, 2009: 125-32. 555 2010 Naxal ambush killed 75 Indian paramilitary troops-the Indian state's largest troop fatality count in a single combat event-and raised calls for the use of Indian air power, the Air Chief reiterated, "Unless we have 120 per cent intelligence that they (Naxals) are enemies, it is not fair to use air force within our borders. The basic thing is Naxals are our own citizens." 494 A senior member of the ruling party again rejected the deployment of the Army because, "They are trained to kill. They are not trained to do policing." 495 In 2006, the GOI experimented with arming and empowering some tribals as special police officers to confront the Naxals in Chhattisgarh, known as Salwa Judum (Peace Movement). The move resulted in a slight uptick in violence (though nothing close to the use of renegades in Kashmir) but prompted a major public backlash from Indian citizens and human rights groups that forced the state to back away from this tactic by 2008. In 2011, the Indian Supreme Court formally buried this option striking down the state's authority to employ such tactics, a landmark decision by a court that rarely challenges the national security perogatives of the state.496 Going forward, there is little reason to expect this to change as Naxals' moral worth and some representation within state leadership will continue to restrain the BJP-led government. Despite some remarks about zero-tolerance, even on the 2014 campaign trail where he had every incentive to embrace more hard line policies, Prime Minister Modi expressed a compassion for the Naxal base due to their historical identity, having "existed since the time of Lord Ram and Lord Krishna" 497 and called for dialogue with "our people.A 98 "Air Force Not Meant to Fight Naxals: Chief," JBN Live/Press Trust ofIndia, April 7, 2010. "Digvijay Latest: Why Use Paramilitary Against Naxalites,"Indian Express, May 15, 2010. 496 Krishnamurthy Ramasubhu, "How the Salwa Judum Experiment Went Wrong," LiveMint.com, July 10, 2008; J Venkatesan, "Salwa Judum is Illegal, Says Supreme Court," The Hindu, July 5, 2011 4 97 Ashutosh Bhardwaj, "In Naxal Hub, Modi Appeals: 'Make Earth Green Not Red'..." Indian Express, March 29, 2014 (http://indianexpress.com/article/india/polities/in-naxal-hub-modi-appeals-make-earth-green-not-re/ 494 495 556 4.3.2. Trust This perceived group worth and loyalty of the SC/ST Naxal base motivated the state to continue reaching out to dialogue with insurgents, offer accommodations, and dole out development resources despite evidence that it leaked into insurgent coffers. Robust ties between the Naxal base and state leaders enabled both sides to communicate to de-escalate violence and trust the other in some small ways to periodically utilize substitutes to kinetic force such as ceasefires, amnesties, and periodic negotiations. This link, sometimes described as an insurgentpolitician nexus, enables valued constituents to influence state and political leaders, sometimes through electoral means. Critics often malign these interactions as evidence of weak political will but the state's confidence in these substitutes stemmed from the belief that due to the nature of the insurgent identity base, alternative methods were available to ratchet down violence and demobilize insurgents while avoiding killing large numbers of the Naxal base. State ministers occasionally brokered ceasefires and truces with Naxal insurgent groups, opening room for development, amnesty schemes, surrender and rehabilitation packages, and broader negotiations. 499 The same surrender packages yielded little traction in Kashmir or Manipur have garnered far more use in Naxal country amongst uncommitted rebels who trust the outreach from state representatives not far apart, or very close, in identity. Negotiations have also been common. Andhra negotiated a lengthy truce from 2004-05. West Bengal initiated negotiations in 2010 and Karnataka setup a short ceasefire in 2012. The federal center has tolerated this state initiative, expressing a willingness to dialogue with Naxals. The Indian state 4 98 "Naxals Are Our 'Own People,': Narendra Modi," IBN Live, May 20, 2010 (http://ibnlive.in.com/news/naxalsare-our-own-people-narendra-modi/ 15792-37-64.html). 499 "Govt May consider Naxal Truce Offer, If Ceasefire for 72 hrs," Hindustan Times, March 3, 2010; "K'taka Announces Week-long Cease-fire in anti-Naxal Ops," Rediff com, September 15, 2012; Amitabh Srivastava, "Naxal Truce: Too Good to Be True," India Today, February 28, 2010; "Officials Cash in on Truce, to Lay Road in Naxalhit Area," The Hindu, January 2, 2005 (http://www.hindu.com/2005/01/02/stories/2005010208600300.htm); "Amnesty, Rehabilitation Deal for Maoists in Jharkhand," HindustanTimes, August 1, 2010; "Now Maharashtra Offers Amnesty to Naxalites," July 7, 2005. 557 repeatedly entered into negotiations with no prerequisites like under the Congress-led Andhra government in 2004.500 More often than not, more informal ceasefires are established where the local government agree to a deal to limit operations as the state of Jharkhand did in 2009.501 Furthermore, the state has embraced a broad generous surrender and rehabilitation program believing most Naxals can be brought back into the fold of normal politics.502 The vast majority of Indians favored these non-kinetic outreach tactics. Surveys showed Indians preferred negotiation, development and unconditional ceasefires over coercive measures, something their elected leaders and bureaucracies have internalized.0 3 Trust stems again from the group's moral worth and the fundamental Indian state belief that even hardcore rebels are not regarded as "enemies" but "adversaries" who oppose the state "tactically," as if they share some root goal. 504 4.3.3. Information In the absence of a high-effort counterinsurgency campaign against the Naxals, the role of identity's strategic mechanisms are less observable but some evidence indicates information plays a crucial role. With SC/ST base entrenched in both the Naxal base and the state, the Indian state has been able to gather substantial insights on Naxal leadership and goals to inform strategy and improve targeting. Indian state leaders with ties to the Naxals' social base have helped guide 500 Ramana, 2009, p. 753; "Indian Minister Seeks 'Negotiations' With Maoist Rebels," BBC News, December 23, 2009. 501 Bibhu Prasad Routray, 2010: 185 502 "Truce Perks for Maoists Hiked," The Telegraph, March 5, 2013 (http://www.telegraphindia.com/1130305/jsp/frontpage/story_16633459.jsp#.U8VFnKjjfrM) 503 State of the Nation Survey, 2010, Table 3.1 (sympathetic or neutral is 38%, negative 7%, other 6%) and Table 3.5 (47% sympathy, 53% hostility) and Table 3.6 shows fairly even distribution of this sympathy across class groups. Table 5.3 shows those who favor negotiation or development amonts to 44% whole those favoring coercion is 19%. and this is in most affected areas In Table 5.3, 67% supported development and in Table 5.4, 56% unconditional ceasefires. suported s0 Brijesh Pandey and Prakhar Jain, "Why Anti-Naxal Ops Have Become a Terrifying Black Hole, Tehelka, 28 (9), July 14, 2012. 558 strategy to address Naxal demands. This was most recently demonstrated by the new Tribal Affairs Minister Jual Oram, himself a tribal, who described Naxalism as a "social problem" stemming from historic mistreatment of tribals, and who is currently pushing development and dialogue tactics to "woo them back."5 05 Shared identity between some in the state and the Naxal base has not only provided insights on their grievances and leadership but also enabled the state to establish dense networks of informants 506 and at times successfully infiltrate some Naxal branches, which relies on some prior social trust, leading to arrests or discrete targeting.507 The state is even better able to infiltrate and understand urban Naxals of middle class backgrounds, which may explain how the state has practically inoculated itself from serious urban Naxal mobilization. The Indian state opted to use police not only for their lighter force but also their ability to harness local social ties to generate information for more discriminate targeting. For instance, Bihar police rely on chowkidars (village watchmen) from the same social groups as the Naxal base to serve as informers.508 In many Naxal-affected states, auxillary constables from the same tribal groups help inform and guide local police forces. 509 In the few cases when the state police have been overhauled such as Andhra, they have become better at gathering intelligence to narrowly target senior Naxal leaders and restore greater stability.5 10 Government advisors have Subodh Ghildiyal, "Unlike Modi, Oram Wants Soft Steps to Curb Naxalism," Times ofIndia, May 29, 2014 Aman Sethi, "Villagers Trapped as Maoists, Police Play Spy Vs. Spy," The Hindu, June 27, 2010 5 07 Debabrate Mohanty, "Chhattisgarh: NAxal Camp busted, arms seized," Indian Express, March 17, 2014; Aman Sharma, "Rajnath Singh asks Intelligence Bureau to infiltrate CPI (Maoist) ranks," Economic Times, June 7, 2014. (http://articles.economictimes.indiatimes.com/2014-06-07/news/50404839_1_central-armed-police-forces-maoisthome-ministry-official); "Intelligence Agencies May Have Infiltrated Maoist Leadership," Rediff com, November 25, 2010 (http://www.rediff.com/news/special/special-intelligence-agencies-may-have-infiltrated-maoistleadership/20101125.htm).; Sudeep Chakravarti, "Revisiting Ficci's Take on Maoists," Live Mint.com, June 20, 2014 (http://www.livemint.com/Opinion/LOCLy3GOSznsNBzSdOyTrK/Revisiting-Ficcis-take-on-Maoists.html); SA Aiyar, "No Military Solutions for Maoism, Swaminomics Blog, Times ofIndia, May 23, 2010 (http://blogs.timesofindia.indiatimes.com/Swaminomics/no-military-solutions-for-maoism/) 508 VK Shahikumar, "Police Cannot Take on Maoists," Indian Defence Review, 25 (4), Oct-Dec 2010. 509 Debaashish Bhattacharya, "After the Fire," The Telegraph, June 9, 2013 505 5 06 510 Verma, 2014: 309 559 been acutely aware that they cannot kill their way out of the insurgency as the collateral damage could escalate the conflict and close off existing information networks.' 4.3.4. Vulnerability Like information, vulnerability is a difficult strategic mechanism to observe in low effort campaigns. Because the state has not launched substantial and sustained operations, despite the degree of representation of the Naxal base, it is difficult to identify where the state has faced strategic costs from the excess use of force, but the main aversion to escalating operations stems from the fear of a backlash - both strategically and politically. First, state leaders can balk at political vulnerabilities as much as security vulnerabilites and there is evidence that political constraints have limited government forces from taking more aggressive kinetic action at the state and national levels. Naxals have strong ties to over ground political parties, particularly leftist parties, and their vast discipline and mobilization potential for guerilla warfare can also be harnessed for normal electoral politics. Elected state leaders like in Bihar, Jharkhand, and West Bengal have been particularly reticent to resort to high-violence strategies that might alienate these voting blocs as was the Andhra government during the 1990s. 5 12 Furthermore, since coalition politics has dominated India for the past two decades or more, governments relying on the leftist parties like the CPI or low caste political parties like the Bahujan Samaj Party could fall apart or at least face efforts by these state governments to obstruct federal forces. Ultimately federal forces would prevail but the political messiness might be too costly and embarrassing for any government to contemplate. 511 Shoma Chaudhury, "Bringing on the Army Against the Naxals Will Be a Disaster," Tehelka, 7 (23), June 12, 2010 (http://archive.tehelka.com/storymain45.asp?filename=Ne12061Obringingon.asp) 512 Routray, 2010: 185; Mukherjee, 2009; Author Interviews. 560 Second, besides political vulnerabilities, the hypothesized deployment of the Army violently unleashed on Naxal insurgents and their base could trigger a tremendous outcry from all sectors of government where the presence of lower castes and tribals has grown in substantial numbers amongst civil servants, paramilitary, and police officers. Already, there has been some evidence at a tactical level in Bihar that the large number of dalits in local policing duties has constrained police action against Naxals, 513 as well as evidence of caste-affinity between Naxal combatants and local police. Recently the Indian state has benefitted from relationships with affected Naxal communities that have allowed it to recruit and deploy thousands of local adivasis from Naxal regions to federal policing forces and Army contributing to better targeting, information, and more discrete use of force. However, if the state adopted a high-violence strategy, it would all but foreclose on these opportunities and/or face serious defection and indiscipline. In fact, one of the reasons why state leaders and top brass have been averse to deploying the Army into the Naxal zones is in part because of the increasing number of low caste groups in the Army (some with their own regiments) and the large tribal component, particularly in the Bihar regiment. 516 One retired senior police officer described deploying these forces against their own kin as a potential "recipe for disaster."5 17 Verma, 2014: 313; Satish Kumar, "Failure to Counter Naxalism in Bihar: An Indigenous View," SecurityRisks.com, November 2, 2010 (http://www.security-risks.com/security-issues-south-asia/terrorism/failure-tocounter-naxalism-in-bihar-an-indigenous-view-90 1.html). 514 Shoumojit Banerjee, "Mungar Massacre Underscores Changing Face of Bihar's Naxal Movement," The Hindu, July 26, 2011 (http://www.thehindu.com/todays-paper/tp-national/munger-massacre-underscores-changing-face-ofbihars-naxal-movement/article2294450.ece). 515 "Battalion of Tribal Youths is SRPF to Fight Naxals," Express News Service, March 31, 2012 (http://archive.indianexpress.com/news/battalion-of-tribal-youths-in-srpf-to-fight-naxals/930709/); "CRPF to Deploy Local Recruits in Naxal-hit States," DNA India, September 20, 2010 (http://www.dnaindia.com/india/reportcrpf-to-deploy-local-recruits-in-naxal-hit-states-1440686); Gautam Datt, "Army Scores 6,500 New Recruits with Rallies Targeting Maoist Hotbeds," Mail OnlineIndia, November 21, 2013 (http://www.dailymail.co.uk/indiahome/indianews/article-2511443/Army-scores-6-500-new-recruits-ralliestargeting-Maoist-hotbeds.html). 513 516 517 Cohen, 1990; Wilkinson, 2000. Shoma Chaudhury, "Bringing on the Army Against the Naxals Will Be a Disaster," Tehelka, 7 (23), June 12, 2010 (http://archive.tehelka.com/storymain45.asp?filename=Nel206lObringingon.asp) 561 4.3.5. Contrast to Telangana The counterinsurgency campaign against the recent Naxal (III) rebellion offers a stark contrast to the Indian state's unflinching use of brute force against the very same social identity base that composed the Telangana rebellion more than fifty years prior. State leaders did not speak of rehabilitating misguided children. Insteadm the home minister talked of complete eradication and the IB director advocated "inflicting more severer punishment" against a "gang of outlaws," which they did with indiscriminate punitive encirclement raids against villages. 518 Part of this may have been a result of uncertainty and reputational concerns in India's early years or potentially because the Telangana rebels were suspected to be collaborating for some time with the Muslim Razkars. 519 But the great social distance or polarity at the time between the insurgent base (tribals and lower castes) and the mostly upper caste leadership in India's political and security institutions likely played an equally important role in sanctioning an attrition strategy. The state's deeds betrayed a lack of empathy as much as Indian leaders' words. The GOI did not hesitate to deploy the military in this campaign and suffered no protests besides a few from students in Andhra. Counterinsurgency operations in Telangana arguiably involved more Army troops than deployed to Kashmir in the first war with Pakistan. At least two thousand mostly tribal peasants were killed in direct action and false encounters, and as many as ten thousand more died after being forcibly relocated to the equivalent of concentration camps. The scale of arrests and torture, even if exaggerated by an order of magnitude, was still astonishing.5 2 0 51 Mullik, 5 19 1972, pp. 60; 277-78, 290; Sundarayya, III, 1973, pp. 32-33. Mullik, 1972, pp. 242, 246-47. 520 Biplab Dasgupta, 1974, pp. 17-18; Kennedy and Purushotham, 2012, p. 842-44; Sundarayya (1972) says 300,000 were tortured and 200,000 arrested during this time. 562 Social distance was evident in numerous tactical interactions. The communist rebels appealed to the shared identity of the Indian soldier but this had no traction. Meanwhile, the military suffered a severe information gap as the forces, being "strangers to the area," lacked language skills, struggled to cultivate informants, and had to rely on local guides and interpreters who frequently betrayed them.521 During these earlier campaigns, the state also dismissed dialogue, compromise, or non-kinetic measures that it would later use in the Naxal III campaign to demobilize rebels. Even before the military launched its offensive after taking Hyderabad, Telangana rebels had entreated the Indian state to negotiate a settlement but the military governor dismissed it promising he would liquidate all the communists in the region within six weeks.m When the CPI modified its platform in early 1951 seeking reconciliation and a compromise, the Indian state again rejected these gestures preferring to win by force. 523 Even the subsequent Naxal I campaign against the same social base did not hesitate to deploy the Army in counterinsurgency operations in large parts of rural Bengal, Bihar, and Orissa from 1967-71. Though not as harsh as the Telangana campaign, operations in Andhra still involved forcible relocation, police brutality, mass arrests (almost 1% of the population), and the use of overwhelming firepower against poorly armed or unarmed peasants under even the slightest suspicion. 2 4 The past sixty years of Indian history have witnessed the dramatic decline of social distance between the Indian state elite and tribals/lower castes due to affirmative action policies that have both increased moral worth and embeddedness of these once totally marginalized Mullik, 1972, pp. 251-253. Sundarayya, 1973 part III, p. 30. 523 Mullik, 1972, pp. 292-93. 524 Machine guns were used against peasants armed with axes, spears, and arrows and few firearms between them. See Banerjee, 1980, pp. 103, 159; Biplab Dasgupta, 1974, pp. 47-50 (at least 1500-1641/200,000 people in the Srikaklulum regionarrested); Kennedy and Purushotham, 2012, pp. 848-49. 521 522 563 groups. The rise in the SC/ST/OBC positional status seems to account for at least some of the change in the state's counterinsurgency approach. In the past decade, the Indian state has chosen a far less violent mitigation strategy than it did sixty years ago in Telangana with at least the some effort at development and political redress of grievances, even while it continues with more violent strategies in other contemporary campaigns in Manipur and Kashmir. 4.4. Alternative Explanations Two alternative arguments are offered to explain the dynamics of the Naxal III campaign strategy. The first argues there has been considerably less effort and violence against Naxals because they are not separatists and therefore do not elicit the same fear, concern, or desperation. The second argument contends the central government has expended less effort h specifically because the institutional strictures of Indian federalism limit the role of the center in law and order issues under state jurisdiction. A closer analysis finds both these explanations are insufficient to explain the current mitigation strategy deployed against Naxal rebels. 4.4.1. Separatism There are three reasons to doubt the separatism explanation. First, the argument that separatist insurgencies trigger strategies of high effort and violence is inconsistent with the evidence presented in chapter 6 on the vast array of Indian campaigns against separatist movements that have involved considerably low effort (in all of the Northeast) or low violence (in Assam and Punjab). Second, the logic for why separatist conflicts would elicit more state effort or violencebased on internal or external deterrence-is based on a set of insurgent properties, capabilities, 564 and goals that the Naxals actually possess. Though led by a Maoist organization, Naxals have previously called for separate "tribal states" in central India for a distinct social group, similar to what tribals fought for in the Telangana rebellion. 25 Currently Naxals seek control of territory for a "compact revolutionary zone" and have worked towards this by collaborating with many other separatist organizations.5 2 6 As they gain control of territory, state officials expect the Naxal goal will escalate from revolutionary to positional or conventional war, which would effectively pose the same threat as a separate state.52 7 Moreover, Indian intelligence has regularly documented their links with India's external threats including Pakistan's ISI, China, and the LTTE. One officer argues that Naxals pose "a much greater potential threat to India's national security than either the Jihadis in Kashmir, the Indian Mujahideen or the under grounds in the North East" 29 and many well-regarded Indian security analysts concur they are "more dangerous," because of the uncertainty around what potential compromises can be made. 53 0 All these activities set dangerous precedents that reputation theories expect states to actively prevent with substantial violence, and yet India persists with a limited strategy of mitigation. Third, the extent an insurgency's separatist goals shape a state response is deeply intertwined with the state's perception of the rebel identity group. Sikh insurgents officially Kennedy and Purushotham, 2012, p. 854. Police Inspector General (Andhra) Anjana Sinha, Slides from "The Naxalite Insurgency in India," Carnegie Endowment for International Peace, Washington DC, Nov 12, 2012; T Sunil Reddy, "PWG Links with Kashmir Ultras Bared," Indian Express, Sept 29, 1998 9980929/27250304.html). 5(http://expressindia.indianexpress.com/ie/daily/l 27 Anjana Sinha, 2012. 528 "Home Secretary: ISI aiding Naxalites," Tribune India, March 7, 2000 (http://www.tribuneindia.com/2000/20000307/nation.htm#6); See N. Manoharan, "No Longer 'Indigenous': External Linkages of Indian Maoists," Naxal Violence: The Threat Within, Ed. Gurmeet Kanwal and Dhruv C. Katoch, New Delhi: CLAWS, 2012, pp. 121-134; Sumir Kaul, "ISI Taking Help of Separatists to Woo Naxals," Outlook India, October 24, 2010, (http://www.outlookindia.com/news/article/ISI-Taking-Help-of-KashmiriSeparatists-to-Woo-Naxals/698536); "BJP Alleges Nexus Among ISI, Naxals, J-K Separatists, Indian Express, October 25, 2010 (http://archive.indianexpress.com/news/bjp-alleges-nexus-among-isi-naxals-jkseparatists/702279/). 5 Purushottam Singh, March 2011, p. 25. 53 0 Rajagopalan, 2009. 525 526 565 declared they were fighting for a separate state but many Indian officials and commanders didn't believe they were serious about it with all they stood to lose given how embedded Sikhs were in the Indian state. Meanwhile, though many Manipuri insurgent organizations are revolutionary and socialist in nature, the state principally views them through the lens of separatism. Though the Naxals employ strategies and espouse goals resembling separatists, the identity links make the Indian state disinclined to acknowledge that. 4.4.2. Federalism A second argument sometimes used to explain the limited effort by the Indian state is that Indian federalism, particularly after a Supreme Court ruling in 1994 limiting the usage of President's Rule, has made it institutionally difficult for the Indian federal government to impose its strategy on its states, particularly across multiple states. 53 1Federalism certainly poses some obstacles to comprehensive strategies but it cannot sufficiently explain longstanding and current ambivalence. When sufficiently threatened, the Indian central government has historically ignored the constraints of federalism and deployed coordinated counterinsurgency campaigns across multiple states as it did with the Army's 1971 Operation Steeplechase (in West Bengal, Bihar, and Orissa) or even some coordinated operations during the Punjab campaign (across Haryana, Chandigarh, and Delhi). By contrast, in the period of ample Presidential authority-when coalition governments became the norm as Congress party dominance waned in the 1980s-90s and national leaders liberally invoked President's Rule-rising Naxal violence never precipitated such a move in Bihar and Andhra Pradesh. This suggests low effort had more to do with perceptions of what was threatened than with institutional constraints. Federalism has also not 531 I thank Devesh Kapur and Anit Mukherjee for bringing this argument to my attention. 566 prevented the operational presence and defacto control of Army and federal paramilitary in Kashmir or across many states of the Northeast even after 1994 despite their presence being resented by popular majorities as well as elected politicians. Federalism also did not technically inhibit Operation Green Hunt in 2009 across multiple Naxal-affected states, but disinterest may explain why it was poorly planned, under resourced, unsustained, and overextended. When it comes to imposing federal control, the Indian state has proved the old adage "where there's a will, there's a way," but there has simply been no will for the Naxal belt. Some have pointed out the central government could still invoke Article 355 to protect states from internal threats and exercise its powers of coordination and inducement for a more comprehensive campaign.532 This is why some seasoned Indian counterinsurgency experts and practitioners argue the state should deploy the Army. 3 The reason the state hesitates to do this is less about the institutional constraints and more a function of the stakes-the Naxal belt is not considered to be worth the material and political costs or strategic tradeoffs. V. Conclusion This chapter has provided detailed evidence from India's campaigns against the Punjab, Kashmir, and Naxal III insurgencies validating the mechanisms of the core-periphery theory. The territory contested by insurgents shaped state strategies of effort through active consideration of the costs and incentives. The positional status of the rebels' social identity base influenced state strategies of violence through affective and strategic considerations of empathy, vulnerability, information, and trust. The territorial incentives that motivated high-effort strategies in Punjab 532 N. Manoharan, "Naxal Violence: What Should be Done to Counter?" IPCS brief #3967, May 31, 2013 (http://www.ipcs.org/article/peace-and-conflict-database-naxal/naxal-violence-what-should-be-done-to-counter- 3967.html). 533 G.D. Bakshi, 2010; D'Souza, 2009 567 and Kashmir were absent in central India's rural corridors where the Naxal insurgents have fought the state for decades. The state's value of the Sikh community due to their worth and embeddedness in state institutions forced the GOI to quickly adjust to a strategy of population control primarily led by empathetic co-ethnics with relatively low violence due to intelligence, public goods, rehabilitations, and bolstering of moderates. The same cannot be said of Kashmiri insurgency whose Muslim base had no standing within state institutions and was considered suspect if not foreign agents evoking contempt and distrust. The almost total lack of empathy drove a strategy of attrition with relentless brute force against civilians and insurgents alike. The Naxal case does double duty revealing how state views of territorial value relative to expected costs shapes low effort strategy because of it, and it provides an interesting look at how the shifting positional status of an identity group over prolonged periods can prompt drastic changes in the state's approach to rebellion against the same social base. While the state had no qualms using brute force against the tribals and lower caste base of the Telangana rebellion, fifty years later this has been made much more difficult as these groups' position has improved with greater social status and institutional ties to state elites. The Indian government has struggled to use serious force even as it has fought insurgents, and displayed tremendous restraint emphasizing non-lethal and non-kinetic tools in the campaign. Visualizing the data from these campaigns' strategy as cumulative returns also shows how different the trajectory of these campaigns look, and also how they adhere to the expected path of the strategy the state adopted. Chapter 2 (the dependent variable) detailed the ideal type counterinsurgency strategies in the literature and based on their properties and tactics, speculated on the functional form of each strategy. The ideal functional form (Figure 7.3) has attrition producing linear returns, population control producing increasing returns, and both containment 568 Figure 7.3: Cumulative Effort - Ideal Type II 6 E Increasing ReturnsopultionControl -~ -Attrition near Returns Attrition 0. -Population Control 2 D mnishing em -Enfeeblement Enfeeblement 46 -Mitigation Diminishiing Returns. Mitigation 01 2 5 4 3 6 9 a 7 10 11 Time (T+1) Figure 7.4: Cumulative Effort in Punjab, Kashmir, and Naxal III 14000 12000 Kashmir 10000- 20% Attrition At-on Kashmir Insurgents Killed (1990-2000) E -- 8000 La Punjab Insurgents Killed (1983-93) 6000 -------- -Naxal Insurgents Killed .-..--. & Surrendered Naxals (Killed & Surrendered) 0% Attrtion 0oo r - (2003-13) Naxal Insurgents Killed (2003-13) 5% Attrition SWr) 2000 1- .Naxals 1 2 3 4 5 6 7 8 9 10 11 Time (T+1) strategies of enfeeblement and mitigation producing diminishing returns but at different slopes. If returns are measured in insurgents neutralized over time, plotting the three campaigns described above in their first 11 years shows the actual trajectory of the campaign adhering 569 closely to the expected functional fonn-Kashmir as attrition, Punjab as population control, and the Naxal III campaign as a form of containment (Figure 7.4).534 Admittedly the exercise is based are fraught with data validity problems, but nevertheless, it provides another piece of evidence confirming the theory's expectations. In the course of this chapter, I was also able to evaluate specific alternative explanations for these cases that typically align with theories of insurgent behavior. I was able to demonstrate there was not a clear link between state strategy and either external support given that Punjabi and Kashmiri insurgents benefited from remarkably similar Pakistani support yet the state chose different strategies. The argument that separatists evoke more effort and violence is also problematized by the state's minimalist approach to a Naxal insurgency that is larger, more centralized, stronger, commands a vast popular base, and display a number of similar goals and tactics as separatists. Furthermore, close evaluation of the Kashmir case reveals much more about the weakness of regime type explanation of state violence. While attrition in Telangana might be partially written off as a state's desperation in its infancy, the Kashmir case reveals regime type to be a distinctly incorrect explanation for strategy. The regime type explanation would expect the Indian state, as a time-tested and stable democracy, to have responded with much less violence. This should have been bolstered by domestic developments of more competitive national elections rather than single party dominance as well as international trends like the ideational ascendance of democratic rights and humanitarian norms, increased media scrutiny, and an increased risk of international pressure or intervention in the post-Cold War era. The fact I include insurgents surrendered in a second observation of Naxal III on the graph since this is considered to play a more prominent role in the campaign. 5 570 is regime type cannot account for a state that is, on average, democratic, but willfully suspends all democratic institutions in a sub-region. In Kashmir, regime type could not constrain the military because it had unfettered freedom with AFSPA no real civilian oversight.5 35 The first seven years of the counterinsurgency were conducted under President's rule with no popular sovereignty. When elections were held in 1996 to finally install a democratic state government, actual turnout averaged about 25%,536 and this figure was suspected inflated by renegades coercing villagers to vote.537 Scholarly work by former Indian military officers has revealed the tenuousness of civilian control throughout the Kashmir campaign despite ostensibly ideal civil-military relations. The Army and intelligence organizations had no trouble shrugging off local accountability, civilian oversight, or political control. 