When Good Fences Make Bad Neighbors When Good Fences Boaz Atzili Make Bad Neighbors Fixed Borders, State Weakness, and International Conºict

Since the end of World War II, the norm of ªxed borders—the proscription against foreign conquest and annexation of homeland territory—has gained prevalence in world politics. But have ªxed borders made international conºict less frequent? Ob- servers might assume they have, given that territorial issues have historically been a major cause of war.1 However, among sociopolitically weak states (i.e., states that lack legitimate and effective governmental institutions), ªxed borders can actually increase instability and conºict. Good fences can make bad neighbors. Until the late 1980s, the scholarly literature had devoted little attention to theories regarding the role of territory and borders in international relations.2 Since then, however, a growing body of work on this subject has emerged.3 One promising line of inquiry has focused on international norms concerning changes in borders. Mark Zacher and Tanisha Fazal, for example, have found that post–World War II cases of foreign conquest and annexation are a rarity and that the norm of ªxed borders has grown stronger over the years.4 The ef-

Boaz Atzili is a Research Fellow in the Belfer Center for Science and International Affairs at the John F. Ken- nedy School of Government at Harvard University.

The author thanks Ericka Albaugh, Michal Ben-Josef Hirsch, Patrick Boyd, Naomi Chazan, Thomas Christensen, Tanisha Fazal, Rachel Gisselquist, Arie Kacowicz, Sarah Lischer, Galia Press- Bar Nathan, Kenneth Oye, Roger Petersen, Jessica Piombo, Jeremy Pressman, Stephen Van Evera, and two anonymous reviewers for helpful comments on earlier drafts of this article.

1. John A. Vasquez, The War Puzzle (New York: Cambridge University Press, 1993), pp. 123–152; and Kalevi J. Holsti, Peace and War: Armed Conºicts and International Order, 1648–1989 (New York: Cambridge University Press, 1991), pp. 306–311. 2. Noting this is John G. Ruggie, “Territoriality and Beyond: Problematizing Modernity in Interna- tional Relations,” International Organization, Vol. 47, No. 1 (Winter 1993), pp. 139–174. 3. See, for example, Friedrich Kratochwil, “Of Systems, Boundaries, and Territoriality: An Inquiry into the Formation of the State System,” World Politics, Vol. 39, No. 1 (October 1986), pp. 27–52; Brendan O’Leary, Ian S. Lustick, and Thomas Callaghy, eds., Right-sizing the State: The Politics of Moving Borders (New York: Oxford University Press, 2002); and Monica Duffy Toft, The Geography of Ethnic Violence: Identity, Interests, and the Indivisibility of Territory (Princeton, N.J.: Princeton Uni- versity Press, 2003). 4. Mark W. Zacher, “The Territorial Integrity Norm: International Boundaries and the Use of Force,” International Organization, Vol. 55, No. 2 (Spring 2001), pp. 215–250; and Tanisha M. Fazal, “The Origins and Implications of the Territorial Sovereignty Norm,” paper presented at the annual convention of the American Political Science Association, San Francisco, California, August 29– September 2, 2001. Rather than referring to this norm as “territorial integrity” or “territorial sovereignty,” I use “ªxed borders” to distinguish between territorial and nonterritorial intervention.

139 International Security 31:3 140

fects of this norm on interstate relations, however, have yet to be analyzed—an omission this article aims to address. The article posits that, in many regions of the world, adherence to the norm of ªxed borders has led to international conºicts and growing instability by perpetuating and exacerbating state weakness. Three factors account for these negative effects. First, an international system of states with ªxed borders deprives states of what were historically their greatest incentives to develop strong institutions: external threats to their territorial integrity and opportunities for territorial expansion. Second, without such territorial threats, a coher- ent in-group identity and loyalty to the state are difªcult to establish. Third, without a mechanism through which weak states can be overtaken by stronger ones, the former may persist and perhaps become even weaker. Sociopolitically weak states in a world of ªxed borders may be more prone to internal conºict or even civil war because the incentives for excluding whole groups of citizens are greater, and because there is a higher likelihood of the emergence of an internal security dilemma. Such conºict may spill over as neighboring states feel obliged to come to the assistance of threatened coethnics within the weak state, and as refugee ºows create breeding grounds for insurgency and counterinsurgency, which in turn can lead to foreign intervention or even full-ºedged war. Neighbors might also view the state’s weakness as providing an opportunity to seek economic gains or political inºuence, including regime change.5 The argument that a norm that seeks to make the world a more peaceful place may instead cause it to become more conºict prone is both counter- intuitive and theoretically new. In addition, given that the phenomenon of weak and failed states is widespread, the argument potentially has important empirical implications. In 2006 the Failed States Index listed twenty-eight countries as being in a state of “alert” and seventy-eight more as in a state of “warning” with regard to their prospects for becoming failed states. These include countries in Africa, the Middle East, Asia, the former Soviet Union, Latin America, and the Balkans.6 This article employs a single case study—the war in Congo, a country that was known as Zaire from 1971 to 1997 and since then as the Democratic

5. Although outside the scope of the article, it is useful to note that states in an international system of ªxed borders create more favorable conditions for global terrorist organizations to emerge (e.g., in Lebanon and Somalia). See Robert I. Rotberg, ed., “Failed States, Collapsed States, Weak States: Causes and Indicators,” in Rotberg, ed., State Failure and State Weakness in a Time of Terror (Cambridge, Mass.: World Peace Foundation, 2003), pp. 1–25. 6. Fund for Peace, “Failed States Index,” http://www.fundforpeace.org/programs/fsi/fsindex .php. See also Robert I. Rotberg, “The New Nature of Nation-State Failure,” Washington Quarterly, Vol. 25, No. 3 (Summer 2002), p. 85. When Good Fences Make Bad Neighbors 141

Republic of Congo (DRC)—as a plausibility probe to determine the theory’s validity and applicability.7 The conºict produced tremendous carnage: as many as 3.8 million dead and many more injured or displaced. Both phases of the war (1996–97 and 1998–2002) involved domestic militias, a massive foreign invasion, and shifting alliances—with Angola, Rwanda, Uganda, and Zimbabwe playing major roles. Even though the war has ofªcially ended, peace remains elusive. I selected the Congo war as my case study for two reasons. First, Congo is an extreme (or most likely) case, because as a country in Africa—a continent where the norm of ªxed borders is strongest and where the states are weakest—it presents high values on both the independent variable (the norm of ªxed borders) and the intervening variable (a very weak state). A probe into the dynamics relating ªxed borders to state weakness and to international conºict should thus offer clear, discernable results.8 Second, the Congo conºict is multifaceted; that is, many observation points can be generated from what is ostensibly a single case. Because this study is an initial probe into a newly identiªed causal mechanism, however, and because it involves only one case, the ability to generalize from its ªndings should not be exaggerated. The ªrst section of the article deªnes the norm of ªxed borders and dis- cusses its development and growing inºuence. The second section explicates the effects of this norm on the strength or weakness of states. The third section examines the effects of the combination of state weakness and the norm of ªxed borders on relations between neighboring countries. The fourth section considers the case of the war in Congo as a preliminary probe of the plausibility of the article’s theoretical model. The article concludes with suggestions for future research.

The Norm of Fixed Borders

State borders are social constructs. Different international systems have historically maintained different types of international borders. This section deªnes and analyzes the inºuence of one such construct: the international norm of

7. I use “Congo” when referring to the country in general, and Zaire and Democratic Republic of Congo (DRC) when it was ofªcially termed so. Plausibility probes are “preliminary studies on relatively untested theories and hypotheses to determine whether more intensive and laborious test- ing is warranted.” See Alexander L. George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences (Cambridge, Mass.: MIT Press, 2005), pp. 75–76. 8. On the justiªcation and logic of using an extreme case, see ibid., pp. 120–123. For an example of a study that uses a similar approach, see Stephen Van Evera, “The Cult of the Offensive and the Origins of the First World War,” International Security, Vol. 9, No. 1 (Summer 1984), pp. 58–107. International Security 31:3 142

ªxed borders,9 or the notion that state borders should not be changed and that annexation of a neighbor’s territory should be viewed as illegitimate.10 Historically, conquest and annexation were common. (See Figure 1.) When Frederick the Great of Prussia invaded the Austrian province of Silesia in 1740, for example, he knew that as long as he maintained control over it, other states would not challenge his rule. Today, however, as Charles Tilly writes, “With a few signiªcant exceptions, military conquest across borders has ended, states have ceased ªghting each other over disputed territory.”11 Indeed, there have been only ten cases of foreign military conquest of homeland territory since 1950.12 In a similar vein, Donald Horowitz has shown that cases of irredentism have become rare, although the potential for irredentist claims based on eth- nicity have greatly increased.13 The norm of ªxed borders is the product of several material and ideational factors. Material factors include a decline in the importance of land as a means of production and the rise of technology in the developing world, which together have greatly decreased the incentives for territorial conquest;14 the potential costs of conquest, which have risen exponentially with the development of nuclear weapons;15 and the inability of weak central governments to project power over vast, scarcely populated tracts of land.16 Without the ideational “glue,” however, these material factors would have

9. For most actors, a norm would mean “collective expectations for the proper behavior of actors with a given identity.” See Peter J. Katzenstein, “Introduction,” in Katzenstein, ed., The Culture of National Security: Norms and Identity in World Politics (New York: Columbia University Press, 1996), p. 5. Some actors, however, adhere to a norm simply because it is common practice, or out of fear of retribution should they breach it. See Robert Axelrod, “An Evolutionary Approach to Norms,” American Political Science Review, Vol. 80, No. 4 (December 1986), pp. 1095–1111. 10. For similar arguments, see Zacher, “The Territorial Integrity Norm”; and Fazal, “The Origins and Implications of the Territorial Sovereignty Norm.” 11. Charles Tilly, Coercion, Capital, and European States, 990–1992 (Cambridge, Mass.: Black- well, 1992), p. 203. 12. I use the term “homeland territory” to control for cases of decolonization and the breakup of empires, which involve territorial changes but are not relevant to the norm of ªxed borders. The ten cases are Israel from Syria, 1967; Israel from Jordan, 1967; Israel from Egypt, 1967; Iran from the United Arab Emirates, 1971; India from Pakistan, 1971; Libya from Chad, 1973; Israel from Syria, 1973; Turkey from Cyprus, 1973; China from South Vietnam, 1974; and Armenia from Azerbaijan, 1991–94 (a border case between foreign conquest and secession). I do not consider the U.S. occupation in Iraq as one of these cases, because it is not intended to be permanent occupation or annexation. Should the occupation continue much longer, however, it might be considered as such. 13. Donald L. Horowitz, Ethnic Groups in Conºict (Berkeley: University of California Press, 1985), pp. 281–288. 14. Carl Kaysen, “Is War Obsolete? A Review Essay,” International Security, Vol. 14, No. 4 (Spring 1990), pp. 42–64. 15. Moreover, John E. Mueller argues that even without nuclear weapons, war has become un- thinkable because of its devastating costs for advanced industrial societies. See Mueller, Retreat from Doomsday: The Obsolescence of Major War (Rochester, N.Y.: University of Rochester Press, 1996). 16. Jeffrey Herbst, States and Power in Africa: Comparative Lessons in Authority and Control (Prince- ton, N.J.: Princeton University Press, 2000). When Good Fences Make Bad Neighbors 143

Figure 1. Conquest and Annexation of Homeland Territory per Contiguous Dyad, 1820–2000

