0 Exiting Anarchy: Explaining Militia Disarmament in Georgia and Tajikistan Jesse Driscoll Stanford University Political Science October 8, 2007 **This is a draft. Comments are welcome.** 1 Tajikistan and Georgia both fell prey to massive bouts of civil violence in the immediate aftermath of independence from the USSR. In both countries, the government changed hands more than twice in the space of a single year. With no meaningful remnants of the Soviet army to draw upon, both the incumbents and the insurgents fought over rural territory with armies of impromptu militia groups, organized into ad-hoc and shifting coalitions. There were dozens of these groups fighting in both countries in the early 1990s, and while they sometimes organized into large umbrella organizations, giving journalists the impression of unified opposition or principled collaboration, the practical size of a fighting force was 40-100 men. When the “conventional” phases of the civil war were concluded, these heavily armed paramilitary coalitions returned to the capital city and turned their guns on civilians. These militia factions had more social legitimacy, better organization, and heavier weapons than the remnants of the Soviet police, and they had demonstrated that they could replace the president at will.1 Social order collapsed for months, and a terrible epidemic of looting, extortion, rape, and murder was set loose upon Dushanbe and Tbilisi. And then the violence stopped. There is great consensus that the urban situation in both states became much more peaceful when the militias departed, and they departed almost overnight. In Georgia, Mkhedrioni and other militia factions finally cleared out of their roadblocks in early September of 1995, and youth groups no longer walked the streets shooting machine guns into the air. In Tajikistan there were two periods of closure and disarmament – April of 1993 and November 1994. Stability and order did not emerge seamlessly in Dushanbe, as there was territory just outside the city limits where government forces could not safely traverse until the late 1990s. But the departure of the militias removed the feeling of an urban war zone. Commerce became possible midway through 1995, and humanitarian aid organizations began to open their doors. This paper explores the puzzle of the relatively sudden emergence of order and stability in Dushanbe and Tbilisi, the capitals of Tajikistan and Georgia. In both cases, a number of wartime militias managed to merge their core memberships into state organs at the end of the civil war, taking over the Ministry of Defense, Interior, State Security, and other so-called “power ministries.” The paper hypothesizes that the anarchic violence in both capital cities during this period was actually the result of competition between militia groups that were bargaining over these spoils of the civil war victory. Militias were an important way for insecure members of the ruling coalition to guarantee that they would not be pushed out of power. The process of moving from a “winning wartime coalition” to a “minimum winning coalition” therefore created perverse incentives among urban militia groups, leading to the rapid expansion of militia groups and a deluge of violent crime. While many existing studies of militia formation tend to focus on factors relating to the supply of militias in a society, such as how they are organized and financed, in these urban environments it is equally important to focus on the demand for militias, which in this case originated in the political system. When the political system was closed and its members were relatively secure, militias disappeared organically, 1 The first democratically elected president of Georgia, Zviad Gamsakhurdia, was removed in a bloody coup carried out by the Mkhedrioni and the National Guard. The first post-Soviet president of Tajikistan, Rahman Nabiev, was also kidnapped at gunpoint by the “Youth of Dushanbe” and forced to sign his own resignation papers at the airport. 2 despite the extreme weakness of the governments and the absence of an external intervener to carrying the weight of urban policing. The first section of this paper will present a quick primer on both states’ civil wars. The second section will briefly discuss my research methods. The third section expands upon the theoretical framework sketched above, and presents data from the Georgian and Tajik cases suggesting the plausibility of the mechanisms. The fourth section discusses possible alternative hypotheses, and the fifth section concludes. SECTION ONE: BACKGROUND ON THE CIVIL WARS The Armies Within a few weeks of independence from the Soviet Union, both Georgia and Tajikistan were embroiled in civil war.2 Though it is beyond the scope of this paper to present a comprehensive history of the events that culminated in the outbreak of these wars, two general points should be raised at the outset of the narrative. First, the militia groups that emerged