3 8 Today, democratic practices in Kashmir and the Northeast are still subverted by corruption, coercion, the retention of AFSPA even amidst civilian opposition, and intimidation during elections.539 535 536 Author Interview, Kashmiri academic, Srinagar, India Dec 8, 2012 (#110). Joshi, 1999, p. 438. Levy and Scott-Clark, 2012, p. 409; Habibullah, 2008, p. 156; "Explains an army official: 'The strategy helps in bringing the administration a little closer to its aim of increasing the poll percentage when ever an election is held: at least in the areas where the surrendered militants are operating, the voter turn out will be slightly more respectable."' See Baweja and Vinayek, 1996 1 3 Ali Ahmed, 2009, pp. 66-70; Anit Mukherjee, 2009; Levy and Scott-Clark, 2012 reaffirm this. 539 Author Interviews; Staniland, 2013. 537 571 CHAPTER 8 Conclusion "Ifrebellion is as old as authority, then the problems of restoring authority are just as old."' For states that inherited a post-colonial patchwork of political order, this challenge of managing, fighting, and staving off rebellions has been a central feature of their political development. State responses to rebellion provide insights into how they govern and wield authority under duress-that is, how they relate to different groups of their citizens, exercise juridical and disciplinary power, abide by their own rules, and politically integrate and consolidate control over contested territories. Episodes of resistance and rebellion provide a unique window into the mechanisms by which the sovereign, the leviathan, imposes, enforces, and guarantees a particular political order. Because not all violent internal challenges constitute existential threats-in fact in the modern era, few initially do-the leviathan therefore posses a flexibility to selectively apply different strategies, involving a different balance of tactics. Restoring authority requires navigating between difficult choices. States must mobilize sufficient resources to defend their writ, but not too much that they foreclose on their future with a pyrrhic victory. Consequently, states must husband resources to face a variety of future threats and challengers, but not so stingily that they allow the one in front of them to metastasize and overwhelm them. States must administer sufficient violence to deter or disarm challengers but not so much that they invite moral and strategic backlashes within their ranks that undermine their cohesion. Therefore, they must at times restrain violence and work through non-kinetic means, but not to the point of appeasement that invites escalated rebel activity. 1Jackson, 2008. 572 This project has fundamentally sought to explain some of the factors that guide states through these tough choices. I have argued throughout that the selection of strategy depends on state incumbents' judgments within the conflict of what is at stake, what is required, and what is permissible-judgments heavily shaped by the contested territory and rebel identity. Roadmap. This chapter sums up what we have learned in the study of counterinsurgency, how much the core-periphery theory contributes to our understanding of strategy, how far the argument travels, and what this means for those scholars and practitioners engaged with civil conflict. The first section summarizes the empirical evidence presented in the previous four chapters on Indian and Pakistan's counterinsurgencies and evaluates how much support it offers for the core-periphery theory. The second section assesses whether the evidence offers any support for some of the leading alternative explanations. Third, the chapter probes the external validity of the theory in a number of counterinsurgency cases outside South Asia. Finally, it outlines the theoretical contributions of this argument and its implications for policymakers. I. Empirical Findings This project has advanced a theory of core-periphery relations to explain state incumbent strategies within asymmetric civil wars. The theory has contended that for states' facing rebellion, the value of contested territory and the positional status of rebels' social base interact to shape the type of strategy the incumbent will deploy to counter the rebellion. Core territory in terms of material, strategic, or ideational importance, will incentivize higher effort than peripheral territory. Core social identity groups with high group worth and embeddedness within the state will motivate greater restraint than peripheral groups. Combined, these two variables predict a counterinsurgency strategy that marshals sufficient material and political resources as 573 well as leverages it towards a particular balance of kinetic and non-kinetic tactics to produce a distinct theory of victory. These variables can explain more sub-national variation in counterinsurgency strategies along these two dimensions (effort and violence) than any of the more prominent structural or strategic interaction theories. The four empirical chapters (4-7) presented evidence on 29 cases from South Asia, and of these, 23 of the 29 cases studied (~80%) offer very strong support for the core-periphery theory described in chapter 3 and confirm its predictions on the state's strategy of effort and violence. The theory's independent variables of territory and identity were at least able to correctly explain either the level of effort or the level of violence in all 29 cases. Rebellions in highly-valued core territory generally elicited high effort strategies and rebels with a social base part of or in close proximity to the core group, due to their high worth and embeddedness, generally elicited lowviolence strategies. Support for the core-periphery theory exceeds the evidence found for some of the major competing accounts of counterinsurgency strategy. These alternative explanations such as structural conditions (state strength and regime type), reputational stakes, and insurgent strength and goals can at most garner strong support from one-third of the cases studied. Most of these theories are unable to explain or predict counterinsurgency campaigns involving a level of effort uncorrelated with the level of violence, i.e. enfeeblement (low effort/high violence) or population control (high effort/low violence). Furthermore, the intensive case studies reveal that state leaders and strategic decision makers were often gauging their strategies based on their value of the contested territory and their empathy or disdain for the rebelling identity group. Of the cases "missed" by my theory, in three of them (Naxal I-Andhra and Naxal IIAndhra, and FATA-II SWA), the theory did a reasonably effective job predicting the strategy, 574 but some inconsistencies and discrepancies remain due to intermediate categorization of a dimension of strategy (moderate) rather than a clear "high" or "low." In three others (Hyderabad, Balochistan IV, and FATA II-Other), there was only partial support for the theory as it failed to predict one dimension of the strategy. Each of these "missed" cases was further probed in the analysis of the results and identified as the product of unique, case-specific factors, rather than being systematically accounted for by a rival theory. Furthermore, the in-depth case studies in chapters 5 and 7 that explore four cases from Pakistan and India respectively, provide more robust evidence beyond surface-level accounts and figures to confirm the character of the campaigns and link them to predicted mechanisms of territory and identity. In these cases, a number of mechanisms and detailed observations anticipated by the theory are detected, reflecting the state's pre-war relationship to the contested territory and contesting identity group. These include: speech evidence directly expressing public or private concerns about the stakes and strategy; decision making processes, including dissent, pushback, and course corrections; accepted trade offs such as force redeployments, doctrinal shifts, political investments, and concessions made; ground-level tactical choices such as the quantity and quality of personnel and materiel deployed, deployment patterns, rules of engagement and behavior, and metrics for success; and, opportunities and institutions leveraged for collecting information from the population for selective targeting and building trust with the rebel base to trade public goods and deals for non-kinetic demobilization. The Pakistani and Indian states deliberately chose to invest very little effort and material or political resources into counterinsurgency campaigns in FATA and the Naxal-belt, and the debates and public statements surrounding these campaigns and territories illustrate why. The FATA and the Naxal-affected regions simply lack any material, strategic, or ideational 575 importance to the state. Combined with the prohibitive costs of power projection in difficult mountainous or jungle terrain and even more complex human terrain, resistant to centralized authority of any kind, the game was judged to not be worth the candle. The chapters also illustrated state leaders actively estimating the importance of contested territories and active deliberations over corresponding high-efforts the state would take to regain control of them. The material and strategic value of East Pakistan and Punjab was such that rebellion in these regions would undoubtedly compel a significant high-effort strategy by the Indian and Pakistani states as they did. What is striking is that even regions of modest material value but significant ideational value would draw significant state efforts and energies. Indian leaders emphasized how the threat to Kashmir was a threat to the idea of a secular nation, while Pakistani leaders distinguished the hazards of Islamic militant threats in "settled areas" and the doorsteps of the capital as opposed to the wild, western frontiers. Case studies of both these campaigns illustrate how leaders down to ground-level commanders mustered tremendous effort, as they were convinced the territorial stakes not only involved immediate material costs but also impacted the country's core due to the region's interconnectedness and ideational significance. The in-depth case studies also draw out the mechanisms through which the rebel group identity influenced the state's strategy of violence. Punjab demonstrates the intense strategic vulnerability states can evince when a rebel identity group is also a core component of their security forces, requiring adjustments in the normal patterns of state violence. East Pakistan traces how the ethnic disdain infiltrates decision making at the strategic, tactical and operational level, all permitting if not requiring greater use of violence. In Kashmir, the Indian state's dearth of co-ethnics within the state at all levels limited their information flows for strategic decisions and tactically handicapped them from distinguishing militants from civilians or effectively 576 interfacing with the population. Along with other mechanisms, this opposing team disadvantage nudged the state in the direction of indiscriminate tactics, heavy violence, and unaccountable renegades even when leaders recognized the flaws in this approach. In Pakistan's Northwest, the state maintained a consistent level of trust in the character of the Pashtun social base, and even the rebels themselves, that most armed insurgents were in fact still loyal to the state and not truly irreconcilable. Consequently, in FATA the state made heavy use of non-kinetic tools to demobilize insurgents, such as deals with rebel commanders and tribal leaders, which continued even as they escalated military offensives.2 And in Malakand II, the combination of trust and effort allowed more comprehensive non-kinetic tactics, such as rehabilitation, public goods, policing, and political reforms, to play a central role in the counterinsurgency strategy. While exogenous theories like my argument have a difficult time explaining change over time, the empirical chapters also depict some instances when the value of territory and identity gradually shifted across decades, therefore enabling different approaches. Rebellions over the same real estate elicited different counterinsurgency strategies because the net value of the territory had significantly increased over decades. As a result of the Malakand region's upgraded value and importance, the Pakistani state mobilized far more effort in the second campaign in KPK in 2009 than in 1994. Rebellions with the same social base also elicited dramatically different strategies between the Telangana and Naxal III campaigns because the rebel social base of lower caste dalits and tribals had grown significantly in worth and embeddedness, constraining the state's use of violence. Though the core-periphery theory is not built to explain shifting micro-dynamics in conflict, it can explain different strategies over broader periods of time driven by shifts in territorial importance and group positional status. 2 Sameer Lalwani and Paul Staniland, "The Politics of Pakistani Strategy on the Northwest Frontier," Paper Presented at the American Political Science Association Annual Conference, Washington DC, August 2014. 577 II. Assessing Competin2 Theories Despite some of its weaknesses, the theory possesses greater validity and explanatory range of strategy across the cases explored in this dissertation than any of the competing theories. The limited explanatory power of each of these major competing theories is summarized below in Table 8.1 and described in detail in this section. Table 8.1: Overview of Findings 3 Core-Periphery Theory 2 3 4 5 6 7 8 Observation India Hyderabad Telangana Nagaland I Mizoram Naxal I - AP NaxalI-WB Nagaland II Manipur 9 Punjab 10 Naxal II -AP 11 Assam 12 13 14 Kashmir I4 Naxal III Kashmir II 1 Pakistan 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 3 Baloch I Baloch II Baloch III East Pakistan Baloch IV Waziristan Sindh-MRD Karachi-MQM Malakand I FATA I BalochV Malakand II FATAII-SWA FATAII-Other FATAII-NWA Total Strong Support Moderate Support Partial Support Alternative Theories (strong support) Relaxed Structure (State Strength & Reputational Rebel Regime Type) Stakes Size Rebel Aims _ _ _ 4 4 44 4 4 4 4 4 4 4 8 10 _____________________ _____________ 4 4 4 23 3 3 10 8 Full evaluation of alternatives along each dimension of strategy located in Appendix Table A8.1. 578 State Strength From the outset, this project sought to hold state strength constant by comparing counterinsurgencies conducted by the same state, because undoubtedly, the capabilities of a state certainly bound its expenditure of effort to fight rebellion. Common sense tells us that in its struggle with Tuareg rebels, the sub-Saharan government of Mali cannot hope to muster a fraction of what more powerful states like Russia or China can deploy against insurgent groups. Furthermore, states on the brink of failure or near total collapse possess no capacity to take on armed challengers and deploy any sort of high effort strategy, whether attrition or population control, and perhaps no capacity for strategy at all. Nevertheless it remains an open question whether state capacity just installs a ceiling on counterinsurgency options, or whether shifts in this capacity can offer meaningful traction on a state's varying strategies across its sub-regions and over time. In the cases studied throughout this project, I find very little support for a functionalist story of periodic increases in state capacity determining counterinsurgency strategies. For this theory to hold, we would need to observe evidence of a state deploying high-effort strategies when it was flush with resources, preferably from exogenous sources. In fact the capacity argument fails in about half the cases (15) to explain the state's deployment of effort. Instead, states exercise tremendous choice, even within the confines of capacity, and often choose to fight rebellion with less than what they are fully capable of harnessing. The in-depth case studies of Indian and Pakistani counterinsurgency demonstrate that state leaders are routinely matching strategies to fit the territorial stakes rather than what the state is capable of overall. Furthermore, in cases where exogenous shocks altered state capacity, this did not have any noticeable effect on the state's choice of strategy. 579 For Pakistan, the spigot of U.S. foreign aid flows turned on in the 1950s-60s, 1980s, and again in the 2000s, but in most of these periods, this added capacity did not alter the military's approach to counterinsurgency in low valued regions. Despite these additional resources, the military displayed no interest in change and reprised its enfeeblement strategy against Baloch rebels in the Baloch 11 (1958-59) and Baloch III (1963-69) campaigns. After aid was resumed in 2001, driving a 7% economic growth rate, the Pakistani state still chose a low-effort strategy to fight insurgents in its peripheral territory of FATA. Even when aid escalated between 2006-09 and was implicitly conditioned on higher commitment, the state persisted with low-effort strategies in the FATA campaigns. Only one exception was made in a campaign targeting a particular group based in a small piece of South Waziristan that had wreaked havoc on Pakistan's urban centers. Even if one attributed the low effort in FATA to a broader strategy of hedging vis-a'-vis Afghanistan, the state capacity theory still cannot explain the continued loweffort in its fifth major campaign against Baloch insurgents. Tightly controlled comparisons with its Malakand 11 (2007-13) campaign reveals the Pakistani state is certainly capable of and proficient in high-effort counterinsurgency strategies like population control; it just chooses not to conduct these in low-valued regions. What further weakens the functionalist state strength argument is Pakistan's Baloch IV campaign strategy following a significant loss in capacity. After Pakistan suffered its greatest exogenous shock-the humiliating loss to India in the 1971 war that also cost it a fifth of its territory, many of its major ports, half its GDP, and significant losses of military personnel and materiel-the government still marshaled significant resources for its only high-effort attrition campaign in Balochistan from 1973-77. Though my argument cannot explain this choice of effort, the case poses a major problem for the capacity argument. In a period when Pakistan's 580 strength was at its lowest ebb and we should most expect its limited capacity to constrain costly and exhausting counterinsurgency campaigns, Pakistan deployed significant manpower and effort in a reckless, open-ended campaign to fight Baloch rebels, despite the absence of a serious threat. In this case, agency overcame structural conditions due to the political ambitions of Prime Minister Bhutto, but in nearly all the others, counterinsurgency agency would be exercised based on the importance territory rather than as a function of state capacity. The Indian cases pose further problems for the capacity argument. India is by most estimations a stronger state than Pakistan, but it has shown no greater overall predilection for high-effort counterinsurgency strategies. For the past twenty-five years, India has grown in strength due to dramatic economic strides, which have also allowed the country to insulate itself somewhat from intense external threat pressures that it dealt with for decades. Nevertheless, over this same period the state has persisted with low-effort campaigns against Maoist Naxalite insurgents despite their growth and expansion. India has certainly leveraged the increased capacity from its economic takeoff for much greater counterinsurgency efforts in the highly valued Punjab and Kashmir, but it has been unwilling to apply the same effort in the peripheral zones of the Naxal belt. The capacity argument simply cannot reconcile with the fact that the only time India deployed a high-effort strategy in the Naxal belt was in the Telangana campaign of 1948-52 when the newly independent Indian government was likely at its weakest point. Regime Type On face, regime type fails to anticipate the level of violence the state is willing to use. This is evident with a consistently democratic India that still made use of high-violence strategies in almost half its campaigns. It also over-predicts violence by non-democratic states like 581 Pakistan, which exhibited considerable restraint in 40% of its campaigns. The shadow cases in the following section confirm this in their depictions of democracies like Britain and Sri Lanka using high violence attrition strategies and non-democracies like Indonesia and Iran employing a high degree of restraint or working closely with the population. There is simply too much variation the regime type theory cannot explain. The problems of regime type persist when evaluating temporal variation within the same country. As the character of Indian democracy varied over time around the 1980s amidst greater political competition and a shift away from one-party dominance and rule, the state continued with high-violence counterinsurgency strategies in the Northeast and Kashmir, specifically directed by civilians, even as it proved willing to act with restraint in Punjab and the Naxal belt. The few instances when Pakistan's civilians also ruled did not usher in greater restraint against rebellion and detailed qualitative evidence shows that civilian leaders played a key role in encouraging the brutal strategy in 1971 East Pakistan even if they were not directly involved in its execution. There may be some truth that the upper and lower bounds of permissible violence to counter rebellion are in part mediated by the states' prior relationship with its citizens, often measured in regime type. India never conducted a campaign involving the level of violence that Pakistan did in East Pakistan, though for a variety of geographic and geopolitical reasons, this was a particularly unique case. Some argue India could never be capable of such brutality though the Kashmir campaign renders this claim questionable. Even if it offers some cross-national traction, the regime type argument cannot account for the sub-national variation exhibited by multiples states. Without exploring the state's particular pre-war relationships with specific groups, institutional theories are not equipped to explain when or why democratic states 582 sometimes use brutal force against their own rebelling citizens or when and why autocratic states are sometimes restrained from employing the levels of violence they are ordinarily accustomed to deploying. It is in this gap that the theory of core and peripheral identity groups makes a valuable contribution. Reputation One benefit of the medium-N approach is that it allows for a longitudinal analysis of counterinsurgency campaigns by the same state across multiple decades to evaluate reputational arguments, which are not easily assessed in narrow temporal windows or with just a few in-depth examinations of particular campaigns. While the reputation argument does a reasonable job of explaining why states fight groups that are mobilizing to challenge state authority, it does not offer much purchase in explaining how India and Pakistan would fight different rebel groups. The cases in this study expose too many inconsistencies with the reputation theory's expectations of early accommodations, persistence, publicity, and continued violence even after signaling toughness. The reputation theory not only expects states to use significant violence against early rebels but it also expects deliberate unwillingness to accommodate groups and sustained commitments to fighting. Both India and Pakistan did use high levels of violence in some early campaigns but also betrayed a number of inconsistencies. Early in its history, India made concessions to violent ethnic group mobilizations (e.g. Marathis) and even vocally secessionist groups (e.g. the Tamils and DMK). One of the reasons for the proliferation of insurgencies in India's Northeast is that alongside its sporadic brutality, India cultivated a reputation for cutting side deals with sections of rebel groups that incented many more to take up arms against the 583 state. These halfhearted accommodations, beginning in Telangana and Nagaland, were always too shallow to offer any redress of grievances or substitutes for brute force but just accommodating enough to incentivize new groups (like the Mizos, Meiteis, and Assamese) to rebel in pursuit of club goods. Meanwhile, prior to conflicts, Pakistan seems to have been unconcerned with its reputation, making some early constitutional concessions to Bengalis and turning a blind eye to Pashtun secessionist activity. The fact that the state never used high violence even against early Pashtun separatists suggests that the state's relationship with particular identity groups trumped the strategic need to cultivate a tough reputation. At times, Pakistan certainly cultivated a reputation for vindictiveness, but not for toughness. The Pakistani military would routinely punish Baloch rebels but then continually walk away from the conflict prematurely before seriously degrading insurgent capacity, signaling a lack of persistence to deter future uprisings. Additionally, neither India nor Pakistan displayed a penchant for publicizing their violence to deter future challengers as the reputation theory expects. Both states closely guarded details about their early counterinsurgency campaigns, recognizing these actions ran contrary to the norms and aspirations leaders had espoused in the independence from Britain and the founding of their states, but they did publicize their early accommodations. The reputation theory also expects a decline in the reliance on violence after a state has sufficiently signaled its toughness, but neither India nor Pakistan have exhibited this decline over time in their more recent campaigns. India persisted with high-violence campaigns in Kashmir and its Northeast and Pakistan continues this in Balochistan even after both states have demonstrated their willingness and ability over multiple decades to crush rebellions with brute force. 584 Overall, the Indian and Pakistani states simply did not perceive the reputational costs of managing insurgents in peripheral territory to be on par with actually losing or forfeiting the territory. The reputational consequences of tolerating some non-state armed actors challenging state authority were likely mitigated by asymmetric power dynamics where the state always retains the capacity to thrash them when it so chooses. On average, the use and permissibility of certain types of tactics has changed, and involves less violence and coercion than the days when the state did not blink when bombing civilians or herding hundreds of thousands of people into prison camps. This change, however, does not stem from predictions of reputation or regime types but rather from overall shifts in norms of war and greater media scrutiny. Nevertheless, India and Pakistan are still quite capable of discrete but nasty violence, particularly when fighting against groups with peripheral, low-worth social base as demonstrated by the Kashmir and Baloch V campaigns. In other examples around the world of recent, high violence counterinsurgencies by states that have already acquired a reputation for toughness (e.g. Sudan, Sri Lanka, and Syria), it is difficult to believe that the identity of the insurgents is not a factor in the intensity of violence. Insurgent Capabilities and Goals It is difficult to deny that states consider insurgent capabilities and goals in formulating strategy, but because these variables are extremely difficult to pin down, especially early on, and are endogenous to conflict strategies, objective measures of rebel attributes encounter significant limitations explaining strategy. Theories of insurgent capacity and goals are better at explaining certain macro-features of a conflict (such as its intensity and duration) in hindsight, after the conflict is over. They are far less useful at explaining or anticipating the state's strategy within 585 that conflict. The 29 India and Pakistan cases only provided strong support for the theories of rebel size and aims in 8 and 10 cases respectively. The rest of the cases could only draw partial support and in a handful of cases, rebel size (6) and rebel aims (4) would anticipate the complete opposite of the actual campaign. Rebel Size/Strength. In many ways, we can control for the role of insurgent capabilities because as is the case with asymmetric conflicts, nearly every single rebel group is judged much weaker than their state adversary except for a handful. The difference of a few thousand fighters then does not alter this fundamental military balance with the Pakistani and Indian states that each command between 500,000 to over 1 million well-trained military and security forces, which is why the relative power of nearly all of the 37 Indian and Pakistani insurgent groups listed in the violent non-state actor database are coded as "weaker" or "much weaker." 4 A major hurdle for insurgent strength arguments is that within this range of weak to very weak insurgent groups, it is difficult for states to assess a group's strength ex ante when formulating a strategy. Most assessments of the figures collected in large-N datasets are based on data discovered expost.5 Furthermore, insurgent size is in constant flux and often endogenous to a state's strategic choices. This makes estimates of size difficult to pin down long after the conflict has ended, leave alone in the midst of the conflict. The medium-N analysis and in-depth studies reveal the Indian and Pakistani states both made wildly inaccurate estimations of 4 I am referring to the Non-state actor dataset (version 3.3, January 24, 2012) presented in David E. Cunningham, Kristian Skrede Gleditsch, and Idean Salehyan, "It Takes Two: A Dyadic Analysis of Civil War Duration and Outcome," Journalof Conflict Resolution, 53 (4), 2009, pp. 570-597. The Mukti Bahini, TNSM, and TTP are coded as having "parity" with the Pakistani state but even these codings are highly problematic. The Bengali insurgents were hardly at parity at the onset of the counterinsurgency campaign in March 1971. Even after intensive training by the Indian state had boosted their capacity by August or September, they still may not have reached parity without the backing of regular Indian Army units. Meanwhile, the TNSM and TTP judgments made are likely a result of a recency bias but, while formidable, no experts on Pakistan militancy would describe them as having parity with the Pakistani military. 5 Valentino et al, 2004; Cunningham et al, 2009. 