IN THE SECOND HALF of the twentieth century, civil wars eclipsed inter–state war as the "far greater scourge" in terms of death toll, duration, and occurrence and with it, a host of new conflict puzzles emerged.1 One puzzling feature is that many states seem to defy normative, political, or strategic incentives and choose costly, heavy–handed, counterinsurgency methods, producing high levels of violence and civilian casualties.2Recent work finds that only 11% of 20th century counterinsurgency campaigns sought to protect civilians,3 and even incumbents that provided the model for Western counterinsurgency doctrine regularly departed from strategies of minimal force.4 A separate but related puzzle is the seemingly casual, haphazard, and limited manner with which states go about fighting rebellion, evidenced by the high number of civil conflicts left as "draws"5 or with enduring "low–activity," particularly in Asia.6 While outcomes are generally murky, states will often persist with costly, unsuccessful strategies. Why then do states choose brutal or minimalist strategies to fight rebellion despite persistent or repeated failure? Understanding the strategic logic of these incumbents in civil war—like when they escalate and when they under react—is essential for policymakers to anticipate states' destabilizing strategies, stem spillovers, and restructure incentives to mitigate violence.
Due to limited prior work on the subject, my research seeks to explain why states choose a certain type of strategy to fight 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 state elites and the identity group rebelling.
CONCEPTUALIZING COUNTERINSURGENCY STRATEGIES
The first step of this project was to theorize types of strategy or the range of values an incumbent's counterinsurgency strategy can take on. Much of current work focuses on whether states fight rather than how,7 or by treating all incumbent counterinsurgency strategies the same.8 But the persistence of the empirical puzzles described earlier suggests more than one approach to fight rebellion, a variation found not only between states but also within states.
Based on a review of military strategy, counterinsurgency, and numerous cases of states fighting rebellion, I identified two critical dimensions of strategy—the level of effort invested and the level of violence employed9 —and from high/low combinations of these two dimensions, I derived four ideal type strategies: attrition,population security, enfeeblement, and cooptation. These strategies can be distinguished by their theories of victory, signature tactics, selection of tradeoffs, and the relative levels of casualties, manpower, money and materiel involved.
CORE–PERIPHERY RELATIONS: EXPLAINING VARIED STATE STRATEGIES
What then might explain varied state strategies in different cases of rebellion? Prevailing theories of strategy rely on state strength, regime type, or culture but have little purchase in explaining sub–national variation—that is, the same state fighting differently over time or across regions. To explain state strategy, I develop a theory based on a country's core–periphery relations—that is, a state's relative value of territory and the embattled rebel identity group's positional status within the state.
First, I argue the variation in the state incumbent's level of effort is explained by the state's value of the contested territory. The incumbent will spare no effort when rebels contest or threaten core territory—a region of high value and strategic importance measured by the level of integration, productivity, and physical and social infrastructure present. Rebels in under–valued, non–strategic peripheral zones will generally elicit low levels of state effort in either an enfeeblement or cooptation strategy intended to contain but not defeat rebellion. One exception is if rebels manage to project force on to core areas, then incumbents will be sufficiently motivated to exert greater effort and defeat them since valued territory is threatened. A state has plenty of incentives to confront rebels with force, but peripheral territories do not provide strong incentives for decisive strategies. The costliness of sustaining a lengthy, resource–intensive campaign and building up a dominant state presence to control peripheries can be prohibitive and cause a state to balk at such a venture, conserve its resources, and deploy a minimalist strategy that merely limits conflict intensity below a certain threshold.
Second, I draw from social identity theory and ethnic conflict studies of inter–group bias10 to argue that the positional status of the rebels' identity group can shape and constrain the level of violence the state is willing to employ. An identity group esteemed by the state with high positional status—most discernable through state elites' assessments of that group's character as well as its representation within state institutions like the security forces and civil bureaucracy—will incentivize the state to choose a less violent and more discriminate strategy. Both affective and strategic mechanisms guide this choice of restraint. State officials advising or leading the counterinsurgency campaign are more likely to empathize with valued groups, even if they are minorities, rendering the state sensitive to the groups' grievances, concerned about indiscriminate violence, and willing to embrace military restraint and address the "demand–side" of rebellion. If security forces are composed of high numbers from the rebels' identity groups, the incumbent will be more vulnerable to backlashes—like desertion or mutiny—that pose strategic threats to cohesion. Co–ethnic representation in state institutions can also offer the state greater information and trust to substitute more discriminate targeting, credible bargaining, and the redressing of grievances for high violence within a military campaign.
CONTEXT: SOUTH ASIA
South Asia offers a compelling context within which to examine these questions and test my theory of state counterinsurgency strategy. It contains a significant number of rebellions, both historic and ongoing, with tremendous subnational variation that exhibits many of the puzzles detailed above. There are also observable differences in strategies employed by the same state incumbent, and enough observations to control for a host of other factors that might influence strategy. Moreover, South Asia has become far more consequential for U.S. foreign policy, not only since 9/11, but also in the "pivot" towards Asia.
I tested my argument in 13 campaigns from India, 15 campaigns from Pakistan, and 7 campaigns from Sri Lanka. The approach allows for tightly paired comparisons, longitudinal analysis in the same conflict region, and in–depth case studies to process–trace decision–making, but also sufficient observations to establish distinct patterns. My analysis was based on a close scrutiny of the case literature, extensive data collection on the regions and rebel identity groups, and nearly half a year of fieldwork in Pakistan, India, the Kashmir Valley, and the British archives. I also drew on new sources of information, such as micro–level and time–series data, local journals and newspapers, government reports, military service journals and doctrinal works, officials' memoirs, and over 140 semi–structured interviews, more than half with retired high–level officials from the civil service, military, or police.