during the period of state collapse had far more social legitimacy than the rump state institutions inherited from the Soviet period. This partially explains why Georgia and Tajikistan collapsed fully in the wake of independence, while most of their neighbors were spared violent state breakdown. Second, the opening phases of both civil wars were fought between improvised private armies, usually pieced together by mobilizing pre-existing family, criminal, or social ties. This contrasts with the wars fought in Chechnya, Transdenistria, and Nagorno-Karabakh during the same period, where remnants of the Soviet army were deployed against irregular militias.3 In both of these states, there were active militia organizations forming alongside political parties during the immediate lead-up to independence. The Mkhedrioni in Georgia, for example, had been actively recruiting in Tbilisi since 1989. The various Islamic and ethnic parties in Tajikistan were recruiting as early as 1987, according to 2 Depending on what sorts of coding rules one whishes to employ, it is possible to identify more than two civil wars in both states during this time period. This sort of “counting” problem is always most acute in situations of state failure, when the identities of incumbents and insurgents are rapidly changing. For excellent analyses on the factors leading to the breakout generally, see Georgia, see James Fearon and David Laitin, “Random Narratives: Georgia,” 2006, available online at , (accessed January 2, 2007); Ronald Suny, “Elite Transformation in Late-Soviet and Post-Soviet Transcaucasia, or What Happens When the Ruling Class Can’t Rule” In Timothy J. Colton and Robert C. Tucker, eds., Studies in Post-Soviet Leadership (Boulder, CO: Westview Press, 1995); Derluguian, G. M. 2005. Bourdieu's Secret Admirer In The Caucasus: A World-System Biography, Chicago: University of Chicago Press. ch. 5-7; Wheatley, J. 2005. Georgia: From National Awakening to the Rose Revolution: Delayed Transition in the Former Soviet Union. Hampshire: Ashgage; Muriel Atkin, “Thwarted Democratization in Tajikistan,” In Conflict, Cleavage, and Change in Central Asia and the Caucasus, edited by Karen Dawisha and Bruce Parrott (Cambridge: Cambridge University Press, 1997): 277-311; Shirin Akiner, Tajikistan: Disintegration or Reconciliation? (London: Royal Institute for International Affairs 2001): 25-50; Olivier Roy, The New Central Asia: The Creation of Nations, (London: I.B. Tauris Publishers, 2000). 3 For an overview of these conflicts see Lynch, D. 1999. Peacekeeping Strategis in the CIS: The Cases of Moldova, Georgia, and Tajikistan. London: Royal Institute of Internaitonal Affairs Press; King, C. 2001. The Benefits of Ethnic War: Understanding Eurasia's Unrecognized States. World Politics 53:524-552; Lieven, A. 1998. Chechnya: Tombstone of Russian Power. New Haven: Yale University Press. Gall, C., and Thomas de Wall. 1998. Chechnya: Calamity in the Caucasus. New York: New York University Press. 3 some sources, and the criminal groups that served as the social basis for the Popular Front have ties that extend back to the 1970s.4 The rise of militia politics had unexpected consequences, contributing further to state disintegration. As nationalist militia groups rose to political prominence in Tbilisi, minority ethnic enclaves began to organize militias of their own and to defend themselves (a process that eventually spiraled into full-blown wars of secession). The same basic dynamic played out in Tajikistan, where the rise of Islamic activists threatened established state patronage networks. The threatened elite from the communist era cannibalized local military garrisons military and police forces to build “defense” militias.5 To summarize: As various social forces lost faith in the ability of the post- Soviet government to secure their interests or fairly, different groups found themselves seeking protection more locally, from mafia protectors or kin networks. The remnants of the Soviet police and military infrastructure were generally ineffective at suppressing militias before this cycle of violence escalated out of control. Though anti-state attitudes were prevalent during this period, is possible that security institutions had particular legitimacy problems in Georgia and Tajikistan.6 The massacre of Georgians by Soviet police in April of 1989 was a focal anti-Soviet event in Georgia in the late-Gorbachev period, but for generations Georgians had been notorious for draft evasion and half-hearted soldiering in the Soviet army.7 In Tajikistan, there were strong perceptions that clan networks originating in the northern town of Leninobod (now Khojand) dominated all practical dimensions of politics.8 Although they may have disagreed about who was ultimately pulling the strings, the larger point is that most Tajiks intuitively understood the ministries to be mechanisms for clan or family dominance, not impersonal institutions likely to secure