586 insurgent sizes at different points during the various campaigns, estimations that sometimes varied by an order of magnitude and are still disputed today. The introductory accounts from chapters 4 and 6 set up controlled tests where the same state faced similarly sized insurgencies yet adopted different strategies against each. But even if this study did not attempt to control for this variable, insurgent size would not be able to account for the variation in strategy. Even taking insurgent strength at face value, in a number of cases, the Indian and Pakistani state acted with little sensitivity to the insurgent size. The state was willing to use brute force against sparse numbers of weak insurgents like the Baloch in 1948, 1958-59, and 1963-69 and even conducted a high-effort attrition campaign to crush the few Sindhi militants amidst what was mainly a large non-violent political movement in 1983. Against relatively stronger and larger groups such as against the better equipped, organized, and capable MQM and TNSM insurgents, the state managed to restrain its violence, and in some cases, effort. This is to say nothing of the patience, timidity, limited effort, and restraint the state exhibited towards the very large, very capable Pakistani Taliban during the first and even the second campaign in FATA. India has also unleashed tremendous violence against small, unskilled, poorly armed tribal-based insurgencies in its Northeast, but stills exercises restraint against an expanding and improving Naxal insurgency that today numbers in the tens of thousands, has a huge potential social base, and its readying themselves for a people's war and eventually a positional war. Rebel Aims. A theory of rebel aims performs slightly better than rebel strength, primarily in anticipating the level of state violence, but in many ways, this explanation suffers from similar problems of states being unable to formulate strategies based on rebel aims they cannot divine ex ante, or even during a rebellion. Instead, they more often than not inferred insurgent aims and 587 estimated insurgent threat levels based on their subjective assessments and prior assumptions, often derived from beliefs about the insurgents' ethnic identity, group character, and loyalty. Though rebel aims in most of the Baloch rebellions (including the recent Baloch V during the early years) and in the rebellions of Sindh did not actively seek separatism, their ethnic social base and state judgments about their group character and loyalty led the state to infer or anticipate their separatist intentions. The state never did this with more proximate identity groups and consistently acted with restraint, for instance, even when it faced Pashtun separatist uprisings and rebellions intertwined with demands for an independent Pashtunistan. Calls for Pashtun separatism were never as alarming to the state as those by Baloch or Sindhis because the Pakistani leadership had confidence that Pashtuns' conservative group character and Islamic identity ensured loyalty while their ties to the state afforded channels of communication, bargaining, and leverage. The Pakistani Army also displayed restraint against insurgents from the well-entrenched Mohajir community, and downplayed the MQM's separatist aims or plans for an independent city-state, which was later discredited. In all these cases, rather than rebel aims dictating strategy, the rebel group's identity mediated the state's perceptions of their aims. Indian leaders have also inferred intentions from identity. The separatist ambitions of the more embedded Sikh and Assamese rebels were and continue to be discounted, resulting in more restraint in the use of violence. Meanwhile the separatist ambitions of Manipuris and Nagas, who for decades have primarily sought meaningful empowerment or autonomy, continue to be overexaggerated allowing for continued brutality, particularly in Manipur. In Kashmir, even organizations with more moderate demands were still targeted with brute force, and despite demands for separatism diminishing considerably, the Indian state continues to wield tremendous repressive force in the region. By contrast, the state has been relatively restrained and unmoved 588 by a Naxal insurgency that openly works towards an independent revolutionary zone as an intermediate goal, tantamount to separatist ambitions. External Deterrence/Support Theories of deterrence based on external backing proved problematic because of the ubiquity of perceived external support for insurgent groups in nearly all of these South Asian conflicts. Even if some of this evidence is suspect, it matters less than the volume of evidence suggesting that both India and Pakistan routinely suspected their adversaries (primarily each other but sometimes others like Russia, China, and Afghanistan) had a hand in supporting various insurgencies, many of which had some truth to them. Differentiating degrees of support was also difficult because this could escalate endogenously for various reasons. Furthermore, the theory fails on a number of important cases. In rare instances where there was zero documented external support in the Telangana and Naxal I-AP campaigns, the Indian state deployed strategies of high violence in both. In the Punjab and Assam cases where there was clear evidence of Pakistani support arming, training, and harboring insurgents, the Indian state acted with tremendous restraint. Similarly, the Pakistani state demonstrated visible restraint against the MQM in Karachi despite publicly declaring it had received arms and training from India's external intelligence. The Pakistani government displayed the same restraint against Pashtun militants throughout its history but particularly in Waziristan in the 1970s despite an escalation in support for armed Pashtun nationalists, including basing and training by the Indianbacked Afghan government. 589 Sons of the Soil Despite accounting for persistent rebellion and conflict duration stemming from resistance to mainland settlers, the problem with theories "sons of the soil" (SOS) rebellions is that they offer few theoretical accounts or specific predictions of strategy. When examined, this explanation found little support as plenty of SOS rebellions met with decisive, high-effort campaigns such as Punjab and Karachi-MQM, while many non-SOS rebellions still elicited loweffort campaigns and protracted conflict (in FATA and the Naxal belt). Overall, the SOS theory could only explain state strategies of effort in about half (15) of the cases while offering zero purchase on state strategies of violence. Temporal Theories Other temporal theories of state counterinsurgency behavior such as international norms and learning are not wrong but are too generalized to offer specific predictions that explain subnational variation. On average, most India and Pakistan's counterinsurgency campaigns today are less brutal than in the 1940s and 1950s, but the norms theory cannot explain why highviolence strategies are regularly revisited in some campaigns, even by democratic regimes that have demonstrated a capacity and willingness for more restrained approaches elsewhere. Similarly, state counterinsurgency learning is occasionally evident over time across campaigns but more often within campaigns like Punjab, Kashmir, Malakand, and (to a degree) FATA. Nevertheless learning is still bounded by the specific state motivations stemming from the conflict's contested territory and contesting identity group. This explains why India and Pakistan have respectively persisted with strategies in the Northeast and Balochistan, which have failed to 590 adapt to defeat rebellion after more than six decades, even as they have developed the tools for high-effort campaigns to regain full control of other embattled territories. III. External Validity After demonstrating the theory's internal validity as well as its superiority to alternative theories in explaining conflicts from India and Pakistan, the next question to address is whether the argument travels outside of India and Pakistan. Evaluating the external validity of my argument is difficult because testing the theory requires close examination of multiple dimensions of a counterinsurgency campaign along with the evaluation of politico-economic topography and social hierarchies within a country. These onerous demands for in-depth case studies are the reason why this project confined itself to two countries from a single region. Nevertheless, some evidence suggests that different parts of the argument travel to other regions and incumbents-including Britain, the Soviet Union, Indonesia, Sri Lanka, Colombia, Iran, Iraq, and Uganda-and that the variables of territory and identity shape the counterinsurgency strategies of multiple types of incumbents: strong, weak, domestic, and foreign. Britain Earlier accounts offered in chapter two reveal patterns by other foreign incumbents, notably the British and Germans, which support the theory's predictions of rebelling peripheral groups being the target of more intensive violence. The Germans employed greater violence and brutality against the more reviled Slavic rebels on the Eastern front than against rebels more proximate in ethnic identity on the Western front, notably French resistance. However, the British are the most interesting incumbents because despite professing a single model of 591 counterinsurgency best practices, both the territory and identity seemed to drive significant variations in the British strategic approach. Evidence presented in figure 2.1 from chapter two confirms the British adhered to the theory's predictions of core territory. Higher British efforts appear to have been invested in regions of greater import to British strategic interests such as Palestine and the Canal Zone, or areas that were a core part of British territory like Northern Ireland. Regions like Oman, Nyasaland (Malawi), and British Guiana drew less effort because they offered little value to British interests relative to the costs of sustaining control. The British experience also comports with my argument's expectations of identity and violence. Though not far apart in time to allow for an evolution in doctrine, British counterinsurgency practices in Northern Ireland against social groups much closer in identity to British security forces were relatively less violent and much more restrained compared to the type of violence it used in counterinsurgency campaigns against nationalist struggles in Malaya and Kenya. Focusing for a moment on its foreign expeditionary campaigns, the British demonstrated a remarkably varying appetite for brutality, permitting a considerable amount against black Mau Mau rebels of Kenya while permitting far less violence against Zionist insurgents of an Ashkenazi Jewish social base. Greater restraint in Palestine may have stemmed from an emergent sympathy for the Jewish people in the aftermath of the Holocaust, as well as some influence of the Jewish community within the British government, notably Baron Rothschild, to support a homeland for the Jewish people.6 For instance, the 1917 declaration issued by British Foreign Secretary Arthur Balfour supporting a Jewish homeland, was in part influenced by the Zionist community and the prominent Rothschild banking family. 6 592 Soviet Union In a harder test of the theory, the Soviet Union (USSR), not known for restraining its use of force, also offers some confirmation of the identity variable in its campaigns against Ukrainian and Lithuanian insurgencies between 1946-52. The Soviets dispensed totally indiscriminate violence against the Lithuanians, like they had with many other ethnic minority groups, but the state eventually adopted more moderate and selective approaches against the Ukrainian insurgents. The normal Soviet counterinsurgency template was deemed untenable where Ukrainians were eastern Slavs and close ethnic kin of Russians. Ukrainians were also amongst the best-integrated and most loyal ethnic groups to the USSR. Composing 3.5 million in the Red Army (including 750,000 from Western Ukraine), more than twice the number of Ukrainians served in the Red Army than fought in the resistance. They also held key posts within the Soviet military leadership including the Commander in Chief of Soviet Ground Forces who later served as Defence Minister.7 This pre-war relationship to the people appears to have produced a different style of Soviet counterinsurgency strategy. One scholar writes, "Despite the dangers the separatists posed, Ukrainians as a nation were not singled out as traitors. ... A special effort was made to link Ukraine to the Soviet nation even as the war intensified divisions within the friendship of peoples." 8 Initially, the Soviets had begun with an approach based on high levels of violence but soon adapted to one based on propaganda, infiltration, selective violence, amnesties, and economic development. 9 At first, the Ukrainian People's Army's (UPA) reckless conventional offensives against dug-in Soviet positions drove an initial high-violence state approach, inflicting 7 Egor Eviskov, "War Against the UPA, 1944-1954: Soviet Counterinsurgency Challenge in Western Ukraine," Unpublished Paper, 2014, pp. 23-24. 8 Kenneth Slepyan, Stalin's Guerrillas:Soviet Partisansin World War H, Lawrence, KS: University Press of Kansas, 2006, p. 214. 9 Eviskov, 2014, pp. 34-35. 593 roughly 13,000 fatalities.10 Learning quickly from the consequences of raids and indiscriminate violence, in 1946 the Communist party began to crack down on security force excesses and changed up their methods to ensure legal compliance, prevent illegal searches, and limit mass detentions. It also began to emphasize selective violence through dispersed operations, intelligence cultivation, infiltration, and informant networks, 1 which in turn drove up the ratio of insurgents arrested to killed while reducing the number of unintended victims.' 2 As much as half the force fighting the counterinsurgency campaign was local Ukrainian militia matched by village defense forces.' 3 In what was "a major reversal in the Soviet counterinsurgency strategy," in 1946 the Soviets also shifted considerable attention to the demand side of insurgency. First, they sought to empower and motivate the lesser peasants to turn on the insurgents and their kulak base.' 4 Agrarian policies offered peasants material incentives in the form of tax breaks, and collectivization denied insurgents access to material supplies.' 5 Second, they inundated the region with resources and development. In the four-year plan of 1946-50, about 20% of the Soviet budget was invested in Ukraine. From 1944-54, factories were transferred out of the Russian Republic to Western Ukraine and 25 new factories were built resulting in employment and urbanization to undercut the insurgent base. 16 Third, the campaign involved a number of major political concessions to Ukraine that were both substantive and symbolic.' 7 Finally, the Soviets offered insurgents off-ramps from the conflict. Six rounds of amnesties were offered 10 Serhiy Kudelia, "Choosing Violence in Irregular Wars: The Case of Anti-Soviet Insurgency in Western Ukraine," East EuropeanPolitics and Societies, 27 (149), 2013, p. 159; Alexander Statiev, The Soviet Counterinsurgencyin the Western Borderlands,New York: Cambridge University Press, 2010, pp. 108-110. " Kudelia, 2013, p. 163. 12 Kudelia, 2013, p. 174; Statiev, 2010, pp. 110, 251-52. 13 Kudelia, 2013, p. 168; Statiev, 2010, pp. 220-24; Clodfelter, 2008. 1 Kudelia, 2013, p. 161. 15 Kudelia, 2013, p. 162. 16 Eviskov, 2014, pp. 32-33. 17 Evsikov, 2014, pp. 24-25. 594 between 1944-45 and roughly 40,000 insurgents took them.' 8 A final amnesty offer in December 1949 drew in another 8,000-14,000 insurgent surrenders.19 While collaboration amongst Ukrainians steadily increased over the course of the insurgency, this was simply not the case in the Forest Brothers rebellions of the Baltics due to the economic, political, and principally ethno-cultural differences. 2 0 While the Soviets were able to co-opt the Orthodox Church in Ukraine and bring the Uniate Church into the fold, they had a much harder time with the Catholic Church in Lithuania.2 ' The same indiscriminate forces used against the Kalmucks, Chechens, and Tartars were now used against the Lithuanians and this was reflected in the relative death toll (see table 2.1). In quite the opposite direction of the Ukrainian campaign, the leadership dispatched to Lithuania suspended normal procedures for arrest and investigation for being too "sentimental" and demanded "total war tactics" involving any means necessary. 22 Because initial attempts at cultivating militias were quashed, the Soviets had to rely on regular internal and state security forces and eight veteran combat infantry Red Army divisions as well as the air force for a more conventional style battles. Over 100,000 troops were stationed here for roughly 2.5 million people. Amnesties were introduced but with the threat of deportations, which occurred on a massive scale with some figures as high as between 100,000300,000 or 4-12% of the population relocated out of the region (often to Siberia). Furthermore Eviskov, 2014, p. 30. 19 Statiev, 2010, p. 199; Kudelia, 2013, p. 174. 20 Statiev, 2010, p. 116. 21 Statiev, 2010, pp. 258, 264-65. 18 2 V. Stanley Vardys, "The Partisan Movement in Postwar Lithuania," LituanusLithuanian JournalofArts and Sciences, 15 (1), Spring 1969. 23 This included Internal Affairs (NKVD) and Ministry for State Security (MGB) forces. See Daniel J. Kaszeta, "Lithuanian Resistance to Foreign Occupation 1940-1952," Lituanus LithuanianJournalofArts and Sciences, 34 (3), Fall 1988. 24 Kaszeta, 1988; Statiev, 2010, p. 177. 595 the destruction of assets and infrastructure visited upon the region during the guerilla war caused it to regress economically compared to the other Baltic countries. 25 Indonesia Counterinsurgency campaigns from Indonesia also seem to fit the predictions of the theory. As briefly described in chapter 3, Indonesia put tremendous effort into a campaign to defeat the Darul Islam insurgency, which threatened Western and Central Java, home to its capital and territorial core. During the campaign, the state sunk significant material resources and adapted its military organizational practices to develop a counterinsurgency strategy to defeat the rebels and regain control of the region. The same state would respond to a rebellion in the peripheral region of West Papua, but it would invest comparatively little effort in force size or force quality during that campaign and settle for a strategy of containment and enfeeblement of the rebel organization, the OPM. 2 6 The Indonesian campaign against the rebel organization Darul Islam pitted the Indonesian forces against Muslim insurgents, who had strong co-ethnic ties to the state and had previously fought alongside them against Dutch colonial rule. The state's strategy entailed less violence than the campaigns against ethnic minorities in East Timor or West Papua and involved a more selective, targeted campaign, improvements in governance, massive development investments, village defense forces and human intelligence, tactics to isolate and deny insurgents access to their support base, and de-radicalization for surrendered insurgents. In East Timor and West Papua, indiscriminate tactics prevailed including large-scale sweep, "search and destroy" 25 26 Vardys, 1969. Kilcullen, 2010, p. 85; Martin O'Hare, The Indonesian Military in IrianJaya, MA Thesis, Australian National University, 1991, p. 53; Reynders, "Consular Trip to West Irian," Confidential Airgram, May 10, 1968, National Security Archives; Richard Chauvel and Ikrar Nusa Bhakti, The PapuaConflict: Jakarta'sPerceptionsandPolicies, Policy Studies 5, Washington DC: East-West Center, 2004, p. 23. 596 - missions, and covert terrorism. 27 A quick estimation of accessible data on manpower and civilian casualties corroborates the theory's expectations (see figure 8.1). The Indonesian state would exert greater effort to regain control over the core provinces of Central and Western Java, as well as East Timor for its strategic value and material wealth, but it would employ less effort in the peripheral region of West Papua/Irian Jaya. Additionally, rebels from peripheral social groups like the mostly Christian Papuans and Timorese would be targeted for much greater violence than Muslim Sundanese and Javanese rebels.2 8 Figure 8.1: Distribution of Selected Indonesian Campaigns 0 East Timor - Fretilin West Papua - OPM 0 $3 Java Darul Islam 06 1 1 -1 -3 Log(Troops/Area of Op (AO)) Sri Lanka Though Sri Lanka may be a country whose expanded state capacity eventually enabled it to change strategy against Tamil insurgents in its embattled North and East between 2006-09, Kilcullen, 2010, pp. 87-95; Christopher Paul, Colin P. Clarke, Beth Grill, and Molly Dunigan, Pathsto Victory: DetailedInsurgency Case Studies, Santa Monica, CA: RAND, 2013, pp. 118-124; David J. Kilcullen, "The Political Consequences of Military Operations in Indonesia, 1945-99," PhD Dissertation, University of New South Wales, 2001, chapters 3-4 (pp. 28-160). 2 Data collected from Friedman, 2011; O'Hare 1991; Ulfelder and Valentino, 2008; Fearon, 2004; Kilcullen, 2010; "Battle Deaths Dataset 1946-2008" (version 3.0 September 2009) from Bethany Lacina & and Nils Petter Gleditsch, "Monitoring Trends in Global Combat: A New Dataset of Battle Deaths," European JournalofPopulation,21(2-3), 2005, pp. 145-166. 27 597 prior to that, it had only projected limited efforts to regain control of the embattled region (other than the urbanized Jaffna peninsula). For nearly two decades, the state largely sought to contain rather defeat the Tamil insurgency in its northern and eastern regions, 2 9 while it summoned the full power of the state to crush and dismantle two People's Liberation Front (JVP) insurgencies in 1971 and again in 1988-90. The JVP insurgency had operated in the country's core-the southern and western regions that contained the country's economic base, population concentrations, and its capital-and "shook the government to its foundations in a manner the Tamil insurrection in the North and East was never able to do.... The theory's differential predictions of violence against core and peripheral groups were also borne out in the Sri Lankan cases. The Sinhalese-dominated government of Sri Lanka conducted a much more brutal counterinsurgency strategy against rebels of the highly loathed and excluded Tamil minority, particularly between 2006-09, than against JVP rebellions in 1971 and 1988-90, composed primarily of Sinhalese Buddhists. The state counterinsurgency campaign against the Tamil insurgents employed an extraordinarily brutal attrition strategy, particularly from 2006 onwards including a final five-month phase that may have killed as many as 40,000 civilians.31 In contrast, the 1971 strategy against the first JVP uprising was quite mild compared to later strategies.3 2 While the campaign against the second JVP rebellion from 198890 involved a non-trivial amount of state-directed violence, the rebels, despite their nasty tactics, were reported to have evoked some sympathy within the security forces, hundreds of whom were 29 Ahmed S. Hashim, When CounterinsurgencyWins: Sri Lanka'sDefeat of the Tamil Tigers, Philadelphia, PA: University of Pennsylvania Press, 2013. C. A. Chandraprema, Sri Lanka: The Years of Terror: The J. V.P. Insurrection, 1987-1989, Colombo: Lake House Bookshop, 1991, pp. 4. Also see pages 134-37, 316 as well as Rohan Gunaratna, Sri Lanka: A Lost Revolution? The Inside Story of the JVP, Kandy, Sri Lanka: Institute of Fundamental Studies, 1990. 31 Hashim, 2013; Jon Lee Anderson, "Death of the Tiger: Sri Lanka's Brutal Victory Over its Tamil Insurgents," The New Yorker, Jan. 17, 2011; Gordon Weiss, The Cage: The Fightfor Sri Lanka and the Last Days of the Tamil Tigers, London: Bodley Head, 2011. 32 Laksiri Fernando, "April 1971 JVP Uprising: Not to Make the Same Mistakes," Colombo Telegraph, April 2, 2013 598 members of the JVP, and soldiers and officers lacked the stomach to seriously fight their own Sinhala brethren described as "our own boys." 33 Consequently, the police-led strategy involved more intelligence-driven selective targeting, less violence against civilians, non-kinetic tactics, protection of villages to secure the confidence of the local population, insurgent surrenders, broad negotiations, and rapid rehabilitation of the JVP as a legitimate political party and eventual member of the ruling coalition.34 Colombia The Colombian government also faced a series of insurgencies following the period of intense civil conflict, La Violencia, from a variety of groups, but for nearly 50 years, the state security forces concentrated their counterinsurgency efforts on fending off and driving out the urban-based insurgent organization. Rather than project comprehensive counterinsurgency efforts nationwide, the state conducted periodic strikes in rural regions, but never consolidated its hold. Insurgent organizations and components were left free to regroup and operate in the eastern plains and rural countryside known as the Lianos Orientales, and local municipalities forced to fend for themselves. Thus while the M- 19 guerilla organization was targeted for dismantlement and defeat, other insurgent organizations like the National Liberation Army (ELN) and Revolutionary Armed Forces of Colombia (FARC) were merely contained for decades until they grew to seriously rival state power. 33 Quote from Chandraprema, 1991, p. 186; Gunaratna, 1990, p. 327. 3 Chandraprema, 1991, pp. 287-292, 297, 300; Gunaratna, 1990, pp. 317-18, 327; Author interview, Former Prime Minister Ranil Wickremesinghe, Cambridge, MA, May 2, 2013. 3 Roman D. Ortiz and Nicola's Urrutia, "A Long Road to Victory: Developing Counterinsurgency Strategy in Colombia," Countering Terrorism andInsurgency in the 21st Century: Lessons from the FightAgainst Terrorism, Vol 3, Ed. James J.F. Forest, Westport, CT: Praeger Security International, 2007, particularly pages 312, 319-323, 328. Also based on Author conversation with consultant and advisor to Colombian Ministry of Defense Roman D. Ortiz, Toronto, Canada, March 28, 2104. 599 This differential treatment might be chalked up to state capacity and organizational learning, but detailed accounts argue this approach was partly a deliberate choice that stemmed from a distinction in the nature of threatened territory. Insurgents were only targeted for management but not defeat because their territorial focus and base-the 66% of the country east of the Andes mountain range-was peripheral to state interests and elites. Marks writes: ... The lack of concern by the ruling elite played a key role. For decades following La Violencia, the insurgents remained largely "out there," out of sight, out of mind, patiently building an alternative society. No one much cared. ... As long as the guerrillas were revolutionary homesteaders in areas no one else wanted, the government bothered with them only when their actions forced a response. 36 Iran Iran's bouts with rebellion in the 1970s to 1980s also exhibit elements of my theory's identity predictions, as the social composition of a resistance movement appears to have shaped the Iranian state's use of violence. The Iranian state and security forces had never encountered an uprising it could not crush such as the Kurdish uprisings prior to World War II, in 1946, and in 1967, but when the Shah's Iranian regime state faced urban guerilla movement from a Persian, Shia-based bazaari middle class, the state was hamstrung from using significant force, even in the early days of the protests, and it cycled between martial law, covert violence, and outright concessions. 3 7 The 1978-79 revolution amounted to an urban civil war that threatened the entire regime and involved violence against agents and institutions of the state. Bizarrely, the Iranian military, the Artesh, exercised considerable restraint despite being the main target of guerilla Thomas Marks, Colombian Army Adaptation to FARC Insurgency, Carlisle, PA: Strategic Studies Institute, January 2002: 4. 3 Said Amir Arjomand, The Turbanforthe Crown: The Islamic Revolution in Iran, New York: Oxford University Press, 1988, pp. 103-133; Gholam Reza Afkhami, The Life and Times of the Shah, Berkeley, CA: University of California Press, 2009, pp. 498-544; Stephen R. Ward, Immortal:A MilitaryHistory ofIran and Its Armed Forces, Washington DC: Georgetown University Press, 2009, pp. 137-138, 178-79, 181-185. 36 600 violence, only killing between 2,000-3,000 people in a year-long nation-wide uprising (potentially because of the social ties between the revolutionaries and agents of the state). The post-1979 Iranian revolutionary regime may have retained this approach of differentiated violence against ethnic minorities. As the regime consolidated its grip on the country in the aftermath of the revolution and during its war with Iraq in the 1980s, it faced a considerable number of rebellions by groups once aligned with it during the revolution against the Shah. During these campaigns, the same military that had been restrained during the revolution, brutally crushed rebellions by the ethnic "others," involving the Kurds, Turkoman, and Baloch, even more so than the urban leftist but ethnically Persian Fedayan-e Khalq.3 9 Iraq Recent developments from contemporary conflicts in Iraq showcase the continued relevance of the theory's territory predictions. In 2008, the ruling Iraqi regime embarked on an aggressive military offensive, Operation Knight's Charge, and comprehensive campaign to wrest control of Southern Iraq from Shia insurgent militias, principally led by Moqtada al Sadr's Jaysh al-Mahdi (JaM), as well as hold and reconstruct it. The state launched this ambitious campaign even despite the concerns raised by U.S. advisors about its capacity, and poured in an estimated 30,000 troops to retake the region from the militias threatened valuable territory including Basra, the second largest city and "Iraq's economic capital" which contained its only ports and maritime access critical to its shipping, logistics, transport, and petro-chemical industries.4 0 Arjomand, 1988; Ward, 2009, pp. 211-230. For estimates on casualty figures based on research by Emadeddin Baghi, see Cyrus Kadivar, "A Question of Numbers," Rouzegar-Now, August 8, 2003, (accessed August 21, 2014). 39 Ward, 2009, pp. 230-236; Arjomand, 1988, pp. 135-36; Amir Taheri, The PersianNight: Iran Under the Khomeinist Revolution, New York: Encounter Books, 2009, pp. 87, 96, 346-47, 349-59. 40 Quote from "Southern Iraq," Institute for the Study of War. See also "Operation Knight's Charge (Saulat AlFursan)," Institute for the Study of War; Marisa Cochrane, "The Battle for Basra," Iraq Report #9, Institute for the Study of War, June 23, 2008. All Accessed August 22, 2014. 38 601 Despite demonstrating such capacity, the Iraqi state remained relatively nonchalant over the resurgence of Sunni militancy in relatively peripheral desert regions of Western Iraq and Anbar province between 2011-13, largely ignoring it or doing just enough to contain it. 41 In 2013, despite possessing much larger and more capable security forces, the state began to deploy small numbers of troops (-8,000) over a vast area for limited combing operations along the border, and would pair this with another limited effort to back local tribal proxies rather than commit large numbers of their own troops. 42 Even when the Sunni Islamist insurgency expanded into major towns in northwestern Iraq, the state was willing to contain them there, until they began to march on major cities like Mosul and Baghdad.43 While there were certainly other factors driving these choices, including dysfunctional leadership of Prime Minister Nuri al Maliki, the neglect of the insurgency extended beyond him. Most state leaders only took notice and called for serious effort once the insurgents began taking urban population centers and advancing towards the capital in May 2014. Uganda The behavior of some African states in civil wars also lends support to my theory. Based on campaigns conducted by the Ugandan government, recent research concludes that the state has chosen "the least coercive and most co-optive counter-insurgency strategies" against those insurgent groups with linkages to the regime, drawn from the same political networks, while it 4' Firas Al-Atraqchi, "Did Iraq Government's Neglect Lead to Mosul's Downfall?" A Jazeera, June 11, 2014. Mohammed al-Qaisi, "Iraq Launches Counter-Terrorism Campaign Along Syria Border," Al Shorfa.com, May 23, 2013; Hamza Mustafa "Iraq Launches Operation Against al Qaeda in Anbar," Asharq Al-Awsat, May 27, 2013; Tim Arango and Kareem Fahim, "Iraq Again Uses Sunni Tribesmen in Militant War," New York Times, January 19, 2014. 43 Michael Knights, The ISIL's Stand in the Ramadi-Falluja Corridor," CTC Sentinel, 7 (5), May 2014, pp. 8-12; Ali A. Nabhan and Matt Bradley, "Islamist Insurgents Advanced Toward Baghdad," Wall Street Journal, June 11, 2014. 42 602 has employed more militarized and coercive responses towards true political outsiders." In the case of the Lord's Resistance Army, the state pursued something closest to a population control strategy because the rebel base had a strong lineage within Ugandan politics. In contrast, the state aggressively targeted the Sudanese backed ADF with impunity, pursuing a total "insurgent elimination" strategy, particularly after its ties to al Qaeda were revealed.4 5 Preliminary Quantitative Tests After identifying critical indicators in case studies, both to measure my independent and dependent variables and test their validity on a number of shadow cases, I offer some preliminary quantitative data to test whether the causal mechanisms "travel" or offer any "contingent generalizations" for a broader universe of counterinsurgency cases.46 To assess the impact of identity on violence, I collected data on the ethnic representation within a number of militaries to assess what impact it might have on the intensity of civil conflict. I drew on data from two major sources of ethnic composition within militaries 47 and supplemented it to generate 41 observations of civil wars where I could estimate the level of representation of an ethnic group and create a dummy variable for whether the insurgent group's base was at least represented in the military proportionate to its relative size. Using this measure, I estimated the difference in average intensity (average battlefield deaths per year) between conflicts where insurgents' identity group was represented in the military, and where it was not. On its own, representation seems to correlate with intensity. Even when using cross-tabulations to control for important variables like 44 Christopher R. Day and William S. Reno, "In Harm's Way: African Counter-insurgency and Patronage Politics," Civil Wars, 16 (2), 2014, pp. 105-126. Quotation from p. 107. 45 Day and Reno, 2014, pp. 116-117. 46 George and Bennett, 2005, p. 32. 47 Roger Petersen, "The Organization of the Military in Multiethnic States," Unpublished MA Thesis, University of Chicago, 1988; Chandra and Wilkinson, 2008. Of the 19 countries, I sought to supplement this with further information on 9 of them. I hope to expand this dataset and subsequent statistical tests. 