SOURCES: For data on the number of conquests and annexations, see Jaroslav Tir, Philip Schafer, Paul F. Diehl, and Gary Goertz, “Territorial Changes, 1816–1996,” Conflict Man- agement and Peace Science, Vol. 16, No. 1 (Spring 1998), pp. 89–97 (adjusted to include 1997–2000 data and to fit author’s definitions). For data on contiguous dyads, see Douglas M. Stinnett, Jaroslav Tir, Philip Schafer, Paul F. Diehl, and Charles Gochman, “The Corre- lates of War Project Direct Contiguity Data, Version 3,” Conflict Management and Peace Science, Vol. 19, No. 2 (Summer 2002), pp. 58–66.

been unlikely to produce the now internationally accepted norm of ªxed borders. The idea that conquest does not entitle one state to annex the territory of another has its roots in the eighteenth century, with the notion of popular sovereign rights. But only at the conclusion World War I, with U.S. President Woodrow Wilson’s promotion of the concept of self-determination, did this idea begin to show tangible results.17 The 1919 Covenant of the League of Nations and the Kellogg-Briand Pact of 1928 both sought to lend strength to the norm of ªxed borders, as did changes in the laws of war, which in-

17. Wilson’s words in 1919 to the Senate Committee on Foreign Relations expressed precisely the idea of ªxed borders: “I understand that article [article 10 of the League covenant] to mean that no nation is at liberty to invade the territorial integrity of another. Its territorial integrity is not destroyed by armed intervention; it is destroyed by the retention of territory, by taking territory away from it.” “A Conversation with Members of the Foreign Relations Committee,” August 19, 1919, in Arthur S. Link, The Papers of Woodrow Wilson, quoted in Fazal, “The Origins and Implica- tions of the Territorial Sovereignty Norm,” p. 25. International Security 31:3 144

creasingly viewed belligerent occupation to be acceptable only if it was temporary.18 The 1945 United Nations charter and many of its subsequent resolutions gave the norm powerful legal backing. As General Assembly resolution 2625 of 1970 declares, “No territorial acquisition resulting from the threat of use of force shall be recognized as legal.”19 The charter does not prohibit war or the use of force per se: under certain circumstances, notably for purposes of self-defense and collective security, war is permissible. It does, though, prohibit territorial annexation, even as a result of a just war (i.e., a war of self-defense).20 During the Cold War, U.S. hegemony over the Western world played an important role in the institutionalization of the norm of ªxed borders, as well as in its promotion and enforcement.21 The Soviet Union played a similar role in its sphere of inºuence. Despite frequently intervening in other states’ affairs and sending troops overseas, both superpowers refrained from territorial ag- grandizement and did not annex any territory after 1945. In essence, both found it beneªcial to safeguard the international territorial status quo.22 In some cases, such as the 1956 Suez crisis, they even intervened to prevent other states from violating the norm of ªxed borders. The norm, however, is not merely a Cold War artifact.23 Since the collapse of the Soviet Union, no state has acquired territory through force (Armenia’s role in Nagorno-Karabakh is ambiguous). The response of the international community to violations of the norm, such as the nonrecognition of Israel’s 1967 conquests, and the military rollback of Iraq following its invasion of Kuwait underscore the strength of the norm. Indeed, as U.S. President Bill Clinton noted, “[The current] era does not reward people who struggle in vain to redraw borders with blood.”24

18. On the development of these ideas, see Sharon Korman, The Right of Conquest: The Acquisition of Territory by Force in International Law and Practice (Oxford: Clarendon, 1996), pp. 133–199; and Zacher, “The Territorial Integrity Norm,” pp. 216–221. 19. United Nations General Assembly, “Declaration on Principles of International Law Concern- ing Friendly Relations and Cooperation among States in Accordance with the Charter of the United Nations,” Resolution 2625, October 24, 1970. In addition, some regional organizations em- phasize the sanctity of borders in their conventions and resolutions. For the institutionalization of the norm in the United Nations and regional organizations, see Korman, The Right of Conquest, pp. 199–214; and Zacher, “The Territorial Integrity Norm,” pp. 221–223. 20. For this interpretation, see Korman, The Right of Conquest, pp. 209–214. 21. Fazal, “The Origins and Implications of the Territorial Sovereignty Norm.” 22. John Lewis Gaddis, “The Long Peace: Elements of Stability in the Postwar International Sys- tem,” International Security, Vol. 10, No. 4 (Spring 1986), pp. 99–142. 23. Michael C. Desch, “War and Strong States, Peace and Weak States?” International Organization, Vol. 50, No. 2 (Spring 1996), pp. 237–268, could be interpreted this way. 24. William Jefferson Clinton, “Remarks by the President in Greeting to the Peoples of Pakistan,” Islamabad, Pakistan, March 25, 2000. When Good Fences Make Bad Neighbors 145

Africa’s borders are particularly intriguing. Despite the arbitrariness with which many state borders in Africa were drawn, they have remained largely ªxed.25 From its inception in 1963, the Organization of African Unity (OAU) has endorsed the norm in accordance with the principle of preserving the colonial territorial status quo.26 In practice, as Jeffery Herbst notes, “the vast major- ity of [African borders] have remained virtually untouched since the late 1800s, when they were ªrst demarcated.” The OAU’s determination to uphold the norm was demonstrated, for instance, in the 1967–70 civil war in Nigeria, when the organization sought to prevent Biafra’s attempts to secede.27 Common sense might suggest that the norm of ªxed borders should increase the stability and security of states. In the next two sections, however, I discuss how it can have the opposite effect, particularly in regions that com- prise mainly weak states. The norm is likely to perpetuate state weakness, and weak states with ªxed borders are often a source of international conºict.

Fixed Borders and State Weakness

Below are four hypotheses regarding the role of ªxed borders in promoting conºict between neighboring states. Figure 2 offers a graphic illustration of this process.

Hypothesis 1: Fixed borders can perpetuate or exacerbate the weakness of already weak states.

Hypothesis 2: Weak states in a ªxed-borders world can create conditions that can give rise to violent internal conºicts.

Hypothesis 3: Refugee movements, insurgencies, and kin connections across international borders can cause civil conºicts in weak states in a

25. Saadia Touval, The Boundary Politics of Independent Africa (Cambridge, Mass.: Harvard Univer- sity Press, 1972). 26. Michel-Cyr Djiena Wembou, “The OAU and International Law,” in Yassin El-Ayouty, ed., The Organization of African Unity after Thirty Years (Westport, Conn.: Praeger, 1994), p. 16. 27. Jeffery Herbst, “The Creation and Maintenance of National Boundaries in Africa,” International Organization, Vol. 43, No. 4 (Autumn 1989), pp. 683–687, at p. 673. See also Herbst, States and Power in Africa; Jeremy M. Weinstein, “Africa’s ‘Scramble for Africa’: Lessons of a Continental War,” World Policy, Vol. 17, No. 2 (Summer 2000), pp. 11–20; and Robert H. Jackson, Quasi-states: Sover- eignty, International Relations, and the Third World (New York: Cambridge University Press, 1990), pp. 87–90. The only case that possibly contradicts the norm of ªxed borders is Morocco’s annexation of the Western Sahara in 1976. Note that the African norm against changing borders goes beyond even that of the wider international community by prohibiting secessions, as well as conquests. I consider this only a variant of the international norm of ªxed borders, not a separate phenomenon. From the point of view of the individual state, moreover, this norm is still “international,” regardless of whether it originates in the region or beyond it. International Security 31:3 146

Figure 2. Potential Effects of Fixed Borders on Weak States

ªxed-borders world to spill over their borders and become international conºicts and possibly full-ºedged wars.

Hypothesis 4: State weakness promotes the possibility of international conºict by creating opportunities for neighbors to intervene to exploit the weak states economically or politically.

In regions where states lack sociopolitical strength, the norm of ªxed borders may perpetuate or exacerbate state weakness because it deprives the state of a key historical factor in state building: an external threat to its borders and, in some instances, to its very survival. I deªne “sociopolitical strength” as the state’s capacity to maintain a monopoly over the legitimate use of force, its ability to govern the population (including extracting revenues through taxes and providing public goods), and its ability to maintain a reasonable level of social cohesiveness and identiªcation of its residents with the state.28 Fixed borders do not make states weak, but they can make those that are already weak even weaker by (1) denying them the incentives and coercive capabilities that have traditionally accompanied the threat of territorial war and

28. For an example of a similar use of state strength, see Kalevi J. Holsti, The State, War, and the State of War (Cambridge: Cambridge University Press, 1996), pp. 82–98. See also Barry Buzan, Peo- ple, States, and Fear: An Agenda for International Security Studies in the Post–Cold War Era, 2d ed. (Boulder, Colo.: Lynne Rienner, 1991), pp. 57–111. When Good Fences Make Bad Neighbors 147

opportunities for territorial expansion; (2) supplying counterincentives to state building; and (3) denying them the mechanism that weeded out weak states and bolstered stronger ones. territorial war and state building Like states in early modern Europe, many countries in the developing world are confronting the daunting challenge of building strong, viable states. The former, however, faced different external pressures and incentives than do states today.29 It is these differences that help to explain the perpetuation of state weakness in the contemporary international system. Contrary to the con- ventional wisdom, many European states in the past were confronted with conditions similar to those faced by weak states today, including a lack of effective institutions and a socially cohesive citizenry. The relative ethnic and linguistic homogeneity of contemporary European states is a product, rather than a precondition, of the state-building process. In 1640, for instance, Brandenburg-Prussia was a disjointed collection of territories with deep reli- gious and regional cleavages; it could claim no collective identity; and it was loosely controlled by a weak center. A century later, its successor—the king- dom of Prussia—had become a stronger, socially cohesive state.30 International war, and territorial war in particular, forced this transforma- tion for two reasons—both of which involved territory. First, the development of large, centrally supplied and ªnanced standing armies profoundly increased the need for states to extract resources from their societies, and at the same time greatly enhanced their ability to coerce their populations into providing these resources. This need, in turn, required the creation of new, more efªcient bureaucracies to manage the collection of taxes and the distribution of resources. Where these efforts succeeded, they strengthened the state and, in the long run, served much more than the narrow military function for which they were created.31

29. Mohammed Ayoob, The Third World Security Predicament: State Making, Regional Conºict, and the International System (Boulder, Colo.: Lynne Rienner, 1995), pp. 21–46, 193–194. 30. See, for example, Samuel E. Finer, “State- and Nation-Building in Europe: The Role of the Mili- tary,” in Charles Tilly, ed., The Formation of National States in Western Europe (Princeton, N.J.: Prince- ton University Press, 1975), pp. 134–144. France is another example. See Eugen Weber, Peasants into Frenchmen: The Modernization of Rural France, 1870–1914 (Stanford, Calif.: Stanford University Press, 1979). 31. Otto Hintze, The Historical Essays of Otto Hintze, Felix Gilbert, ed. (New York: Oxford Univer- sity Press, 1975); Charles Tilly, “Reºections on the History of European State-Making,” in Tilly, The Formation of National States in Western Europe, pp. 3–83; Tilly, Coercion, Capital, and European States, 990–1992; Brian M. Downing, The Military Revolution and Political Change: Origins of Democracy and Autocracy in Early Modern Europe (Princeton, N.J.: Princeton University Press, 1992); and Thomas Ertman, Birth of the Leviathan: Building States and Regimes in Medieval and Early Modern Eu- rope (New York: Cambridge University Press, 1997). International Security 31:3 148