public goods for competing factions. With this as background, it is not surprising that two years after the collapse of the state, these institutions could mount only a feeble response to determined militias. As was mentioned above, the very large paramilitary groups that emerged during this period were never very well organized. Since these groups were operating in the shadow of the law in the late Gorbachev period, membership was rarely formalized. Groups tended to rely on some mix of charismatic authority and established local networks (kin and criminal) for recruitment.9 After the threat of police repression 4 Olimova, S. a. M. O. 2001. "The Islamic Renaissance Party," in Accord: An International Review of Peace Initiatives. Edited by K. A. a. C. Barnes. London: Conciliation Resources.; See also Shirazi, H. A. 1997. Political Forces and Their Structures in Tajikistan. Central Asian Survey 16:611-622. 5 Whitlock, M. 2001. The Land Beyond The River. New York: St. Martin's Press, p. XX. 6 Mark Beissinger, Nationalist Mobilization and the Collapse of the Soviet State (Cambridge: Cambridge University Press, 2002): pages XXX-XXX [police protests]. 7 See Blauvelt, T. "Military Mobilization and National Identity in the Soviet Union", War & Society, May 2003. 8 Kathleen Collins suggests that this is one of the primary factors that explains why Tajkistan fell prey to civil war in the aftermath of independence while the other Central Asian republics managed to basically preserve peace with inherited Soviet institutions. See Kathleen Collins, Clan Politics and Regime Transformation in Central Asia, (Cambridge: Cambridge University Press, 2005), esp. ch. 3-4.; For a similar view, see also Olivier Roy, “Is The Conflict In Tajikistan A Model For Conflicts Throughout Central Asia?” in Tajikistan: The Trials of Independence, edited by Mohammad-Reza Djalini, Frederic Grare and Shirin Akiner (Richmond, Surrey: Curzon): 132-50. 9 See Ro’I, Y. “Nationalism in Central Asia in the Context of Glasnost & Perestroika” in Z. Gitelman (ed), The Politics of Nationalisty & The Erosion of the USSR (1990); ch. 3; Nourzhanov, K. 2005. Saviors of the 4 disappeared, the groups’ size expanded exponentially, and the top leadership had strong incentives to pretend to be in control of a hierarchically organized army. In reality, however, these groups were beset with internal divisions that persevered from initial recruitment patterns. There were no mechanisms in place to screen new members or discipline individuals that misbehaved. Almost none of the recruits had any military experience, but groups still found that they could easily access cheap weapons through collusion with unpaid Soviet army units “beached” in the newly independent republics. There was no shortage of handguns, grenades, machine guns, ammunition -- or even tanks, mortars, and heavy weapons -- for those that could generate hard currency or generate favors with the local Russian military commanders.10 The general proliferation of cheap armaments into society lowered the costs of militia formation, and had a snowballing effect by magnifying local security dilemmas. Local violence entrepreneurs recruited militia members based on the looming threat of annihilation by the enemy.11 The Wars The Georgian Civil Wars of the early 1990s were multi-front clashes between various local Georgian, Abkhaz, Adjarian, and Ossetian militia formations, all using weapons looted from Soviet military garrisons in the late Gorbachev period. In the chaotic aftermath of independence, Russian troops sided with ethnic/religious minorities in South Ossetia and Adjara, leading to the quiet secession of these territories from the Georgian state. After only eight months, the recently elected president of Georgia, Zviad Gamsakhurdia, was removed from power by two urban paramilitary warlords (Dzabba Ioseliani, the head of the Mkhedrioni and Tengiz Kitovane, head of the National Guard). These men invited Eduard Shevernadze to return from Moscow to his native homeland and assume the presidency, thinking to make him their puppet. Supporters of the previous democratically elected Gamsakhurdia government (“Zviadists”) re-grouped in their traditional homeland in Western Georgia and organized for guerilla war. Under the pretext of counterinsurgency, Kitovane’s militias then “invaded” Abkhazia, another de-facto secessionist province. This period of fighting was substantially bloodier than others, and ended with massive population displacements against Georgian civilians by the Abkhaz militias (who were aid by regular and irregular Nation or Robber Barons? Warlord Politics in Tajikistan. Central Asian Survey 24:109-130; Naiyazi, A. 1998. "Tajikistan I: The Regional Dimension of the Conflict," in Conflicting Loyalties and the State in Post-Soviet Russia and Eurasia, vol. 140-170. Edited by B. C. a. A. M. M. Walker. London: Frank Cass. 10 The belief by different militia groups that they could draw Russian military and political aid to their cause, adding men and material to aid in their struggle, was instrumental in the decision to escalate bargaining into war. Laitin, D. 2001. Secessionist Rebellion in the Former Soviet Union. Comparative Political Studies 34: 839-861. [[+interviews ##68 (Dushanbe), 22 (Tbilisi), 71 (Dushanbe), 112 (Kalikhum), 14 (Tbilisi).]] 