603 regime type, insurgent aims, or type of rebellion, representation still seems to have some influence (Table 8.2). Though more sophisticated tests are required with a wider range of observations, this preliminary evidence suggests representation plays a role in the intensity, and potentially the strategy of an incumbent. Table 8.2: Relationship Between Representation in the Armed Forces on Conflict Intensity Avg. Intensity with Representation 6,854 Obs (16) Avg. Intensity with NO Representation 48,912 Obs (25) Difference in Intensity 42,058 1,230 (3) 1,404 (5) 174 8,152 (13) 60,789 (20) 52,637 1,631 (4) 2,383 (7) 752 8,595 (12) 67,007 (18) 58,412 Separatist Center-Seeking 5,933 7,775 (8) (8) 56,766 32,201 (17) (8) 50,833 24,426 Irregular "Regular" Conventional Symmetric NonConventional Conventional recoded 48 Irregular recoded 7,340 5,786 7,125 (11) (5) (4) 8,128 121,418 136,345 16 (9) (8) 788 115,632 129,220 429 (1) 2,000 (1) 1,571 7,125 7,340 (4) (11) 97,892 39,489 (6) (18) 90,767 32,149 (N=41) All Cases Democracy (Polity Score of 7+) Non-Democracy (Polity Score of <7) Democracy (Polity Score of 5+) Non-Democracy (Polity Score of <5) I conducted a similar "smell test" of the other axis to determine whether relative value might predict effort on a broader sample of cases. Taking urban territories to be a rough proxy for high value areas, my theory would expect that incumbents would deploy a greater percentage of their military relative to the size of the area when urban territories were the fundamental sites 48 1 disagree on the coding of the 1971-East Pakistan case and the 1983 Sri Lanka case, and changing their coding shows the influence of representation even more pronounced. Kalyvas and Balcells (2010) acknowledge the uncertainty of the1983 Sri Lanka case. The East Pakistan case is one where, despite some defections from paramilitaries and training from the Indian government, the conflict prior to Indian entrance in Dec of 1971 was highly asymmetric and most of the rebel fighting irregular. 604 of contestation. Data was collected from published counterinsurgency datasets, and combined with unpublished data on urban and rural insurgencies to assess the levels of military effort employed. I found that on average, this prediction held up to be accurate, particularly when holding aside mixed territory and comparing the effort exerted by states in only urban-based insurgencies versus rural ones.4 9 Table 8.3: The Relationship Between Territorial Value and State Incumbent Effort Trichotomous Coding Avg Effort in Urban Obs Avg. Effort in Territory Mixed Territory 4.24E-05 25 Obs Avg. Effort in Obs Difference in Effort Rural Territory 2.71E-05 26 3.28E-05 51 29.28% Avg. Effort in Obs Avg. Effort in Obs Difference in Effort 51 Rural Territory 3.28E-05 51 4.98% Dichotomous Coding Mixed Territory 3.45E-05 While these quantitative assessments are merely suggestive and demand much more inquiry, combined with the set of shadow cases above, they offer further support that my argument possesses broader generalizability beyond India and Pakistan. In future developments of this project, I intend to develop the quantitative analysis into more meaningful statistical tests by employing regression analysis on an expanded set of observations and control variables. Beyond Counterinsurgency Since the theory is as much about projection of state authority, it makes sense that the key variables might have some explanatory power outside the realm of counterinsurgency and civil wars. For instance, sociologists have observed that in some parts of urban areas, notably housing projects, the police generally defer day-to-day enforcement of political order to local, informal 49 Friedman, 2011; Nicholas T. Calluzzo, "The Urbanization of Insurgency: Shifts in the Geography of Conflict," Unpublished Masters Thesis, Massachusetts Institute of Technology, 2010. 605 authorities. Political authority is not only ceded to the informal complementary institutions like civic and community organizations that social capital theorists praise, but administration of order is often yielded to ostensibly antagonistic, rival political institutions like armed gangs or criminal networks.50 Though a regular feature of urban centers throughout the underdeveloped world, this tacitly accepted practice in capable, developed countries like the United States indicates there are real limits to theories of state capacity and control. Certainly the thresholds of acceptable violence are lower in developed countries but the fact that one of the most powerful and developed states consciously accepts the limits of its reach and disciplinary power suggests segmented sovereignty or ungoverned spaces may be more commonplace than anomalies. 1 Beyond rebellion, some of the theory's identity mechanisms are also found in other types of state-subgroup interactions. Social and ethnic ties between the state and distinct subgroups can account for: Protestant self-defense militias evading the same degree of targeting or coercion as Catholic militias during Northern Ireland's "Troubles," 52 Italian-Americans being subject to less suspicion and internment than suffered by Japanese-Americans during World War II,53 and an Israeli state's struggling, "halfhearted response" to manage Zionist settler violence when they have never lacked the capability or will to deal with Arab Israelis or Palestinians. 54 The common thread is that the identity of a particular sub-group mediates its relationship with the state, 50 Sudhir Venkatesh, GangLeaderfor a Day: A Rogue Sociologist Takes to the Streets, New York: Penguin, 2008. 51 Graham Denyer Willis, The Killing Consensus: Police, OrganizedCrime and the Regulation ofLife andDeath in UrbanBrazil, Berkeley, CA: University of California Press, forthcoming; Laurent Gayer, Karachi:Ordered Disorderand the Strugglefor the City, New York: Oxford University Press, 2014; Paul Staniland "Cities on Fire: Social Mobilization, State Policy, and Urban Insurgency." ComparativePoliticalStudies, Vol. 43, No. 12 (December 2010), pp. 1623-1649; Naseemullah, 2011; Staniland 2012. 52 Owen Bowcott, "MoD took Softer Line on Loyalist Paramilitaries, Secret Files Reveal," The Guardian,October 10, 2010. 5 Mary Elizabth Basile Chopas, "Law, Security, and Ethnic Profiling: Italians in the United States during World War II," PhD Dissertation, University of North Carolina, Chapel Hill, 2013. 54 Daniel Byman and Natan Sachs, "The Rise of Settler Terrorism: The West Bank's Other Violent Extremists," ForeignAffairs, 91 (5), September/October, 2012: 73-86. 606 particularly when it comes to issues of loyalty and national security. In cases of civil war or rebellion, this differential treatment becomes even more evident. IV. Theoretical and Policy Contributions This project has made a number of theoretical contributions to the study of civil conflict by improving the conceptual refinement and measurement of important variables and mechanisms, identifying the limits of structure and agency, corroborating and refining other theories of conflict dynamics, and adding to our empirical knowledge of South Asia. It also offers a number of insights for policymakers interested in conflict prevention and stabilization. Theoretical Contributions New Variables, Mechanisms, and Measures The first theoretical contribution made at the outset of the project is chapter two's distinction between fighting hard and fighting violently, which forms the crux of the dependent variable. Theoretically and empirically, these are revealed to be very different dimensions of counterinsurgency strategy, which helps scholars refine a typology that can include the theoretically novel but empirically frequent combination of low effort and high violence. However, the distinction of the two dimensions of counterinsurgency also poses a larger theoretical implication: war making need not necessarily involve state-building (or nationbuilding) as state leviathans can restore sufficient authority through cheaper options. Furthermore, there is nothing sacrosanct about retaining an even grasp throughout the state's inorganic territorial boundaries or its polity. Many of the regions that comprise a country are artificially bestowed and maintained through international events, structures, and rules, but the 607 state may not have ever exercised infrastructural or disciplinary power throughout this territory. This distinction between the Westphalian boundaries of the state and the Weberian practice of the state, between a state's territory and its polity, may lead it to distinguish between core and periphery, and therefore exercise caution to avoid overextension. This is not dissimilar to the case of American political development during the antebellum period when the United States purposefully chose not to expand its administrative power to its territorial frontiers, instead proceeding selectively and at times employing cheaper forms of control. Additionally, the theory provides an explanation for a dependent variable, strategy, that is often cast as an independent variable to explain outcomes. Elements of strategy, such as 58 manpower5 6 or violence5 7 or even the interaction of the two with the concept of force have been used to explain counterinsurgency and civil war outcomes, but these arguments have remained muted on where these strategies come from and why these best practices are not emulated by security-seeking states. The core-periphery theory illuminates when and why states might choose what some consider "suboptimal" strategies-limited in effort or excessive in violence-because they are instrumental to a distinct theory of victory.59 Second, the theory also contributes to our understanding of mechanisms that link identity to strategy via a particular set of tactical choices and opportunities. We know from a lot of recent scholarship that despite the low expectations of large-N cross-national studies, identity plays an important role in conflict dynamics. Now we have developed ways to assess how, within conflict, identity mediates the assignment of motives and suspicion, the attribution of 55 Paul Frymer, "'A Rush and a Push and the Land is Ours': Territorial Expansion, Land Policy, and U.S. State Formation," Perspectives on Politics, 12 (1), March 2014, pp. 119-144. 56 Friedman, 2011. 57 Kalyvas, 2006; Johnston, 2009. 58 Arreguin-Toft, 2005. 59 The term "suboptimal strategy" is used by Caverly, 2009/10, p. 121. 608 responsibility, the estimation of the threat, the scope of targeting, and the permissibility of different degrees of violence. The mechanisms are particularly versatile and robust because they can explain behavior at the individual level but also aggregate up to the tactical, operational, and strategic level. Particular individuals, units, or the national public and leadership can all harbor empathy or trust for a group, experience vulnerability because of that group, or leverage information about that group and members within it. Third, the argument also offers a more precise method for estimating the territorial stakes based on material, strategic, and ideational or historic value. This argument does not take issue with the rationalist theoretical approach to estimating territorial value but rather the empirical approach generally employed, which has routinely done a lazy job of estimating worth (relying on dummy variables of extractable resources rather than focusing on production), and discounted costs of an operation for sustainment (which are profound in counterinsurgency amongst different types of physical and human terrain). Estimating the net worth of a territory requires more than super-aggregated, macro-level information used in most cross-national datasets to sharpen our understanding of regional variation within a country. One cannot simply rely on the presence of natural resource deposits or the distance from the capital, which treat Sindh and Balochistan or Tamil Nadu and Manipur as co-equally important. My method for assessing the strategic and ideational value based on a robust understanding of the institutional history and patterns of contestation in a region provides a more precise assessment than counting up the number of bordering countries. 609 Defying Structure, and Calibration Fourth, my argument challenges structural theories of constraints-state capacity, regime type, and organizational culture-by both showing how these features either cannot account for sub-national variation or poorly predict cross-temporal variation. While weak states certainly face a ceiling on their capability without external assistance, even states like India and Pakistan, which fall within the bottom third of most cross-national measures of state strength, still possess moderate capacity to engage in comprehensive, high-effort, or "ideal-type" counterinsurgency campaigns. These states with moderate strength also possess the agency to choose not to extract and deploy sufficient resources for a campaign, instead preferring to manage or contain violence. Similarly, organizations can actively evolve out of certain pathologies if pressured by the territorial stakes, as the Pakistani military did under severe pressure in Malakand II. Even if these pathologies prevail, many sufficiently motivated weak states still have options-military, paramilitary, police, intelligence, and irregulars-to supplement, substitute or compete with each other to field a high-effort strategy, as India did with its array of security forces in Punjab. Fifth, my argument demonstrates the limits of rationalist strategic interaction theories that fully embrace endogeneity; while they reveal important dynamics in retrospect, they cannot anticipate the counterinsurgency strategies of the incumbent. Strategic interaction theories presume that both violence and effort are carefully calibrated to threats posed by an insurgency, whether the reputational stakes, external support, insurgent strength or aims, or local balance of power. This seriously neglects the cognitive biases and pathologies that filter perceptions and shape intelligence, 60 and assumes perfect information and attribution that incumbents can only wish they had. My argument demonstrates how the lens of identity, laden with emotional 60 Robert Jervis, Why IntelligenceFails: Lessonsfrom the IranianRevolution and the Iraq War, Ithaca: Cornell University Press, 2010; Richard K. Betts, Enemies ofIntelligence: Knowledge and Power in American National Security, New York: Colombia University Press, 2007. 610 baggage, ideational priors, and bankrupt assumptions, filters what limited information trickles in and fills the void so that goals are misinterpreted, actions are misattributed, and strength, reputational stakes, and external links are all misestimated to fit a conflict narrative that state leaders already possessed. The limitations of my argument are that it has a harder time anticipating fine-grained shifts over time or variation between different militant organizations from the same identity groups within the same conflict, though this was always beyond the scope of my argument. This project has intended to predict strategy at a higher level of analysis-the generally broad and long-term strategy of a campaign-even if there are periodic tactical shifts. Future work in the field could begin to explain this fine-grained variation in time and across groups with greater precision by moving beyond the cues offered by the broad identity of groups to the specific political relationships between particular militant organizations and the state.61 OrganizationalLearning Sixth, the empirical work supports other scholarship on what drives organizational learning and adaptation from the military's "operational code" to a more complex strategies of two-way politics and violent negotiation. 62 Task pressure, and civilian intervention to a degree, prompted changes and refinements in strategy over time, but this pressure and intervention was still intertwined with territory and identity. Task pressure, stemming from a belief that the state's overall security or survival was at stake, still depended on the territorial stakes. Insurgency and the risk of failure in core regions of Kashmir and Punjab applied high task pressure to the Indian state, prompting major adaptations in both campaigns, but in the Northeast and the Naxal zones this did not occur. Similarly, the insurgent threat to core territories in East Pakistan and the 61 62 Lalwani and Staniland, 2014; Day and Reno, 2014. Jackson, 2008. 611 "settled areas" of KPK during the Malakand II campaign pressed the military to institute significant and costly organizational changes because what was threatened was too important to risk. At the same time, the threat to peripheral territory in Balochistan and FATA, even during periods overlapping with Malakand II, did not prompt this organizational adaptation or its application to those campaigns. While civilian leaders generally held the reigns of power in India, they actively intervened when they were most concerned about their own are being killed, forcing changes in strategy to less violent approaches. This contributed to the organizational changes in Punjab, and to less violent strategies in the Naxal belt, both of which principally relied on policing forces rather than the Army. However, even when civilians are institutionally empowered, this does not mean they will always intervene. The type of intervention in Punjab was starkly absent in the Northeast and Kashmir where peripheral identity groups did not arouse the empathy of civilian leaders to compel changes. Indian civilian leadership not only punted to the military on Kashmir but empowered intelligence agencies to indulge in even more brutality. In Pakistan, civilian intervention was less potent, but the chorus of civilians pressing for restraint against Pashtun rebels was muted with regards to the more scorned and less embedded Baloch insurgents. The one area where this departs from the operational code thesis is the role of resources for the military. While Jackson contends that limited resources propel organizational learning, this project finds no definitive effect of resources alone. Resources can fuel organizational learning as they did in Punjab and Malakand and their denial can limit strategies to repetitive containment strategies that never resolve a conflict. Resources can also be applied to some campaigns but poorly disciplined and squandered as they were in India's Northeast or in Pakistan's Baloch IV campaign. The resources variable may then be more salient to foreign 612 incumbents incurring costly overseas force deployments, but it may not have the same effect on domestic incumbents that can cheap sustain containment strategies. Limits of Control Seventh, this theory confirms Kalyvas's insight that territorial control plays a pivotal role in state strategy, 63 but it also departs from his work by revealing how control has a more complex relationship to violence. Though my project concerns a slightly different level of analysis, focusing on meso-level variation in violence while Kalyvas is principally concerned with explaining micro-level violence, my argument offers a number of theoretical caveats and challenges to his theory of control determining the distribution of selective violence. Kalyvas treats control as exogenous to strategy, and a given based on relative force presence, while the first half my argument sets out to endogenize levels of control by arguing it is a strategic choice based on the level of effort a state is willing to exert. Empirically, I find that there are a number of states that purposefully do not seek control over a region contested by rebels, even though they are willing to fight. This makes a larger contribution by examining the intermediate zone of cases, previously ignored or discounted, between fighting hard and not fighting where an incumbent turns to violence against rebellion, but in a halfhearted manner. My argument then disputes the relationship between control and violence by demonstrating state agency within both spheres of limited and strong control. In pockets of limited control, states can choose heavy, indiscriminate violence of enfeeblement but they can also work through localized informal agents like tribes, militias, warlords, or even rival insurgent factions, not fully allied with the state, to manage rebellion while still limiting violence. The corollary is that while a state may possess ostensible control in terms of overwhelming 63 Kalyvas, 2006. 613 superiority, its use of selective violence can still involve brutality and produce heavy casualties. Group culpability means that selective violence can still purposefully target more people and produce more violence. In other words, the relationship between selective and indiscriminate violence is not dichotomous but a continuum within which states are more or less discriminate in their targeting. The Indian state's use of violence was not random in Kashmir nor did it employ indiscriminate area weapons. Nevertheless, it killed thousands of innocent people because it defined insurgents and culpability broadly due to perceived disloyalty of Kashmiri Muslims. Even in the 1971 Bengali rebellion, Pakistan employed some selection criteria to target insurgent participants and supporters, but because the state perceived Bengali civilians as disloyal and supporting the rebels, the state selection criteria demanded it kill more Bengalis. Furthermore, good information that enables selectivity does not solely derive from area control with overwhelming numbers but also depends on potential access points to information sources. Despite a massive force presence in Kashmir, the Indian state could never replicate the scale and quality of intelligence it accessed in Punjab due to the identity of its personnel that lacked pre-war links and networks of information. Similarly, the Pakistani state never accessed the same information sources in East Pakistan or Baloch IV that it would later collect in Karachi and Malakand II because it held much weaker pre-existing links to the Bengalis and Baloch. Empirical Contributions Finally, this project has contributed to our empirical knowledge of a number of understudied counterinsurgency campaigns. Most of the South Asian campaigns explored in this project have never made it into the anthologies and edited volumes of counterinsurgency, and some have even been neglected by historians of South Asian conflict. The frequency with which 614 the early campaigns in Hyderabad, Telangana and Balochistan are glossed over is astonishing, even though they were some of the earliest clues as to how the Indian and Pakistani states would assert their authority. Understanding South Asian conflicts and violence management is particularly important because, unlike Southeast Asia that has consumed most of the attention, many civil conflicts in South Asia have persisted or recently emerged, making it a likely theater for future strategic developments and experiments. Another striking feature of this research is how similar India and Pakistan behave. Much of the South Asian security and comparative politics literature builds off of the divergence in South Asian capabilities, institutions, civil-military relations, doctrine, and strategic culture. Nevertheless, the parallels between India and Pakistan's counterinsurgency strategies towards different sets of groups and territories suggests similar incentives are at work despite varied macro-structures and environments, also demonstrating the robustness of the theory. Policy Contributions The findings of this research point to a number of important insights for policymakers, principally those with an interest in understanding, anticipating, and influencing the behavior of states fighting rebellion. First, by generally correcting a bias against studies of domestic counterinsurgents, this project contributes to policymakers' understanding of domestic counterinsurgents and allows them to prepare for and anticipate when they are likely to under balance against insurgent groups, or violate international norms on the use of violence. This will prove important as great powers look to work with local counterinsurgent partners and support their foreign internal defense instead of directly intervening in conflicts under conditions of increasing multipolarity. These indirect interventions still require a stronger understanding of 615 what structures local incentives and constraints in order to guard against free rider, principalagent, or moral hazard problems. Second, political institutions and regime type are not sufficient to restrain states from resorting to strategies of high violence. International institutions and actors invested in conflict prevention and stabilization need to pay more attention to the composition of enduring bureaucratic institutions like the civil service and security forces that control day-to-day functions rather than focusing principally on representation within legislatures and local government. 64 Concerned parties that condition incentives on socio-political reforms to promote stability (e.g. aid organizations and institutions) might consider encouraging societal mirroring and minority integration in these government agencies in periods priorto war so that in the event of armed conflict, states have stronger reasons for restraint, at least towards a broader population. Though it may involve a multi-decade time horizon, shifting this ethnic balance towards proportionality within the state and bureaucracy can potentially reduce the costs of discriminate force in the near term and raise the costs of brute force in the long-term when dealing with violent, non-state challengers to state authority. Empathy is extremely difficult to manufacture and should not be expected overnight, but ethnic integration can more easily facilitate other mechanisms of restraint such as links for better information and nodes of mutual trust. In statebuilding efforts in Iraq and Afghanistan, the United States has placed considerable efforts to ensuring free, fair, and representative elected institutions and building up the size of security forces. However, it has spent considerably less time working to ensure ethnic balance and proportionality in Iraqi or Afghan security institutions (particularly the leadership), which are bound to play a leading role in the ongoing civil war and anticipated escalation. This inattention to ethnic composition in central civil and security bureaucracies, as opposed to the legislature 64 There is a reason the security forces and bureaucracy are often described as the "deep state." 616 and local government, not only sewed the conditions for a renewed Sunni-based insurgency in Iraq but has also constrained the ways in which the state will fight insurgents in the coming years, particularly due to a scarcity of linkages for transmitting information and trust. Third, the study demonstrates how and why states can endure prolonged conflict, casting doubt on expectations of self-correction through failure as well as the inevitability of hurting stalemates. Despite the claims of ideal type counterinsurgency advocates, who contend that restraint, politics, and public goods is the only successful path, victory can be achieved through other means. States can "win" with high levels of unremitting violence to attrite insurgents and wear down public support, or by lowering their bar for victory and simply holding its ground, despite insurgent organizations remaining in tact and fostering low-levels of rebel activity. In the latter case, not losing is tantamount to winning, and given the advantages afforded to states by the international system, this approach may be perfectly tolerable. States need not emulate counterinsurgency "best practices" and can entertain multiple strategic options, including attrition and various containment strategies, by operating with theories of victory and politics that depart from Western approaches to total victory, sovereignty, and Weberian authority. We need to take seriously this "conceptual refinement of victory and defeat" to improve strategic design and evaluation, and should not then be surprised if states pursue these options. 65 The "give war a chance" school of thought presumes that as war runs its course, either belligerent will achieve a decisive victory or that they will reach a mutually hurting stalemate that compels them to resolve their dispute through peaceful means.66 In fact, this study reveals an alternative outcome might prevail-that many incumbents are able to prevail by avoiding losing to rebels. Rather than forcing a resolution, this stalemate could result in chronic or enduring low- 65 66 Zhukov, 2012, p. 292. Edward N. Luttwak, "Give War a Chance," ForeignAffairs, 78 (4), July/August 1999, pp. 36-44. 617 level conflict, though with the latent potential to flare up or escalate into something bigger. A steady state of disorder could result from state choice and insurgent durability amidst difficult, peripheral terrain that offers no incentive for a decisive, sustained counterinsurgent campaign. Fourth, by identifying the role of agency and domestic-level motives, the study identifies some of the limits to international intervention or assistance. Counterinsurgency variation is not strictly a capabilities problem and increased capacity and resources (from internal or external sources) may not cause a state to leverage that additional capacity in its counterinsurgency campaigns. It may instead choose to husband or deploy these resources for other contingencies, for instance to manage its non-violent political opposition, or it may simply operate with a different theory of victory that demands less effort. Similarly, since strategic variation is not just a product of learning, international investment in professional training and doctrine can improve a security force's capacity for restraint, but it may not have its intended effect on actual state behavior in conflict. Domestic incumbents might work around the professionalized force with cheaper options, assessing the fight is not worth the cost or that their adversary only understands the crude language of brute force. At the same time, ideology and strategic culture are not strait jackets and under the right conditions with sufficient motivation, additional inputs from external actors can boost the capacity of medium strength states to conduct decisive, comprehensive, and restrained counterinsurgencies. The challenge for suppliers of international assistance is honestly assessing the motivation of the domestic incumbent rather than projecting their own motives and goals. 618 Violence Management and Containment Fifth, in between these two poles of intervening to resolve conflicts, and pulling back with the expectation of natural dynamics of conflict activating to resolve itself, international policymakers should allocate greater attention to techniques of containing and managing violence, especially those which states have developed for themselves. This means coming to terms with some armed non-state actors and fragmented sovereignty in "ungoverned spaces" and helping weak states manage these zones. Material assistance that reduces the costs of physical control may be helpful to this management process. This does not necessarily mean the wholesale transfer of remotely piloted aircraft technology, though there is evidence that these tools helped Pakistan regain some control over contested territory. Assistance can involve support for fencing borders, durable road construction, and helicopters for improved state reach and mobility. Upgrades in social infrastructure and human development are harder and economic transformations in production the most difficult. In the long run, incremental economic transformation change might gradually seep out of core regions into the periphery, but in the meantime, states and international institutions need to get used to the idea of coping with unproductive, ungoverned spaces where illicit actors and activity thrive in different equilibrium. An equally important part of violence management is a state cultivating localized capillaries of power and influence, even if it outsources some sovereignty. Most counterinsurgency campaigns, regardless of strategy, involve leveraging some informal local authorities-local tribes, militias, warlords, village defense forces, turned insurgents, and even terrorist and criminal networks-but their role as a complementary or substitute force and the degree of accountability varies. Rather than compelling states to dismantle all these informal management tools or eschew them for formal ones, international policymakers need to begin 619 accepting the reality of working within these localized instruments of social stability and control, even if they are unorthodox or depart from Western liberal values. Deftly networking these local instruments to the state might actually enhance state authority rather than compromise it. In sum, the core periphery theory offers an important contribution to the scholarship on civil conflict and counterinsurgency. The argument improves our understanding of a state incumbent, the prime mover in asymmetric conflicts, by disaggregating its varied capabilities and interests within its domain that are mediated by territory and identity. Additionally, the theory of these sub-national forces at the meso-level of civil conflict helps to qualify the relationship between macro-structure and endogenous micro-dynamics and larger conflict strategies. Robust empirical support for this theory is marshaled with a number of understudied, domestic incumbent cases from South Asia, but evidence from conflicts around the world seems to support parts of the theory and its mechanisms. Nevertheless, further study is required to closely test this the predictions and mechanism on a larger set of cases. The project has not only improved our understanding of the incentives and constraints that shape state strategic decision making in war, but it also makes theoretical contributions on the larger state project of building, extending, and sustaining its authority. However, the implications for policymakers are disconcerting. Stable equilibriums of instability and conflict may not self-correct and continue to produce international negative externalities. At the same time international interventions to bolster or alter state institutional structures may not yield desired effects on these conflicts. Instead, policymakers need to embrace working within a paradigm of violence management rather than conflict resolution while supporting broader, longterm economic and political changes that can eventually improve stability. 