Second, a common external threat forged a common internal identity. Be- longing to a particular group (the in-group) can help individuals fulªll some of their psychological needs (e.g., the need for self-esteem and affection). Yet membership in one group may mean nonmembership in another (the out- group). Hostility toward the out-group often increases in-group cohesiveness.32 The more threats that are posed to both its existence and its boundaries, the more likely the group is to develop internal cohesion.33 Opportunities for territorial expansion through warfare can also strengthen society’s bonds and enhance its self-perception.34 Wars in early modern Europe were effective in promoting state building precisely because they threatened the seizure of territory. Wars that do not seek territorial gains do not endanger the survival of the state, nor do they endanger the population as a whole. Portraying an attack as an assault on the in- group is much harder when only some segments of the population feel threatened. Nonterritorial wars are less likely to result in public and elite acceptance of a heavier ªscal burden, increased central control, and greater social cohesion.35 For the same reasons, civil war is unlikely to signiªcantly strengthen the state.36 Developing states today thus lack the very incentives that allowed most European states to succeed in the state-building project.37 To be sure, the mechanism that causes territorial pressures to result in stronger states is not deterministic. Although wars over territory provide states

32. See, for example, Henry Tajfel and John C. Turner, “The Social Identity Theory of Intergroup Behavior,” in Stephen Worchel and William G. Austin, eds., Psychology of Intergroup Relations, 2d ed. (Chicago: Nelson-Hall, 1986), pp. 7–24. For a sociological viewpoint, see Lewis A. Coser, The Functions of Social Conºict (New York: Free Press, 1956), pp. 87–110. 33. Coser, The Functions of Social Conºict, pp. 87–110; Janice Gross Stein, “Image, Identity, and the Resolution of Violent Conºict,” in Chester A. Crocker, Fen Osler Hampson, and Pamela Aall, eds., Turbulent Peace: The Challenges of Managing International Conºict (Washington, D.C.: United States Institute of Peace Press, 2001), pp. 189–208; and Desch, “War and Strong States, Peace and Weak States?” pp. 247–248. 34. See Jack S. Levy, “Domestic Politics and War,” Journal of Interdisciplinary History, Vol. 18, No. 4 (Spring 1988), pp. 653–673; and Max Weber, “The Prestige and Power of the Great Powers,” in H.H. Gerth and C. Wright Mills, eds., From Max Weber: Essays in Sociology (New York: Oxford Uni- versity Press, 1946), pp. 159–162. 35. Here I diverge from the argument of Herbst, “War and the State in Africa”; and Desch, “War and Strong States, Peace and Weak States?” Both authors argue that the absence of international wars (of any kind) is what weakens third world states, whereas I maintain that only the absence of a particular kind of war (i.e., wars of territorial conquest) has this effect. 36. See Cameron G. Thies, “State Building, Interstate and Intrastate Rivalry: A Study of Post- colonial Developing Country Extractive Efforts, 1975–2000,” International Studies Quarterly, Vol. 48, No. 1 (March 2004), pp. 53–72. 37. In most cases in which a state in the post–World War II era faced a real threat to its territorial integrity and its survival—such as in Israel, South Korea, and Taiwan (though not ofªcially a state)—it engaged relatively successfully in state building. See, for instance, Joel S. Migdal, Strong Societies and Weak States: State-Society Relations and State Capabilities in the Third World (Princeton, N.J.: Princeton University Press, 1988), pp. 142–205. Although some states that faced territorial threats did not grow stronger (e.g., Jordan and Pakistan), they were not wiped off the map, as they probably would have been in previous eras. When Good Fences Make Bad Neighbors 149

with incentives to become stronger, the decision to engage in state building rests with policymakers. These individuals often do not correctly interpret their strategic environment, or they are unable to engage in the costly project of state building due to objections from powerful actors in their societies. Still, war is often the ªnal arbiter between correct (self-empowering) decisions and wrong (self-defeating) ones. In the past, states that built strong institutions and mobilized their societies not only survived but thrived. Those that did not eventually crumbled.38 By contrast, states today do not “die,” because this selection mechanism no longer operates, thus practically guaranteeing the survival of weak states.39 juridical statehood, moral hazard, and state building By embracing the idea of ªxed borders, the international community agreed, in essence, to preserve the shell of the state regardless of its overall weakness— a phenomenon that Robert Jackson called “juridical statehood.”40 Juridical statehood provides rulers in already weak states with strong incentives to abandon investment in and control of their peripheries, thus making the state even weaker. The logic of Jackson’s argument is simple: to achieve a minimally strong state, rulers must invest resources and take signiªcant political risks by threatening the interests of entrenched elites. This includes creating a monopoly over the means of violence (i.e., policing) and developing investment strat- egies to meet the population’s basic needs (e.g., building infrastructure and establishing viable educational, judicial, and bureaucratic institutions). In the absence of such steps, rulers will have great difªculty convincing their citizens of the legitimacy and authority of the state. Yet juridical statehood guarantees states membership in the community of nations, with all the advantages this entails. It also guarantees that as long as a ruler is able to control the capital, his status as head of state is assured. It is, Jackson concluded, “like an insurance policy: The policy holders, and conse- quently the main beneªciaries, are the rulers and regimes—not the people”— hence, the moral hazard. In the absence of a juridical or military threat to the existence of a state and its territory, a ruler of a weak state has few incentives to engage in the costly and often dangerous enterprise of controlling the state’s

38. Tilly, Coercion, Capital, and European States; and Hendrik Spruyt, The Sovereign State and Its Com- petitors (Princeton, N.J.: Princeton University Press, 1994), pp. 32–33. For discussion of a similar process in ancient China, see Victoria Tin-bor Hui, “Toward a Dynamic Theory of International Politics: Insights from Comparing Ancient China and Early Modern Europe,” International Organi- zation, Vol. 58, No. 1 (January 2004), pp. 175–205. 39. See Tanisha M. Fazal, “State Death in the International System,” International Organization, Vol. 58, No. 2 (April 2004), p. 320. 40. Robert H. Jackson, “Juridical Statehood in Sub-Saharan Africa,” Journal of International Affairs, Vol. 46, No. 1 (Summer 1992), pp. 1–16. International Security 31:3 150

periphery.41 Moreover, well-trained armies and efªcient bureaucracies—two central institutions of state building—are not only expensive to create and maintain, but given their potential to compete for power, they may be perceived as potential threats to the ruler. In addition, rulers of weak states are likely to marginalize peripheral regions both politically and economically without fear of their seceding or being annexed. Neither secession nor annexation by a neighbor would gain support from an international community that highly values the territorial status quo.42 The consequences, again, are a further weakening of the state and, in some cases, its complete disintegration— the ultimate form of which is a collapsed state. Examples include, at various times, Afghanistan, Cambodia, Chad, Lebanon, Liberia, Sierra Leone, Somalia, and Zaire/DRC.43 The next section details the mechanisms that make many weak states in a world of ªxed borders constant sources of conºict and instability.

Fixed Borders, Weak States, and International Conºict

Insecurity and opportunism remain important factors in the instigation of conºict in regions with weak states. The mechanisms through which these two factors operate, however, are profoundly different from those typically discussed in the international relations literature, which has concentrated instead on the developed world, in general, and on great power relations, in particular. In the case of weak states, insecurity can lead to the spillover of civil conºict into neighboring states, while opportunism may be manifested through nonterritorial intervention.44 weak states, civil conflict, and spillover The spillover of internal strife is the ªrst mechanism through which weak states can become a source of international conºict. Since the end of World

41. Jackson, Quasi-states, pp. 1–12, 40–59, 74–81. 42. William Reno, Warlord Politics and African States (Boulder, Colo.: Lynne Rienner, 1998), pp. 1– 40, at pp. 8–9. 43. I. William Zartman, “Introduction: Posing the Problem of State Collapse,” in Zartman, ed., Col- lapsed States: The Disintegration and Restoration of Legitimate Authority (Boulder, Colo.: Lynne Rienner, 1995), pp. 1–11; René Lemarchand, “Patterns of State Collapse and Reconstruction in Cen- tral Africa: Reºections on the Crisis in the Great Lakes,” African Studies Quarterly, Vol. 1, No. 3 (1997), http://web.africa.uº.edu/asq/v1/3/2.htm; and Reno, Warlord Politics and African States, pp. 45–79. 44. Brian L. Job, “The Insecurity Dilemma: National, Regime, and State Securities in the Third World,” in Job, ed., The Insecurity Dilemma: National Security of Third World States (Boulder, Colo.: Lynne Rienner, 1992), pp. 17–19; and Steven R. David, “Explaining Third World Alignment,” World Politics, Vol. 43, No. 2 (January 1991), pp. 233–256. When Good Fences Make Bad Neighbors 151

War II, internal wars have been much more common than wars between states.45 One study, for example, identiªed 126 such conºicts out of a total of 164 wars fought from 1945 to 1995.46 Another found that 92 of the 108 armed conºicts in the 1990s were wars between organized communal groups or between such groups and their governments.47 Communal wars are a common feature in regions with relatively young states, such as Africa, Asia, and the Middle East.48 That some regions are more prone to civil and communal war than others cannot be explained solely by the presence of ethnic or other minority groups. Although many states contain such groups, the distribution of civil/communal conºict remains uneven. The strength of the state often accounts for this disparity: the weaker the state, the more likely it is to be involved in a civil or communal war. The outbreak of communal ªghting can be the result of one of two pro- cesses, both of which are characteristic of weak states. The ªrst is “emerging anarchy”: that is, as a state’s ability to enforce order and provide internal security decreases, the potential danger that one group poses to another increases.49 At the same time, the weakness of state authority encourages some groups to think they can achieve victory in a civil war or otherwise gain from predatory behavior.50 Second, the lack of institutional means for gaining legitimacy is inherent in weak states. To compensate for this deªciency without having to resort to the expensive practice of state building, leaders might choose to incite or promote internal conºict in hopes of riding an ethnic wave to political and economic gain. They might engage in divide-and-rule tactics or side explicitly or implic- itly with one group against another—for instance, by branding members of a group as “foreign agents.” This practice is likely to be more common in an era

45. Ted Robert Gurr, “Minorities and Nationalists: Managing Ethnopolitical Conºict in the New Century,” in Crocker, Hampson, and Aall, Turbulent Peace, pp. 163–188. 46. Holsti, The State, War, and the State of War, p. 22. 47. Peter Wallensteen and Margareta Sollenberg, “Armed Conºict, 1989–98,” Journal of Peace Re- search, Vol. 36, No. 5 (September 1999), pp. 593–606. 48. Holsti, The State, War, and the State of War, pp. 19–21. Civil wars were common in Latin America as well, but their motives—at least during the Cold War era—were more often ideological. 49. See Barry R. Posen, “The Security Dilemma and Ethnic Conºict,” Survival, Vol. 35, No. 1 (Spring 1993), pp. 27–47; James D. Fearon, “Commitment Problems and the Spread of Ethnic Conºict,” in David A. Lake and Donald Rothchild, eds., The International Spread of Ethnic Conºict: Fear, Diffusion, and Escalation (Princeton, N.J.: Princeton University Press, 1998), pp. 115–125; David A. Lake and Donald Rothchild, “Containing Fear: The Origins and Management of Ethnic Conºict,” International Security, Vol. 21, No. 2 (Fall 1996), pp. 48–52; and Sarah K. Lischer, “Causes of Communal War: Fear and Feasibility,” Studies in Conºict and Terrorism, Vol. 22, No. 4 (November 1999), pp. 331–335. 50. On greed as a motive for violence in failed states, see Nelson Kasªr, “Domestic Anarchy, Secu- rity Dilemmas, and Violent Predation: Causes and Failures,” in Robert I. Rotberg, ed., When States Fail: Causes and Consequences (Princeton, N.J.: Princeton University Press, 2004), pp. 53–76. International Security 31:3 152

of ªxed borders because targeted groups have no exit option (i.e., secession or irredentism). It is also likely to further increase communal fear and communal opportunities for predation. Thus, the weaker the state, the greater is the likelihood of civil violence.51 Civil conºict has the potential to lead to international conºict in two ways,52 both of which are related to fear and insecurity. The ªrst involves refugee ºows and cross-border insurgency. Refugee ºows are often the vehicle through which internal ªghting spreads to neighboring countries. Internal wars, most notably ethnic conºicts, tend to produce more refugees than do interstate conºicts.53 These refugees, especially those living in camps close to the borders of their homeland who enjoyed relatively high group cohesion prior to ºeeing, may engage in cross-border inªltrations and attacks. When their home country retaliates, the conºict can escalate into a full-scale international war. Refugee ºows are more likely to produce international conºict when weak states are involved. The governments of such states lack the capacity either to resettle the refugees (and thus provide them with incentives to integrate into local populations rather than to continue to ªght against their home countries) or to force them to abandon their armed struggle to avoid retaliation by their state of origin.54 The case of Palestinian refugees in Lebanon is a classic example: the Lebanese government, too weak to integrate the refugees, to force its authority over them, or to effectively counter Israeli retaliation against attacks by the Palestinian Liberation Organization, could do little to prevent Israel’s 1982 invasion. The second way in which civil strife can lead to international war is through the “kin-country syndrome”: that is, when ethnic afªliations do not corre-