My research confirmed that core–periphery relations had a major effect on incumbent strategy, even when controlling for macro–structural features like state–strength, regime type, and organizational culture, which would undoubtedly have some influence on strategic choices. I found that both threatened territory and rebel identity were consequential in shaping a state incumbent's strategy. High valued or strategic territories that were contested or threatened by rebels elicited high effort responses by the state (attrition or population security). The Indian state exerted far greater effort when rebellions threatened its economically and strategically valuable regions of Punjab and Kashmir than nearly all other counterinsurgency campaigns. The Pakistani state did the same in Sindh and Bangladesh and the Sri Lankan state in its Western and Southern provinces. Additionally, low positional status elicited strategies of high violence (enfeeblement or attrition) against rebels and their civilian base. The highest levels of violence were employed by India against Northeast tribal and Kashmiri Muslim rebels, by Pakistan against Bengali and Baluch rebellions, and by Sri Lanka against Tamil rebels, all groups with very low status and poor representation within the state.
Extensive qualitative evidence also validates a number of the theory's mechanisms. Even when rebelling, esteemed identity groups elicited greater empathy from state policy elites and sympathy for their grievances. These elites expressed their conviction in the group's loyalty save for hardcore insurgents, a desire for rehabilitation of many rebel combatants, and concern for the moral and strategic consequences of indiscriminate violence. They also held greater confidence in their ability to acquire better information or deploy tactical substitutes for violence to mitigate rebel support. This was the Indian state's response to Sikh rebels in Punjab, the Pakistani state's response to Pashtun rebels in the northwest, and the Sri Lankan state's response to Sinhalese rebels in the southwest. Meanwhile, state elites' disdain and broad suspicion of groups with low positional status led to perceptions of disloyalty and guilt by association allowing for higher levels of violence against combatants and civilians alike. Neither a group's positional status nor a territory's value are easily malleable, but both can gradually shift over extended periods of time as has been the case with the schedule castes and tribes of India or parts of Pakistan's northwest frontiers.
My research offers some unique but policy–relevant conclusions applicable beyond South Asia. First, states are more likely to deploy a minimalist approach against rebellions in undervalued regions that merely contains conflict below a certain threshold, sometimes yielding a steady–state of chronic, low–level 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, achieving a monopolization of violence is the far more daunting task. Additionally, despite numerous cross–national studies concluding that identity does not affect civil war dynamics, in these sub–national studies I find the positional status of an identity group is critical to explaining state strategy and levels of violence or restraint. The findings, moreover, are not limited to South Asia as my initial probes have found some evidence for this theory in conflicts from the Philippines, Colombia, Turkey, Russia, Indonesia, Iraq, and even Britain.
Furthermore, I find that democracy alone is not sufficient to restrain states from unleashing high levels of violence. India in 1990, Pakistan in 2010, and 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, South Waziristan, and Tamil Eelam, respectively.
Finally, despite the often–theorized divergence of India's and Pakistan's state capacity, institutions, and democratic culture, it is striking how similar patterns of core-periphery relations have shaped their respective counterinsurgency strategies and patterns of conflict. It suggests that their colonial legacy of the British Raj^—both its treatment of 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 shapes Indian and Pakistani state behavior to be far similar than generally acknowledged.
1 James D. Fearon and David Laitin, "Ethnicity, Insurgency, and Civil War," American Political Science Review, 97, (1), February 2003: 75–90.
2 Recent examples include Turkey against Kurdish insurgents, Sudan against Darfur insurgents, and Sri Lanka against the Tamil Tigers.
3 Patrick Johnston, "The Treatment of Civilians in Effective Counterinsurgency Operations," PhD Dissertation, Northwestern University, December 2009.
4 David French, The British Way in Counter-Insurgency, 1945–1967, (New York: Oxford University Press, 2011).
5 Staniland observes this of Lyall and Wilson III. See Paul Staniland, "States, Insurgents, and Wartime Political Orders," Perspectives on Politics, 10 (2), June 2012: 243–264; Jason Lyall and Isaiah Wilson III, "Rage Against the Machines: Explaining Outcomes in Counterinsurgency Wars," International Organization 63 (1), 2009: 67–106.
6 James Fearon "Why Do Some Civil Wars Last So Much Longer Than Others,"Journal of Peace Research, 41 (3), May 2004: 275–302; David E. Cunningham, Kristian Skrede Gleditsch, and Idean Salehyan, "It Takes Two: A Dyadic Analysis of Civil War Duration and Outcome," Journal of Conflict Resolution, 53 (4), August 2009: 570–597.
7 Barbara Walter, Reputation and Civil War: Why Separatist Conflicts are so Violent, (New York: Cambridge University Press, 2009).
8 The U.S. Army/Marine Corps Counterinsurgency Field Manual, (Chicago: University of Chicago Press, 2007); David J. Kilcullen, Counterinsurgency, (New York: Oxford University Press, 2010); Daniel Marston and Carter Malkasian, Counterinsurgency in Modern Warfare, (New York: Osprey Publishing, 2008).
9 These dimensions are similar to the way Lobell describes security strategies as both "the extraction of manpower and resources" and "the mobilization of resources" where effort is akin to extraction and force employment is akin to mobilization. See Steven E. Lobell, "War is Politics: Offensive Realism, Domestic Politics, and Security Strategies," Security Studies, 12 (2), Winter 2002/03: 165–95.
10 Donald Horowitz, Ethnic Groups in Conflict, (Berkeley: University of California Press, 1985); Jason Lyall, Graeme Blair, and Kosuke Imai, "Explaining Support for Combatants During Wartime: A Survey Experiment in Afghanistan," American Political Science Review 107 (4), November 2013: 679–705.