620 Appendix: Table A8.1: Evaluation of Main and Alternative Theories (X = incorrect) Observation (IV) (DV) India 1. Hyderabad 2. Telangana 3. Nagaland I 4. Mizoram 5. Naxal I - AP 6. Naxal I - WB 7. Nagaland II 8. Manipur 9. Punjab 10. Naxal II -AP 11. Assam 12. Kashmir I 13. Naxal III 14. Kashmir II Pakistan 15. Baloch I 16. Baloch II 17. Baloch III 18. East Pakistan 19. Baloch IV 20. Waziristan 21. Sindh-MRD 22. Karachi-MQM 23. Malakand I 24. FATA I 25. Baloch V 26. Malakand II 27. FATAII-SWA 28. FATAII-Other 29. FATAII-NWA Total _____________Combined: Territory (Effort) Identity (Violence) State Strength (Effort) Regime Type (Violence) Reputation (Violence) (Effort) X X X X X X X X X X X X Rebel Size (Effort) (Violence) X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Rebel Aims/ Separatist (Effort) (Violence) External Support (Violence) X X X X X X X X X X '-4 X X X X X X X X X X X X X X X X X X 2X 26 23 X X X 26 14 JCombined: 10 X X X x8 X X X X X X x7 X X X X X X 15 1818 -- [Combined: 8 177 *10 X X X X X X X X X X 1Combined: 8 X X X X I3I 14 tombined:-10 X X X 21 X 17 J_____ Bibioaraphy "2006: Who's Who in Balochistan," Dawn, May 28, 2011. 25 Years ofPakistan Statistics, 1947-72, Karachi: Government of Pakistan, 1972. "900,000 Hyderabad Refugees Return to State, Government's Plan for Speedy Rehabilitation," Times ofIndia, Nov 29, 1948. Abbas, Hassan, "Musharraf Contends with the Pashtun Element in the Pakistani Army," Terrorism Monitor, 3 (42), Nov 1, 2006. Abbas, Hassan, "Pakistan's Volatile North-West Frontier: Pashtun Tribes, Taliban, and alQaeda," Presentation to the Jamestown Foundation, Washington DC, December 14, 2006 (http://www.jamestown.org/single/?txttnews[swords]=8fd5893941 d69d0be3f378576261 ae3e &txttnews[anyof the-words]=Scheuer%20&txttnews[pointer]=3&tx ttnews[tt news]=1 32 &txttnews[backPid]=7&cHash=030de6a7al90a421b86435630fbl3cfd#.U9uTMahhK68). Abbas, Hassan, "Transforming Pakistan's Frontier Corps," Terrorism Monitor, 5 (6), March 30, 2007. Abbas, Hassan, "A Profile of the Tehrik -i-Taliban," CTC Sentinel, 1 (2), January 2008, pp. 14. Abbas, Hassan, "The Black Turbaned Brigade: The Rise of the TNSM in Pakistan," Terrorism Monitor, 4 (23), November 30, 2006. Abbas, Hassan, and Shehzad H. Qazi, "Rebellion, Development and Security in Pakistan's Tribal Areas, CTC Sentinel, June 25, 2013. Abbasi, Ansar, "Where PPP, PML-N and MQM stood on Jinnahpur in 1992," The News International, September 3, 2009 (http://www.thenews.com.pk/TodaysPrintDetail.aspx?ID=196297&Cat=2&dt=9/3/2009). Abouzeid, Rania, "Taliban Gone, Pakistan Area Still Wants Islamic Justice," Time, April 25, 2010 (http://www.time.com/time/world/article/0,8599,1984067,00.html). Abraham, A.S., "Optimism on the North-East Some Signs of Pacification," Times ofIndia, March 29, 1985, p. 8. Adhopia, Ajit, "Offspring Feel Threatened by Mother Faith," Toronto Star, February 17, 2001. Adinarayan, S.P., "Before and After Independence: A Study of Racial and Communal Attitudes in India," British JournalofPsychology, 44 (2), 1953, pp. 108-115. Afkhami, Gholam Reza, The Life and Times of the Shah, Berkeley, CA: University of California Press, 2009. Afridi, Tayyeb, "Pakistani Province Reaches Out Through Radio," KUT News, July 8, 2011 (http://kutnews.org/post/pakistani-province-reaches-out-through-radio). 622 Ahmad, Aijaz, "The Rebellion of 82.625 A Balance Sheet," South Asia Bulletin, 4 (1), Spring 1984, p 27-44 Ahmad, Lt. Col. Syed Iqbal, Balochistan and its StrategicImportance, Karachi: Royal Book Company, 1992. Ahmad, Mavish, "Balochistan: Middle Class Rebellion," Dawn, June 5, 2012. Ahmad, Naveed, "Trouble in Pakistan's Energy-rich Balochistan," ISN Security Watch, January 30, 2006. Ahmed, Akbar S., Resistance and Control in Pakistan, London: Routledge, 2004. Ahmed, Ali, "Countering Insurgency in Jammu and Kashmir: Debates in the Indian Army," Confronting Terrorism, Ed. Maroof Raza, New Delhi: Penguin, 2009, 58-79. Ahmed, Ali, "Mizo Hills: Revisiting the Early Phase," Scholar Warrior,Autumn 2011. Ahmed, Feroz, Ethnicity and Politics in Pakistan, Karachi: Oxford University Press, 1998. Ahmed, Khaled, "Zahid Hussain's counter-narrative," Express Tribune, Dec 11, 2010. Ahmed, Siraj, "Police Reform in Balochistan," Stabilizing Pakistan Through Police Reform, Report by the Independent Commission on Pakistan Police Reform, Ed. Hassan Abbas, New York: Society, July 2012, pp. 113-119. Ahuja, Amit and Susan L. Ostermann, "Blurred Lines: Intermarriage and the Weakening Caste Cleavage in Urban India," Presented at Annual International Studies Aassociation Conference, Toronto, April 2014. Aiyar, SA, "Declasify Report on the 1948 Hyderabad Massacre," Swaminomics Blog, Times of India, Nov 25, 2012 (http://blogs.timesofindia.indiatimes.com/Swaminomics/declassifyreport-on-the-i 948-hyderabad-massacre/). Aiyar, SA, "The Unexpected Rise of Dalit Millionaires," Swaminomics Blog, Times ofIndia, July 31, 2011. (http://blogs.timesofindia.indiatimes.com/swaminomics/entry/the-unexpectedrise-of-Dalit-millionaires). Aiyar, Swaminathan S. Anklesaria, "Why Punjab Has Suffered Long, Steady Decline," Indian Express, Nov 8, 2012. Akbar, Malik Siraj, "The Disloyal Baloch," January 22, 2007 (http://gmcmissing.wordpress.com/2007/01/22/the-disloyal-baloch/). Akbar, Malik Siraj, "The Ifs and Buts of Negotiating with Insurgents in Balochistan," The Baloch Hal, August 1, 2013. Akbar, Malik Siraj, Refined Dimensions ofBaloch NationalistMovement, Xlibris, 2011. Akhund, Iqbal, Memoirs ofa Bystander: A Life in Diplomacy, Karachi: Oxford University 623 Press, 1997. Al-Atraqchi, Firas, "Did Iraq Government's Neglect Lead to Mosul's Downfall?" Al Jazeera, June 11, 2014 (http://www.aljazeera.com/indepth/opinion/2014/06/iraq-government-neglectled-mosu-20146117235750482.html). al-Qaisi, Mohammed, "Iraq Launches Counter-Terrorism Campaign Along Syria Border," Al Shorfa.com, May 23, 2013 (http://mawtani.alshorfa.com/enGB/articles/iii/features/2013/05/23/feature-01). Albertus, Michael, and Oliver Kaplan, "Land Reform as a Counterinsurgency Policy: Evidence from Columbia," Journal of Conflict Resolution, 57 (2), 2012, pp. 198-23 1. Ali, Arshad, "Internal instability in Pakistan - Ideological and Socioeconomic Perspectives," Strategic Studies, (ISSI), 31 (1-2), Spring 2011, p. 97 Ali, Lubna Abid and Naveed Iqbal Khan, "The Rise of the Tehreek-e-Nifaz-e-Shariat-eMohammadi in Malakand Division, NWFP: A Case Study of the Process of 'State Inversion,"' Pakistan 11 (1), 2010, pp. 89-119. Ali, Zulfiqar, "Kashmir Militants Give Up Fight and Head Home," BBC Urdu, May 31, 2012. Alikuzai, Hamid Wahed, A Concise History ofAfghanistan in 25 Volumes, Volume 14, S.I.: Trafford, 2013. Alvi, Hasan Asad, "Police and Counterterrorism in Khyber Pakhtunkhwa," Stabilizing Pakistan Through Police Reform, Report by the Independent Commission on Pakistan Police Reform, Ed. Hassan Abbas, York: Asia Society, July 2012, pp. 104-112. Amin, Tahir, "Pakistan in 1994: The Politics of Confrontation," Asian Survey, 35 (2), Feb 1995, pp. 140-146. Amin, Tahir, Ethno-NationalMovements ofPakistan, Islamabad: Institute of Policy Studies, 1988. Analysis based on Brig. A.R. Siddiqui, "Faqir of Ipi's Cross Border Nexus," Khyber.org, September 16, 2005 (http://www.khyber.org/articles/2005/Faqir-ofIpisCrossBorderNex.shtml) Accessed August 4, 2014. Anant, Arpita, Counterinsurgencyand 'Op Sadbhavana'in Jammu andKashmir, New Delhi: IDSA Occasional Paper, 2011. "Anarchy in Assam," Hindustan Times, January 16, 1998, p. Factiva, accessed 3/5/14 Anderson, John, "India's Muslims Fear New Physical Threat," Washington Post, March 7, 1994. Anderson, Jon Lee, "Death of the Tiger: Sri Lanka's Brutal Victory Over its Tamil Insurgents," The New Yorker, Jan. 17, 2011. 624 Anderson, Perry, "Why Partition?" London Review of Books, Vol. 34, No. 14, July 19, 2012 "Andhra Chief Differs with Delhi on Naxalites," Delhi Domestic Service, September 3, 1989, Daily Report. Near East & South Asia, FBIS-NES-89-170 on 1989-09-05 p. 50, Accessed via Newsbank Feb 19, 2014. Andres, Richard B., Craig Wills, and Thomas E. Griffith Jr., "Winning with Allies: The Strategic Value of the Afghan Model," International Security, 30 (3), Winter 2005/06, 124160. "Anemic Assam," The Hindu, November 1, 1998, p. Factiva, accessed 3/5/14. Anthony, P.J. "K.R. Narayanan Dies at 85; 'Untouchable' India President," New York Times, November 10, 2005. Arango, Tim, and Kareem Fahim, "Iraq Again Uses Sunni Tribesmen in Militant War," New York Times, January 19, 2014. Arbab, Taimur, "Wearing a Sindhi Topi and Ajrak Won't Rebuild a Backward Province," Express Tribune Blog, December 10, 2013 (http://blogs.tribune.com.pk/story/2000 1/wearinga-sindhi-topi-and-ajrak-wont-rebuild-a-backward-province/). Arif, Ghulam Mohammad, "Pakistan Poverty Assessment Update," Asian Development Bank, Project # 37717, Feb. 2006. Arjomand, Said Amir, The Turbanfor the Crown: The Islamic Revolution in Iran, New York: Oxford University Press, 1988. Arora, Kim, "Five Throw Helpline to North East Bretheren," Times ofIndia, October 17, 2013. Arpi, Claude, "The Blunder of the Pandit," Rediffcom, June 16, 2004 (http://www.rediff.com/news/2004/jun/I6spec3.htm). Arreguin-Toft, Ivan, "How the Weak Win Wars: A Theory of Asymmetric Conflict," International Security, 26 (1), Summer, 2001, pp. 93-128. Arreguin-Toft, Ivan, How the Weak Win Wars: A Theory ofAsymmetric Conflict, Cambridge: Cambridge University Press, 2005. Ashraf, Tariq Mahmud, "Pakistan's Frontier Corps and the War Against Terrorism - Part Two," Terrorism Monitor, 6 (16), Aug 11, 2008. Ashraf, Tariq Mahmud, "The Pakistan Frontier Corps in the War on Terrorism- Part I," Terrorism Monitor, 6 (15), July 25, 2008. "Assam Government to Recruit Surrendered Militants," The Pioneer, May 21, 2008. "Assam Rifles Likely to Replace Army, BSF to Guard Manipur's border," Times ofIndia, March 5, 2013 (http://timesofindia.indiatimes.com/city/guwahati/Assam-Rifles-likely-toreplace-army-BSF-to-guard-Manipurs-border/articleshow/18804621.cms) 625 "Assam Timeline - Year 1993" South Asia Terrorism Portal, (http://www.satp.org/satporgtp/countries/india/states/assam/timelines/YearI 993.htm). Athale, Anil A., Counter Insurgency& Questfor Peace: A ComparativeStudy, New Delhi: Vij Books, 2012. Atzili, Boaz, Good Fences, Bad Neighbors: BorderFixity and InternationalConflict, Chicago: University of Chicago Press, 2011. Awan, A.B., Baluchistan:Historicaland PoliticalProcesses, London: New Century Publishers, 1985. Axmann, Martin, Back to the Future: The Khanate ofKalat and the Genesis of Baloch Nationalism 1915-1955, Karachi: Oxford University Press, 2009. Azfar, Kamal, "Constitutional Dilemmas in Pakistan," Pakistan Under the Military: Eleven Years of Zia ul-Haq, Ed. Shahid Javed Burki and Craig Baxter, Boulder, CO: Westview 1991, pp. 49-85. Aziz, Sartaj, "The Economic Cost of Extremism," Pakistan's Quagmire: Security, Strategy, and the Future of the Islamic-nuclear Nation, Ed. Usama Butt and N. Elahi, New Continuum International Publishing, 2010, pp. 75-91. Bahadur, Kalim, Democracy in Pakistan: Crises and Conflicts, New Delhi: Har-Anand Publications, 1998 Baixas, Lionel, "Thematic Chronology of Mass Violence in Pakistan, 1947 - 2007," Online Encyclopedia of Mass Violence, June 24, 2008 (http://www.massviolence.org/ThematicChronology-of-Mass-Violence-in-Pakistan- 1947-2007). Bakshi, G.D., Kishtwar Cauldron: The Struggle Against ISI's Ethnic Cleansing, New Delhi: Pentagon Press, 2013, pp. 45, 48. Baldauf, Scott, "Afghan Army Gets Ahead by Getting Along," Christian Science Monitor, February 11, 2001 (http://www.csmonitor.com/2003/0211/pO7sOl-wosc.html). "Baloch in Karachi Ask Not for Entitlement, Only Equal Rights," Express Tribune, September 24, 2012. "Baloch Missing Persons: 621 or 23,000? A Matter of Contested Figures," Express Tribune, March 9 2013. Baloch, Sanaullah, "Baloch Demands Still Unmet," Dawn, February 19, 2009 (http://www.dawn.com/news/940332/baloch-demands-still-unmet). Bandyopadhyay, Sanghamitra, "Why is Economic Growth Across Indian States Uneven?" London School of Economics Blog February 18, 2013 (http://blogs.lse.ac.uk/indiaatlse/2013/02/18/why-is-economic-growth-across-indian-statesuneven/). Bandyopadhyaya, J., The Making ofIndia's ForeignPolicy, New Delhi: Allied Publishers, 626 1970. Banerjee, Chandrani, "A Sieve of a Scheme?" Outlook India, April 19, 2010 (http://www.outlookindia.com/article/A-Sieve-Of-A-Scheme/264994). Banerjee, Sumanta, India's Simmering Revolution: The Naxalite Uprising, London: Zed Books, 1984. Bansal, Alok, "Factors Leading to Insurgency in Balochistan," Small Wars & Insurgencies, 19 (2), June 2008, p. 182-200. Bansal, Alok, Balochistan in Turmoil: Pakistanat Crossroads,New Delhi: Manas Publishing, 2010. Barbora, Sanjay, "Rethinking India's Counter-Insurgency Campaign in North-East," Economic and Political Weekly, 41 (35), September 2-8, 2006, p. 3805-3812. Barman, Abheek, "Air Attacks in Mizoram, 1966: Our Dirty, Little Secret," Economic Times, February 19, 2013 (http://articles.economictimes.indiatimes.com/2013-0219/news/37179679_1_mna-chhinga-veng-air-force) Bartholomees, Jr., J. Boone, "The Issue of Attrition," Parameters, Spring 2010, pp. 5-19. Barua, Pradeep P., "Ethnic Conflict in the Military of Developing Nations: A Comparative Analysis of India and Nigeria," Armed Forces & Society, 19 (1), Fall 1992, pp 123-137. Baruah, Sanjib Kr, "Manipur CM is Corrupt: Wikileaks," Hindustan Times, March 21, 2011 Baruah, Sanjib, "A New Politics of Race: India and its North-east," India InternationalCentre Quarterly, 32 (2/3), Winter 2005, pp. 163-176. Baruah, Sanjib, "Confronting Constructionism: Ending India's Naga War," JournalofPeace Research, 40 (3), 2003, p. 335. Baruah, Sanjib, Beyond Counter-Insurgency:Breaking the Impasse in NortheastIndia, New Delhi: Oxford University Press, 2009. Baruah, Sanjib, DurableDisorder:Understandingthe Politics ofNortheastIndia, New Delhi: Oxford University Press, 2005. Bass, Gary J., The Blood Telegram: Nixon, Kissinger, and a Forgotten Genocide, Knopf, 2013. Bassok, Moti, and Haim Bior, "Israel's Cabinet Creates Two Panels on Haredi, Arab Integration," Haaretz, May 26, 2014 (http://www.haaretz.com/news/national/.premium1.592752). Baviskar, Amita, "Indian Indigeneities: Adivasi Engagements with Hindu Nationalism in India," Indigenous Experience Today, Ed. Marisol de la Cadena and Orin Starn, Berg Publishers: Oxford, 2007. Baxter, Craig, Dairiesof FieldMarshallMohammadAyub Khan: 1966-72, New York: Oxford 627 University Press, 2008. Bedi Rahul, "Need a Cop? Just Call your Local Guerilla," Courier-Mail,July 7, 1993; Routray, 2007. Benbabaali, Dalal, "Questioning the Role of the Indian Administrative Service in National Integration," South Asia MultidisciplinaryAcademic Journal, September, 2008. Bengali, Kaiser and Mahpara Sadaqat, Regional Accounts of Pakistan:Methodology and Estimates 1973-2000, Karachi: Social Policy and Development Centre, 2006. Bennett, Huw, "The Other Side of the COIN: Minimum and Exemplary Force in British Army Counterinsurgency in Kenya," Small Wars and Insurgencies, 18 (4), December pp. 638-664. Bercovitch, Jacob and Richard Jackson, InternationalConflict: A ChronologicalEncyclopedia of Conflicts and Their Management, 1945-1995, Washington, DC: Congressional Quarterly, 1997. Berman, Eli, Jacob N. Shapiro, and Joseph Felter, "Can Hearts and Minds Be Bought? The Economics of Counterinsurgency in Iraq," Journal of Political Economy, 119 2011, pp. 766819. Betts, Richard K., Enemies ofIntelligence: Knowledge and Power in American National Security, New York: Colombia University Press, 2007 Bhan, Mona, Counterinsurgency,Democracy, and the Politics ofIdentity in India, New York: Routledge, 2013. Bhat, Anil, Assam: Terrorism and the DemographicChallenge, New Delhi: Centre for Land Warfare Studies, 2009. Bhatt, Sheela, "Is Narendra Modi Relaly an OBC?" Rediff com, June 5, 2014 (http://www.rediff.com/news/column/ls-election-sheela-says-is-narendra-modi-really-anobc/2014051 0.htm). Bhaumik, Subir, "Ethnicity, Ideology and Religion: Separatist Movements in India's Northeast" Religious Radicalism & Security in South Asia, Ed. Satu P. Limaye, Honolulu, HI: Asia Pacific Center for Security Studies, 2004, p. 237. Bhowmick, Nilanjana, "What the Death of Nido Taniam Tells Us About Racism in India," Time, February 6, 2014 (http://time.com/4876/nido-taniam-india-racism/). Biddle, Stephen D., "Allies, Airpower, and Modern Warfare: The Afghan Model in Afghanistan and Iraq," International Security, 30 (3), Winter 2005/06, pp. 161-176. Biddle, Stephen D., "The New U.S. Army/Marine Corps Counterinsurgency Field Manual as Political Science and Political Praxis," Perspectives on Politics, 6 (2), June 2008, pp. 347-350. Biddle, Stephen D., Military Power: Explaining Victory and Defeat in Modern Battle, Princeton: Princeton University Press, 2004. 628 Biddle, Stephen, Jeffrey A. Friedman, and Jacob N. Shapiro, "Testing the Surge: Why Did Violence Decline in Iraq in 2007?" International Security, 37 (1), Summer pp. 7-40. Bidwai, Praful, "Northeast Erupts, But New Delhi is Not Panicking," Inter Press Service, January 27, 1994, Factiva accessed 3/5/14 Biswas, Bidisha, Managing Conflicts in India: Policies of Coercion andAccommodation, Lanham, MD: Lexington Books, 2014. Biswas, Soutik, "Living on a Razor's Edge," India Today, October 15, 1993 (http://indiatoday.intoday.in/story/kuki-naga-conflict-recent-massacres-in-manipur-take-statevirtually-to-brink-of-civil-war/1/303114.html). Blitz, James, James Lamont, and Farhan Bokhari, "Swat Outlook 'Pretty Bleak' for Pakistan," Financial Times, May 13, 2009 (http://www.ft.com/cms/s/0/46c20ab0-3f59-1 Ide-ae4f00144feabdcO.html). Blood, Peter, "Baloch," Pakistan: A Country Study, Washington: GPO for the Library of Congress, 1994 (http://countrystudies.us/pakistan/35.htm). Blood, Peter, "Sindhis," Pakistan: A Country Study, Washington: GPO for the Library of Congress, 1994 (http://countrystudies.us/pakistan/34.htm). Boggs, Robert, "Pakistan's Pashtun Challenge: Moving from Confrontation to Integration," Strategic Analysis, 36 (2), 2012, pp. 206-216. Boone, Catherine, "Decentralization as Political Strategy in West Africa," Comparative Political Studies, 36 (4), May 2003, pp. 355-380. Boone, Catherine, PoliticalTopographies of the African State: TerritorialAuthority and InstitutionalChoice, New York: Cambridge University Press 2003 Borhi, Laszlo, "Rollback, Liberation, Containment, or Inaction? U.S. Policy and Eastern Europe in the 1950s," Journal of Cold War Studies, 1 (3), 1999, pp. 67-110. Bose, Sarmila, DeadReckoning: Memories of the 1971 Bangladesh War, London: Hurst and Co., 2011 Bose, Sumantra, Kashmir: Roots of Conflict, Paths to Peace, Cambridge, MA: Harvard University Press, 2003. Boulding, Kenneth, Conflict and Defense: A General Theory, New York: Harper and Brothers, 1962 Bowcott, Owen, "MoD took Softer Line on Loyalist Paramilitaries, Secret Files Reveal," The Guardian, October 10, 2010. Boyini, Deepak Aneel, "Explaining Success and Failure: Counterinsurgency in Malaya and India," Unpublished masters thesis, Naval Postgraduate School, December 2010. Brass, Paul R., Ethnicity andNationalism: Theory and Comparison,New Delhi: Sage 629 Publications, 1991. Brown, Vahid, and Don Rassler, FountainheadofJihad: The HaqqaniNexus, 1973-2012, New York: Oxford University Press, 2013, p. 31-36. Brubaker, Rogers, and David D. Laitin, "Ethnic and Nationalist Violence," Annual Review of Sociology, 24, 1998, pp. 423-452. Budania, Rajpal, India's National Security Dilemma: The PakistanFactorand India's Policy, New Delhi: Indus Publishing Company, 2001. Buhaug, Halvard, Scott Gates, and Paivi Lujala, "Geography, Rebel Capability, and the Duration of Civil Conflict," Journal of Conflict Resolution, 53 (4), August 2009. Buncombe, Andrew, "Victims of India's Dirty War in Manipur Poised for Justice," The Independent, August 5, 2013. Butt, Ahsan Ishaq, "Goodbye or See You Later: Why States Fight Some Secessionist But Not Others," PhD Dissertation, University of Chicago, 2012. Byman, Daniel L., "Friends Like These: Counterinsurgency and the War on Terrorism," International Security, 31 (2), 2006, pp. 79-115. Byman, Daniel L., and Kenneth M. Pollack, "Let Us Now Praise Great Men: Bringing the Statesman Back In" International Security, 25 (4), Spring 2001, pp. Byman, Daniel, and Natan Sachs, "The Rise of Settler Terrorism: The West Bank's Other Violent Extremists," Foreign Affairs, 91 (5), September/October, 2012: 73-86. Calluzzo, Nicholas T., "The Urbanization of Insurgency: Shifts in the Geography of Conflict," Unpublished Masters Thesis, Massachusetts Institute of Technology, 2010 Cantwell Smith, Wilfred, "Hyderabad: Muslim Tragedy," Middle EastJournal,4 (1), January 1950. Catignani, Sergio, "Israeli Counterinsurgency: The Never-ending 'Whack-A-Mole,"' The Routledge Handbook of Insurgency and Counterinsurgency, Ed. Paul B. Rich and Isabella Duyvesteyn, New York: Routledge, 2012, 263-275. Caverley, Jonathan D., "The Myth of Military Myopia: Democracy, Small Wars, and Vietnam," International Security, 34 (3), Winter 2009/10, pp. 119-157. Caverly, Jonathan D., "The Myth of Military Myopia: Democracy, Small Wars, and Vietnam," International Security, 34 (3), Winter 2009/10, pp. 119-157. Cederman, Lars-Erik, Andreas Wimmer, and Brian Min, "Why Do Ethnic Groups Rebel? New Data and Analysis," World Politics, 62 (1), 2010, pp. 87-119. Center for Research and Security Studies, Balochistan'sMaze of Violence, Islamabad: Center for Research and Security Studies, 2011. 630 "Center Warns States About Naxalite Activities," Delhi Domestic Service, March 2, 1988, Daily Report. Near East & South Asia, FBIS-NES-88-041 on 1988-03-02 p. 54, Accessed via Newsbank Feb 19, 2014. Chadda, Maya, Ethnicity, Security, and Separatism in India, New York: Columbia University Press, 1997. Chadha, Vivek, "India's Counterinsurgency Campaign in Mizoram," India and Counterinsurgency:Lessons Learned, Ed. Sumit Ganguly and David P. Fidler, New York: Routledge, 2009. Chadha, Vivek, Low Intensity Conflicts in India: An Analysis, New Delhi: Sage, 2005. Chakrabarty, Bidyut and Rajat Kumar Kujur, Maoism in India: Reincarnationof Ultra-Left Wing Extremism in the Twenty-First Century, New York: Routledge, 2010, p. 27. Chakravarti, Sudeep, Red Sun: Travels Through Naxalite Country, New Delhi: Viking, 2008. Chandan, Rahul, KPS Gill: The ParamountCop, Noida: Maple Press, 2013. Chandio, Amir Ali, Mughis Ahmad, and Fouzia Naseem, "Struggle for Democracy in Sindh: A Case Study of Movement for Restoration of Democracy (1983)," Berkeley Journal of Sciences, 1 (1), Jan 2011 (http://www.berkeleyjournalofsocialsciences.com/). Chandola, Harish, "The Much Maligned Nagas," The Times ofIndia, May 13, 1956, p. 8. Chandra, Kanchan and Steven Wilkinson, "Measuring the Effect of 'Ethnicity'," Comparative Political Studies, 41(4-5), 2008, pp. 515-63. Chandraprema, C. A., Sri Lanka: The Years of Terror: The J. VP. Insurrection, 1987-1989, Colombo: Lake House Bookshop, 1991. Chatterjee, Siddharth, "Flaws in India's Strategy to Counter the Maoist Insurgency," Global Observatory, International Peace Institute, August 27, 2012 (http://www.theglobalobservatory.org/analysis/338-flaws-in-indias-strategy-to-counter-themaoist-insurgency.html). Chauchard, Simon, "Why Representation Matters: The Meaning of Ethnic Quotas in Rural India," Book Manuscript Under Review, p. 24. Chaudhuri, J.N. (Maj. Gen), 1 ArmoredDivision in Operation "Polo", New Delhi: Indian Army Headquarters, 1949. Chaudhuri, J.N., GeneralJN Chaudhuri:An Autobiography, Written (through narration) by BK Narayan, New Delhi: Vikas, 1979. Chaudhury, Shoma, "Bringing on the Army Against the Naxals Will Be a Disaster," Tehelka, 7 (23), June 12, 2010 (http://archive.tehelka.com/storymain45.asp?filename=Ne 12061 Obringingon.asp). 631 & Chaudhury, Shoma, "Life in a Shadow Land," Tehelka, 32 (6), August 15, 2009. Chauvel, Richard and Ikrar Nusa Bhakti, The Papua Conflict: Jakarta'sPerceptionsand Policies, Policy Studies 5, Washington DC: East West Center, 2004. Cheema, Pervaiz Iqbal, "Challenges Facing a Counter-Militant Campaign in Pakistan's FATA," NBR Analysis, 19 (3), August 2008, pp. 21-29. Cheema, Pervaiz Iqbal, The Armed Forces ofPakistan, New York: New York University Press, 2002. Chima, Jugdep S., The Sikh SeparatistInsurgency in India:PoliticalLeadershipand EthonationalistMovements, Los Angeles: SAGE, 2010. Chitralekha, "'Committed, Opportunists and Drifters': Revisiting the Naxalite narrative in Jharkhand and Bihar," Contributionsto Indian Sociology, 44 (3), 2010, pp. 299-329. Chopas, Mary Elizabeth Basile, "Law, Security, and Ethnic Profiling: Italians in the United States during World War II," PhD Dissertation, University of North Carolina, Chapel Hill, 2013. Chowdhry, Arjun and Ronald Krebs, "Making and Mobilizing Moderates: Rhetorical Strategy, Political Networks, and Counterterrorism," Security Studies, 18 (3), 2009, pp. 371-399. Christia, Fotini, Alliance Formationin Civil Wars, New York: Cambridge University Press, 2012. Cline, Lawrence E., "The Insurgency Environment in Northeast India," Small Wars Insurgencies, 17 (2), 2006, pp. 126-147. Clodfelter, Michael, Warfare andArmed Conflicts: A StatisticalEncyclopedia of Casualtyand OtherFigures, 1494-2007, Jefferson, NC: McFarland & Company, 2008. Cloughley, Brian, "Insurrection, Terrorism and the Pakistan Army," Pakistan's Quagmire: Security, Strategy, and the Future of the Islamic-nuclear Nation, Ed. Usama Butt and N. Elahi, York: Continuum International Publishing, 2010, p. 108. Cloughley, Brian, A History of the PakistanArmy: Wars andInsurrections,New York: Oxford University Press, 2006. Cloughley, Brian, War, Coups, and Terror: Pakistan'sArmy in Years of Turmoil, New York: Skyhorse, 2008. Clunan, Anne L. and Harold A. Trinkunas, Ungoverned Spaces: Alternatives to State Authority in an Era ofSoftened Sovereignty, Stanford, CA: Stanford University Press, 2010. Cochrane, Marisa, "The Battle for Basra," Iraq Report #9, Institute for the Study of War, June 23, 2008 (https://www.understandingwar.org/report/battle-basra). Cohen, Stephen P., "The Military and Indian Democracy," India'sDemocracy: An Analysis of ChangingState-Society Relations, Ed. Atul Kohli, Princeton, NJ: Princeton University Press, 632 1988, pp. 99-143. Cohen, Stephen P., The PakistanArmy, Berkeley, University of California Press, 1994, p. 45. Cohen, Stephen, India: EmergingPower, Washington DC: Brookings Institution Press, 2002. Cohen, Stephen, The Idea ofPakistan, Washington DC: Brookings Institution Press, 2004. Cohen. Stephen P., The Indian Army: Its Contributionto the Development of a Nation, New York: Oxford University Press, 1990. Coll, Steve, "The Back Channel: India and Pakistan's Secret Kashmir Talks," The New Yorker, March 2, 2009. Collier, Paul and Anke Hoeffler, "Greed and Grievance in Civil War," World Bank Policy Research Working Paper 2355, World Bank, 2000. Corbridge, Stuart, "Competing inequalities: The Scheduled Tribes and the Reservations System in Jharkhand, India," Jharkhand:Environment, Development, Ethnicity, Ed. S. Corbridge, S. Jewitt, and S. Kumar. New Delhi: Oxford University Press, 2004. Correll, John T., "What Happened to Shock and Awe?" Air Force Magazine, 86 (11), November 2003, pp. 52-57. "Countering Terror: 'Militants Are trying to Regroup in Swat,"' Express Tribune, Jan 16, 2013 (http://tribune.com.pk/story/494541/countering-terror-militants-are-trying-to-regroup-inswat/). "Country's 'Biggest-Ever' Drug Seizure Reported," The News, January 29, 1995, p. 1. FBIS Daily Report. Near East & South Asia, FBIS-NES-95-019 on 1995-01-30, p. 71 (Accessed via Newsbank, February 19, 2014). Cunningham, David E., Kristian Skrede Gleditsch, and Idean Salehyan, "Corrigendum for Cunningham, Gleditsch, Salehyan, 2009," January 13, 2012 (http://privatewww.essex.ac.uk/~ksg/papers/CorrigendumCunninghametal2009.pdf). Cunningham, David E., Kristian Skrede Gleditsch, and Idean Salehyan, "It Takes Two: A Dyadic Analysis of Civil War Duration and Outcome," Journal of Conflict Resolution, (4), August 2009 pp.570 - 597. D'Souza, Shanthie Mariet, "Naxalites: Is there a need to 'bring in' the Army?" Journalof Defence Studies 3 (3), July 2009, pp. 125-32. Dalmia, Yashodhara, "History Text-Books Foster Prejudice," Times ofIndia, March 9, 1993, p. 3. Dalyrmple, William, The Age ofKali: Indian Travels andEncounters, London: HarperCollins, 1998. Darden, Keith and Harris Mylonas, "The Promethean Dilemma: Third-party State-building in 633 Occupied Territories," Ethnopolitics, 11 (1), March 2012, pp. 85-93. Das, Bijoyeta "Racism on the Rise in India," Asia Times Online, March 26, 2014 (http://www.atimes.com/atimes/SouthAsia/SOU-03-260314.html). Das, Shaswati and Mallica Joshi, "Delhi Won't Lend a Home to Students from the Northeast," Hindustan Times, August 9, 2013. (http://www.hindustantimes.com/Indianews/NewDelhi/Delhi-won-t-lend-a-home-to-students-from-northeast/Articlel-11 05466.aspx) Das, Suranj an, Kashmir andSindh: Nation-Building,Ethnicity, and Regional Politics in South Asia, London: Anthem Press, 2001 Dasgupta, Biplab, The Naxalite Movement, Bombay: Allied Publishers, 1974. Dastarac, Alexandre and Robert Dersen, "Baloutchistan: La Guerre Oubliee," Le Monde Diplomatique, August 1976, pp. 38-39. Datta, Saikat and Bhavana Vij-Aurora, "Muslims and Sikhs Need Not Apply," Outlook India, November 13, 2006. Davis, Anthony, "Pakistan: State of Unrest," Jane's Intelligence Review, 11 (1), January 1, 1999. Day Christopher R. and William S. Reno, "In Harm's Way: African Counter-insurgency and Patronage Politics," Civil Wars, 16 (2), 2014, pp. 105-126. Deka, Harekrishna, "The Assamese Mind: Contours of a Landscape," India International Centre Quarterly, 32 (2/3), Winter 2005, pp. 189-202. "Delhi Gang-rape: Unprecedented Protests Erupt," Press Trust ofIndia, Dec 22, 2012. "Demonstrations, Riots, Clashes Reported," Agence France- Presse, August 21, 1983. FBIS DAILY REPORT. South Asia, FBIS-SAS-83-163 on 1983-08-22, p Fl (Accessed via Newsbank on Feb 19, 2014). Denyer, Simon, "Muslims Excluded from India's Spy Agency," Reuters, November 6, 2006. Desch, Michael C., Civilian Control of the Military: The Changing Security Environment, Baltimore: Johns Hopkins University Press, 1999. Deshpande, S.H., "Evaluating Hindutva I - The Muslim Question & Sarvarkar," Times of India, January 4, 1994, p. 14 Deutsch, Karl, Nationalism and Social Communication, Cambridge: MIT Press, 1966. Development Challenges in Extremist Affected Areas: Report of an Expert Group to Planning Commission, New Delhi: Government of India, April 2008 (http://planningcommission.nic.in/reports/publications/repdce.pdf). Dhar, M.K., "The Lake of Fire: Manipur Burns, Delhi Bungles," Maloy Krishna Dhar Blog, 634 May 20, 2010 (http://maloykrishnadhar.com/the-lake-on-fire-manipur-burns-delhi-bungles/). Dhar, Maloy Krishna, "Balochistan - Cruces of History," M.K. Dhar Blog, April 25, 2008 (http://maloykrishnadhar.com/balochistan-cruces-of-history/). Dhar, Maloy Krishna, Open Secrets: India's Intelligence Unveiled, New Delhi: Manas Publications, 2005. Dhillon, K. S., Identity and Survival: Sikh Militancy in India, 1978-1993. New Delhi: Penguin Books, 2006. Dholabhi, Nishit, "Schools Shut, CRPF Suffers," The Telegraph, March 14, 2014. DIG of Police Muhammad Azeem, "The Softer Option," Dawn, March 25, 2013 (http://www.dawn.com/news/797860/the-softer-option). Dixit, J.N., Anatomy ofa