51. Stein, “Image, Identity, and the Resolution of Violent Conºict,” p. 193; and Lake and Rothchild, “Containing Fear,” pp. 53–56. Of course, protracted civil war might further magnify the state’s weakness. Yet cases in which a strong state becomes embroiled in civil war and emerges from it as a weak state are few. Yugoslavia is perhaps one, but it was considerably weak even before the war. 52. Michael E. Brown, “Introduction,” in Brown, ed., The International Dimensions of Internal Conºict (Cambridge, Mass.: MIT Press, 1996), pp. 1–32. 53. Myron Weiner, “Bad Neighbors, Bad Neighborhoods: An Inquiry into the Causes of Refugee Flows,” International Security, Vol. 21, No. 1 (Summer 1996), pp. 5–42. See also Job, “The Insecurity Dilemma,” pp. 3–7; and Kathleen Newland, “Ethnic Conºict and Refugees,” in Michael E. Brown, ed., Ethnic Conºict and International Security (Princeton, N.J.: Princeton University Press, 1993), pp. 143–163. 54. See especially Sarah K. Lischer, Dangerous Sanctuaries: Refugee Camps, Civil War, and the Di- lemmas of Humanitarian Aid (Ithaca, N.Y.: Cornell University Press, 2005), pp. 28–31; but also Ted Robert Gurr, “The Internationalization of Protracted Communal Conºicts since 1945: Which Groups, Where, and How?” in Manus I. Midlarsky, ed., The Internationalization of Communal Strife (London: Routledge, 1992), pp. 4–5; and I. William Zartman, “Internationalization of Communal Strife: Temptations and Opportunities of Triangulation,” in Midlarsky, The Internationalization of Communal Strife, pp. 27–42. When Good Fences Make Bad Neighbors 153

spond to international borders, and when ethnic groups in one country become alarmed by the grievances of their brethren across the border.55 This syndrome can increase tensions between neighboring states and ultimately lead to an international war or intervention.56 Turkey’s 1974 intervention in Cyprus, for instance, could be considered a consequence of the kin-country syndrome. The potential for this syndrome to occur exists everywhere. Yet it is much more likely to emerge in weak states, because the people have less afªnity for those states, and because these states are unlikely to achieve greater social cohesiveness in the absence of external territorial pressures. opportunism and greed The second mechanism through which weak states can spur international conºict is opportunity and greed. In a world of ªxed borders, opportunities for territorial revisionism are greatly reduced. Still, in some cases greed can play a signiªcant role. Although states cannot legally annex a neighbor’s territory, they may be able to exploit this territory for economic or political gain, including regime change.57 Weak states are more likely to be victims of neighboring states’ greed for two reasons. First, neighbors are more likely to ªnd potential allies in weak states than in strong ones. Second, many sociopolitically weak states are also militarily weak, because maintaining a strong, cohesive military and collecting taxes for this purpose represent huge challenges.58 Examples of states that have pursued predatory policies at the expense of a weak neighbor include Israel and Syria. Since the 1970s, both countries have sought (sometimes successfully) to exert control over Lebanese politics and (in Syria’s case) to exploit Lebanon’s economy through direct military intervention. Never, however, has either country sought to annex Lebanese territory. The effect of these nonterritorial threats differs signiªcantly from that of territorial threats because the former are often perceived as threats to a particular

55. Samuel P. Huntington, “The Clash of Civilizations,” Foreign Affairs, Vol. 72, No. 3 (Summer 1993), pp. 35–39; Lemarchand, “Patterns of State Collapse and Reconstruction in Central Africa,” pp. 5–6; and Stuart J. Kaufman, “An ‘International’ Theory of Inter-ethnic War,” Review of Interna- tional Studies, Vol. 22, No. 2 (April 1996), p. 153. 56. Stephen M. Saideman, The Ties That Divide: Ethnic Politics, Foreign Policy, and International Conºict (New York: Columbia University Press, 2001), pp. 1–35, 203–222. For a counterargument, see Michael E. Brown, “The Causes and Regional Dimensions of Internal Conºict,” in Brown, The International Dimensions of Internal Conºict, pp. 603–627. 57. On the proposition that territorial conquest was replaced by intervention for regime change, see Tanisha M. Fazal, “From Conquest to Intervention: State, Regime, and Leader Exit,” paper presented at the annual meeting of the International Studies Association, Honolulu, Hawaii, March 1– 5, 2005. 58. Buzan, People, States, and Fear, p. 113. International Security 31:3 154

group within the weak state, not to the state as a whole. Therefore they do not generate the usual in-group/out-group dynamics. As Lewis Coser notes, “The relations between outer conºict and inner cohesion does not hold true where internal cohesion before the outbreak of the conºict is so low that the group members have ceased to regard preservation of the group as worthwhile, or actually see the outside threat to concern ‘them’ rather than ‘us.’”59 The next section analyzes the case of the war in Congo in light of the theory suggested here and the hypotheses derived from it.

The Case of Congo

The war in Congo illustrates the mechanism through which adherence to the norm of ªxed borders can lead to international conºict. After offering a brief review of Congo’s postindependence history, I describe how the norm perpet- uated and exacerbated the state’s weakness. I then show how this weakness triggered mechanisms that expanded the conºict between Congo and its neighbors. from independence to war When Congo achieved independence from Belgium in 1960, it displayed all the signs of a weak state: it suffered from regional fragmentation, a crisis in central authority, and a high level of political violence.60 Political divisions along ethnic lines were prevalent, though this fact could be misleading. Ethnic- ity became the primary source of political mobilization in Congo not because of the salience of rigid ethnic identities, but because the state lacked legitimacy, a unifying idea around which the new nation could rally, and the institutions to promote it.61 In 1965 Chief of the Army Mobutu Sese Seko seized power in a coup and established, in the state he renamed Zaire, what became one of Africa’s most enduring regimes.62 Throughout Mobutu’s rule, Zaire frequently interfered in the internal affairs of neighboring states, backing guerrilla movements in, for example, Angola, Burundi, Rwanda, and Uganda. Such actions often produced counterinter- ventions, including Angolan support of the Katangese rebellions in Shaba Province in the 1970s.

59. Coser, The Functions of Social Conºict, pp. 93–95, at p. 93. 60. Crawford Young and Thomas Turner, The Rise and Decline of the Zairian State (Madison: Univer- sity of Wisconsin Press, 1985), pp. 41–42. 61. In fact, scholars of Congo note the considerable degree of ºuidity of Congolese ethnic identities at the time. See ibid., pp. 40–42; and John Breuilly, Nationalism and the State, 2d ed. (Chicago: University of Chicago Press, 1994), pp. 200–204. 62. Young and Turner, The Rise and Decline of the Zairian State, pp. 42–43. When Good Fences Make Bad Neighbors 155

In 1996 an armed rebellion against Mobutu’s regime, led by Laurent Désiré Kabila of the Alliance of the Democratic Forces for the Liberation of Congo- Zaire (ADFL), erupted in Kivu Province. The ADFL, backed by Rwandan forces and supported (ªnancially and politically) by most states in the region, rapidly defeated Congo’s armed forces. In May 1997 Mobutu ºed into exile, allowing Kabila to take power and change the name of the country from Zaire to the Democratic Republic of Congo. The rebellion represented the ªrst phase in a protracted war that ended only in 2002. Soon after taking Kinshasa, Kabila began to alienate many of his former allies, both domestic and foreign. Lending support to a new rebellion in eastern Congo, Burundian, Rwandan, and Ugandan forces undertook a massive intervention in August 1998 and pushed rapidly toward the Congolese capital. To save his collapsing regime, Kabila requested military support from Angola, Chad, Namibia, Sudan, and Zimbabwe. Their counterintervention prevented Kabila’s downfall, but the war raged on as foreign troops and local rebels established control over much of Congo’s vast territory. Casualties ran as high as 3.8 million dead as a direct or indirect result of the war.63 The country’s already fragile economic situation and poor health conditions deterio- rated even further. Laurent Kabila was assassinated by his bodyguard in January 2001. Replacing him was his son, Joseph Kabila. In July 2002 the DRC and Rwanda signed the Pretoria accord, which led to the withdrawal of all foreign troops from Congo and the creation of a transitional government that included rebels and opposition parties. What has been labeled “Africa’s Great War” reached its end. The ªghting between the Congolese army and various local militias continues, however, while armed clashes between political parties in the aftermath of the August 2006 elections illustrate the country’s ongoing vulnerabil- ity. Moreover, Rwanda has threatened renewed intervention should the Congolese government continue its support of the anti-Rwandan Hutu insurgency mounted from the DRC. The disastrous consequences of Africa’s Great War—including the huge death toll, tremendous human suffering, massive displacement, and signiªcant damage to Congo’s infrastructure will probably take decades to overcome. In addition, the country’s ongoing instability serves as a powerful impediment to prospects for a brighter future.64

63. This ªgure comes from a survey conducted by the International Rescue Committee in 2004; it includes military and civilian casualties, as well as war-related deaths caused directly by violence and indirectly as a consequence of famine and epidemics. 64. Comprehensive sources on the war are still scarce. For a decent beginning, see John F. Clark, ed., The African Stakes of the Congo War (New York: Palgrave Macmillan, 2002). International Security 31:3 156

fixed borders and state weakness in congo Congo’s emergence from Belgian rule as a weak state is not surprising; most states are weak at this early stage. Like many other states in Africa and Asia, however, Congo has remained so; indeed, it is weaker today than it was in the 1960s. The explanation for this ongoing weakness lies largely with the norm of ªxed borders as stipulated in hypothesis 1. This section examines Congo’s state-building efforts (or lack thereof) in ªve core areas: (1) monopoly over the legitimate use of violence; (2) revenue extraction; (3) state bureaucracy and institutions; (4) distribution of public goods; and (5) social cohesion and ident- iªcation with the state. monopoly over the legitimate use of force. Absent an external territorial threat, governments can afford to have inefªcient, weak, and corrupt militaries. Writing about Zaire under President Mobutu, Edgar O’Ballance noted, “The main role of the armed forces remained internal security, with little thought or energy going to national defense.”65 Militaries of this kind re- quire little investment or sacriªce by either the state or its citizens. Moreover, a weak military reduces the threat of a coup. In Zaire’s case, Mobutu systemati- cally rechanneled resources intended for the army to his Presidential Guard, whose members were largely recruited from his home province, Equator.66 “The very importance of the military in Zairian politics,” argued Michael Shafer, resulted “in Mobutu’s extraordinary efforts to divide, control, manipu- late, politicize, and otherwise deinstitutionalize and de-professionalize it.”67 When faced with internal rebellion, as for example in the two Shaba (Katanga) rebellions in the 1970s, Mobutu had to rely on troops from Morocco and France to compensate for Zaire’s deªcient army.68 By the 1990s, rather than paying Zaire’s military forces with state funds, Mobutu encouraged them to seek compensation through illegal activities, such as looting and kidnapping for ransom.69 Laurent Kabila’s record in this respect was no different. Despite the country being torn by civil and interstate war, the DRC’s military expenditures in 2003 stood at only 1.4 percent of the country’s gross domestic product (GDP).70 Yet