11 For local security dilemmas, and especially the importance of offense-defense indistinguishability in situations when there are incentives to deploy population displacement strategies, see Barry Posen, “The Security Dilemma and Ethnic Conflict,” Survival 35, no.1 (1993): 27-47. It is important to note that the emotional appeal to “fear” is often a psychologically satisfying substitute for “greed.” It’s much easier to justify taking something you want from someone on the pretense that she was about to steal from you first. I found this sort of reflection somewhat lacking in my interview respondents, but I believe it is an underappreciated aspect to what was motivating conflict participants. For an alternative perspective, see Petersen, R. (2002). Understanding Ethnic Violence: Fear, Hatred, and Resentment in 20th Century Eastern Europe. Cambridge University Press. 5 Russian detachments). This military victory by the Abkhaz rebels led to the complete de- facto secession of Abkhazia, and the defeated Georgian units returned to the capital of Tbilisi along with 200,000 displaced civilians fleeing the conflict zone.12 Between the secessionist provinces and the territory still in the hands of the Zviadists, the government in Tbilisi controlled only a small fraction of the state’s territory at the end of 1994. The civil war in Tajikistan was equally complex. The fighting began over control of the post-Soviet state apparatus rather than by various attempts to secede, so it is difficult to sub-divide the civil war violence into smaller conflicts. From the beginning, the paramilitaries organized into two large coalitions. “The Old Guard” was primarily made up of regional-clan organizations that occupied privileged government and black- market positions in the latter period of the USSR, mostly based in the regions of Khojand and Kulob. “The Opposition” was made up of a mix of democratic, Islamic, and Pamiri intellectuals, and rural clans from the disenfranchised agricultural regions of Gharm and Khatlon. The war began in full when militias loyal to the opposition removed the first post-Soviet president (Rahman Nabiev), and implemented a power-sharing agreement that was backed by the remnants of the Soviet Army in Tajikistan. During this period, Akbarshah Iskanderov served as a figurehead for the gridlocked power-sharing government. The warring factions inside the capital jockeyed for influence, while militias from regions controlled by “The Old Guard” (Khojand and Kulob) left the capital to regroup. These rebellious militia groups renamed themselves as the Popular Front for Tajikistan (PFT), refused to recognize the legitimacy of the new government, returned to their regional homelands, and organized for war. Safarali Kenjayev and his militia seized the capital in October, and he briefly declared himself president; Russian forces blocked reinforcements from PFT forces in the south, and Kenjayev was forced to retreat. In the following weeks, Russians and Uzbeks withdrew critical support for the coalition government of Iskanderov, forcing them to surrender the presidency to the Popular Front. Kulobi militias under the general command of Sangak Safarov captured Dushanbe in December of 1993 and installed the presidency of Emomali Rakhmonov. To cement their victory, they engaged in massive population displacement in the densely populated lowlands of Khatlon, where citizens were thought to be sympathetic to the opposition. UNHRC estimates that at least 100,000 civilians forded the icy Amu Daria River in midwinter, fleeing to the relative safety of Northern Afghanistan. Fighting was sporadic and disorganized, with criminal hooliganism and mob violence far more common than pitched battles.13 Interview data confirm the impression that groups survived this period through a mix of looting, racketeering, and charitable donations from friends and family until they became the de-facto state army. The Aftermath It was not until the conclusion of the “conventional” stage of the civil war (e.g., when Kulobi militias took Dushanbe in December of 1992 or when the Abkhaz forces routed the Georgians and ethnically cleansed Sukhumi midway through 1993) that intellectuals and journalists had any experience with the war at all. That experience came in the form of the influx of refugees and militia fighters to the capital city. These men 12 See Nadareishvili, T. 1997. Genocide in Abkhazia. Tbilisi: Samshoblo. 13 Mueller, J. (2000). "The Banality of 'Ethnic War'." International Security 25(1): 42-70; Charles Fairbanks, “The Postcommunist Wars,” Journal of Democracy 6 (October 1995): 18-34. 