FlawedInheritance:Indo-PakRelations, 1970-94, New Delhi: Konark Publishers, 1995. Dixon, Paul, "'Hearts and Minds'? British Counter-Insurgency Strategy in Northern Ireland," Journal of Strategic Studies, 32(3), 2009, pp. 445-474. Dorronsoro, Gilles, "Pakistan and the Taliban: State Policy, Religious Networks and Political Connections," Pakistan: Nationalism Without a Nation, Ed. Christophe Jaffrelot, New York: Zed Books, pp. 161-178. Downes, Alexander B., "'The Problem with Negotiated Settlements to Ethnic Civil Wars," Security Studies, 13 (4), 2004, pp. 230-279. Downes, Alexander B., "Draining the Sea by Filling the Graves: Investigating the Effectiveness of Indiscriminate Violence as a Counterinsurgency Strategy," Civil Wars, 9 (4), December 2007, 420-444. Downes, Alexander B., Targeting Civilians in War, Ithaca: Cornell University Press, 2008. Duparcq, Emmanuel, "Swat Rebuilds Year After Pakistan Floods," Agence France Presse, July 24,2011 (http://www.google.com/hostednews/afp/article/ALeqM5iinvhomnj06EdQxJfDKhEnNmOm A?docld=CNG.e53b2b929df6alfba2dd39f4 1 alfec 11.51). Dupuy, T.N., UnderstandingWar: History and Theory of Combat, Leo Cooper, London, 1992. Dutta, Sujan, "Clamour to Let Army Fight Maoists Rises," The Telegraph, Sept. 27, 2009. Dye, Peter, The Jebel Akhdar War: The Royal Air Force in Oman 1952-1959, Air Power Review, 11 (3), Winter 2008, pp. 18-32. "Editorial: 3Ds Dialogue, Development, Deterrence?" Daily Times, February 3, 2009. Elliott, Carolyn M., "Decline of a Patrimonial Regime: The Telengana Rebellion in India, 635 1946-51," JournalofAsian Studies, 34 (1), Nov. 1974. Embree, Ainslee T., "Pakistan's Imperial Legacy," Ed. Ainslee T. Embree, Pakistan's Western Borderlands: The Transformation of a Political Order, Karachi: Royal Book Co., 1979, pp. 2441. "The Enfeebled Internal Security," DainikJagran(Delhi,), November 28, 1996, p. 6. Daily Report. Near East & South Asia, FBIS-NES-94-232 on 1994-12-02, p. 34. Accessed via Newsbank, February 19, 2014. Engineer, Asghar Ali, "Quotas for Muslims," Times ofIndia, October 22, 1994, p. 16. Enloe, Cynthia H., Ethnic Soldiers: State Security in Divided Societies, Athens: University of George Press, 1980. Epstein, Susan B. and K. Alan Kronstadt, "Pakistan: U.S. Foreign Assistance," Congressional Research Service, R41856, October 4, 2012, p. 40, "Ethnic Discrimination Infests Afghan Army, Soldiers Say," Agence France-Presse, January 3, 2011. Eviskov, Egor, "War Against the UPA, 1944-1954: Soviet Counterinsurgency Challenge in Western Ukraine," Unpublished Paper, 2014, pp. 23-24 (http://www.academia.edu/6387532/War_againstthe_UPA_19441954_SovietCounterinsurgencyChallenge inWestern_Ukraine). Fair C. Christine and Seth G. Jones, "Pakistan's War Within," Survival, 51 (6), December 2009-January 2010, pp. 161-188. Fair, C. Christine, "Islamist Militancy in Pakistan: A View from the Provinces," World Public Opinion, July 10, 2009, Fair, C. Christine, "Lessons from India's Experience in the Punjab, 1978-93," India and Counterinsurgency:Lessons Learned, Ed. Sumit Ganguly and David P. Fidler, New York: Routledge, 2009, pp. 107-126. Fair, C. Christine, Fightingto the End: The PakistanArmy's Way of War, New York: Oxford University Press, 2014. Fair, C. Christine, UrbanBattle Fields of South Asia: Lessons Learnedfrom Sri Lanka, India, and Pakistan, Santa Monica: RAND, 2004. Fair, Christine, "Pakistan Loses Swat to Local Taliban, Terrorism Focus, 4 (37), November 13, 2007. Fallaci, Oriana, Interviews with History and Conversationswith Power, New York: 2011. Fazal, Tanisha M., State Death: The Politics and Geography of Conquest, Occupation, and Annexation, Princeton, NJ: Princeton University Press, 2007. Fearon, James D. and David D. Laitin, "Explaining Ethnic Cooperation," American Political 636 Science Review, 90 (4), 1996 pp. 715-735. Fearon, James D. and David D. Laitin, "Sons of the Soil, Migrants, and Civil Wars," World Development, 39 (2), 2011, pp. 199-211. Fearon, James D. and David Laitin, "Ethnicity, Insurgency, and Civil War," American Political Science Review, 97 (1), February 2003, pp. 75-90. Fearon, James, "Why Do Some Civil Wars Last So Much Longer Than Others," Journal of Peace Research, 41 (3), May 2004: 275-302. Fernando, Laksiri, "April 1971 JVP Uprising: Not to Make the Same Mistakes," Colombo Telegraph, April 2, 2013. Fick, Nathaniel, "Fight Less, Win More," Washington Post, August 12, 2007 Finnemore, Martha, The PurposeofIntervention: ChangingBeliefs about the Use ofForce, Ithaca: Cornell University Press, 2003. Food Insecurity in Rural Pakistan, 2003, Ed. Amanur Rehman Khan and Shafqat Munir, Islamabad: World Food Program (WFP)/Sustainable Development Policy Institute (SDPI), 2004. Foucault, Michel, Discipline and Punish: The Birth of the Prison, Trans. Alan Sheridan, New York: Vintage, 1995. French, David, "Have Options Really Changed? British Defence Policy in the Twentieth Century," Journal of Strategic Studies, 15 (1), 1992, pp. 50-72. French, David, The British Way in Counter-Insurgency, 1945-1967, New York: Oxford University Press, 2011. Friedman, Jeffrey A., "Manpower and Counterinsurgency: Empirical Foundations for Theory and Doctrine," Security Studies, 20 (4), 2011, pp. 1-36. Friedman, Thomas L., "No Way, No How, Not Here," New York Times, February 18, 2009. Frymer, Paul, "'A Rush and a Push and the Land is Ours': Territorial Expansion, Land Policy, and U.S. State Formation," Perspectives on Politics, 12 (1), 2014, pp. 119-144. Galhotra, Sumit, "Blacklisted in Balochistan" Committee to Protect Journalists, March 28, 2014 (http://www.cpj.org/blog/2014/03/blacklisted-in-baluchistan.php). Gall, Carlotta, "Peace Talks with Taliban Top Issue in Afghan Vote," New York Times, August 17, 2009 Galula, David, Counter-Insurgency Warfare: Theory and Practice, London: Pall Mall Press, 1964 Gandhi, Rohit, "Deradicalization: A Tall Order," Asia Pacific Defense Forum, July 1, 2012 637 (http://apdforum.com/enGB/article/rmiap/articles/print/features/2012/07/0 1/feature-pr- 10). Ganguly, Sumit and David P. Fidler, India and Counterinsurgency:Lessons Learned, New York: Routledge, 2009. "Gas Reserves Discovered in Sindh: Report," Express Tribune, June 19, 2013. Gates, Robert M., "Helping Others Defend Themselves: The Future of U.S. Security Assistance," Foreign Affairs, 89 (3), May/June, 2010, pp. 2-6. Gautam, P.K., Composition and Regimental System of the Indian Army: Continuity and Change, Delhi: Shipra Publications, 2008. Gayer, Laurent, "The Khalistan Militias: servants and Users of the State," Armed Militias of South Asia: Fundamentalists,Maoists and Separatists,Ed. Laurent Gayer and Christophe Jaffrelot, New York: Columbia University Press, 2009. Gayer, Laurent, Karachi: OrderedDisorderand the Strugglefor the City, New York: Oxford University Press, 2014 Gazdar, Haris, "'Counter-Insurgencies' in Pakistan," Economic and Political Weekly, May 20, 2006, pp. 1952-53. George, Alexander and Andrew Bennett, Case Studies and Theory Development in the Social Sciences, Cambridge: MIT Press, 2005. George, Alexander L., Case Studies and Theory Development: The Method of Structured, Focused Comparison, in Diplomacy: New Approached in History, Theory, and Politics, ed. Paul G. Lauren, New York: Free Press, 1979. Ghildiyal, Subodh, "Steady Rise in Inter-Caste Marriages Involving Dalits," Times ofIndia, September 19, 2013 (http://timesofindia.indiatimes.com/india/Steady-rise-in-inter-castemarriages-involving-Dalits/articleshow/22728944.cms). Ghildiyal, Subodh, and Swati Mathur, "Silent Revolution: Rise in Marriages of Others to Dalits," Times ofIndia, January 31, 2010 (http://timesofindia.indiatimes.com/india/Silentrevolution-Rise-in-marriages-of-others-to-Dalits/articleshow/5518307.cms); Ghosh, Palash, "Blood Red Corridor," InternationalBusiness Times, Feb. 28, 2014 (http://www.ibtimes.com/blood-red-corridor-maoist-rebels-kill-seven-policemen-chhattisgarh- india-1558617). Gianetti, Candice, Gaye Facer, Andy McCord, Smita Patel, and Vikram Singh, "North Central India," India, Ed. Diane Mehta, New York: Fodors LLC, 2005. Gilady, Lilach and Joseph MacKay, "Bringing the Insurgents Back In: Early Wars in British India," Terrorism and Political Violence, 26, 2014, pp. 1-2 1. Gilani, Ijaz, Voice of the People: Public Opinion in Pakistan, 2007-2008, Karachi: Oxford University Press, 2010. 638 & Gill, K.P.S., "Endgame in Punjab: 1988-1993," Faultlines: Writings on Conflict and Resolution, 1, 1999 Gilpin, Robert, War and Change in World Politics, New York: Cambridge University Press, 1981. Glancey, Jonathan, Nagaland:A Journey to India's ForgottenFrontier,London: Faber Faber, 2011. Glaser, Charles L., "The Security Dilemma Revisited," World Politics, 50 (1), October 1997, pp. 171-201. Goddard, Stacie E., "Uncommon Ground: Indivisible Territory and the Politics of Legitimacy," International Organization, 60 (1), Winter 2006, pp. 35-68. "Gogoi Seeks Special Fund from PM to Rehab Former Ultras," Rediff com, February 7, 2012 (http://www.rediff.com/news/eport/gogoi-seeks-special-fund-from-pm-to-rehab-formerultras/20120207.htm). "Gogoi urges PM to revamp Brahmaputra Board," Press Trust ofIndia, Feb 7, 2012. Gomes, Joseph Flavian, "The Political Economy of the Maoist Conflict in India: An Empirical Analysis" Unpublished Paper, 2011 (http://www.uclouvain.be/cps/ucl/doc/core/documents/Gomes.pdf). Gopal, Anand, Mansur Khan Mahsud, and Brian Fishman, "The Taliban in North Waziristan," Talibanistan:Negotiatingthe Borders Between Terror, Politics, and Religion, Ed. Peter Bergen Katherine Tiedemann, New York: Oxford University Press, 2013, pp. 128-163. Gordon, Michael R., "Military Hones a New Strategy On Insurgency," New York Times, October 5, 2006. Gough, Kathleen, Indian Peasant Uprisings, Economic and PoliticalWeekly, 9 (32/34), August 1974, pp. 1391-1412. "Government was Blocking Access to Balochistan IDPs," Dawn, January 17, 2007. Graham, Bradley, "Calls for Shift in Iraq Strategy Growing; Lawmakers, Experts Urge Military to Focus More on Protecting Population Centers," Washington Post, November 11, 2005. Grare, Frederic, "Balochistan: The State Versus the Nation," Carnegie Paper, April, 2013. Green, Matthew, "Hidden War Embodies Pakistan's Struggle," Financial Times, May 25, 2012. Griffiths, Ryan D., Catch and Release: Expansion, Contraction, and the Downsizing of States, PhD dissertation, Columbia University, 2010. Guha, Ramachandra, "Radical American Environmentalism and Wilderness Preservation: A 639 Third World Critique." EnvironmentalEthics 11 (1), 1989, pp. 71-83. Gul, Imtiaz, The Most DangerousPlace: Pakistan'sLawless Frontier, New York: Viking, 2011. Gunaratna, Rohan, Sri Lanka: A Lost Revolution? The Inside Story of the JVP, Kandy, Sri Lanka: Institute of Fundamental Studies, 1990. Guruswamy, Mohan, "Call for Pashtun Nation is Not Far Away," Rediff.com, July 28, 2009. Guruswamy, Mohan, "Armed Might Won't Defeat the Naxals," Rediff com, January 3, 2008 (http://www.rediff.com/news/2008/jan/03guest2.htm). Guruswamy, Mohan, "There Once Was a Hyderabad!" India Seminar, 585, 2008 Habibullah, Wajahat, My Kashmir: Conflict and the Prospectsfor EnduringPeace, Washington DC: U.S. Institute of Peace Press, 2008, p. 70. Habyarimana, James, Macartan Humphreys, Daniel N. Posner, and Jeremy M. Weinstein, "Why Does Ethnic Diversity Undermine Public Goods Provision?" American Political Science Review, 101 (4), 2007, pp. 709-725. Haider, Ejaz, "Counterinsurgency: The Myth of Sisyphus?" Pakistan's Counterterrorism Challenge, Ed. Moeed Yusuf, Washington, DC: Georgetown University Press, 2014, pp. 6382. Haider, Ejaz, "Intelligent Intelligence," Daily Times, July 26, 2007. Haider, Ejaz, "Time for Plan B," Daily Times, April 21, 2009. Haines, David R., "COIN in the Real World," Parameters, Winter 2008-09, pp. 43-59. Hammond Jeremy R., "The Situation in Swat: An Interview with Shahid R. Siddiqi," Foreign Policy Journal, July 9, 2009. Haqqani, Hussain, Pakistan:Between Mosque and Military, Washington, DC: Carnegie Endowment for International Peace, 2005. Haque, Saiful and Sujit Nath, "Operation Green Hunt Fizzles Out in Red Zone," India Today, Feb 22, 2010 (http://indiatoday.intoday.in/story/Operation+Green+Hunt+fizzles+out+in+Red+zone/l/85117. html). Hardin, Russell, Onefor All: The Logic of Group Conflict, Princeton, NJ: Princeton University Press, 1995. Harris, Gardiner, "For India's Persecuted Muslim Minority, Caution Follows Hindu Party's Victory," New York Times, May 16, 2014. Harrison, Selig S., "Baluch Nationalism and Superpower Rivalry," International Security, 5 640 . (3), Winter 1980-81, pp. 15 2 - 16 3 Harrison, Selig S., "Discarding an Afghan Opportunity," Washington Post, January 30, 2007 (http://www.washingtonpost.com/wpdyn/content/article/2007/01/29/AR2007012901448.html). Harrison, Selig S., "Ethnic Conflict in Pakistan: The Baluch, Pashtuns, and Sindhis," Conflict and Peacemaking in Multiethnic Societies, Ed. Joseph V. Montville, Lexington Ma: Lexington Books: pp. 301-325. Harrison, Selig S., "Ethnic Hostility is a Big, Maybe the Biggest, Part of the Afghan War," The Globe and Mail, Nov 2, 2009 (http://www.theglobeandmail.com/globe-debate/ethnic-hostilityis-a-big-maybe-the-biggest-part-of-the-afghan-war/articlel348333/). Harrison, Selig S., "Ethnicity and the Political Stalemate in Pakistan," Regional Imbalances and the National Question in Pakistan, Ed. Akbar S. Zaidi, Lahore: Vanguard Books, 1992, 226-260. Harrison, Selig S., "The Pashtun Time Bomb," New York Times, August 1, 2007 (http://www.nytimes.com/2007/08/01/opinion/01iht-edharrison. 1.6936601.html). Harrison, Selig S., In Afghanistan's Shadow: Baluch Nationalism and Soviet Temptations, New York: Carnegie Endowment for International Peace, 1981 Hart, B.H. Liddell, Strategy: the IndirectApproach, London: Faber and Faber Limited, 1967. Hasan, Mohammed, "Communalism in Indian Literature," Man andDevelopment, 1984, 6, pp. 144-49. Hasan, Qamar, Muslims in India: Attitudes, Adjustments andReactions, New Delhi: Northern Book Centre, 1988. Hashim, Ahmed S., When CounterinsurgencyWins: Sri Lanka's Defeat of the Tamil Tigers, Philadelphia, PA: University of Pennsylvania Press, 2013. Hassan, Ahmad, "Operation in Malakand, Waziristan is No Solution," Dawn, Nov 8, 2009. Hassan, Ali, Pakistan:Generals aur Siyasat, Urdu (Pakistan: Generals and Politics), Lahore: Vanguard Books, 1991, p. 167. Hassan, Khalid, "Zia Formed the MQM: General Beg," Daily Times, May 27, 2007 (http://archives.dailytimes.com.pk/national/27-May-2007/zia-formed-the-mqm-general-beg). Hassan, Nadir, "The Faqir of Ipi and the Taliban," Express Tribune, April 18, 2012 (http://tribune.com.pk/story/366427/the-faqir-of-ipi-and-the-taliban/). Hassner, Ron E., "'To Halve and to Hold': Conflicts over Sacred Space and the Problem of Indivisibility," Security Studies, 12 (4), 2003, pp. 1-33. Hazarika, Sanjoy, "Insurgents in India's Northeast Call for Peace Talks," New York Times, 641 February 20, 1991. Hazelton, Jacqueline L., "Compellence and Accommodation in Counterinsurgency Warfare," PhD Dissertation, Brandeis University, 2011. Hechter, Michael, Internal Colonialism: The Celtic Fringe in British National Development, 1536-1966, Berkeley, University of California Press, 1975. Henriksen, Thomas H., Afghanistan, Counterinsurgency, and the IndirectApproach, Hurlburt Field, Fl: Joint Special Operations University Press, 2010. Herbst, Jeffrey, States and Power in Africa: ComparativeLessons in Authority and Control, Princeton, N.J.: Princeton University Press, 2000. Hewstone, Miles Mark Rubin, and Hazel Willis, "Intergroup Bias," Annual Review of Psychology, 53, 2002, pp. 575-604. Hicks, Madelyn Hsiao-Rei and Michael Spagat, "The Dirty War Index: A Public Health and Human Rights Tool for Examining and Monitoring Armed Conflict Outcomes," PLOS Medicine, 5 (12), December 2008, pp. 1658-1664. Hines, John G. and Donald Mahoney, Defense and Counteroffensive Under the New Soviet Military Doctrine, Santa Monica: RAND, 1991. Hirschman, Albert 0., Exit, Voice, Loyalty: Responses to Decline in Firms, Organizations,and States, Cambridge, MA: Harvard University Press, 1970 Hnatkovska, Viktoria, Amartya Lahiri, and Sourabh B. Paul, "Breaking the Caste Barrier: Intergenerational Mobility in India," JournalofHuman Resources, 48 (2), Spring 2013, pp. 435-473. Hoffman, Bruce, "We Cant Win if We Don't Know the Enemy," Washington Post, March 25, 2007. Hoffhian, Bruce, British Air Power in PeripheralConflict, 1919-1976, Santa Monica: RAND, 1989. Hoffman, Elizabeth, Kevin McCabe and Vernon L. Smith, "Social Distance and OtherRegarding Behavior in Dictator Games," The American Economic Review, 86 (3), June 1996. Hoover, Amelia, Romesh Silva, Tamy Guberek, and Daniel Guzman, "The 'Dirty War Index' and the Real World of Armed Conflict," Unpublished Paper, Human Rights Data Analysis Group, May 23, 2009. Horowitz, Donald L., "Ethnic Identity," Ethnicity: Theory and Experience, Ed. Nathan Glazer, Daniel Patrick Moynihan, Corinne Saposs Schelling, Cambridge, MA: Harvard University Press, 1975. Horowitz, Donald L., Ethnic Groups in Conflict, Berkeley: University of California Press, 1985. 642 Hosain, Quatrina, "Paramilitary troops to be increased in Karachi," Agence France Presse, June 18, 1995. Hourani, Albert, "Ottoman Reform and the Politics of Notables," The Modem Mddle East: A Reader, Ed. Albert Hourani, Philip S. Khoury, and Mary C. Wilson, CA: University of California Press, 1993, pp. 83-109. "How Large if the NREGS?" Research Unit for Political Economy, March 23, 2012 (http://rupeindia.wordpress.com/2012/03/23/how-large-is-the-nregs/). Howard, Roger, "Evolving Rather than Receding, the Killing in Kashmir Continues," Jane's IntelligenceReview, 11 (1), Jan 1, 1999, p. 39. Human Rights Commission of Pakistan (HRCP), Annual Report, 1992 to 2006, Conflict in Balochistan: HRCPFact-FindingMissions, December 2005-January 2006. Hussain, Akmal, PakistanNationalHuman Development Report 2003 Poverty, Growth and Governance, Karachi: UNDP, 2003. Hussain, Brig. Javed, "The Hornet's Nest," The News, January 19, 2011 2011. Hussain, Hamid, "Demons of December - Road from East Pakistan to Bangladesh," Defence Journal (Karachi), December 2002. Hussain, Hamid, "Back in the Game - Tribes and Society in Pakistan," Defence Journal (Karachi), April 2007. Hussain, Jamal, "PAF in the Countrinsurgency Role in the Tribal Areas of Pakistan, Defence Journal (Karachi), November 2012. Hussain, Wasbir, "Insurgency in India's Northeast: Cross-border Links and Strategic Alliances," Faultlines, 17, 2004. Hussain, Wasbir, "Manipur: Impending Collapse of Governance," 2004. Hussain, Wasbir, "Manipur: Impending Collapse of Governance," South Asia Intelligence Review, 3 (4), August 9, 2004). Hussain, Wasbir, "Multi-force Operations in Counter Terrorism: A View from the Assam Theatre," Faultlines,9, July 2001 (http://www.satp.org/satporgtp/publication/faultlines/volume9/Article2.htm). Hussain, Zahid and Matthew Rosenberg, "Pakistani Peace Deal Gives new Clout to Taliban Rebels," Wall Street Journal, April 14, 2009 (http://online.wsj.com/article/SB123963706622913745.html). Hussain, Zahid, "It's War Now," Newsweek, January 15, 2006. Hussain, Zahid, The Scorpion's Tail: The Relentless Rise ofIslamic Militants in Pakistan-And How It ThreatensAmerica, New York: Free Press, 2010. 643 Hussein Mullick, Haider Ali, "Lions and Jackals: Pakistan's Emerging Counterinsurgency Strategy," Foreign Affairs (online), July 15, 2009 Hyder, Mohammed, October Coup, A Memoir of the Strugglefor Hyderabad,Roli Books, Delhi, 2012. "Hyderabad Police Force to be Reorganised," Times ofIndia, Nov. 8, 1948. "Hyderabad Refugees Problem, Government Plans for Resettlement," Times ofIndia, Nov 8, 1948, p. 7 "Hyderabad Relieved of Acute Shortage of Essential Goods: Second Phase of Military Government Begins," Times ofIndia, Nov 7, 1948, p. 3. "Hyderabad Settling Down Rapidly," Times ofIndia, Nov 25, 1948, p.1. "I.A.S. as on 1.1.1981," Muslim India, 1 (1), 1983, p. 29. Ilan, Shahar, "The Myth of Haredi Moral Authority," Haaretz, July 12, 2012 (http://www.haaretz.com/the-myth-of-haredi-moral-authority-1.450670). "Imran Khan's Statement on Acts of Violence," Pakistan Press International Information Services, Aug 6, 2004 (Factiva). Inbar, Efraim and Eitan Shamir, "'Mowing the Grass': Israel's Strategy for Protracted Intractable Conflict," Journal of Strategic Studies, 37 (1), 2014, pp. 65-90. "Indian Foreign Minister on Summit, Says Kashmir 'Core' of Indian Nationhood," BBC MonitoringSouth Asia, July 12, 2001. "An Indian Summary: Comparing Indian States and Territories with Countries" Economist, 2011 (http://www.economist.com/content/indian-summary) "India's Naxalite Insurgency: Not a Dinner Party," Economist, Feb 25, 2010 (http://www.economist.com/node/15580130). "India's Ungoverned Spaces: Out of the Trees," The Economist, June 1, 2013 (http://www.economist.com/news/asia/21578686-murderous-attack-highlights-neglect-indiasoutland-out-trees). Inter-Services Public Relations (Pakistan), "COAS Key Note Address - National Seminar Deradicalization at Mingora Swat," ISPR No PR157/2011, July 6, 2011. Inter-Services Public Relations (Pakistan), "COAS Key Note Address - National Seminar Deradicalization at Mingora Swat," ISPR No. PR157/2011, July 6, 2011. International Crisis Group, Pakistan:CounteringMilitancy in PA TA, Asia Report No. 242, Brussels/Islamabad: ICG, January 15, 2013. International Crisis Group, Pakistan: The Forgotten Conflict in Balochistan, ICG Asia 644 Briefing No. 69, Islamabad/Brussels: ICG, October 22, 2007. International Crisis Group, Pakistan: The Worsening Conflict in Balochistan, Asia Report No. 119, Islamabad/Brussels: ICG, September 14, 2006. International Crisis Group, Pakistan's TribalAreas: Appeasing the Militants, Asia Report No. 125, Islamabad/Brussels: ICG, Dec. 11, 2006. "Introduction," Central Reserve Police Force website, (http://crpf.nic.in/crpb.htm), last accessed April 25, 2014. "I Was One of India's Unclean Dalits... Now I Am a Millionaire," AssociatedPress, December 27, 2011 (http://www.thenational.ae/news/world/south-asia/i-was-one-of-indias-uncleanDalits-now-i-am-a-millionaire). Iyer, Lakshmi, "Direct versus Indirect Colonial Rule in India: Long-term Consequences," Review of Economics and Statistics 92 (4), November 2010, pp. 693-713. Jabar, Faleh A., "Why the Uprisings Failed," Middle East Report, 22 (176), May/June, 1992 (http://www.merip.org/mer/merl 76/why-uprisings-failed). Jackson, Brian A, "Counterinsurgency in a 'Long War': The British Experience in Northern Ireland, Military Review, Jan/Feb 2007, pp. 74-85. Jackson, Colin, "Defeat in Victory: Organizational Learning Dysfunction in Counterinsurgency," PhD Dissertation, Massachusetts Institute of Technology, June 2008. Jackson, Robert H., Quasi-states:Sovereignty, InternationalRelations, and the Third World, New York: Cambridge University Press, 1990. Jackson, Robert, "Juridical Statehood in Sub-Saharan Africa," Journal of International Affairs, 46, 1992, pp. 1-16. Jafa, Vijendra Singh, "Counterinsurgency Warfare: The Use & Abuse of Military Force," Faultlines3, 1999. Jaffrelot, Christophe, A History ofPakistanand Its Origins, London: Anthem Press, 2002. Jaffrelot, Christophe, India'sSilent Revolution: The Rise of the Lower Castes in North India, New York: Columbia University Press, 2003. Jaffrelot, Christophe, Religion, Caste, and Politics in India, New York: Columbia University Press, 2011. Jalal, Ayesha, The State ofMartialRule: The Origins ofPakistan'sPoliticalEconomy of Defense, Cambridge: Cambridge University Press, 1990. James, Sir Morrice, Pakistan Chronicle, New York: St. Martin's Press, 1993. Jan Orakzai, Ali Mohammad, "Situation in FATA: Causes, Consequences and the Way Forward," Policy Perspectives (published by Institute for Policy Studies), 6 (1), Jan-June, 645 2009. Jan, Delwar, "Malakand Division on Road to Civilian Control," The News International, Aug 8,2012 (http://www.thenews.com.pk/Todays-News-7-125321-Malakand-Division-on-road-tocivilian-control). Jan, Reza and Sam Worby, "Limited Goals, Limited Gains: The Pakistan Army's Operation in Kurram, Critical Threats, September 6, 2011. Jan, Reza, "Pakistani Military Operations in Orakzai," Critical Threats, May 13, 2010. Jan, Reza, "Paradise Regained: Swat One Year On," Critical Threats, May 25, 2010. Jan, Reza, "Trickling Home to South Waziristan," Critical Threats, December 13, 2010. Janine, Laishramcha "The Militarisation of Manipur," Article 2, 5 (6), Dec 2006. Janyala, Sreenivas, "When They Talked And Failed," Indian Express, March 14, 2010. Jervis, Robert, Why Intelligence Fails:Lessons from the IranianRevolution and the Iraq War, Ithaca: Cornell University Press, 2010. Jha, Saumitra and Steven Wilkinson, Does Combat Experience Foster Organizational Skill? Evidence from Ethnic Cleansing during the Partition of South Asia," American Political Science Review, 106 (4), November 2012, pp. 883-907. Jha, Shiv Puj an, "Police Finds Muslims an Easy Target" CNN-IBN, March 18, 2006. Jilangamba, Yengkhom, "Let's Stop Pretending There's No Racism in India," The Hindu, May 29, 2012. Jinnah, Akbar Ahmed, PakistanandIslamic Identity: The Searchfor Saladin, London: Routledge, 1997. Joes, Anthony James, America and Guerilla Warfare, Lexington: University of Kentucky Press, 2000. Joes, Anthony James, ResistingRebellion: The History and Politicsof Counterinsurgency, Lexington, KY: University Press of Kentucky, 2004. Johnson, Thomas H. and M. Chris Mason, "No Sign Until the Burst of Fire: Understanding the Pakistan-Afghanistan Frontier, International Security, 32 (4), Spring 2008, pp. 41-77. Johnston, Patrick B., "Does Decapitation Work? Assessing the Effectiveness of Leadership Targeting in Counterinsurgency Campaigns," International Security, 36 (4), Spring 2012, pp. 47-79. Johnston, Patrick, "The Treatment of Civilians in Effective Counterinsurgency Operations," PhD Dissertation, Northwestern University, December 2009. Jones, Owen Bennett, Pakistan:Eye of the Storm, New Haven: Yale University Press, 2002. 646 Jones, Seth G. and C. Christine Fair, Counterinsurgencyin Pakistan, Santa Monica, CA: Rand Corporation, 2010. Joshi, Manoj, The Lost Rebellion, New Delhi: Penguin, 1999. "Judicial Reform in Swat," Dawn, March 25, 2010 (http://archives.dawn.com/archives/29660). "Justice Minister Denies Army Attack on Balochistan," Agence France-Presse, October 3, 1974. FBIS DAILY REPORT. Middle East & North Africa, FBIS-MEA-74-194 on 1974-1004 p. TI. Accessed via Newsbank on February 19, 2014. Kadivar, Cyrus, "A Question of Numbers," Rouzegar-Now, August 8, 2003, found on http://www.emadbaghi.com/en/archives/000592.php (accessed August 21, 2014). Kagan, Frederick W., "The Two-Front War; Pakistan is Finally Doing its Part. Now We Need to Do Ours," The Weekly Standard, 15 (8), November 9, 2009. Kagan, Frederick W., Reza Jan, and Charlie Szrom, "The War in Waziristan: Analysis of Operation Rah-e-Nijat," Critical Threats, November 18, 2009 Kahl, Colin H., "How We Fight," Foreign Affairs, 85 (6), November-December 2006, pp. 83101. Kaiser, Dr Fawad, "Swat Deradicalisation Project - II," Daily Times, February 3, 2014 (http://www.dailytimes.com.pk/opinion/03-Feb-2014/swat-deradicalisation-project-ii). Kakakhel, Shakeel, "Integration of FATA into NWFP: An Analysis," Forum of Federations Project, Feb. 2010, p. 7 [http://www.forumfed.org/en/where/pakistan.php]. Kalita, Rakhee, "Writing Terror: Men of Rebellion and Contemporary Assamese Literature," Beyond Counter-Insurgency:Breaking the Impasse in NortheastIndia, Ed. Sanjib Baruah, New Delhi: Oxford University Press, 2009. Kalyanraman, Sankaran, "The Indian Way in Counterinsurgency," Democracies and Small Wars, Ed. Efraim Inbar, Portland: Frank Cass and Company, 2003, pp. 85-100. Kalyvas, Stathis N. and Laia Balcells, "International system and Technologies of Rebellion: How the End of the Cold War Shaped Internal Conflict," American Political Science Review, (3), August 2010, pp. 415-429. Kalyvas, Stathis N., "Ethnic Defection in Civil War," Comparative Political Studies, 2008, 41 (8), pp. 1043-68. Kalyvas, Stathis N., "The Urban Bias in Research on Civil Wars," Security Studies, 13 (3), Spring 2004, pp. 1-3 1. Kalyvas, Stathis N., "Wanton and Senseless? The Logic of Massacres in Algeria," Rationality and Society, 11 (3), 1999, pp. 243-285. Kalyvas, Stathis N., The Logic of Violence in Civil War, Cambridge: Cambridge University 647 & Press, 2006. Kanwal, Gurmeet (Brig.), "Strategic Stalemate in Assam," FinancialExpress, April 12, 2009, (http://www.financialexpress.com/story-print/445908). Kaplan, Fred, "CT or COIN? Obama Must Choose This Week Between Two Radically Different Afghanistan Policies," Slate (War Stories), March 24, 2009. Kapur, Ashok, Pakistanin Crisis, New York: Routledge, 1991 Kapur, Devesh, Chandra Bhan Prasad, Lant Pritchett, and D. Shyam Babu, "Rethinking Inequality: Dalits in Uttar Pradesh in the Market Reform Era," Economic & Political Weekly, 45 (35), August 28, 2010. Kapur, Devesh, Kishore Gawande, and Shanker Satyanath, "Renewable Resource Shocks and Conflict in India's Maoist Belt," Center for Global Development Working Paper #302, August 2012. Kardar, Shahid, "Polarization in the Regions and Prospects for Integration," Regional Imbalances and the National Question in Pakistan, Ed. Akbar S. Zaidi, Lahore: Vanguard Books, pp. 306-333. "Kashmir Muslim boy rules India, tops Civil Services exam," TwoCircles.net, May 6, 2010 (http://twocircles.net/2010may06/kashmirmuslimboyrules-india-tops_civilservicesexam .html). Kaszeta, Daniel J., "Lithuanian Resistance to Foreign Occupation 1940-1952," Lituanus Lithuanian Journal of Arts and Sciences, 34 (3), Fall 1988, (http://www.lituanus.org/1988/88_3_01.htm). Katoch, Lt. Gen. Prakash, "Why Maoist Insurgency Will Flourish in India," Indian Defence Review, June 13, 2013. Katznelson, Ira, "Flexible Capacity: The Military and Early American Statebuilding," Shaped by War and Trade: International Influences on American Political Development, Eds. Ira Katznelson and Shefter, Princeton, N.J.: Princeton University Press, 2002. Kaufmann, Chaim, "Possible and Impossible Solutions to Ethnic Civil Wars," International Security, 20 (4), Spring 1996, pp. 136-175. Keller, Arthur, "Propaganda and Peace Deals: The Taliban's Information War in Pakistan," CTC Sentinel, 1 (8), July 2008. Kennedy, Charles H., "Managing Ethnic Conflict: The Case of Pakistan," Regional Politics Policy, 3 (1), Spring 1993, pp. 123-143. Kennedy, Charles H., "Pakistan: Ethnic Diversity and Colonial Legacy," The Territorial Management of Ethnic Conflict, Ed. John Coakley, 2003, pp. 143-172. Kennedy, Charles H., "Policies of Redistributional Preference in Pakistan," Ethnic Preference and Public Policy in Developing States, Ed. Neil Nevitte and Charles H. Kennedy, Boulder, 648 CO: Rienner, 1986, pp. 63-93. Kennedy, Jonathan and Lawrence King, "Adivasis, Maoists and Insurgency in the Central Indian Tribal Belt," EuropeanJournalofSociology, 54 (1), April 2013, pp. 1-32. Kennedy, Jonathan and Sunil Purushotham, "Beyond Naxalbari: A Comparative Analysis of Maoist Insurgency and Counterinsurgency in Independent India," ComparativeStudies in Society andHistory, 54 (4), 2012. Kessler, Glenn, "Clinton Says Pakistan is Giving in to Extremists," Washington Post, April 23, 2009. Khalidi, Omar, Khaki and the Ethnic Violence in India, New Delhi: Three Essays Collective, 2003. Khalidi, Omar, Muslims in Indian Economy, Gurgaon, India: Three Essays Collective, 2006. Khaliq, Fazal, "Much Changed for Better Last Year, More Still to Come: GOC," Express Tribune, January 18, 2013 (http://tribune.com.pk/story/495525/much-changed-for-better-lastyear-more-still-to-come-goc/). Khaliq, Fazal, "New Beginnings: After Six Years of Closure, Saidu Sharif Airport Ready to Become Operational," Express Tribune, July 12, 2012 (http://tribune.com.pk/story/407078/ew-beginnings-after-six-years-of-closure-saidu-sharifairport-ready-to-become-operational/). Khaliq, Fazal, "Nizam-e-Adl Regulation: Top Sharia Court Set Up in Swat," Express Tribune, January 19, 2011 (http://tribune.com.pk/story/105998/nizam-e-adl-regulation-top-sharia-courtset-up-in-swat/); Khaliq, Fazal, "Peace Maintainable Only Through Public's Commitment," Express Tribune, Dec 29, 2012 (http://tribune.com.pk/story/486061/peace-maintainable-only-through-publicscommitment/). Khaliq, Fazal, "Reclaiming Swat: Influential Peace Jirga Member Escapes Attack," Express Tribune, Nov 14, 2011 (http://tribune.com.pk/story/291506/reclaiming-swat-influential-peacejirga-member-escapes-attack/). Khaliq, Fazal, "Speedy Justice: Mobile Courts for Swat Announced," Express Tribune, Nov. 26, 2011 (http://tribune.com.pk/story/297746/speedy-justice-mobile-courts-for-swatannounced/). Khaliq, Fazal, "Terrorists Behind Swat Unrest to Face Justice," Express Tribune, Nov 27, 2011 (http://tribune.com.pk/story/298244/terrorists-behind-swat-unrest-to-face-justice/). Khaliq, Fazal, "Timely Justice: Community Body Gives Quick Solutions for Small Disputes," Express Tribune, March 13, 2013 (http://tribune.com.pk/story/519924/timely-justicecommunity-body-gives-quick-solutions-for-small-disputes/). Khan, Adeel, "Pakistan's Sindhi Ethnic Nationalism: Migration, Marginalization, and the 649 Threat of Indianization Asian Survey, 42 (2), March/April 2002, p. 213-229. Khan, Adeel, Politics ofIdentity: Ethnic Nationalismand the State in Pakistan, New Delhi: Sage Publications, 2005 Khan, Afzal, "Revolt in Pakistan's NWFP: A Profile of Maulana Fazlullah of Swat," Terrorism Focus, 4 (38), November 20, 2007. Khan, Akbar, "Early Days of Freedom: Reminiscences of a Solider," Dawn (Independence Day Supplement, Karachi), August 14, 1960, p 1-2, 8. Khan, Akbar, Raiders in Kashmir, Lahore: Jang Publishers, 1992 Khan, Amina, "FATA: Voice of the Unheard - Path-Dependency and Why History Matters," Strategic Studies (Islamabad), 31 (1 and 2), Spring 2011, pp. 40-74. Khan, Commander Adnan Mazher, "Tribal Areas - Need for Integration into National Mainstream - Recommendations, Pakistan Navy War College Review, 2005-06. Khan, Commander Adnan Mazher, "Tribal Areas - Need for Integration into National Mainstream - Recommendations," Pakistan Navy War College Review, 2005-06, pp. 63-68. Khan, Gul Hassan, Memoirs of Lt. Gen. Gul Hassan Khan: The last Commander-in Chief of the Pakistan Army, Karachi: Oxford University Press, 1993. Khan, Ijaz, "Challenges Facing Development in Pakistan's FATA" NBR Analysis, 19 (3), August 2008, pp. 13-20. Khan, Ismail, "Militant Leader Threatens to Attack Forces, Dawn, July 30, 2005. Khan, Ismail, "Talibanisation Imperils Security, NSC Warned: Immediate Action Urged," Dawn, June 22, 2007 Ithaca: Cornell University Press, 2004. Khan, Javed Aziz, "Soldier Who Led Operation Against TNSM Passes Away," The News International, April 20, 2009 (http://www.thenews.com.pk/TodaysPrintDetail.aspx?ID=173321&Cat=7&dt=4/19/2009). Khan, Liquat Ali, "Understanding Pakistani Geopolitics," Huffington Post, Nov 11, 2013 Khan, Lt. Gen. Jahan Dad, PakistanLeadership Challenges, Karachi: Oxford University Press, 1999. Khan, M. Ilyas, "Balochistan 'Troops Numbers Cut,"' BBC News, May 14, 2008. Khan, M. Ilyas, "Taleban Spread Wings in Pakistan," BBC News, March 5, 2007. Khan, Nichola, Mohajir Militancy in Pakistan: Violence and Transformation in the Karachi Conflict, London: Routledge, 2010. Khan, Rina Saeed, "Swat Valley NGO Finds a Solution to the Assault on Education," The Guardian, May 10, 2011. 