65. Edgar O’Ballance, The Congo-Zaire Experience, 1960–98 (New York: St. Martin’s, 2000), pp. 114– 115. 66. David Shearer, “Africa’s Great War,” Survival, Vol. 41, No. 2 (June 1999), pp. 89–106. 67. Quoted in Young and Turner, The Rise and Decline of the Zairian State, p. 274. 68. Ibid., pp. 74–75, 248–275; and Thomas M. Callaghy, The State-Society Struggle: Zaire in Compara- tive Perspective (New York: Columbia University Press, 1984), pp. 206–209. 69. Reno, Warlord Politics and African States, pp. 159–162; and René Lemarchand, “The Democratic Republic of Congo: From Failure to Potential Reconstruction,” in Rotberg, State Failure and State Weakness in a Time of Terror, pp. 40–41. 70. Central Intelligence Agency, “The World Factbook, 2004.” When Good Fences Make Bad Neighbors 157

even when foreign troops controlled much of Congo’s territory, Kabila’s government was still the recognized sovereign of the entire country. The norm of ªxed borders prevented external challenges to Congo’s juridical sovereignty, and no internal secession attempts could muster the international support needed to succeed.71 revenue extraction. The ability of the state to tax its citizenry is a sign of a strong state, and the revenues derived from taxation help to sustain it. Under Mobutu, Zaire’s tax collection system was inefªcient and corrupt.72 Tax revenues throughout the 1970s and 1980s amounted to a mere 6–11 percent of the country’s GDP, of which only 25–35 percent consisted of taxes on income and capital revenues. In 1995 tax revenue plunged to about 5 percent of GDP; in 2000 it fell to 4 percent.73 Given that by the 1990s many economic transactions were being conducted outside the ofªcial Zairian economy, the actual tax burden on Zaire’s citizens was probably much smaller. Instead of tax revenues, Mobutu relied on short-term policies, external support, and Zaire’s increasing foreign debt to sustain the patrimonial network through which he controlled the state. In 1973, for example, he oversaw the state’s seizure of vast tracts of land and commercial enterprises owned by foreign nationals in a process called “Zairianization.” As Crawford Young and Thomas Turner write, “The cement of clientage was access to resources. The sudden takeover of this huge zone of the economy offered a vast new pool of goods for patrimonial distribution to deserving members of the political class.”74 state bureaucracy and institutions. Early in his reign, Mobutu did take some steps toward building Zaire’s institutions. He founded the Popular Movement of the Revolution (known by the French acronym MPR) as Zaire’s sole political party in an attempt to generate legitimacy for his regime and provide it with the organizational tools needed to penetrate the civil society. In the late 1960s, Mobutu sought to use the MPR to expand his institutional control over the country, including the education system (hitherto controlled by the Church), the military, and the regional authorities.75

71. Jeffrey Herbst, “Let Them Fail: State Failure in Theory and Practice—Implications for Policy,” in Rotberg, When States Fail, pp. 306–308; and Reno, Warlord Politics and African States, pp. 172–173. 72. Callaghy, The State-Society Struggle, pp. 194–204. 73. For these data, see World Bank, Development Indicators, http://devdata.worldbank.org.ezp2 .harvard.edu/dataonline. Compare this, for example, with tax revenues of 30–38 percent for the same years in France and 16–19 percent in the United States, which in the latter case are considered to be very low levels of taxation by Western standards. 74. Young and Turner, The Rise and Decline of the Zairian State, p. 328. See also Reno, Warlord Politics and African States, p. 152; and Callaghy, The State-Society Struggle, pp. 191–192. 75. It is at least plausible to assume, though, that it took Mobutu and his regime a few years to rec- ognize the extent of the new norm of ªxed borders, and that only after the unequivocal position of International Security 31:3 158

The appearance of state building, though, is misleading. In fact, Mobutu’s policies and rhetoric were geared primarily toward achieving three aims. First, Mobutu aspired to be an African—or even a third world—leader, an aspiration that sometimes conºicted with his pro-Western policies and dependence on Western support and loans. His nationalist rhetoric and extravagant projects were instead meant to cast him in a more positive light in the eyes of other African leaders. Two public relations campaigns—the “Authenticity” cam- paign of 1971 and “Mobutism” in 1974—attempted to achieve this objective through the use of nothing more than cheap rhetoric. “Zairianization” in 1973 and “Radicalization” in 1974 sought to fulªll the same aim through economic means. In this way, Mobutu believed that he could claim African leadership without abandoning his dependence on Western assistance.76 Second, Mubutu’s policies aimed at sustaining and enhancing his patrimonial network. Zaire, as Young and Turner write, possessed “a dual character: formally institutionalized, in party and administration, but informally patrimonial and personal.”77 By the mid-1980s, Zaire had become “an early modern leviathan, but a lame one.”78 On the one hand, the state was highly centralized and authoritarian, much like the absolutist states of seventeenth- and eighteenth-century Europe. On the other hand, its reach was extremely limited. The state of Zaire, as an entity separate from its ruler, did not exist in any meaningful sense.79 By 1975 Mobutu had abandoned even the pretense of state building. Faced with a growing foreign debt and a declining revenue base (due to the falling price of copper and the colossal economic failure of Zairianization), “Mobutu rejected the pursuit of policies or the building of institutions that would have served the collective good. Instead, he consolidated the international community on the Nigerian civil war (in 1970) became clear did they fully grasp the new norm’s implications. 76. “Authenticity” emphasized the need to replace colonial and Western names and institutions with authentic African ones. “Mobutism” was, in essence, the elevation of Mobutu’s policies and words over all others. “Zairianization” and “Radicalization” were economic policies that allowed the state to seize private assets belonging to foreign nationals and redistribute them among its supporters. See Young and Turner, The Rise and Decline of the Zairian State, pp. 54–71, 185–247, 326– 362, at p. 327; Callaghy, The State-Society Struggle, pp. 233–276; Kevin C. Dunn, “Imagining Mobutu’s Zaire: The Production and Consumption of Identity in International Relations,” Millen- nium: Journal of International Studies, Vol. 30, No. 2 (June 2001), pp. 245–258; and Reno, Warlord Poli- tics and African States, pp. 151–153. 77. Young and Turner, The Rise and Decline of the Zairian State, p. 397. Max Weber uses the term “patrimonial” to denote a regime that is, for most practical purposes, an extension of the ruler’s household. See Weber, Economy and Society: An Outline of Interpretive Sociology (New York: Bedminster, 1968), pp. 231–232, 1010–1044. 78. Callaghy, The State-Society Struggle, p. 409. See also Jean-Claude Willame, Patrimonialism and Political Change in the Congo (Stanford, Calif.: Stanford University Press, 1972), p. 2. 79. Young and Turner, The Rise and Decline of the Zairian State, pp. 164–184. When Good Fences Make Bad Neighbors 159

his own authority by monopolizing resources,” ushering in an era of state de- cay and collapse.80 Third, Mobutu deferred from creating a stronger state apparatus because an efªcient bureaucracy, much like a strong army, could have served as an independent power base that might have threatened his rule, particularly in times of crisis. Instead, he chose to shrink Zaire’s state bureaucracy and rely on a network of clients (especially regional strongmen) to fulªll state functions.81 When Laurent Kabila took power in 1997, he “inherited less a state than a ªefdom. Normal state functions had been replaced by patronage.”82 Like Mobutu, Kabila faced structural disincentives to state building, including the international community’s support for the territorial status quo and the moral hazard that attended its uncritical recognition of Congo’s juridical statehood. Indeed, Kabila’s brief reign revealed many of the same tendencies discussed above, such as constructing a clientele network instead of efªcient bureaucracies and playing ethnic groups against each other.83 As a result, Laurent Kabila’s Congo remained “an institutional clone of its predecessor.”84 distribution of public goods. The ability of the state to provide its citizens with public goods is an important indicator of its overall strength. Mobutu’s government spent little on public goods, and when it did the focus was mostly on extravagant, highly visible, and often misguided projects—for example, the building of the Ingha-Shaba Dam and construction of a high- power grid to transfer the energy produced by the dam to the Shaba mines, which actually could have relied on much cheaper energy sources. At the same time, state spending on social services shrank from 17.5 percent in 1972 to virtually nothing in 1992. This pattern of state expenditure points again to a policy aimed at gaining cheap legitimacy (both internal and external) at the expense of creating real, though politically less beneªcial, public goods.85 In addition, an ever increasing portion of the state’s budget was put under the sole discretion of the president. These funds, which in 1992 amounted to 95 percent of Zaire’s budget, were funneled either to Mobutu’s personal accounts or to his cronies. During this period, government spending targeted some parts of the coun-

80. Ibid., pp. 71–77; and Weinstein, “Africa’s ‘Scramble for Africa,’” pp. 14–15. 81. Reno, Warlord Politics and African States, pp. 149–152. 82. Shearer, “Africa’s Great War,” p. 92. 83. Reno, Warlord Politics and African States, pp. 172–173; and Lemarchand, “The Democratic Re- public of Congo,” pp. 44–45. 84. Lemarchand, “The Democratic Republic of Congo,” p. 52. 85. Ibid., p. 154; and Callaghy, The State-Society Struggle, pp. 184–194. International Security 31:3 160

try while neglecting others. Because control of the capital city is the only re- quirement the international community maintains for recognizing state sovereignty, it was the only asset that Mobutu’s regime needed to invest in (apart from sustaining the stream of cash from Zaire’s natural resources through local strongmen). Between 1969 and 1972, for example, Kinshasa— home to about 6 percent of Zaire’s population—was allocated 31 percent of the budget for government-approved investment projects. Shaba Province received about 47 percent (all of which went to the mineral mining industry), despite being home to less than 13 percent of Zaire’s population. In contrast, Kivu Province, with approximately 15 percent of the population, was allocated only 1.5 percent of the budget. And Kasai Province, with almost 20 percent of the state’s population, received virtually nothing.86 Had the government needed to enhance the legitimacy of the state and assure, in particular, the alle- giance of inhabitants along Zaire’s borders (so as to prevent irredentism), the distribution of funds would have looked quite different. social cohesion and identification with the state. In many African states, social cohesion and identiªcation of the population with the state are tenuous at best.87 Congo is no different, although determining the level of loyalty to the state is difªcult in the absence of multiparty elections or reliable polling. Moreover, with authoritarian regimes such as Mobutu’s Zaire, ascer- taining whether resistance is directed at the dictator or at the state more gener- ally can be extremely difªcult. Still, one can glimpse some signs of the low level of legitimacy the population of Zaire accorded their state. One was the outbreak of rebellions that explicitly sought secession from the state. Three such rebellions occurred in the Shaba region in the early 1960s and in 1977 and 1978. A second indicator was the size of the informal (or parallel) economy. Smuggling, tax evasion, a robust black market, unregulated trade, and withdrawal to noncommercial agriculture can all be considered forms of social disengagement from and disapproval of the state—and in Zaire, all were common practices that became more popular over time.88 “The informal sec- tor,” a U.S. State Department analysis concluded in 2005, “now dominates the