6 claimed to be patriots and war heroes, and having “bled for their nation” they now demanded recognition and special privileges as patriots. These men could easily justify their continued mobilization by noting that the lull in fighting was very fragile, and that the next phase of the civil war could erupt at any time. In the interim, these men set up armed checkpoints inside the capital city, and began raiding neighborhoods of minority ethnic groups who were seen as residual “collaborators” in the civil war (Abkhaz, Ossetians, and Muslims in Georgia; Pamiris, Garmis, and Karateginis in Tajikistan). Various militia captains and high-profile wartime field commanders were immediately incorporated into the new state as heads of ministries or deputy heads of ministries. Incorporating the militias in this manner was obviously a case of dividing the spoils of victory to the men with guns, but it was also the only obvious way for the state to have any coercive power at all. Recruiting, training, and arming an entire new police force was a task that would take years. In the interim, as a Georgian intellectual suggested, “better the wolf in your living room, where you can hope to tame him, then hungry in the yard outside.”14 So in the initial aftermath of the war, many militia members found themselves as agents of the state, often in newly created ministries headed by their former field commanders. Militia members were, for the most part, initially placed in the same ministry as their regimental lieutenant or field commander. The explicit hope would be that this would quietly push the hooligans off the streets. [Insert Figure 1 About Here] In defiance of expectations, however, the number of urban militias in both Georgia and Tajikistan was paradoxically expanding during this time. The risk of dying in the civil war had deterred many youth from joining; now that the worst phases of the violence had passed, many young men came to realize that joining a militia offered excitement, social recognition, opportunities for easy money, and now potentially government jobs, as well. As the rank-and-file absorbed new members, even the most disciplined wartime cadres began to take on criminal characteristics. Worst of all, since the state had no salaries to pay any of its employees, these “incorporated” militias were still supporting themselves through extortion, roadblocks, exploiting economic bottlenecks, and simple theft and looting. But now they were in uniforms. It was an awful time. It did not take long for public order to collapse completely. Well-armed units from the war could credibly threaten the president with a coup if he did not implement policies to their liking, which removed all pretense of a rule of law. Multiple uniformed agents, representing different branches of different ministries, could steal with absolute impunity. Violent crime continued to escalate during this period. Rape, racketeering, and murder became commonplace. In response to this seemingly uncontrollable urban violence, many of the citizens closed their businesses, stayed indoors, or fled the capital all together.15 14 Author interview with Alexander Rondeli, 10/28/07 15 For Georgia, see Gia Nodia, 2002. "Putting The State Back Together in Post-Soviet Georgia," in Beyond State Crisis?: Post-Colonial Africa and Post-Soviet Eurasia in Comparative Perspective. Edited by M. B. a. C. Young, pp. 413-439. Washington DC: Woodrow Wilson Center Press. For Tajikistan, Atkin, M. 1995. "Islam As Faith, Politics, and Bogeyman," in The Politics of Religion in Russia and the New States of Eurasia. Edited by M. Bourdeaux, pp. 247-272. London: Sharpe. Akiner, S. 2001. Tajikistan: Disintegration of Reconciliation. London: Royal Institute of International Affairs. 7 This period of intense violence and predation did not last forever. There is a great deal of consensus that the streets of Tbilisi became relatively safe again in September of 1995, in the immediate aftermath of the assassination attempt against the president of Georgia, Eduard Shevardnadze. This was not the end of racketeering or corruption in the police forces, but it was the end of heavily armed bands of men wandering the streets of Georgia. In Tajikistan, the worst period of urban violence (including the pogrom of the Pamiris and Karateginis) concluded in April of 1993, and there was a second-tier drop-off in urban violence in late 1994. Although Dushanbe continued to have an unusually high violent crime rate for the Former Soviet Space throughout the 1990s, it was nothing out of the ordinary for a large urban South American city.16 NGOs and Christian Missionaries were able to safely open their doors by 1995. Two puzzles emerge from these empirical patterns. First, what explains the upsurge in violence inside the capital after the end of the civil war, and second, why did that violence end so suddenly? Most descriptions of this period are quick to dismiss this period as essentially anarchic, which implies