650 Khan, Roedad, Pakistan-A Dream Gone Sour, Karachi: Oxford University Press, 1997 Khan, Sohail, "Disband Agencies 'Death Squads', Stop Mly Action in Balochistan: SC," The News Sept 12, 2012. Khan, Zahid Ali, "Military Operations in FATA and PATA: Implications for Pakistan," Strategic Studies (Islamabad), 31/32 (4/1), Winter 2011 and Spring, 2012, pp. 129-146. Khanna, Gaurav and Laura Zimmerman data, "Guns and Butter? Fighting Violence with the Promise of Development," Working Paper, SSRN, April 1, 2014. Khanna, K.C., "The Troubled North-East: Impermanent Frontiers," Times ofIndia, March 14, 1974. Khattak, Daud Khan, "The Battle for Pakistan: Militancy and Conflict in the Swat Valley," Counterterrorism Strategy Initiative Policy Paper, New America Foundation, April 2010. Khattak, Daud, "'Clearing' Kurram," Foreign Policy's AfPak Channel, August 25, 2011. Khattak, Daud, "Reviewing Pakistan's Peace Deals with the Taliban," CTC Sentinel, 5 (9), September, 2012, pp. 11-14. Khattak, Iqbal, "Kabul Trained Baloch, Pushtoon Youth in 1970s: Ex-ANP Leader," Daily Times, February 15, 2010. Khoury, Philip, "Interview: Philip Khoury, Associate Provost," Precis (MIT Center for International Studies), Fall 2012. Kilcullen, David J., "Terrain, Tribes, and Terrorists: Pakistan, 2006-2008," Brookings Institution Counterinsurgency and Pakistan Paper Series, No. 3, September 10, 2009. Kilcullen, David J., "The Political Consequences of Military Operations in Indonesia, 194599," PhD Dissertation, University of New South Wales, 2001. Kilcullen, David J., Counterinsurgency, New York: Oxford University Press, 2010. Kilcullen, David, The Accidental Guerrilla:FightingSmall Wars in the Midst of a Big One, New York: Oxford University Press, 2009, pp. 48-49. Knights, Michael, "The ISIL's Stand in the Ramadi-Falluja Corridor," CTC Sentinel, 7 (5), May 2014, pp. 8-12. Kocher, Matthew Adam, Thomas B. Pepinsky, and Stathis N. Kalyvas, "Aerial Bombing and Counterinsurgency in the Vietnam War," American Journal of Political Science, 55 (2), 2011, 201-218. Kotwal, Ashok, Bharat Ramaswami, and Wilima Wadhwa, "Economic Liberalization and Indian Economic Growth: What's the Evidence?" JournalofEconomic Literature,49 (4), 2011, pp. 1152-99. Krishna, Ashok (Maj. Gen. (Retd)), "Counter Insurgency and Internal Security," The Indian 651 Army: A BrieffHistory, Ed. Ian Cardozo, New Delhi: United Service Institution, 2007, pp. 192227. Kudelia, Serhiy, "Choosing Violence in Irregular Wars: The Case of Anti-Soviet Insurgency in Western Ukraine," East European Politics and Societies, 27 (149), 2013, pp. 149-181. Kudelia, Serhiy, "Choosing Violence in Irregular Wars: The Case of Anti-Soviet Insurgency in Western Ukraine," Eastern European Politics and Societies and Cultures, 27 (1), February pp. 149-181. Kumar, Satish, "How Anti-Naxal Forces Weakened in Bihar," IPCS brief, #2821, February 27, 2009 (http://www.ipcs.org/article/naxalite-violence/how-anti-naxal-forces-weakened-in-bihar2821 .html). Kumar, Suneel, "Sikh Ethnic Uprising in India and Involvement of Foreign Powers," Faultlines, 18, January 2007. Kundi, Mansoor Akbar, "Insurgency Factors in Balochistan," Central Asia Journal, No. 64, Spring 2010. Kundi, Mansoor Akbar, "Insurgency Factors in Balochistan," The Nation, Oct 20, 2004. Kwama, Kenneth, "Why the Colonialists Banned Kikuyu, Meru from Joining the Army," The Standard (Kenya), Ocotber 8, 2013 (http://www.standardmedia.co.ke/?articleID=2000095098#). Kydd, Andrew H. and Barbara F. Walter, "The Strategies of Terrorism," International Security, 31 (1), Summer 2006, pp. 49-80. Lacina, Bethany and Nils Petter Gleditsch, "Monitoring Trends in Global Combat: A New Dataset of Battle Deaths," European Journal of Population, 21(2-3), 2005, pp. 145-166. Lacina, Bethany, "Explaining the Severity of Civil Wars," Journal of Conflict Resolution, 50 (2), April 2006, pp. 276-289. Lacina, Bethany, "How Governments Shape the Risk of Civil Violence: India's Federal Reorganization, 1950-56," American JournalofPoliticalScience, 58 (3), July 2014, pp. 720738. Lahiri, Amartya, and Kei-Mu Yi, "A Tale of Two States: Maharashtra and West Bengal," Review ofEconomic Dynamics, 12 (3), July, 2009, pp. 523-542. Laitin, David D., Hegemony and Culture: Politicsand Religious Change among the Yoruba, Chicago: University of Chicago Press, 1986. Lalwani, Sameer and Paul Staniland, "The Politics of Pakistani Strategy on the Northwest Frontier," Paper Presented at the American Political Science Association Annual Conference, Washington August 2014. Lalwani, Sameer, "Pakistani Military Capabilities for a Counterinsurgency Campaign: A Net Assessment," Counterterrorism Strategy Initiative Publication, New America Foundation, 652 September 2009. Lalwani, Sameer, "Pakistan's Counterinsurgency Strategy," Talibanistan: Negotiating the Borders Between Terror, Politics, and Religion, Ed. Peter Bergen and Katherine Tiedemann, New York: Oxford University Press, 2013. LaPorte, Jr., Robert, "Direct Development: An Analysis of the Approach and Methodology of the Narcotics Affairs Section/Peshawar of the U.S. Embassy Islamabad," Unpublished Paper, August 16, 2008. LaPorte, Jr., Robert, "Pakistan in 83.125 The Continuing Crisis," Asian Survey, 36 (2), Feb 1996, p. 179-89. Larson, Gerald James, India's Agony Over Religion, Albany: State University of New York Press, 1995. Lavoy, Peter R., Pakistan'sStrategic Culture, Washington, DC: SAIC, October 31, 2006 Lawson, Alastair, "'First' Kashmir Survey Produces 'Startling' Results," BBC News, May 27, 2010. Lebovic, James, The Limits of U.S. Military Capability:Lessonsfrom Vietnam andIraq, Baltimore: The John Hopkins University Press, 2010 Lee, Alexander and Kenneth A. Schultz, "Comparing British and French Colonial Legacies: A Discontinuity Analysis of Cameroon," Quarterly Journal of Political Science, 7 (4), 2012, 365410. Leites, Nathan and Charles Wolf, Jr., Rebellion andAuthority: An Analytic Essay on Insurgent Conflicts, Chicago: Markham, 1970. Levi, Margaret, OfRule and Revenue, Berkeley, CA: University of California Press, 1988. Levy, Adrian and Cathy Scott-Clark, The Meadow: Kashmir 1995 - Where the TerrorBegan, London: Harper Press, 2012. Liberman, Peter, Does Conquest Pay? The Exploitation of Occupied IndustrialSocieties, Princeton NJ: Princeton University Press, 1996. Licklider, Roy, "The Consequences of Negotiated Settlements in Civil Wars, 1945-1993," American Political Science Review, 89 (3), September 1995, pp. 681-90. Lieb, Peter, "Few Carrots and a Lot of Sticks: German Anti-Partisan Warfare in World War Two," Counterinsurgency in Modem Warfare, Ed. Daniel Marston and Carter Oxford: Osprey Publishing, 2008, pp 57-77. Lieberman, Evan S. and Prema Singh, "Conceptualizing and Measuring Ethnic Politics: An Institutional Complement to Demographic, Behavioral, and Cognitive Approaches," Studies in Comparative International Development, 47 2012, pp 255-286. Lieberman, Evan S. and Prerna Singh, "The Institutional Origins of Ethnic Violence," 653 Comparative Politics, 45 (1), October 2012, pp. 1-24. Lieberson, Stanley and Arnold R. Silverman, "The Precipitants and Underlying Conditions of Race Riots," American Sociological Review, 30 (6), Dec. 1965, pp. 887-898. Lieven, Anatol, "Pakistan's Paradoxical Survival," Pakistan's Counterterrorism Challenge, Ed. Moeed Yusuf, Washington, DC: Georgetown University Press, 2014, pp. 187-202. Lieven, Anatol, Pakistan:A Hard Country, New York: Public Affairs, 2011 Lindholm, Charles, "Justice and Tyranny: Law and the State in the Middle East," Journal of the Royal Asiatic Society, 9 (3), Nov 1999, p. 375-388. Lindsay, Jon and Roger Petersen, Varieties ofInsurgency and Counterinsurgencyin Iraq, 2003-2009, Newport, RI: US Naval War College, 2011. Linz Juan J. and Alfred Stepan, The Breakdown ofDemocraticRegimes, Baltimore, MD: Johns Hopkins University Press, 1978. Lobell, Steven E., "War is Politics: Offensive Realism, Domestic Politics, and Security Strategies," Security Studies, 12 (2), Winter 2002/03, pp. 165-95. Lodhi, Maleeha, "A Nation At Risk," South: The Third World Magazine, London, December 1983, p. 22. Long, Austin, "Small is Beautiful: The Counterterrorism Option in Afghanistan," Orbis, Spring 2010, pp. 199-214. Long, Austin, "The Anbar Awakening," Survival, 50 (2), 2008, pp. 67-94. Long, Austin, Doctrine of EternalRecurrence: The U.S. Military and Counterinsurgency Doctrine, 1960-1970 and2003-2006, Santa Monica, CA: RAND, 2008. Loshak, David, PakistanCrisis, New York: McGraw Hills Book Company, 1971, p. 112 Louis, Prakash, People Power: The Naxalite Movement in CentralBihar, New Delhi: Wordsmiths, 2002. Luttwak, Edward N., "Give War a Chance," Foreign Affairs, 78 (4), July/August 1999, pp. 3644. Luttwak, Edward, "Dead End: Counterinsurgency Warfare as Military Malpractice," Harper's, February, 2007, pp. 33-42. Lyall, Jason and Isaiah Wilson III, "Rage Against the Machines: Explaining Outcomes in Counterinsurgency Wars," International Organization 63 (1), 2009, pp. 67-106. Lyall, Jason, "Are Coethnics More Effective Counterinsurgents? Evidence from the Second Chechen War," American Political Science Review, 104 (1), February 2010, pp. 1-20. Lyall, Jason, "Do Democracies Make Inferior Counterinsurgents? Reassessing Democracy's 654 Impact on War Outcomes and Duration," International Organization, 64 (1), Winter 2010, pp. 167-92. Lyall, Jason, Graeme Blair, and Kosuke Imai, "Explaining Support for Combatants During Wartime: A Survey Experiment in Afghanistan," American Political Science Review, 107 (4), November pp. 679-705. Magnier, Mark, "In India, Bengalis Seek to Recapture Their Glory as Intellectuals," Los Angeles Times, September 8, 2012. Maher, Ajaz, "Pakistan's Army Steps Up Radio Wars," BBC, August 14, 2012 (http://www.bbc.co.uk/news/world-asia- 18766713); Author Interview with U.S. advisor involved in the program, April 2013 (#120). Mahsud, Mansur Khan, "The Taliban of South Waziristan," Talibanistan: Negotiating the Borders Between Terror, Politics, and Religion, Ed. Peter Bergen and Katherine Tiedemann, New York: Oxford Press, 2013, pp. 164-201. Mahurkar, Uday, "IB Chief Does What the Congress Didn't Want, Counters CBI on Ishrat Case," India Today, June 14, 2013 (http://indiatoday.intoday.in/story/asif-ibrahim-ib-chiefpolitical-move-haunt-congress-ishrat-jahan-encounter/1/280045.html). Malhotra, Jyoti, "The Road to Delhi Runs through Uttar Pradesh," Foreign Policy's South Asia Channel, May 6, 2014 (http://southasia.foreignpolicy.com/posts/2014/05/06/theroadtodelhiruns-throughuttar_p radesh). Malik, Iftikhar H., State and Civil Society in Pakistan:Politics ofAuthority, Ideology, and Ethnicity, New York: St. Martin's Press, 1997. Mallah, Samina, "Two-Nation theory Exists," Pakistan Times, August 6, 2007. Manipur:A Memorandum on Extrajudicial,Summary or ArbitraryExecutions, Guwahati: Civil Society Coalition on Human Rights in Manipur and the UN, March 28, 2012. "Manipur remains the second most conflict ridden state in the NE" Access Northeast(NENA), 3 (41), Jan 1-15, 2008 (http://www.nenanews.com/ANE%20Jan%20115,%2008/special%20reportl.htm) Mann, Michael, "The Autonomous Power of the State: Its Origins, Mechanisms and Results," European Journal of Sociology, 25 (2), 1984, pp. 185-213. Markey, Daniel S., No Exitfrom Pakistan:America's TorturedRelationship with Islamabad, New York: Cambridge University Press, 2013, p. 128. Markey, Daniel, SecuringPakistan'sTribal Belt, Council Special Report No. 36, Washington DC: Council on Foreign Relations, August 2008, p. 11 Marks, Thomas, ColombianArmy Adaptation to FARC Insurgency, Carlisle, PA: Strategic Studies Institute, January 2002 655 Marston, Daniel and Carter Malkasian, "Introduction," Counterinsurgency in Modem Warfare, Ed. Daniel Marston and Carter Malkasian, New York: Osprey Publishing, 2008, pp. 13-19. Marten, Kimberly, Warlords: Strong-arm Brokers in Weak States, Ithaca, NY: Cornell University Press, 2011. Marwah, Ved, "Naxal Gambit State Response," Tehelka, 45 (6), Nov. 14, 2009 (http://www.tehelka.com/naxal-gambit-state-response/). Marwah, Ved, India in Turmoil: Jammu & Kashmir, the Northeast andLeft Extremism, New Delhi: Rupa, 2009. Mascarenhas, Anthony, "Genocide," The Sunday Times, June 13, 1971. Mascarenhas, Anthony, The Rape ofBangladesh, New Delhi: Vikas Publications, 1971. Mazari, Sherbaz Khan, A Journey to Disillusionment, Karachi: Oxford University Press, 1999. Mazari, Shireen, "India's Unconventional Strategy," Defence Journal, January, 1999 Mazetti, Mark, The Way of the Knife, New York: Penguin Press, 2013. Mazumder, Rajit K., The Indian Army and the Making ofPunjab, Delhi: Permanent Black, 2003. McChrystal, Stanley, "Generation Kill," Foreign Affairs, 92 (2), March/April, 2013, pp. 2-8. McDuie-Ra, Duncan, "The 'North-East' Map of Delhi," Economic and PoliticalWeekly, 47 (30), July 28, 2012. McGeough, Paul, "Inside the Taliban's Heart of Darkness: The Violent Frontier," Sydney Morning Herald, May 22, 2007. McGirk, Tim, "Code of the Frontier," Time, January 30, 2005. McKelvey, Tara, "A Return to Hell in Swat," Foreign Policy, March 2, 2011. Mearsheimer, John J., ConventionalDeterrence, Ithaca: Cornell University Press, 1985. Mearsheimer, John J., Lidell Hartand the Weight offHistory, Ithaca, NY: Cornell University Press, 1988. Mehdudia, Sujay, "Sonia Extends Olive Branch to Maoists," The Hindu, April 22, 2009 (http://www.hindu.com/2009/04/22/stories/2009042262501000.htm). Mehra, Ajay K., "Naxalism in India: Revolution or Terror?" Terrorism and Political Violence, 12 (2), 2000. Menezes, Maj L, "Intelligence - Weapon of a Counter Insurgent," Pratividrohi,4 (2), September 1992, pp. 15-18 - this quote from page 16. Metcalf Barbara D. and Thomas R. Metcalf, A Concise History ofIndia, New York: 656 Cambridge University Press, 2002. Meyerle, Jerry, Unconventional Warfare and Counterinsurgencyin Pakistan:A BriefHistory, Alexandria, VA: Center for Naval Analysis, November 2012 Michael, S.M., Untouchable:Dalits in Modern India, Boulder, CO: Lynne Rienner, 1999. Migdal, Joel S., Strong Societies and Weak States: State-Society Relations andState Capabilitiesin the Third World, Princeton, N.J.: Princeton University Press, 1988. Miles, William F. S., HausalandDivided: Colonialism andIndependence in Nigeriaand Niger, Ithaca, NY: Cornell University Press, 1994. The Military Balance, London: IISS, 1971. "Minister Urges Plan to Solve 'Naxalite Problems,"' Delhi All India Radio Network, August 3, 1991, Daily Report. Near East & South Asia, FBIS-NES-91-150 on 1991-08-05, p. 60. Accessed via Newsbank on Feb 19, 2014. Mir, Hamid, "The Tragedy of Journalists in Balochistan," The News, October 2, 2012 (http://www.thenews.com.pk/Todays-News- 13-17846-The-tragedy-of-journalists-inBalochistan). Mishra, Trinath, Barrelof the Gun: The Maoist Challenge andIndian Democracy, New Delhi: Sheriden Book Company, 2007. Mitra, D.M., Genesis andSpread ofMaoist Violence andAppropriateState Strategy to Handle It, Sponsored by Bureau of Police Research and Development, New Delhi: Institute of Social Sciences, 2011. Mitra, Durga Madhab (John), UnderstandingIndian Insurgencies:Implicationsfor CounterinsurgencyOperations in the Third World, Carlisle, PA: Strategic Studies Institute, February 2007. Mitra, Sahib K., "Social Prejudice and Adolescent Personality," A Perspective on Psychology in India: Dr. S. M Mohsin Felicitation Volume, Ed. Simaleswar De and Durganand Sinha, Allahbad, India: Sinha, 1977, pp. 256-79. "Mizoram: False Hopes," Economic andPolitical Weekly, 17 (27) July 3, 1982, p. 1083. Mohan, Saumitra, "Naxalism: The Enemy Within," Institute for Peace and Conflict Studies, #2284, May 7, 2007 (http://www.ipcs.org/article/naxalite-violence/naxalism-the-enemywithin-2284.html). Mohan, Vishwa, "Over One-fifth of the Country's Police Force Just on Paper," Times of India, Dec. 24, 2012 (http://timesofindia.indiatimes.com/india/Over-one-fifth-of-countrys-policeforce-just-on-paper/articleshow/17737603.cms). Moitra, Sumit, "Pakistan Seeks Coal India Help to Mine Thar," DNA India, April 29, 2013. Monroe, Kristen Renwick, Ethics in an Age of Terror and Genocide: Identity and Moral 657 Choice, Princeton: Princeton University Press, 2011. "More Intensified Anti-Naxalite Drive in Bengal," Times ofIndia, October 5, 1970, p. 9. Moza, Capt. Yudhishthir, "A Perspective on Jammu and Kashmir," Pratividrohi,Autumn 1996. Mukherjee, Anit, "India's Experience with Insurgency and Counterinsurgency," The Routledge Handbook ofAsian Security Studies, Ed. Sumit Ganguly, Andrew Scobell and Joseph Liow, New York: Routledge, 2009. Mukherjee, Brig J.R., "Perspective of the North East," Pratividrohi,7 (2), September 1995, pp. 52-60. Mukherjee, Rohan, "Promoting Peace Through Development: Assam State, India, 2001-2009," Princeton Innovations for Successful Societies brief, 2010. Mukherjee, Shivaji, "Econometric analysis of spatial variation in Maoist insurgency in India," Paper Presented at American Political Science Association Annual Conference, Chicago, IL, August 2013. Mukherjee, Shivaji, "Why Are the Longest Insurgencies Low Violence? Politician Motivations, Sons of the Soil, and Civil War Duration," Civil Wars, 16 (2), 2014. Mukhim, Patricia, "Life Under Martial Law," The Telegraph, September 21, 2004 (http://www.telegraphindia.com/104092 1/asp/northeast/story_3784644.asp). Mukhim, Patricia, "Where is this North-east?" India InternationalCentre Quarterly, 32 (2/3), Winter 2005, p. 187. Mukhopadhyay, Ashoke Kumar, The Naxalites Through the Eyes of the Police: Select Notificationsfrom the CalcuttaPolice Gazett, 1967-1975, Kolkata: Deys Publishing, 2006. Mullick, Haider A. H., Pakistan'sSecurity Paradox: Counteringand FomentingInsurgencies, Hurlburt Field, FL: Joint Special Operations University, December 2009. Mullick, Haider Ali Hussein, "Holding Pakistan: The Second Phase of Pakistan's Counterinsurgency Operations," Foreign Affairs (online), March 24, 2010. Mullick, Haider Ali Hussein, "Recalibrating U.S.-Pakistan Relations," The Washington Quarterly, 35 (3), Summer 2012, pp. 93-107. Mullick, Rehan, and Joseph Hraba, "Ethnic Attitudes in Pakistan," International Journal of Intercultural Relations, 25 (2), March 2001, pp. 165-179. (http://www.sciencedirect.com/science/article/pii/S0147176700000493#). Mullik, B.N., My Years with Nehru: 1948-1964, Bombay: Allied Publishers, 1972. Munir, Asad, "Narratives on the TTP," The News International, July 6, 2013 (http://www.thenews.com.pk/Todays-News-9-188128-Narratives-on-the-TTP). 658 Munir, Brig. Ased, "The Faqir of Ipi of North Waziritan" Express Tribune, Nov. 15, 2010 (http://tribune.com.pk/story/77388/the-faqir-of-ipi-of-north-waziristan/). Muralidharan, Sukumar, "Alternate Histories: Hyderabad 1948 Compels a Fresh Evaluation of the Theology of India's Independence and Partition," History and Sociology ofSouth Asia, 8 (2),2014: 119-138. Murphy, Gardner, In the Minds of Men: The Study of Human Behaviorand Social Tensions in India, New York: Basic Books, 1953. Murshed, S. Mansoob and Scott Gates, "Spatial-Horizontal Inequality and the Maoist Insurgency in Nepal," Review of Development Economics, 9 (1), 2005, pp. 121-134. Murshid, M. Tazeen, "A House Divided: The Muslim Intelligentsia of Bengal," Ed. A. Donald Low, The Political Inheritance of Pakistan, London: MacMillan, 1991, p. 147. Mustafa Hamza, "Iraq Launches Operation Against al Qaeda in Anbar," Asharq Al-Awsat, May 27, 2013 (http://www.aawsat.net/2013/05/article55303287). Mylonas, Harris, The Politics ofNation-Building: Making Co-Nationals, Refugees, and Minorities, New York: Cambridge University Press, 2013. Nabhan, Ali A. and Matt Bradley, "Islamist Insurgents Advanced Toward Baghdad," Wall Street Journal, June 11, 2014. Nadeem, Azhar Hassan, Pakistan: The PoliticalEconomy ofLawlessness, Oxford University Press, 2002. Nafziger, E. Wayne and William L. Richter, "Biafra and Bangladesh: The Political Economy of Secessionist Conflict," Journal of Peace Research, 13 (2), 1976. Nag Sajal, ContestingMarginality:Ethnicity, Insurgency andSubnationalism in North-East India, New Delhi: Manohar, 2002. "The Naga Regiment" Bharat Rakshak, (http://www.bharat-rakshak.com/LANDFORCES/Units/Infantry/i 17-Naga-Regt.html). Nagchoudhury, Subrata, "Going Back to Blood and Tears" India Today, June 15, 1996. Nagl, John, Learningto Eat Soup With a Knife: CounterinsurgencyLessons From Malaya and Vietnam, Westport, Conn.: Praeger Publishers, 2002. Nambath, Suresh, "Manipur More a Colony of India," The Hindu, March 21, 2011 (http://www.thehindu.com/todays-paper/tp-opinion/manipur-more-a-colony-ofindia/article1 557379.ece); Nandani Sundar, "Spatial Issues and Maoist Insurgency in India," Observer Research Foundation, New Delhi, February 24, 2012 (http://www.orfonline.org/cms/sites/orfonline/modules/report/ReportDetail.html?cmaid=33813 &mmacmaid=33814). 659 Nandy, Ashis, "Engagement and Alienation in Indian Politics," ComparativePoliticalStudies, 7 (3) 1974, pp. 334-356. Naqvi, Lt. Col. Hamid Hussain, "Army and Sindh's Internal Security," Pakistan Army Journal, 35 (3), Autumn 1994, pp. 59-62. Narahari, N.S. (Lt. Gen.), Security Threats to North-EastIndia: the Socio-ethnic Tensions, New Delhi: Manas, 2002, pp. 110-111. Narang, Vipin, Nuclear Strategy in the Modern Era, Princeton, NJ: Princeton University Press, 2014. Naseem, S.M., "A Consistent Series of National Accounts for East and West Pakistan: 194950 to 1969-70," Pakistan Development Review, 1, 1975: 33-52. Naseemullah, Adnan A., "Foundations of the Frontier: Norms, Resources and Limited Sovereignty in Pakistan's Tribal Areas," Unpublished Paper, 2011. Naseemullah, Adnan and Paul Staniland, "Varieties of Indirect Rule: Explaining Governance Beyond Weberian Sovereignty," Governance, forthcoming. Naseemullah, Adnan, "Shades of Sovereignty: Explaining Political Order and Disorder in Pakistan's Northwest," Unpublished Paper, 2013. Nasr, Syyed Vali Reza, Islamic Leviathan: Islam and the Making of State Power, New York: Oxford University Press, 2001. Nasr, Vali, The DispensableNation:American ForeignPolicy in Retreat, New York: Doubleday, 2013. Nawaz, Shuja, "Pakistan's Summer of Chaos," Foreign Policy, June 17, 2009. Nawaz, Shuja, "The Pakistan Army and Its Role in FATA," CTC Sentinel, 2 (1), January 2009, pp. 19-21. Nawaz, Shuja, CrossedSwords: Pakistan, Its Army, and the Wars Within, Karachi: Oxford University Press, 2008. Nawaz, Shuja, FATA-A Most DangerousPlace: Meeting the Challenge of Militancy and Terror in the FederallyAdministered TribalAreas of Pakistan, Washington, D.C.: Center for and International Studies Press, January 2009. Nawaz, Shuj a, LearningBy Doing: The PakistanArmy's Experience with Counterinsurgency, Washington DC: Atlantic Council, February, 2011. "Naxalism: Y2K Problem Sans Solution?" The Hindustan Times, January 29, 2000 [Accessed on Factiva on July 16, 2014]. "Naxal Response has to be More Humane: CRPF DG," The Hindu, September 24, 2013 (http://www.thehindu.com/news/national/naxal-response-has-to-be-more-humane-crpf- 660 dg/article5163926.ece). Nesbitt, Eleanor, "Pitfalls in Religious Taxonomy: Hindus and Sikhs, Valmikis and Ravidasis," Religion Today, 6 (1), 1990, pp. 9-12. Newberg, Paula R., Double Betrayal: Repression andInsurgency in Kashmir, Washington DC: Carnegie Endowment for International Peace, 1995. Niazi, Lt. Gen. Amir Abdullah Khan (Ret.), The Betrayal ofEast Pakistan, Karachi: Oxford University Press, 1999. Nibedon, Nirmal, Mizoram: The DaggerBrigade, New Delhi: Lancer Publishers, 1980. Nibedon, Nirmal, Nagaland, The Night of the Guerillas,New Delhi: Lancer Publishers, 1978. Nichols, Robert, A History ofPashtun Migration, 1775-2006, Karachi: Oxford University Press, 2008. Nomani, Asra Q., "The New Untouchables," Washington Post, November 4, 2007. Noorani, A.G., "A.G. Noorani Blames Patel for Downfall of Hyderabad," Deccan Chronicle, November 30, 2013. Noorani, A.G., The Muslims ofIndia: A Documentary Record, New Delhi: Oxford University Press, 2003. Nunthara C., Impact ofIntroduction of Groupingof Villages in Mizoram, New Delhi: Omsons Publications, 1989. O'Hare, Martin, The IndonesianMilitary in IrianJaya, MA Thesis, Australian National University, 1991. Oetken, Jennifer L., "Counterinsurgency Against Naxalites in India," India and Counterinsurgency:Lessons Learned, Ed. Sumit Ganguly and David P. Fidler, New York: Routledge, 2009, pp. 127-15 1. Omissi, David, "Technology and Repression: Air Control in Palestine, 1922-36," Journal of Strategic Studies, 13 (4), 1990, pp. 41-63. "Operation Knight's Charge (Saulat Al-Fursan)," Institute for the Study of War, (https://www.understandingwar.org/operation/operation-knights-charge-saulat-al-fursan), Accessed August 22, 2014. Oppel, Jr., Richard A., Eric Schmitt and Thom Shanker, "Baghdad Bombings Raise Anew Questions About US Strategy in Iraq," New York Times, Sept 17, 2005. Ortiz, Roman D. and Nicolas Urrutia, "A Long Road to Victory: Developing Counterinsurgency Strategy in Colombia," Countering Terrorism and Insurgency in the 21st Century: Lessons from Fight Against Terrorism, Vol. 3, Ed. James J.F. Forest, Westport, CT: Praeger Security International, 2007, pp. 310-333. 661 "Over 10,000 militants surrendered since '98," Assam Tribune, March 16, 2010. "PAF Conducted 5,500 Bombing Runs in FATA since 2008," Express Tribune, November 14, 2011. Pakistan Air Force Official Website (http://www.paf.gov.pk/activebases.html). "Pakistan and the Taliban: A Real Offensive or a Phoney War?" The Economist, April 30, 2009 (http://www.economist.com/node/13576371). "Pakistan Army Strategy in Question After Attacks," Dawn, Oct. 22, 2012. "Pakistan Army Strategy in Question After Attacks," Dawn, October 22, 2012. "Pakistan Army Uses Bullets and Classrooms to Fight Militancy," Reuters, April 25, 2012 (http://tribune.com.pk/story/369703/pakistan-army-uses-bullets-and-classrooms-to-fightmilitancy/). "Pakistan Development Forum 2010" Reform Agenda-Road to Sustainable Economic Development, Govt of Gilgit-Baltistan, Annual Plan 2009-10. "Pakistan occupies strategic importance: Gen Ehsan," January 14, 2005 (posted at PakistaniDefence.com, Last accessed September 2, 2014). "Pakistani Authorities Close Three Radio Stations for Pro-Taliban Broadcasts," Associated Press, November 1, 2009 (http://www.voanews.com/content/a-13-2008-03-21voa56/349852.html). "Pakistani PM Announces Permanent Base in Swat," RFE/RL, January 16, 2014. Pakistan'sThar Coal Power Generation Potential, Private Power and Infrastructure Board, July 2008 Paluck, Elizabeth and Donald Green, "Prejudice Reduction: What Works? A Critical Look at Evidence from the Field and the Laboratory," Annual Review of Psychology, 60, pp. 339-67. Pape, Robert A., Bombing to Win: Air Power and Coercion in War, Ithaca: Cornell University Press, 1996. Parsons, Timothy H., The 1964 Army Mutinies and the Making of Modern EastAfrica, Westport CT: Praeger, 2003. "Pashto-Language TV Station [Khyber TV] Interviews Pakistan's President Musharraf," BBC Monitoring South Asia, May 21, 2006. "Pashtunistan - 1947-1955," Global Security, (http://www.globalsecurity.org/military/world/war/pashtunistan- 1 947.htm). Accessed August 4, 2014. Paul, Christopher, Colin P. Clarke, Beth Grill, and Molly Dunigan, Paths to Victory: Detailed 662 Insurgency Case Studies, Santa Monica, CA: RAND, 2013 Paul, T.V., The WarriorState: Pakistanin the Contemporary World, New York: Oxford University Press, 2014. Perlez, Jane, "Pressure from U.S. Strains Ties with Pakistan," New York Times, October 26, 2009. Petersen, Roger D., Resistance and Rebellion: Lessons from Eastern Europe, New York: Cambridge University Press, 2001. Petersen, Roger D., Western Intervention in the Balkans: The Strategic Use ofEmotion in Conflict, New York: Cambridge University Press, 2011. Petersen, Roger, Western Intervention in the Balkans: The Strategic Use of Emotion in Conflict, New York: Cambridge University Press, 2011. Petersen, Roger, "The Organization of the Military in Multiethnic States," Unpublished MA Thesis, University of Chicago, 1988. Petersen, Roger, UnderstandingEthnic Violence: Fear, Hatred, andResentment in TwentiethCentury EasternEurope, New York: Cambridge University Press, 2002. Petraeus, General David H., "CENTCOM in 2010: Views from General David H. Petraeus," Event Transcript, Institute for the Study of War, Washington, DC, January 22, 2010 (http://www.understandingwar.org/press-media/webcast/centcom-2010-views-general-david-hpetraeus). Petraeus, General David H., "Striking a Balance: A New American Security," Transcript of Keynote Address, Center for New American Security, Washington DC, June 11, 2009. Phadke, M.D. (Brig), "Countering Insurgency: Thrust in a Broader Dimension," Pratividrohi, 4 (2), September 1992, pp. 5-12. Phanjoubam, Pradip, "Northeast Problems as a Subject and Object," Beyond CounterInsurgency: Breaking the Impasse in NortheastIndia, Ed. Sanjib Baruah, New Delhi: Oxford University Press, 2009, pp. 147-169. Plowright Adam, "Two Thirds in Indian Kashmir Want Independence: Poll," Agence FrancePresse, September 12, 2010 "Police Crush Revolt of Kashmiri Secessionists," APP, August 7, 1980, Daily report, South Asia, FBIS-SAS-80-154 on 1980-08-07, p. E3. Accessed via Newsbank on February 19, 2014. Posen, Barry R., Inadvertent Escalation:Conventional War andNuclear Risks, Ithaca: Cornell University Press, 1991. Posen, Barry R., The Sources of MilitaryDoctrine: France, Britain, and Germany Between the World Wars, Ithaca: Cornell University Press, 1984. Posner, Daniel N., "The Political Salience of Cultural Difference: Why Chewas and Tumbukas 663 Are Allies in Zambia and Adversaries in Malawi," American Political Science Review, 98 November 2004, pp. 529-545. "Profile: The 'Good' Taliban Leader," Express Tribune, January 4, 2013 (http://tribune.com.pk/story/488855/profile-the-good-taliban-leader/). "Q&A: Pakistan's Waziristan Challenge," BBC News, October 19, 2009 (http://news.bbc.co.uk/2/hi/southasia/8296245.stm). "Queen Elizabeth II in Swat," The Friday Times, April 4, 2014 (http://www.thefridaytimes.com/tft/queen-elizabeth-ii-in-swat-1961/). Raghavan, Srinath, 1971: A Global History of the Creationof Bangladesh, Cambridge: Harvard University Press, 2013, p. 50. Rahman, Shaikh Azizur, "Amid Violence, Hamlet Joins Electoral Politics in Bengal," India Ink blog, New York Times, July 24, 2013 (http://india.blogs.nytimes.com/2013/07/24/amidviolence-hamlet-joins-electoral-politics-in-bengal/). Rajagopalan, Rajesh, "'Restoring Normalcy': The Evolution of the Indian Army' Counterinsurgency Doctrine," Small Wars & Insurgencies, 11 (1), Spring 2000, p. 52. Rajagopalan, Rajesh, "Innovations in Counterinsurgency: The Indian Army's Rashtriya Rifles," ContemporarySouth Asia, 13 (1), March 2004, pp. 25-37. Rajagopalan, Rajesh, "Insurgency and Counterinsurgency," India Seminar, 599, July 2009. Raman, B., The Kaoboys ofR&A W: Down Memory Lane, New Delhi: Lancer, 2007. Ramana, P.V., "A Critical Evaluation of the Union Government's Response to the Maoist Challenge," StrategicAnalysis, 33 (5), 2009, pp. 749-750. Ramana, PV, "Vacillating Andhra Naxal Policy," IPCS brief # 1728, May 4, 2005. Rammohan, E.N., "Blue Print for Counter Insurgency in Manipur," USIJournal, 548, AprilJune 2002. Rammohan, E.N., "Manipur: A Degenerate Insurgency," USI Journal, 547, Jan-March 2002. Rammohan, E.N., "Unleash the Good Force," Outlook India, July 16, 2012 (http://www.outlookindia.com/printarticle.aspx?281554). Rammohan, E.N., CounteringInsurgencies in India: An Insider's View, New Delhi: United Service Institution of India, 2011. Rammohan, E.N., Insurgent Frontiers:Essaysfrom the TroubledNortheast. New Delhi: India Research Press, 2005. Rana, Muhammad Amir, "Swat Deradicalization Model..." Conflict and Peace Studies (PIPS journal), 4 (2), April-June, 2011, pp. 5-12. 