86. For the marginalization of provinces beyond Kinshasa and the Copper Belt region in Shaba, see Young and Turner, The Rise and Decline of the Zairian State, pp. 78–91. For data, see ibid., pp. 82– 83. 87. See Naomi Chazan, “Patterns of State-Society Incorporation and Disengagement in Africa,” in Donald Rothchild and Chazan, eds., The Precarious Balance: State and Society in Africa (Boulder, Colo.: Westview, 1980), pp. 121–148. 88. For the general argument, see Victor Azarya, “Civil Society and Disengagement in Africa,” in John W. Harbeson, Donald Rothchild, and Naomi Chazan, eds., Civil Society and the State in Africa (Boulder, Colo.: Lynne Rienner, 1994), pp. 83–100. For the case of Congo, see Janet MacGaffey, “Civil Society in Zaire: Hidden Resistance and the Use of Personal Ties in Class Struggle,” in Harbeson, Rothchild, and Chazan, Civil Society and the State in Africa, pp. 169–189. When Good Fences Make Bad Neighbors 161

[Congolese] economy.”89 Writing in 1984, Thomas Callaghy noted a third sign: Mobutu had “occupied just about every possible ideological position in a whirl of legitimating eclecticism.”90 Mobutu’s decision to focus much of his efforts on the campaigns described above highlights a basic lack of state legitimacy and low levels of public loyalty to the state. Although these signs are indirect measures, they should sufªce in cases where direct measures are unavailable. Kabila’s DRC, like Mobutu’s Zaire, was confronted with threats of external intervention, but none produced efforts at increasing social cohesion because none represented a territorial threat, let alone a threat to the survival of the state. As Coser notes, for an external threat to produce internal cohesiveness, that threat must be perceived as a threat to the entire country.91 In the DRC, the external threat was viewed in ethnic terms: a threat to some, but an opportunity for others to change the internal balance of power, though not the DRC’s borders. Therefore the 1998 invasion by foreign forces, much like previous interventions, did not generate greater in-group cohesion.92 Mobutu’s Zaire was a weak state at its inception and remained so throughout his reign; the same was true of Laurent Kabila’s DRC. Although Mobutu’s and Kabila’s decisionmaking may not directly prove the relationship between state weakness and the norm of ªxed borders, it does point to the absence of concern for the survival of the state (as opposed to the regime) and to the via- bility of their country’s international boundaries. With the norm of ªxed borders in place, Congo was able to endure and remain territorially intact. state weakness, civil conflict, and spillover in central africa If the norm of ªxed borders perpetuates and exacerbates weakness in already weak states, then one should at least be able to expect that it would curb instances of external conºict. In the case of Congo, however, it had encouraged them. This section explores the proposition in hypotheses 2 and 3 that state weakness can lead to international conºict through the spillover of internal conºict into neighboring states. It examines the spread of ethnic war across the borders of Congo and its neighbors and explains why state weakness and ªxed borders facilitated this process and brought about intervention and international war.

89. U.S. Department of State country background proªle at http://www.state.gov/r/pa/ei/bgn/ 2823.htm. 90. Callaghy, The State-Society Struggle, pp. 410–414, at p. 412. 91. Coser, The Functions of Social Conºict, pp. 93–95. 92. On the way various ethnic groups in Congo allied themselves with the different external inter- veners, see Thomas Turner, “The Kabilas’ Congo,” Current History, May 2001, pp. 215–218. International Security 31:3 162

state weakness and civil conflict in congo. Emerging anarchy and the exclusion of certain ethnic groups for political purposes increase the likelihood that weak states will become involved in civil conºicts and internal wars. Both factors played crucial roles in Congo. Low-level, violent civil conºicts simmered throughout Mobutu’s rule, though usually out of the public eye (e.g., the Katangese rebellions in Shaba Province in the 1970s, which were part of an Angolan-assisted secessionist drive that had to be quelled by Belgian and Tunisian troops when it became clear that Zaire’s armed forces were not up to the task). In the early 1990s, tensions among Zaire’s multiple ethnic groups erupted into full-ºedged internal wars. In Shaba Province, pro-Mobutu gangs drove hundreds of thousands of Kasai people from their homes. In North Kivu Province, approximately 10,000 Banyarwanda were killed and another 250,000 became refugees.93 In a strong state, the government and the military would have sought to deter such wars or at least step in once they had begun. But in Zaire, “three months and several thousand deaths [after the breakup of violence,] the authorities had done nothing.”94 The internal wars, the Economist thus concluded, were “a symptom of Zaire’s chronic ungovernment.”95 This, however, is only part of the explanation. Another factor was the use of diversionary political tactics. The ethnic dimension of Zairian politics gained increasing salience as Mobutu lost popular legitimacy and as external powers repeatedly intervened to prevent secessionist attempts. Although Mobutu’s rhetoric played down Zaire’s ethnic divisions, in practice he came to rely in- creasingly on an inner circle of Ngbandi from Equator Province and Lingala speakers, like himself. This was especially true of the security apparatus and the government’s most sensitive ªnancial posts.96 Throughout his reign, Mobutu employed divide-and-rule tactics, particularly in Kivu Province, where in the 1980s and 1990s he sought to portray the Banyarwanda as “for- eigners” in an attempt to gain the support of ethnic groups that competed with the Banyarwanda for land and resources. In 1991 he signed a law that stripped the Banyarwanda people of their citizenship and right to hold ofªce. Jermaine McCalpin attributes this action to Mobutu’s desire “to create a scapegoat for

93. Lemarchand, “The Democratic Republic of Congo,” pp. 38–43; and Chris McGreal, “Zaire Buries the Victims of Hatred and Expediency,” Guardian, September 24, 1993. Banyarwanda is the collective term for groups that migrated from Rwanda and Burundi (ethnically both Hutu and Tutsi) to Congo, mostly in the precolonial era. 94. McGreal, “Zaire Buries the Victims of Hatred and Expediency.” 95. “Zaire: Folly by the Numbers,” Economist, August 7, 1993, p. 41. 96. Young and Turner, The Rise and Decline of the Zairian State, pp. 152–157; and Lemarchand, “The Democratic Republic of Congo,” pp. 40–41. When Good Fences Make Bad Neighbors 163

Zaire’s many problems and to distract the attention of the populace from their real source of misery.”97 Laurent Kabila engaged in similar tactics, targeting in particular the Banyamulenge (i.e., ethnic Tutsi who are part of the larger Banyarwanda group) and accusing them of being spies for Rwanda. Kabila’s policy of exclusion also originated from a lack of alternatives to gain legitimacy in a weak state. By the end of 1997, wrote René Lemarchand, “the choice [Kabila] faced was either to hang on to his Rwandan protectors, and suffer an even greater loss of legitimacy, or to free himself of their embrace and face the consequences.”98 Still, Kabila chose not only to order all foreign troops out of the DRC, but also to encourage his army to take part in a massacre of hundreds of Tutsi living in Kinshasa.99 He then tried to ride this wave of hatred to regain legitimacy. The strategy was successful because, at least in the short run, Kabila’s popularity did rise.100 These new exclusionary policies, however, had tremendous consequences for the second Rwandan invasion. spillover of civil conflict: refugees and insurgencies. Congo was not the only weak state in the neighborhood, and its civil conºicts combined with those of neighboring states to produce an international conºagration in the mid-1990s. Chief among these states were Rwanda, Burundi, Uganda, and Angola.101 In the aftermath of the 1994 genocide in Rwanda—in which roughly 800,000 Tutsi and moderate Hutu were slaughtered—massive numbers of Hutus ºed the country, fearing retribution at the hands of the government, which had come under the control of the Tutsi-dominated Rwandan Patriotic Front. The leadership of the Hutu militants fed this fear and encouraged the exodus.102 The Rwandan refugees settled in camps in Burundi, Tanzania, Uganda, and Zaire. In Zaire, 1 million to 1.5 million Rwandan Hutus settled in huge refugee camps in South and North Kivu, the easternmost provinces of the country, in close proximity to the Rwandan border. Among them were 50,000–100,000 for-

97. Jermaine O. McCalpin, “Historicity of a Crisis: The Origins of the Congo War,” in Clark, The African Stakes in the Congo War, p. 46. See also Kisangani N.F. Emizat, “The Massacre of Refugees in Congo: A Case of UN Peacekeeping Failure and International Law,” Journal of Modern African Studies, Vol. 38, No. 2 (June 2002), pp. 166–167; and Reno, Warlord Politics and African States, p. 161. 98. Lemarchand, “The Democratic Republic of Congo,” p. 45. 99. Ibid. 100. David Shearer, “Lines on a Map,” World Today, Vol. 45, No. 11 (November 1998), p. 295. 101. All of these states are ranked among the weakest in the Failed States Index; their rankings would have been even worse in the early 1900s. See Fund for Peace, “Failed States Index”; and Lemarchand, “Patterns of State Collapse and Reconstruction in Central Africa,” pp. 7–10. 102. Sarah K. Lischer, “Internal Conºict and International Contagion: Refugees, Rebels, and Hu- manitarian Aid,” paper presented at the annual meeting of the American Political Science Associa- tion, Boston, Massachusetts, August 29–September 1, 2002, pp. 13–14. International Security 31:3 164

mer soldiers from the Rwandan Armed Forces (known by its French acronym FAR) and members of the militant Hutu militia known as the Interahamwe, which was largely responsible for the genocide. These groups were highly organized and well armed, making them a classic example of a “state in exile.” The presence of the Hutu refugees in eastern Zaire and their domination by militant groups were likely the primary cause of both phases of the international intervention in this country. Under the leadership of the Interahamwe, “the refugee camps were turned into military bases from which regular cross- border incursions were launched in order to destabilize the new Rwandan government.”103 The exodus of refugees after a military defeat in a civil war is common, as is the presence among them of guerrillas and former soldiers from the losing side. These factors do not always result in border wars of incursion and retaliation, however. If the host state can prevent refugee camps from becoming bases for insurgency, a border war may be avoidable. The Rwandan Hutu refugees who ªlled the camps of western Tanzania, for example, were prevented from using their camps as a base for launching attacks on Rwanda by the Tanzanian military’s forceful policing and sealing off of the border. Moreover, Tanzania extended citizenship to the refugees, thus reducing their level of re- sentment and increasing their chances of peacefully integrating into the local population.104 In sharp contrast, and despite similarities in the composition of the population and the prevailing economic conditions,105 Zaire could not control its borders with Rwanda or the refugee camps. Despite repeated Rwandan requests, Mobutu not only refused to confront the Hutu militants but allowed (and possibly even directed) his army to assist them. Zaire’s army facilitated the arms trade of the Interahamwe, allowed ex-FAR forces to make use of its military camps and headquarters, and was heavily involved in the militants’ extortion and suppression of the camps’ civilian refugees.106 Mobutu met regularly with senior elements of the Hutu militants, including the former Rwandan army chief of staff.107 Although Mobutu never tried seriously to disarm the