that it would be impossible to make predictions about the timing of the intensity of the violence. In order to understand how the armed groups were removed from the streets, it will be necessary to look more closely at the dynamics that sustained the “anarchic” violence in the aftermath of the civil war. SECTION TWO: RESEARCH METHODS Data presented in this paper is based on fifteen months of field research were conducted in Georgia and Tajikistan between January 2006 and August 2007. I conducted a total of 119 formal semi-structured interviews with civil war participants, as well as another 43 interviews with individuals that I will term “opportunistic joiners.”17 Many of the respondents are current or recent members of the police or the army; those who are not tend to be marginally employed (e.g., as farm laborers, construction workers, taxi drivers, migrant/seasonal laborers, store owners, nightclub bouncers, or working generally in the informal/criminal). Respondents do not represent a random sample of the former combatants from either country, but great lengths were taken to find a sample of diverse respondents, believed to be representative. The historical stories generated from interview data were triangulated with local academics, journalist and researchers to ensure accuracy. Due to the sensitive nature of the data, every effort was made to build rapport with interview informants, including extended ethnographic research in a remote village on the Tajiki-Afghan border.18 Interviews took place in the capital cities of Tblisi and Dushanbe, as well as more rural towns and villages. Most interviews were conducted in Russian and the topics covered included pre-war occupations, history of recruitment, activities during and immediately following the civil war, and a variety of topic related to the life trajectories 16 Lafree, G. 2006, “Democracy and Crime: A Multilevel Analysis of Homicide Trends in Forty-Four Countries, 1950-2000,” The Annals of the American Academy of Political and Social Science, Vol. 605, No. 1, 25-49 (2006) 17 In this study, I define “opportunistic joiners” as a men that affiliated themselves with a militia 1) exclusively in the capital city or another major urban area, and 2) after the end of the end of the worst phases of the civil war violence. 18 In order to gather perspectives from the consolidation period’s “losers,” four months of extended ethnographic field research were conducted in the village of Kalikhum in the Darvaz district of Tajikistan. 8 of their co-combatants. Key informants were interviewed multiple times, and often provided a range of introductions.19 Primary interviews usually took place in government offices or public restaurants, but as rapport developed, subsequent interviews often took place in informant’s homes. The following theoretical sketch is drawn almost entirely from conversations with these men. Occasionally I found it useful to triangulate their “street level” perspectives with a more “top down” perspective provided by a local academic or political figure. But in general, it quickly became obvious that the former combatants and postwar militia members watched the week-to-week changes in the emerging power structure with much greater attention than the elites. They realized their role as pawns in a dangerous game of power consolidation, and they had a great deal of incentive to understand the rules of the game. Lucky for me, they were willing to speak about that period with great candor. The fact that these men were often quite capable of speaking in great detail about the micro- politics of consolidation 11-13 years after the fact is itself data. As will become clear, they were trying to position themselves to eventually secure jobs in the next government, and often took great personal risks in the mean time. SECTION THREE: THEORY AND CASE STUDIES A satisfactory theory of militia disarmament should be able to contribute to our understanding of four aspects of the postwar story above. First, the theory must be able to explain the rapid expansion in the number of militia members on the street after the conclusion of the civil war. Second, these militia members began to engage in violent actions against each other and against civilians. Taken together, these are the two facets of the “anarchy” described by most observers in Dushanbe and Tbilisi at the time. Third, the theory must be able to explain a decline in militia membership and a reduction in violent acts over time. Fourth and finally, the theory should ideally be able to comment on the speed of the final transition away from militia politics and towards recognizable state-led order. Phase One: The Winning Coalition Expands The expansion of militias in the immediate aftermath of the war was the easiest aspect of this story for my informants to analyze and comment upon. Two factors seem to have been working simultaneously to explain the proliferation and expansion of militias in Dushanbe and