664 Rao, K. V. Krishna, In the Service of the Nation: Reminiscences, New Delhi: Viking, 2001. Rashid, Ahmed, "Pakistan's Continued Failure to Adopt a Counterinsurgency Strategy," CTC Sentinel, 2 (3), March 2009, pp. 7-10. Rashid, Ahmed, Descent into Chaos, London: Penguin Books, 2008 Rashid, Ahmed, Pakistanon the Brink: The Future ofAmerica, Pakistan, andAfghanistan, New York: Viking, 2012. Rasool, Kamran, "Pakistan's Perspective on the 'War on Terror,"' Military Technology, 32 (11), November 2008, pp. 12-17. Ravi, RN, "Nagaland: Descent into Chaos," The Hindu, January 23, 2014. Raza, Rafi, Zulfikar Ali Bhutto and Pakistan, 1967-1977, Karachi: Oxford University Press, 1997. "Recruit More Nagas in Naga Regiment: Jamir" NortheastNews Agency (Oriental Times), 3 (49-50), May 7-21, 2001 (http://nenanews.com/OT%2OMay%207%20%20%2OMay%202 1,01/oh 13.htm). Reghunath, Leena Gita, "The Believer," Caravan,February 1, 2014. Rehman, Major Ata Ur, "Responsibilities of Pakistan Army in LIC Environment," Pakistan Army Green Book 2002: Low Intensity Conflict, Rawalpindi, Pakistan Army General Headquarters, 2002, pp. 201-208. Rehman, Teresa, "Murder in Plain Sight," Tehelka, August 8, 2009 Reis, Bruno C., "The Myth of British Minimum Force in Counterinsurgency Campaigns during Decolonisation (1945-1970)," Journal of Strategic Studies, 34 (2), April 2011, pp. 245279. Reiter, Dan, and Curtis Meek, "The Determinants of Military Strategy," International Studies Quarterly 43 (2), 1999, pp. 363-387. "Report in Detention of Extremists Refuted," Delhi Domestic Service, September 17, 1974, Daily Report, Middle East & North Africa, FBIS-MEA-74-183 on 1974-09-19, p. U2. Accessed via Newsbank, February 19, 2014. Rettie, John, "India Sends in Troops to Stop Ethnic Slaughter," The Guardian,September 24, 1993, p. 16; Reynders, Thomas, "Consular Trip to West Irian," Confidential Airgram, May 10, 1968, from "Indonesia's 1969 Takeover of West Papua Not by 'Free Choice,"' National Security Electronic Briefing Book 128, July 9, 2004 (http://www2.gwu.edu/~nsarchiv/NSAEBB/NSAEBB 128/). Rhode, David, "Army Enlists Anthropology in War Zones," New York Times, October 5, 665 2007. Rich, Bruce, To Uphold the World: The Message ofAshoka and Kautilyafor the 21" Century, New Delhi: Viking, 2008. Ricks, Thomas E., "US Counterinsurgency Academy Giving Officers a New Mind-Set; Course in Iraq Stresses the Cultural, Challenges the Conventional," Washington Post, February 21, 2006. Rizvi, Hasan Askari, "Understanding the Insurgency," Daily Times, October 5, 2008. Rizvi, Hasan Askari, The Military and Politicsin Pakistan, 194 7-86, Lahore: Progressive Publishers, 1987. Robertson, Nic and Matiullah Mati, "Minority Group Emerges as a Force in New Afghan Parliament," CNN, December 2, 2010 (http://www.cnn.com/2010/WORLD/asiapcf/12/02/afghanistan.elections/). Robinson, Linda, "The Future of Special Operations, Foreign Affairs, 91 (6), Nov/Dec 2012, pp. 110-122. Ross, Michael L., "How Do Natural Resources Influence Civil War? A Medium-N Analysis," UCLA Department of Political Science, January 2, 2003. Ross, Michael L., "How Do Natural Resources Influence Civil War?" International Organization, 58, Winter 2004, pp. 35-67. Routray Bibhu Prasad, "Manipur: The Nexus again," South Asia Intelligence Review, 6 (14), October 15, 2007 (http://www.satp.org/satporgtp/sair/Archives/6_14.htm#assessment2). Routray, Bibhu Prasad, "Manipur: The State Abdicates," South Asia Intelligence Review, 6 (43), May 5, 2008. Rovner, Joshua, "The Heroes of COIN," Orbis, 56 (2), Spring 2012, pp. 215-232. Roy, Kaushik, The Armed Forces ofIndependent India, 1947-2006, New Delhi: Manohar Publishers, 2010. Rubel, Nora L., Doubtingthe Devout: The Ultra-Orthodoxin the Jewish American Imagination, New York: Columbia University Press, 2010. Rubin, Alisa J., "Afghans' Distrust Threatens U.S. War Strategy," New York Times, May 12, 2010. Rubin, Alissa J. and Michael R. Gordon, "Iraq's Military Seen as Unlikely to Turn the Tide," New York Times, June 23, 2014. Rudra, Ashok, "Urban Guerrillas in Calcutta," Economic and Political Weekly, 6 (28), 1971, pp. 1379-82. Rustomji, Nari, Imperilled Frontier:India'sNorth-EasternBorderlands, New Delhi: Oxford 666 University Press, 1983. Sahni, Ajai and Bibhu Prasad Routray, "SULFA: Terror By Another Name," Faultlines,Vol. 9, July 2001. Sahni, Ajai, "'Naxalism': The Retreat of Civil Governance," Faultlines, 5, 2000 (http://www.satp.org/satporgtp/publication/faultlines/volume5/Fault5-7asahni.htm). Sahni, Ajai, "Andhra Pradesh: The State Advances, the Maoists Retreat, South Asia Intelligence Review, 6 (10), September 18, 2007. (http://www.satp.org/satporgtp/sair/Archives/6_10.htm). Sahni, Ajai, "India's Maoists and the Dreamscape of 'Solutions'," South Asia Terrorism Portal, February 2010 (http://www.satp.org/satporgtp/ajaisahni/1OAS-3Seminar.htm). Sahni, Ajai, "Maoists: Enduring Strengths," South Asia Intelligence Review 10 (39), April 2, 2012 (http://www.satp.org/satporgtp/sair/Archives/sair10/10_39.htm#assessmentl) Sahni, Ajai, "Maoists: Surviving Adversity," South Asia Intelligence Review, 12 (37), March 18, 2014 (http://www.satp.org/satporgtp/sair/Archives/sair12/12_37.htm#assessmentl). Saksena, Amit R., "India's Muslim Soldiers," The Diplomat, May 20, 2014 (http://thediplomat.com/2014/05/indias-muslim-soldiers/). Saksena, N. S., "Why Muslims Do Not Fare Well," Times ofIndia, May 31, 1983, p. 8. Salik, Siddique, Witness to Surrender, Karachi: Oxford University Press, 1977. Samad, Yunas, "In and Out of Power but not Down and Out: Mohajir Identity Politics," Pakistan: Nationalism Without a Nation, Ed. Christophe Jaffrelot, New York: Books, 2002, pp. 63-84. Sambanis, Nicholas, "Do Ethnic and Nonethnic Civil Wars Have the Same Causes? A Theoretical and Empirical Inquiry (Part 1)," Journal of Conflict Resolution, 45 (3), 2001, pp. 259-82. Sambanis, Nicholas, Jonah Schulhofer-Wohl, and Moses Shayo, "Parochialism as a Central Challenge in Counterinsurgency," Science, 336 (6083), May 18, 2012, p. 805-808. Sappenfield, Mark, "Afghanistan's Success Story: The Liberated Hazara Minority," Christian Science Monitor, August 6, 2007 (http://www.csmonitor.com/2007/0806/p06s02-wosc.html). Sarbahi, Anoop, "Insurgent-Population Ties and the Variation in Trajectory of Peripheral Civil Wars, ComparativePoliticalStudies, 47 (10), September 2014, pp. 1470-1500. Sartori, Giovanni, "Concept Misformation in Comparative Politics," American Political Science Review 64 (4), 1970, pp. 1033-1053. Sartori, Giovanni, "Concept Misformation in Comparative Politics," American Political Science Review, 64 (4), 1970, pp. 1033-1053. 667 Sarwar, Beena, "Pakistan-Politics: Radical Muslims Ignite Northern Hills," IPS-Inter Press Service, November 7, 1994. Sayah, Reza, "Can the Return of Justic Halt the Taliban?" CNN, Nov 14, 2009 (http://edition.cnn.com/2009/WORLD/asiapcf/ 1/14/pakistan.swat.justice). Sayeed, Khalid Bin, "Pakistan in 82.625 Internal Stresses More Serious than External Problems," Asian Survey, 24 (2), February 1984, pp. 219-228. Schaffer, Teresita C., Kashmir: The Economics ofPeace Building, Washington DC: Center for Strategic and International Studies, 2005. Schelling, Thomas, Arms andInfluence, New Haven, CT: Yale University Press, 1966 Schermerhorn, R.A., Comparative Ethnic Relations: A Frameworkfor Theory and Research, Chicago: University of Chicago Press, 1970 Schermerhorn, R.A., Ethnic Plurality in India, Tuscan: University of Arizona Press, 1978. Schofield, Carey, Inside the PakistanArmy, London: Biteback Publishing, 2011. Schweller, Randall L., "Unanswered Threats: A Neoclassical Realist Theory of Underbalancing," International Security, 29 (2), Fall 2004, pp. 159-201. Scott-Clark, Cathy and Adrian Levy, "How to Defuse a Human Bomb: Rescuing the Taliban's Teenage Recruits," The Guardian, October 15, 2010. Scott, James C., Seeing Like a State: How CertainSchemes to Improve the Human Condition Have Failed, New Haven, CT: Yale University Press, 1998. "Security Mutiny Sparks Fears of Rebel Attacks in Northeast India," Agence France-Presse, December 12, 2000. Seghal, Ikram, "Mission Impossible," Ikram Seghal Blog, September 24, 1994 (http://www.sehgalfamily.com/?p=1548).. Seghal, Ikram, "The Ostrich Syndrome," Ikram Seghal Blog, December 22, 1994 (http://www.sehgalfamily.com/?p=2345). Sen, Avirook, "Assam: Disarming Action," India Today, Aug 10, 1998; Sen, Avirook, "Assam: Jungle Justice" India Today, Nov 8, 1999, Factiva accessed 3/5/14. Sen, Lionel Protip, Slender Was the Thread: Kashmir Confrontation, 1947-1948, New Delhi: Orient Longman, 1969. Sepp, Kalev I., "From 'Shock and Awe' to 'Hearts and Minds': The Fall and Rise of US Counterinsurgency Capability in Iraq," Third World Quarterly, 28 (2), pp. 217-230. Seymour, Kristen, "De-radicalisation: Psychologists' War Against Militants," Express Tribune, July 17, 2011 (http://tribune.com.pk/story/211479/de-radicalising-rehab- 668 psychologists-war-against-militants/). Shah, Aqil, The Army and Democracy: Military Politicsin Pakistan, Cambridge: Harvard University Press, 2014. Shah, Brig. Syed Ithar Hussain, "How to Cope with LIC at National Level," PakistanArmy Green Book 2002: Low Intensity Conflict, Rawalpindi, Pakistan Army General Headquarters, 2002. Shah, Mehtab Ali, The ForeignPolicy of Pakistan:EthnicImpacts on Diplomacy 1971-1999, London: I.B. Tauris, 1997. Shah, Saeed, "In Remote Baluchistan, Pakistan Fights Shadowy War," McClatchy News, March 29, 2012. Shah, Saeed, "Pakistan Plans Air Assault in Waziristan," The Globe and Mail, July 13, 2009 (http://www.theglobeandmail.com/news/world/pakistan-plans-air-assault-inwaziristan/article 1204752/). Shah, Saeed, "Pakistan Taliban Leader Faces Threat form Fellow Tribesman," McClatchy Newspapers, June 13, 2009. Shah, Tayyab Ali, "Pakistan's Challenges in Orakzai Agency," CTC Sentinel, 3 (7), July 2010. Shahid, Saleem, "Action Group on Balochistan Urged," Dawn, September 10, 2012. Shahikumar, V.K., "Police Cannot Take on Maoists," Indian Defence Review, 25 (4), Oct-Dec 2010. Shahzad, Khurram, "What Ails Balochistan?" The Nation, Feb. 14, 2005. Shahzad, Syed Saleem, "Musharrafs moment draws near," Asia Times Online, June 9, 2004 (http://www.atimes.com/atimes/SouthAsia/FF09DfO3.html). Shahzad, Syed Saleem, "Taliban and al-Qaeda: Friends in Arms," Asia Times Online, May 5, 2011. Shahzad, Syed Saleem, "Warning shots for Musharraf," Asia Times Online, June 12, 2004 (http://www.atimes.com/atimes/SouthAsia/FF12Df04.html). Shahzad, Syed Saleem, Inside Al-Qaeda and the Taliban: Beyond Bin Laden and 9/11, London: Pluto Press, 2011. Sharma, A.K. (Lt. Col.), "The Assam Rifles," BharatRakshak Monitor, 4 (5), March-April 2002 (http://www.bharat-rakshak.com/MONITOR/ISSUE4-5/sharma.html) Sharma, A.K. (Lt. Col.), "The Assam Rifles," Indian Defence Review, 16 (4), 2002 Sharma, A.K. (Lt. Col.), "The Assam Rifles," IPCS brief #638, November 22, 2001 (http://www.ipcs.org/article/military/the-assam-rifles-638.html) 669 & Sharma, D.P., The Punjab Story: Decade of Turmoil, New Delhi: APH Publishing Corporation, 1996. Shekatkar, D.B., "India's Counterinsurgency Campaign in Nagaland," India and Counterinsurgency:Lessons Learned, Ed. Sumit Ganguly and David P. Fidler, New York: Routledge, 2009. Shinn, James and James Dobbins, Afghan Peace Talks: A Primer, Santa Monica: RAND, 2011. Shinwari, Naveed Ahmad, UnderstandingFATA 2011: Attitudes Towards Governance, Religion & Society in Pakistan'sFederallyAdministered TribalAreas, Vol. V, Islamabad: Community Appraisal and Motivation Program, 2012. Shinwari, Naveed Ahmad, UnderstandingFATA: Attitudes Towards Governance, Religion Society in Pakistan'sFederallyAdministered TribalAreas, Vol IV, Islamabad: Community Appraisal and Motivation Program, 2010. "The Showdown in Malakand," The Frontier Post, November 7, 1994, p. 10 (FBIS Daily Report. Near East & South Asia, FBIS-NES-94-216 on 1994-11-08, p. 70 (Accessed via Newsbank, February 19, 2014). Siddiqa, Ayesha, "Pakistan's Counterterrorism Strategy: Separating Friends from Enemies," Washington Quarterly, 34 (1), Winter 2011, pp. 149-162. Siddiqa, Ayesha, MilitaryInc.: Inside Pakistan'sMilitaryEconomy, London: Pluto Press, 2007. Siddiqi, Abdul R., East Pakistanthe Endgame: An Onlooker'sJournal 1969-1971, Karachi: Oxford University Press, 2004. Siddiqi, Farhan Hanif, The Politics of Ethnicity in Pakistan: The Baloch, Sindhi and Mohajir Ethnic Movements, New York: Routledge, 2012. Siddique, Osama, Pakistan'sExperience with FormalLaw: An Alien Justice, New York: Cambridge University Press, 2013. Siddiqui, Imran Ahmed, "Shunned, Raped Mizo Girl Quites Delhi," The Telegraph, November 1, 2011 (http://www.telegraphindia.com/1111 101/jsp/nation/story_14693129.jsp). Siddiqui, Taha, "Deradicalizing Boys in Pakistan, Christian Science Monitor, May 25, 2012. Sidhva, Shiraz "Survey of India 5: Delhi Rules with Carrot and Stick," FinancialTimes, 30 September 1993, p. (Accessed on Factiva, March 5, 2014). Silinsky, Mark, "The Problem with Pakistan," inFocus Quarterly, Winter 2010 (http://www.jewishpolicycenter.org/2145/pakistan-problem). Silver, Eric, quoting Salman Kurshid, " India's Other Nation - The Status of the Minority Muslims," The Guardian,August 28, 1986. 670 & Simpson, Erin Marie, "The Perils of Third-Party Counterinsurgency Campaigns," PhD Dissertation, Harvard University, June 2010, 2010. Singh, Col. Purushottam, "To Combat Left Wing Extremism, Tackle its Roots Rather than Merely Treating Symptoms," Pratividrohi, 1 (3), March 2011, pp. 25-29. Singh, Gopal, "Hindu-Sikh Fraternity: Don't Negate History of Five Centuries," Times of India, July 14, 1986, p. 8. Singh, Kanwar Randip, The Nagas of Nagaland:DesperadoesandHeroes ofPeace, Deep Deep Publishing: 1987. Singh, Manmohan, "PM's Valedictory Address at the Seminar on the Occasion of Golden Jubilee of National Defence College," October 22, 2010. Singh, N. Somorendro, "Integration and Development in North-east India: An Assessment," Indian JournalofPoliticalScience, 67 (2), April-June 2006. Singh, Prakash, India's Northeast: The Frontierin Ferment, Hurlbert Field, FL: Joint Special Operations University, 2008. Singh, Prakash, The Naxalite Movement in India, New Delhi: Rupa, 2005, Singh, Pritam, Federalism, NationalismandDevelopment: India and the Punjab Economy, New York: Routledge, 2008. Singha, Lt. Col. Iqbal Singh, "Combating LIC in J&K," Pratividrohi,1995. Sinha, Ashish, "UP and Bihar India's Top Two IAS Churners," India Today, Feb 1, 2010 (http://indiatoday.intoday.in/story/UP+and+Bihar+IndiaE2%AC%22s+top+two+IAS+churn ers/1/81822.html). "Six New Military Bases to Be Set Up in Sindh," The Pakistan Observer, March 10, 1993, p. 1. Slepyan, Kenneth, Stalin's Guerrillas: Soviet Partisans in World War II, Lawrence, KS: University Press of Kansas, 2006. Smitson, Scott, The Road to Good Intentions: British Nation-buildingin Aden, Case Study No. 12, Center for Complex Operations: Naval Postgraduate School, 2010. Snedden, Christopher, The Untold Story of the People ofAzad Kashmir, New York: Columbia University Press, 2012. Social, Economic and EducationalStatus of the Muslim Community in India: A Report, New Delhi: Prime Minister's High Level Committee, Cabinet Secretariat, Government of India, November 2006. "Southern Iraq," Institute for the Study of War, (https://www.understandingwar.org/region/southern-iraq), Accessed August 22, 2014. Spiegel, Peter, "'Oil Spot' Replaces 'Whack-a-Mole' Strategy," Financial Times, November 671 30, 2005. Srinavasan, Nirmala, PrisonersofFaith: A View from Within, New Delhi: Sage Publications, 1989. Srinivasan, Nirmala, Identity Crisis of Muslims: Profiles of Lucknow Youth, New Delhi: Concept Publishing Company, 1980. Srinivasan, Rukmini, "Massive Misuses of Funds Bleeds War on Naxals," The Economic Times (India), April 10, 2010. Stackhouse, John, "Karachi Pays High Price for Peace," The Globe and Mail, Oct. 25, 1996, p. A16. Staniland Paul, "Cities on Fire: Social Mobilization, State Policy, and Urban Insurgency." Comparative Political Studies, Vol. 43, No. 12 (December 2010), pp. 1623-1649. Staniland, Paul, "Explaining Civil-Military Relations in Complex Political Environments: India and Pakistan in Comparative Perspective," Security Studies, 17 (2), 2008, pp. 322-362. Staniland, Paul, "Kashmir Since 2003: Counterinsurgency and the Paradox of "Normalcy," Asian Survey, 51 (5), 2013, pp. 931-957. Staniland, Paul, "States, Insurgents, and Wartime Political Orders," Perspectives on Politics, 10 (2), June 2012: 243-264. Staniland, Paul, Networks ofRebellion: ExplainingInsurgent Cohesion and Collapse, Ithaca: Cornell University Press, 2014 Stanton, Jessica, "Strategies of Violence and Restraint in Civil War," PhD Dissertation, Columbia University, 2009. Statiev, Alexander, The Soviet Counterinsurgencyin the Western Borderlands, New York: Cambridge University Press, 2010. Staub, Ervin, "Ethnopolitical and Other Group Violence: Origins and Prevention," EthnopoliticalWarfare: Causes, Consequences, and PossibleSolutions, Ed. Daniel Chirot and Martin E. P. Seligman, DC: American Psychological Association, 2001. "Is the Strategy Against the Naxals Working?" India Today, Nov. 6, 2009 (http://indiatoday.intoday.in/story/Is+the+Strategy+Against+the+Naxals+Working/1/69648.ht ml). Subramanian, L.N., "Operation Polo: The Liberation of Hyderabad" BharatRakshak, 2(3), November/December 1999. Sundarayya, P., "Telangana People's Armed Struggle, 1946-195 1. Part Three: Pitted Against the Indian Army," Social Scientist, 1 (9), April 1973 Sundaryya, P., TelanganaPeople'sStruggle and its Lessons, Calcutta: Chadha, 1972. 672 "Sunni-Shia Relations in Iraq," Iraq: A Country Study, Ed. Helen Chapin Metz, Washington DC: GPO for the, Library of Congress, 1988 (http://countrystudies.us/iraq/38.htm). Swami, Praveen, "The Colors of Army Green," Frontline, Jan 2, 2004 (http://www.frontline.in/static/html/fl2026/stories/20040102002803800.htm). Swami, Praveen, "Despair and Democracy," Frontline, 21 (12), June 5-18, 2004, (http://www.frontline.in/static/html/fl2112/stories/20040618000106500.htm). Swami, Praveen, India, Pakistan andthe Secret Jihad: The Covert War in Kashmir, 19472004, New York: Routledge, 2007. Swidler, Nina, "Brahui Political Organization and the National State," Pakistan's Western Borderlands: The Transformationof a Political Order, Ed. Ainslie T. Embree, New Delhi: Vikas 1977, pp. 109-125. Taheri, Amir, The PersianNight: Iran Under the Khomeinist Revolution, New York: Encounter Books, 2009. Talbot, Ian, Pakistan:A Modern History, New York: St. Martin's Press, 1998. Talbot, Phillips, "Kashmir and Hyderabad", World Politics, 1 (3), 1999, pp. 324-25. Talukdar, Mrinal, Utpal Borpujari and Kaushik Deka, Secret Killings ofAssam, Guwahati, India: Nanda Talukdar Foundation, 2009. Talukdar, Sushanta, "Mahanta at the helm of secret killings in Assam: Saikia Commission," The Hindu, 16 Nov 2007, (http://www.hindu.com/2007/11/16/stories/2007111654701400.htm). Tandon, Prakash, Punjab Century: 1857-1947, Berkeley, CA: University of California Press, 1968. Tankel, Stephen, Domestic Barriersto Dismantlingthe MilitantInfrastructurein Pakistan, Washington DC: US Institute of Peace, September 2013. Tarrow, Sidney, "Bridging the Quantitative-Qualitative Divide," Rethinking Social Inquiry: Diverse Tools, Shared Standards, Ed. Henry E. Brady and David Collier, Lanham, MD: Rowman and Littlefield, 2004, pp. 171-180. Telford, Hamish, "Counter-Insurgency in India: Observations from Punjab and Kashmir," The Journalof Conflict Studies, 21 (1), 2001. Teltumbde, Anand, "Race or Caste, Discrimination is a Universal Concern," Economic and Political Weekly, 44 (34), August 22-28, 2009, p. 17. Tempest, Rone, "Sikh Unrest: Roots are not Persecution," Los Angeles Times, August 4, 1985 (http://articles.latimes.com/print/I 985-08-04/opinion/op-4492_1_sikh-men). Temple-Raston, Dina, "Pakistan's Ambitious Program to Re-Educate Militants," NPR, April 1, 673 2013. Tewari, Manish, MP Lok Sabha debate, July 8. 2009 "Need to formulate an effective strategy to combat the menace of naxalism in the country", 15t Lok Sabha Tharoor, Shashi, India: From Midnight to the Millennium and Beyond, New York: HarperPerennial, 1998, p. 11 "They're Disgruntled Tribals, Not Naxals: KJ George," DNA India, June 19, 2013 (http://www.dnaindia.com/bangalore/report-theyre-disgruntled-tribals-not-naxals-kj-george1850102). Thomson, Mike, "Hyderabad 1948: India's Hidden Massacre," BBC Magazine, September 23, 2013. (http://www.bbc.com/news/magazine-24159594). Tillin, Louise, "United in Diversity? Asymmetry in Indian Federalism," Publius, 37 (1), Winter 2007. "Timeline: A History of the MQM," Dawn, May 25, 2013 (http://www.dawn.com/news/1027569/timeline-a-history-of-mqm). Titus, Paul and Nina Swidler, "Knights, Not Pawns: Ethno-Nationalism and Regional Dynaics in Post-Colonial Balochistan," International Journal of Middle East Studies, 32, 2000, pp. 4769. Titus, Paul, "Honor the Baloch, Buy the Pushtun: Stereotypes, Social Organization and History in Western Pakistan," Modern Asian Studies, 32 (3), July 1998, pp. 657-687. Toft, Monica Duffy and Yuri M. Zhukov, "Denial and Punishment in the North Caucasus: Evaluating the Effectiveness of Coercive Counter-insurgency," Journal of Peace Research, 49 (6), pp. 785-800. Toft, Monica Duffy, The Geographyof Ethnic Violence: Identity, Interests, and the Indivisibility of Territory, Princeton, NJ: Princeton University Press, 2003. Toor, Saadia, State ofIslam: Culture and Cold War Politics in Pakistan, London: Pluto Press, 2011. Treisman, Daniel, "Rational Appeasement," International Organization, 58 (2), Spring 2004, pp. 345-373. "Tribesmen May Seek more Time," Dawn, October 11, 2003 (http://archives.dawn.com/2003/10/11/top4.htm). Tsai, Lily L., Accountability Without Democracy: Solidary Groups and Public Goods Provision in Rural China, New York: Cambridge University Press, 2007. Tudor, Maya, The Promise ofPower: The Origins of Democracy in India andAutocracy in Pakistan, New York: Cambridge University Press, 2013. 674 Tully, Mark and Satish Jacob, Amritsar: Mrs. Gandhi'sLast Battle, London: Cape, 1985. Tully, Mark, No Full Stops in India, New Delhi, India: Viking, 1991. Twenty-Five Years ofPakistan in Statistics, 1947-72, Central Statistical Office, Economic Affairs Division, Ministry of Finance, Planning and Development, Govt. of Pakistan, Karachi, 1972. Tyagi, Vidya Prakash, MartialRaces of Undivided India, New Delhi: Kalpaz Publications, 2009. The U.S. Army/Marine Corps CounterinsurgencyFieldManual, Chicago: University of Chicago Press, 2007. U.S. Consulate (Dacca) Cable, Extent of Casualtiesin Dacca, March 31, 1971, Confidential, Source: Record Group 59, Subject Numeric File 1970-73, Pol and Def, Box 2530 Accessed via National Security Archive Electronic Briefing Book No. 79, Dec. 16, 2002 (http://www2.gwu.edu/~nsarchiv/NSAEBB/NSAEBB79/). "U.S. Counterinsurgency Strategies in Iraq," Stratfor Global Intelligence, July 7, 2003, (http://www.stratfor.com/weekly/us_counterinsurgencystrategiesiraq). Accessed June 13, 2014. U.S. Department of State, Issue Paper for Vice President of the United States (VPOTUS), "Counterterrorism Activities (Neo-Taliban)," December 9, 2005, Secret, 4 pp. [Excised] (from National Security Archive Electronic Briefing Book 325, document 14 [http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB325/]). Ucko, David H., "Counterinsurgency in El Salvador: The Lessons and Limits of the Indirect Approach," Small Wars and Insurgencies, 24 (4), 2013, pp. 669-695. Ulfelder, Jay and Benjamin Valentino, "Assessing Risks of State-Sponsored Mass Killing," Unpublished Paper, Political Instability Task Force, February 2008. Ullman, Harlan, James Wade Jr., L.A. Edney, Frederick Franks, Jr., Charles Homer, Jonathan Howe, and Keith Brendly, Shock andAwe: Achieving Rapid Dominance, Washington: D.C., Defense University Press, 2002. Vadlamannati, Krishna Chaitanya, "Why Indian Men Rebel? Explaining Armed Rebellion in the Northeastern States of India, 1970-2007," JournalofPeace Research, 48 (5), 2011. Valentino, Benjamin A., FinalSolutions: Mass Killing and Genocide in the 20th Century, Ithaca: Cornell University Press, 2004. Valentino, Benjamin, Paul Huth, and Dylan Balch-Lindsay, "'Draining the Sea: Mass Killing and Guerrilla Warfare," International Organization, 58 (2), 2004, pp. 375-407. Van Evera, Stephen, Guide to Methods for Students ofPoliticalScience, Ithaca: Cornell University Press, 1997. Vardys, V. Stanley, "The Partisan Movement in Postwar Lithuania," Lituanus (Lithuanian 675 Quarterly Journal of Arts and Sciences), 15 (1), Spring 1969 (http://www.lituanus.org/1969/69_1_02.htm). Varma, Subodh, "SC/STs fail to break caste ceiling: No SC in 149 top government officers, 40 pc do menial jobs," Economic Times, September 6, 2012. Varshney, Ashutosh, "Ethnic Conflict and Civil Society: India and Beyond," World Politics, 53 (3), April 2001, pp. 362-98. Varshney, Ashutosh, "Nationalism, Ethnic Conflict, and Rationality," Perspectives on Politics, 1 (1), March 2003, pp. 85-99. Varshney, Ashutosh, Ethnic Conflictand Civic Life: Hindus and Muslims in India, New Haven, CT: Yale University Press, 2002. Varshney, Ashutosh, Ethnic Conflict and Civic Life: Hindus andMuslims in India, New Haven: Yale University Press, 2002. Venkatesan, J., "Salwa Judum is Illegal, Says Supreme Court," The Hindu, August 3, 2011 (http://www.thehindu.com/news/national/salwa-judum-is-illegal-says-supremecourt/article2161246.ece). Venkatesh, Sudhir, GangLeaderfor a Day: A Rogue Sociologist Takes to the Streets, New York: Penguin, 2008. Verghese, Ajay and Emmanuel Teitelbaum, "Colonialism and Armed Conflict in the Indian Countryside," Working Paper, SSRN, July 18,2014. Verghese, B.G., India'sNortheastResurgent: Ethnicity, Insurgency, Governance, Development, Delhi: Konark, 1996. Verma, Arvind, "The Police and India's Maoist Insurgency," PolicingInsurgencies: Cops as Counterinsurgents,Ed. C. Christine Fair and Sumit Ganguly, New Delhi: Oxford University Press, 2014. Vij-Aurora, Bhavana, "Armed and Dangerous," India Today, April 20, 2011 (http://indiatoday.intoday.in/story/ministry-of-home-affairs-report-country-losing-war-againstnaxals/1/135349.html) Vij-Aurora, Bhavana, "Armed and Dangerous," India Today, April 20, 2011. Vogt, Heidi, "Afghan Army Struggles with Ethnic Divisions," Associated Press, July 27, 2010 (http://www.nbcnews.com/id/38432732/ns/worldnews-southandcentral-asia/t/afghanarmy-struggles-ethnic-divisions/#.U4zWLseLGUc). Wallace, Paul, "Countering Terrorist Movements in India: Kashmir and Khalistan," Democracy and Counterterrorism:Lessons from the Past, Ed. Robert J. Art and Louise Richardson, 2007, pp. 425-481. Wallerstein, Immanuel, The Modern World System: CapitalistAgriculture and the Origins of 676 the European World Economy in the Sixteenth Century, New York: Academic Press, 1974. Walsh, Declan, "Pakistan's Secret Dirty War," The Guardian, March 29, 2011. Walsh, Declan, "Pakistani Premier Faces Verdict in Clash with Court," New York Times, April 25, 2012. Walter, Barbara F., "The Critical Barrier to Civil War Settlement," International Organization, 51 (3), Summer 1997, pp. 335-364. Walter, Barbara F., Reputation and Civil War: Why SeparatistConflicts Are So Violent, New York: Cambridge University Press, 2009. Ward, Stephen R., Immortal: A MilitaryHistory of Iran and Its Armed Forces, Washington DC: Georgetown University Press, 2009. Warraich, Omar, "Refugees Head Home After Army Scatters the Taliban," The Independent, Aug 3, 2009 (http://www.independent.co.uk/news/world/asia/refugees-head-home-after-armyscatters-the-taliban- 1 766564.html). Waseem, Mohammad, "Ethnic Conflict in Pakistan: The Case of the Mohajir Qaumi Movement," Regional Security, Ethnicity and Governance: The Challengesfor South Asia, Ed. Justus and Christian Wagner, New Delhi: Manohar, 1998, pp. 73-92. Watts, Stephen, Jason H. Campbell, Patrick B. Johnston, Sameer Lalwani, Sarah H. Bana, Countering Others'Insurgencies:UnderstandingU.S. Small-FootprintIntervention in Local Context, Santa Monica, RAND, 2014. "We Must Root Out Naxal Meance, Build On Our Successes: Manmohan Singh," DNA India, November 23, 2013 (http://www.dnaindia.com/india/report-we-must-root-out-naxal-menacebuild-on-our-successes-manmohan-singh-1923622); "We Only Receive Back the Bodies," The Economist, April, 7, 2012. Weinbaum, Marvin G., "Hard Choices in Countering Insurgency and Terrorism Along Pakistan's North-West Frontier," Journal of International Affairs, Vo. 63, No. 1, Fall/Winter 2009, pp. 73-88. Weinstein, Jeremy, Inside Rebellion: The Politics ofInsurgent Violence, New York: Cambridge University Press, 2007. Weiss, Gordon, The Cage: The Fightfor Sri Lanka and the Last Days of the Tamil Tigers, London: Bodley Head, 2011 Weisskopf, Thomas E., Affirmative Action in the United States andIndia: A Comparative Perspective, New York: Routledge, 2004. Welch Jr., Claude Emerson, Anatomy ofRebellion, Albany: State University of New York Press, 1980. Wellhofer, E. Spencer, "Core and Periphery: Territorial Dimensions in Politics," Urban 677 Studies, 26 (3), June 1989, pp. 340-355. White Joshua T., and Shuja Ali Malik, "Governance Reforms in Pakistan's Tribal Areas: The Long Road to Nowhere?" Peace Brief No. 135, United States Institute of October 15, 2012. White Paperon Baluchistan, Rawalpindi: Government of Pakistan, October 19, 1974. White Paperon Hyderabad,New Delhi: Government of India, 1948. White Paperon Jammu & Kashmir, New Delhi: Government of India, 1948. "Why is this Region Asia's Insurgency Haven?" Tehelka, August 7, 2004. Wilkinson, Steven I., Votes and Violence: ElectoralCompetition and Ethnic Riots in India, New York: Cambridge University Press, 2004. Wilkinson, Steven Ian, "India, Consociational Theory, and Ethnic Violence," Asian Survey, 40 (5), Sept-Oct 2000. Wilkinson, Willis, and Graham Denyer, The Killing Consensus:Police, Organized Crime and the Regulation ofLife andDeath in UrbanBrazil, Berkeley, CA: University of California Press, forthcoming. Winks, Robin W. and Alaine M. Low, The Oxford History of the British Empire: Historiography, Oxford University Press, 2001. Wirsing, Robert G., Baloch Nationalism and the Geopoliticsof Energy Resources: the ChangingContext ofSeparatism in Pakistan, Carlisle, PA: Strategic Studies Institute, April 2008. Wirsing, Robert G., India, Pakistan, andthe Kashmir Dispute: On Regional Conflict and its Resolution, New York: St. Marten's Press, 1994. Woodward, Bob, Obama's Wars, New York: Simon and Schuster, 2010. Worrall, James, "Bringing the Soil Back in: Control and Territoriality in Western and NonWestern COIN," The New Counter-InsurgencyEra in CriticalPerspective, Ed. David Martin Celeste Ward Gventer, and M.L.R. Smith, New York: Palgrave Macmillan, 2014. Worth, Robert F., "Blast Destroys Shrine in Iraq, Setting Off Sectarian Fury," New York Times, February 22, 2006. Wright, Jr., Theodore P., "Center-Periphery Relations and Ethnic Conflict in Pakistan: Sindhis, Mohajirs, and Punjabis," Comparative Politics, 23 (3), April 1991, pp. 299-312. Wucherpfennig, Julian, Nils W. Metternich, Lars-Erik Cederman, and Kristian Skrede Gleditsch, "Ethnicity, the State, and the Duration of Civil War," World Politics, 64 (1), January pp. 79-115. Yong, Tan Tai, The Garrison State: The Military, Government and Society in ColonialPunjab, 678 1849-1947, New York: Vanguard Books, 2005. Yumnam, Jiten and Phulindro Konsam, "Militarisation & Impunity in Manipur," Article 2, 5 (6), Dec 2006. Zaheer, Hasan, The Separation ofEast Pakistan: The Rise andRealization ofBengali Muslim Nationalism, Karachi: Oxford University Press, 1994. Zhukov, Yuri M., "Counterinsurgency in a Non-Democratic State: The Russian Example," The Routledge Handbook ofInsurgency and Counterinsurgency, Ed. Paul B. Rich and Isabella Duyvesteyn, New Routledge, 2012, pp. 286-300. Zhukov, Yuri M., "Population Resettlement in War Theory and Evidence from Soviet Archives ," Journal of Conflict Resolution, (forthcoming). Zhukov, Yuri, "Examining the Authoritarian Model of Counter-insurgency: The Soviet Campaign Against the Ukrainian Insurgent Army," Small Wars and Insurgencies, 18 (3), September 2007, pp. Zurutuza, Karlos, "A Black Hole for Medis in Balochistan," Al Jazeera, February 5, 2014 (http://www.aljazeera.com/indepth/features/2014/02/black-hole-media-balochistan2014238128156825.html). 679