103. Christian R. Manahl, “From Genocide to Regional War: The Breakdown of International Order in Central Africa,” African Studies Quarterly, Vol. 4, No. 1 (Spring 2000), http://web .africa.uº.edu/asq/v4/v4i1a2.htm. See also Lischer, Dangerous Sanctuaries, p. 82; Shearer, “Africa’s Great War,” pp. 90–92; and Emizet, “The Massacre of Refugees in Congo,” p. 165. 104. Lischer, Dangerous Sanctuaries, pp. 91–111; and Herbst, States and Power in Africa, p. 238. 105. See Lischer, Dangerous Sanctuaries, pp. 108–111. 106. For more details, see Human Right Watch Arms Project, “Rwanda/Zaire: Rearming with Im- punity: International Support of the Perpetrators of the Rwandan Genocide” (Washington, D.C.: Human Rights Watch, May 1995). 107. Lischer, Dangerous Sanctuaries, p. 85. When Good Fences Make Bad Neighbors 165

Rwandan Hutu militia, the rapid crumbling of his army in the face of the 1996 rebellion made clear that he would not have succeeded even if this had been an objective. Because the goal of the Hutu militants was to replace Rwanda’s Tutsi regime (and perhaps also continue with their genocide), the Rwandan government viewed the Hutu insurgency in Zaire as a grave threat.108 Thus, Rwanda’s vice president, Paul Kagame, told diplomats in early 1996 that “if the international community was unable to stop the delivery of weapons to the ex-FAR and Interahamwe and the military training in the refugee camps, the Rwandan government could decide to take preventive military action.”109 Similar dynamics, though perhaps on a smaller scale, played out between Mobutu’s Zaire and Burundi, Uganda, and Angola. Refugee camps around Uvira in South Kivu Province contained 150,000 Burundian Hutu refugees and had been a breeding ground for rebels seeking to overthrow Burundi’s Tutsi minority government. Beginning in 1995, the rebel group National Council for the Defense of Democracy mounted attacks using these camps as a rear base. Burundi’s backing of the 1996–97 AFDL war against Mobutu was largely a response to this threat.110 Uganda shared similar concerns. Three of its armed opposition groups, most notably the Allied Democratic Forces (ADF), had also found safe haven in eastern Zaire.111 Again, regardless of whether Mobutu was just allowing these groups to operate from Zairian soil or was actively assisting them, Uganda’s interests in sponsoring a friendlier regime in Kinshasa and the possibility for direct action in eastern Zaire against the rebels were clear. Uganda, which supported Kabila’s war to oust Mobutu, later received permis- sion from Kabila to enter the DRC in pursuit of rebels threatening its borders and population.112 The case of Angola offers another example of intervention in a neighboring state that assisted a domestic foe. The National Union for the Total Independ- ence of Angola (known by the Portuguese acronym UNITA) fought against the Angolan government from independence, in 1975, until 2002. Mobutu supplied the primary lifeline for UNITA both by allowing the smuggling of dia-

108. Manahl, “From Genocide to Regional War,” pp. 1–4; Shearer, “Africa’s Great War,” pp. 90–92; Bruce Baker, “Going to War Democratically: The Case of the Second Congo War (1998–2000),” Con- temporary Politics, Vol. 6, No. 3 (September 2000), p. 266; and Lemarchand, “Patterns of State Col- lapse and Reconstruction in Central Africa,” p. 2. 109. Manahl, “From Genocide to Regional War,” p. 3. See also Shearer, “Africa’s Great War”; and Emizet, “The Massacre of Refugees in Congo,” p. 168. 110. Lemarchand, “Patterns of State Collapse and Reconstruction in Central Africa,” p. 2; and Shearer, “Africa’s Great War,” p. 95. 111. Lemarchand, “Patterns of State Collapse and Reconstruction in Central Africa ,” p. 2; and John F. Clark, “Explaining Ugandan Intervention in Congo: Evidence and Interpretations,” Journal of Modern African Studies, Vol. 39, No. 2 (June 2001), pp. 271–272. 112. Clark, “Explaining Ugandan Intervention in Congo,” pp. 271–273. International Security 31:3 166

monds from Angola to pass through Zaire (estimated at $500 million annually) as well as the transfer of weapons bought with this money, and by permitting UNITA to establish its rear bases on Zairian territory. Zaire also served as a conduit for the movement of money from the United States and South Africa to be used by UNITA (until the early 1990s) against Angola’s socialist government. Mobutu himself amassed huge sums from these transactions by playing the middleman and extracting fees for his services. Not surprisingly, the Angolan government supported Kabila’s effort to oust the friend of its worst enemy—Mobutu.113 Laurent Kabila’s accession to power in 1996 changed the players but not the rules of the game. Congo (now the DRC) was still an extremely weak state, unable to prevent most of the attacks on its neighbors by rebels operating from its territory, an issue that Kabila chose not make a priority of his regime. Kabila’s lack of effort to disarm the Hutu militias operating in the DRC’s eastern provinces remained Rwanda’s largest concern and led to its decision to replace him—just as it had done with Mobutu. Rwanda’s second invasion of Congo in 1998 and its selection of a local ally to lead the march on Kinshasa bore a strik- ing resemblance to the 1996 invasion.114 And, driven by a desire to eradicate Burundian Hutu rebels operating along Lake Tanganyika, Burundi partici- pated in the invasion as well.115 Uganda’s ofªcial explanation of its decision to join the 1998 intervention stressed the need to ºush out guerrilla bases in eastern Congo that served the ADF in its raids across the Ugandan border. At the minimum, Kabila had done nothing to rid his country of these rebels. At the maximum, as the Ugandan government alleged, he allowed Sudan to assist in the arming and training of the ADF, with the hope of securing aid for his army as well. Uganda’s claim of ªnding a “smoking gun” pointing to Sudanese involvement in such activities is thus plausible, given the Sudanese leadership’s policy that any enemy of Uganda was a friend of Sudan and because Uganda was assisting the rebels of southern Sudan in their ªght against the Khartoum government.116 Uganda, though, was not content with conducting limited operations against the ADF or even with establishing a security zone along its border with Congo. It par- ticipated in Rwanda’s initial airlift of troops that was intended to take control of Kinshasa, and it supported rebel groups in Equator Province (more than

113. Thomas Turner, “Angola’s Role in the Congo War,” in Clark, The African Stakes of the Congo War, pp. 75–83; Reno, Warlord Politics and African States, pp. 151–162; Weinstein, “Africa’s ‘Scram- ble for Africa,’” p. 14; and Shearer, “Africa’s Great War,” pp. 96–97. 114. Shearer, “Africa’s Great War,” pp. 93–95; Weinstein, “Africa’s ‘Scramble for Africa,’” p. 12; and Lemarchand, “The Democratic Republic of Congo,” pp. 45–48. 115. Baker, “Going to War Democratically,” p. 269. 116. Shearer, “Africa’s Great War,” pp. 95–96; Baker, “Going to War Democratically,” pp. 268–269, at p. 269; and Clark, “Explaining Ugandan Intervention in Congo,” pp. 271–273. When Good Fences Make Bad Neighbors 167

1,000 miles away from the Ugandan border). Uganda’s president, Yoweri Museveni, apparently concluded that a long-term solution to his country’s security problems required a friendlier regime in Kinshasa.117 In 1998 Angola, an ally of Rwanda and Uganda in their 1996 drive to oust Mobutu, found itself supporting Kabila against them. Angolan troops, with Zimbabwean assistance, saved Kinshasa from falling once again into the hands of a group of Rwandan-backed rebels known as the Congolese Rally for Democracy (known by the French acronym RCD). Considerations regarding Congo’s internal war account for Angola’s actions this time as well. Kabila dis- rupted the ºow of UNITA’s weaponry and diamonds through Congo. There- fore Luanda, though not satisªed with Kabila’s handling of his government, was unwilling to risk creation of a regime in Kinshasa that might renew its support for the Angolan rebels.118 Rumors of UNITA and Mobutu sympathiz- ers’ connections with the RCD and with Kigali fueled Angolan suspicions, eventually bringing about the decision to heed Kabila’s request and send in troops.119 That curbing UNITA’s activity was the primary factor in prompting Angola’s intervention on behalf of Kabila’s government is evident from An- gola’s quick withdrawal from Congo following UNITA’s defeat and the death of its leader, Jonas Savimbi. spillover of civil conflict: the kin-country syndrome. Like Burundi, Rwanda, and western Tanzania, eastern Congo is home to both Hutus and Tutsis. Yet unlike in Burundi and Rwanda, few tensions existed between these two groups until the 1990s. Rather, the Hutus and Tutsis in eastern Congo (again, known collectively as Banyarwanda) were resented by some ethnic groups, which considered them aliens (even though many of the Banyarwanda immigrated to this region more than 200 years ago).120 The Rwandan genocide and the ensuing refugee ºows had consequences for all of Central Africa. Hutu Zairians and so-called indigenous Zairians allied with the Rwandan Interahamwe against the Tutsi, including those native to Zaire as well as relative newcomers from Rwanda (known as Banyamulenge, a term that eventually came to describe all Tutsis in Zaire). In 1996 the

117. Clark, “Explaining Ugandan Intervention in Congo,” pp. 272–273, 278–281; and Shearer, “Af- rica’s Great War,” pp. 95–96. This is consistent with Fazal’s argument regarding the replacement of the practice of territorial conquest with one of regime change. See Fazal, “From Conquest to Intervention.” 118. Weinstein, “Africa’s ‘Scramble for Africa,’” p. 14; and Shearer, “Africa’s Great War,” p. 95. 119. Shearer, “Lines on a Map,” p. 295; and Manahl, “From Genocide to Regional War,” p. 6. On Kabila’s appeal to Angola to intervene, which explicitly promised to continue pursuing anti- UNITA policies, see François Misser and Alan Rake, “An African World War?” New African, Vol. 367 (October 1998), p. 14. 120. Emizet, “The Massacre of Refugees in Congo,” pp. 166–167; Lemarchand, “The Democratic Republic of Congo,” pp. 49–52; and Lemarchand, “Patterns of State Collapse and Reconstruction in Central Africa,” p. 6. International Security 31:3 168

Interahamwe and ex-FAR soldiers in the refugee camps attacked and killed thousands of Tutsis in North and South Kivu Provinces. Given that the Hutu and indigenous Zairian groups were supported by Mobutu’s troops and money, the Tutsi sought assistance from the Tutsi-ruled Rwandan government. In the words of Lemarchand, “The kin-country syndrome...asserted itself with a vengeance, driving Hutu and Tutsi, irrespective of other distinctions, to opposing camps.”121 Mobutu’s exclusion of the Banyamulenge was a primary cause of the rebellion against him and, because of the Banyamulenge’s kin relations to Rwanda, of its intervention in Congo.122 In October 1996, after being ordered to “return” to Rwanda within six days, the Banyamulenge looked to Kigali for military support. This situation only exacerbated the fears of the Tutsi in Rwanda and led to their massive attack on the refugee camps in eastern Zaire (and to the massive slaughter of Hutu refugees), which precipitated the rapid crumbling of Mobutu’s regime.123 Similarly, the DRC’s ethnic policies regarding the Banyamulenge were an important factor in propelling the 1998 Rwandan invasion. First, Kabila failed to fulªll his promise to Rwanda to bestow citizenship on the Banyamulenge. Second, in early 1998 he tried to boost his popularity in Kinshasa by ousting the Tutsi ministers in his government and issuing a de- cree that all foreign (i.e., Rwandan) troops leave the country. Third, his army then encouraged a massacre of the Banyamulenge (and anyone else with Tutsi- like physical features) in Kinshasa.124 In sum, spillover of the civil conºicts between Zaire and many of its neighbors played a primary role in causing both phases of the war in Congo. The spillover was a result of two factors: refugees and insurgencies, and neighboring states coming to assist their excluded and endangered coethnics across the border. Both factors were directly related to state weakness, and both would have been less likely had the norm of ªxed borders not existed. state weakness, opportunism, and greed in the congo war “Africa’s ‘scramble for Africa,’” Jeremy Weinstein argues, “is a primary cause of the rise of the interstate war on the continent.”125 To assess the validity of