Tbilisi. The first was simply that without any functioning state bodies to arrest them, and without any real risk of dying in war to deter them, there was a large influx of opportunistic young men who wanted to take part in the looting, and grab their share of the spoils. While family or clan connections defined the core membership, there was always tertiary milieu of young toughs that could mimic the groups’ symbols and claim to be agents of one field commander or another, and in the chaotic environment, it was relatively easy for newcomers to put on a black leather jacket and “pass’ as a member of almost any group. In a crude economic sense, one might consider this factor as lowering the “supply costs” for a militia seeking to start-up or expand its membership. 19 To protect the identities of respondents, only first names were recorded in these interviews. Often the respondents asked me to record a false name, or no name at all. 9 An underappreciated corollary to this first factor, however, is that there was also a sharply increasing demand for militias. The field commanders, militia captains, paramilitary lieutenants, and other mid-level to high-level members of the wartime coalition that controlled the capital were generating this demand. These men had organized for war on the implicit assumption that the spoils of victory would be eventually split among them, and now that the spoils were about to be divided, everyone understood that “power sharing” between the militias would create winners and losers. Having lived through the excesses of state corruption in the late Soviet period, militia captains intuitively understood that the stakes were high: Winners would get life-long positions in government ministries or exclusive control over gray-market bottlenecks, and those deemed expendable would have to retreat from the capital, land in prison, or worse. For the foot soldiers in these militia wars, the alternative to success was equally grim. Most of these men were marginally employed in the late Soviet period, with no skills and little education. Many were alcoholics, criminals, common hooligans, or serially unemployed; virtually all were men who had worked with their hands their entire lives, and expected their children to do the same. They had come of age in a system where familial and kin networks were necessary for everyday life, and had an intuitive understanding that getting one of the new police jobs would be their best, only chance to increase their life opportunities. The late stages of militia consolidation represented their only chance to convert their violent social capital into a better future. It was also clear that the spoils of the winner’s coalition would be divided up in rough proportion to the size of the militia that was under their command. The most lucrative and secure positions would go to the individuals that were strong enough to threaten the rest of the coalition members – and even credibly threaten to remove the president – if they were pushed aside.20 In practice, this meant that men who had initially helped to organize the paramilitaries in the late Soviet period, and then risked their lives in the civil war, were suddenly at risk of being excluded from the spoils of victory unless they kept a large number of men mobilized, ready to make trouble. Militias in the post- Soviet consolidation period emerged as a form of political insurance – their purpose was not to conquer territory or even to fight other militias, but to remind observers that there would be serious consequences if the warlord were cut out of the distribution of pork. In order to keep their privileged role in the state, then, influential individuals scurried to establish networks of loyal and dependent men, whose career trajectories were directly tied the warlord’s patronage. Whether one imagines that these moves were driven by greed or fear, they had the indirect the effect of splintering the umbrella militia coalitions (e.g., the Popular Front and the Mkhedrioni) from within. Now more than ever before, these coalitions were more historical markers than anything else, signaling a somewhat redundant political affiliation of the prominent leader. The new rank-and-file that emerged in this period were either men with strong familial or friendship ties to a warlord, or men that could be manipulated and discarded at low cost. My interviews with the “opportunistic joiners” revealed that it was common for extended family members of wartime fighters that had stayed with their families during the war to come to the city “in 20 This was a point on which there was surprising consensus among my respondents – that the militias primary purpose was as bargaining leverage and signaling power to secure patronage inside the state, rather than to fight each other or simply raise money. Many men referenced the removal of the previous presidents as proof that large numbers of guns and supporters was the only way to secure ones’ future.