121. Lemarchand, “Patterns of State Collapse and Reconstruction in Central Africa,” p. 6. 122. Herbst, States and Power in Africa, p. 238. 123. Emizet, “The Massacre of Refugees in Congo,” pp. 167–179; Lemarchand, “Patterns of State Collapse and Reconstruction in Central Africa,” pp. 6–7; Manahl, “From Genocide to Regional War,” pp. 2–4; and Weinstein, “Africa’s ‘Scramble for Africa,’” p. 13. 124. Lemarchand, “The Democratic Republic of Congo,” pp. 45–46; and Misser and Rake, “An Af- rican World War?” p. 15. 125. Weinstein, “Africa’s ‘Scramble for Africa,’” p. 18. For a similar explanation of the war, see Marc Lacey, “Congo Tires of War, but the End Is Not in Sight,” New York Times, July 15, 2002. When Good Fences Make Bad Neighbors 169

hypothesis 4, this section explores the exploitation of the collapsed Congolese state by its neighbors as a cause of the Congo war. Three states involved in the war—Namibia, Chad, and Zimbabwe—had few reasons to feel directly threatened by the situation in Congo. Yet all three sent troops to back Kabila’s regime, despite not sharing borders with Congo. Their motives for intervening, therefore, must have laid elsewhere. Namibia’s moti- vations are unclear. While the government claimed to be enraged by Rwanda’s and Uganda’s violation of DRC sovereignty, the political opposition accused the Namibian president of sending troops into Congo to defend his family’s mining interests there.126 The case of Chad is also ambiguous, but there are two possible explanations for its troop deployments. First, Chad supposedly received ªnancial assistance from Libyan President Muammar Gaddhaª, who was eager to exert his inºuence in sub-Saharan Africa. Second, Chad might have been interested in the gold mines of northern Congo.127 Zimbabwe, on the other hand, had three explicit reasons for intervening in the DRC, all having to do with opportunism. First, Kabila and his ADFL movement owed the Zimbabwean government large sums of money that had been lent to them during the rebellion against Mobutu. Zimbabwe feared that if Kabila’s regime were ousted, it would not get its money back.128 Second, although the state of Zimbabwe probably lost money from the Congolese adventure, the extension of patronage politics from Zimbabwe to Congo hugely beneªted some private interests. Zimbabwe’s president, Robert Mugabe, openly discussed the potential economic proªt to be reaped in the DRC as a chief reason for Zimbabwean involvement and, therefore, encouraged businesses to exploit the intervention. Having secured lucrative deals with Kabila’s government, both the army as an organization and his commanders as private businessmen were heavily involved in the mining and trade of diamonds, gold, and copper from parts of the DRC controlled by Zimbabwean troops. Business interests associated with the ruling Zimbabwe African National Union–Patriotic Front and cronies of Mugabe reaped their share of the proªts as well.129 The concentration of Zimbabwean troops around important mining towns in the DRC testiªed to these interests.130 Third, Mugabe hoped to exploit the Congolese conºict to promote his ambition of becoming a regional leader in southern Africa.131

126. Baker, “Going to War Democratically,” p. 274. 127. Ibid., p. 269; and Shearer, “Lines on a Map,” p. 295. 128. Misser and Rake, “An African World War?” p. 14; and Shearer, “Africa’s Great War,” p. 98. 129. Baker, “Going to War Democratically,” pp. 268, 274–275; and Shearer, “Africa’s Great War,” p. 98. 130. Weinstein, “Africa’s ‘Scramble for Africa,’” pp. 15–16. 131. Shearer, “Lines on a Map,” p. 295; and Misser and Rake, “An African World War?” p. 14. International Security 31:3 170

The backers of Kabila’s government, though, did not have a monopoly on the economic exploitation of Congo through military intervention. Rwanda and Uganda also had economic interests in Congo, although for them econom- ics seems to have played a secondary role to security. Sufªce it to say that since 1999 Rwanda and Uganda, like Zimbabwe, have been exporting diamonds, despite the absence of diamond mines on their territory.132 A 2001 UN report documents the great extent to which Rwandan and Ugandan military forces and private companies exploited the territory under their occupation. In addition, the document traces the relationship between these elements and the highest echelons in Kigali and Kampala.133 Although greed played an important role in the escalation of the war and the prolongation of the Rwandan and Ugandan presence in Congo, it was not the main reason for the two countries’ intervention. Their agreement to remove their forces from the DRC in 2002 supports this contention. Congo’s weakness clearly enabled this economic exploitation. And as a 2003 UN panel of experts noted, “In the absence of a strong, central, and democratically elected government that is in control of its territory, illegal exploitation will continue and serve as the motivation and the fuel for continued conºicts in the region.”134 Congo’s weakness also encouraged political predation and played a signiªcant role in facilitating the war. The likelihood of war (in both 1996 and 1998) increased because regime change was a real possibility and because all of the protagonists had political allies within Congo. The aim of Rwanda’s 1996 invasion, writes Lemarchand, was to “wrestle the Mobutist monster to the ground and make the whole of Congo safe for Rwanda.”135 Rwanda succeeded in removing Mobutu and almost succeeded in ousting Kabila precisely because Congo was an extremely weak state. Similarly, Angola’s apparent involvement in the assassination of Laurent Kabila could be viewed as an example of political opportunism—in this case, replacing a disliked leader of a weak state with a more favorable one.136 Of course, absent the security motiva- tions, these predatory actions would not have been necessary. But once a secu-

132. Baker, “Going to War Democratically,” pp. 275, 278. For more evidence of exploitation, see Clark, “Explaining Ugandan Intervention in Congo,” pp. 275–278; Weinstein, “Africa’s ‘Scramble for Africa,’” p. 17; and Turner, “The Kabilas’ Congo,” p. 217. 133. United Nations Security Council, “Report of the Panel of Experts on the Illegal Exploitation of Natural Resources and Other Forms of Wealth in the Democratic Republic of Congo,” S/2001/ 357 (New York: United Nations, April 12, 2001). 134. United Nations Security Council, “Report of the Panel of Experts on the Illegal Exploitation of Natural Resources and Other Forms of Wealth in the Democratic Republic of Congo,” S/2003/ 1027 (New York: United Nations, October 23, 2003). 135. Lemarchand, “The Democratic Republic of Congo,” p. 43. 136. Turner, “The Kabilas’ Congo,” pp. 216–217; and Lemarchand, “The Democratic Republic of Congo,” pp. 52–53. When Good Fences Make Bad Neighbors 171

Table 1. The Norm of Fixed Borders and Weak States: A Summary of Findings in the Congo Case

Findings in the Case of the Congo Hypotheses War

Hypothesis 1: Fixed borders can perpetuate or Confirmed, though indirect exacerbate the weakness of already weak states. Hypothesis 2: Weak states in a fixed-borders world Strongly confirmed can create conditions that can give rise to violent internal conflicts. Hypothesis 3: Refugee movements, insurgencies, and Strongly confirmed for both the kin connections across international borders can 1996–97 and 1998–2002 cause civil conflicts in weak states in a fixed-borders interventions world to spill over their borders and become international conflicts and possible full-fledged wars. Hypothesis 4: State weakness promotes the Confirmed for Zimbabwe’s possibility of international conflict because it creates intervention opportunities for neighbors to intervene to exploit Confirmed as an important reason the weak state economically and politically. for continuation of involvement of other actors, though not as a primary reason for going to war

rity threat exists, the weakness of the state provides relatively cheap opportunities for political predation as a way to offset it. Table 1 summarizes the ªndings in the case of Congo.

Conclusion

“The greatest threats to our security,” U.S. Secretary of State Condoleezza Rice remarked in a 2005 op-ed, “are deªned more by dynamics within weak states than by the borders between strong and aggressive ones.”137 This article offers an explanation for this phenomenon. In regions where states are sociopolitically weak, the norm of ªxed international borders can increase the likelihood of international conºict. In such situations, ªxed borders perpetuate and exacerbate the weakness of the state, which, in turn, is a major cause of internal wars. These internal wars can then spill over and become international conºicts. The

137. Condoleezza Rice, “The Promise of Democratic Peace: Why Promoting Freedom Is the Only Realistic Path to Security,” Washington Post, December 11, 2005. Rice’s remarks were in line with the 2002 U.S. National Security Strategy. See President George W. Bush, National Security Strategy of the United States of America (Washington, D.C.: White House, 2002). International Security 31:3 172

Congo war illustrates what can happen when the norm of ªxed border exacerbates state weakness. The norm of ªxed borders creates constraints and incentives that compel weak states to remain weak or that make them even weaker. With no territorial threats and their juridical statehood assured by the international community, rulers are not pressured to assume the risks and to pay the costs that are involved in the process of state building. And in contrast to the experience of states prior to World War II, failed or collapsed states are no longer weeded out by the system. Scholars widely agree that state weakness can lead to internal strife and civil conºicts. What is less often acknowledged is that such weakness, especially when borders are unchangeable, can also precipitate international wars. This article explored two paths leading from state weakness to international conºict. In the ªrst, weak states create conditions that are rife for internal conºict by giving way to emerging anarchy and by allowing leaders to exploit ethnic divisions to compensate for their own lack of legitimacy. Civil conºicts in weak states under the norm of ªxed borders, in turn, are more likely to spill over and involve neighboring states than are other kinds of civil conºicts because of both the inability of the state to prevent insurgencies and the existence of the kin-country syndrome. In the second path, the weakness of the state creates opportunities for its neighbors to engage in political or economic (though not territorial) predation. The case study of Congo largely conªrms these arguments. The behavior of Congo’s leaders suggests that the norm of ªxed borders was a signiªcant factor in the weakening of the Congolese state. Congo’s growing weakness, in turn, was a major cause of its internal wars, especially during the 1990s. The spillover of these wars (as well as the civil conºicts in neighboring weak states) contributed signiªcantly to the outbreak of war in the Great Lakes region in 1996 and again in 1998. Greed was also an important factor: it spurred Zimbabwe’s decision to go to war, and it contributed to other actors’ pro- longed involvement in the conºict. Yet it was not Rwanda’s, Burundi’s, Uganda’s, or Angola’s primary motivation for going to war. Caution is needed when attempting to draw general conclusions from a single case study. A carefully designed study that compares the Congo war with other cases, past and present, is needed to conªrm the conclusions presented here.138 A study that corroborates my hypotheses on the relationship between the norm of ªxed borders and the likelihood of international conºict would

138. See, for example, Boaz Atzili, “Border Fixity: When Good Fences Make Bad Neighbors,” Ph.D. dissertation, Massachusetts Institute of Technology, 2006. When Good Fences Make Bad Neighbors 173

represent an important step in understanding conºict in regions where sociopolitically weak states predominate, as is the case in most of the developing world. That an international norm designed to enhance stability and peace might actually produce the opposite effect is a subject worthy of scholarly attention. That good fences can make bad neighbors may be bad news, but it is news that we might well have to reckon with.