The Myth of “Ethnic Conflict”

Beverly Crawford and Ronnie D. Lipschutz, (editors)

International and Area Studies Research Series/Number 98

University of California, Berkeley

1998

 

1. The Causes of Cultural Conflict: An Institutional Approach

As the cold war came to a close, the Soviet Union collapsed, and the rhetoric of economic and political liberalization swept the globe, there were more than thirty violent conflicts raging around the world, most of these ethnic and sectarian in nature. ₁ Indeed these conflicts are not new. Some estimates suggest that cultural conflicts have inspired over half of the violent struggles within states between 1945 and 1960. The proportion increased to three quarters from 1960 to 1990. ₂ What marks these conflicts, separates their effects from those of interstate wars, and thus makes them an important subject for investigation on their own is that they have been significantly more devastating to civilians. In World War I, 14 percent of all deaths were civilian. That figure rose to 67 percent in World War II. And in the 1990s, where most wars were within rather than between states, civilian deaths totaled 90 percent of all deaths. ₃ By 1995 deaths in the war in the former Yugoslavia reached over 200,000; over half the population of Bosnia became refugees, and virtually all of the Serb population of Croatia was forced to flee. ₄ By 1993 civilian deaths in the war in Abkhazia were estimated at between 25,000 and 30,000. More than half of the prewar population became refugees. ₅ Five thousand people have been killed in Kashmir since 1990; over 30,000 have died in Algeria, and 18,000 have died in Punjab since the storming of the temple in 1985. While war between states seems to be on the decline, ethnic and sectarian conflict within them is on the rise. Why?

The causes of these conflicts are not immediately obvious. Some minority ethnic and religious groups assimilate and are eventually accommodated in a unitary nation-state. Bulgarian Muslims (Pomaks), unlike Muslims in Bosnia, are peacefully integrated. Malaysia in recent years has maintained interethnic peace under a moderate unitary Muslim majority state, while secular Egypt is coping with rising violence perpetrated by more radical factions of “political Islam.” Among liberal democracies, England has achieved relative political integration of minority ethnic groups, while Germany eschews such integration and has experienced relatively high levels of violence in the form of hate crimes against non-German immigrants. Other Western democracies, such as Switzerland and Belgium, maintain federations that separate ethnic and religious groups into political entities with limited autonomy that peacefully coexist with one another. And the breakup of Czechoslovakia demonstrates that secession does not have to be violent. Among our cases to be considered here, Ajaria has peacefully insulated itself from the Georgian turmoil and gained a separate and more autonomous constitutional status within Georgia, while neighboring Abkhazia was ravaged by war. Why is it, then, that some ethnic and religious problems are resolved peacefully, others remain unresolved but do not erupt in violence, and still others seek resolution in violent conflict?

We argue here that the current round of ethnic and sectarian violence is ironically linked to the apparent triumph of economic globalization and institutional transformation—the opening of new markets for goods, services, capital, and people; the construction of new democracies; and the implementation of “state-shrinking” ideologies that have swept the globe. While, with some important exceptions, developed market economies have experienced relatively low levels of cultural conflict, they have experienced that conflict nonetheless, as they have begun the state-shrinking process of economic liberalization. And transition to the market and the pressures of globalization—increased demands for industrial competitiveness and rising external debt that weakens the state’s capability and willingness to allocate resources—are associated with high levels of conflict and even violence.

While many analysts suspect that there is a link between economic globalization and the current round of cultural conflict, few have investigated potential causal forces that might explain that relationship. ₆ In this study we assess alternative explanations for cultural conflict and attempt to discover the causal mechanisms that might explain its relationship to economic globalization and liberalization. Taken together, the essays in this volume argue that cultural violence erupts most vociferously where secular economic decline, neoliberal economic reforms, and institutional transformation have broken old “social contracts”—that is, where they have broken the rules and norms by which access to political and economic resources was once granted. Globalization and liberalization are thus “triggers” for cultural conflict, but they are not the only triggers, and they are not the underlying causes.

The breaking of old social contracts—by whatever means— leads to shifts in political power. When these power shifts are experienced as ethnic and religious discrimination and privilege, the resulting resentment and opportunity provide fertile ground for modern political entrepreneurs to mobilize support around ethnic and sectarian identities. ₇ But changes in power do not always result in cultural discrimination. Often, for example, they result in economic oppression or discrimination along ideological lines. When, then, do power shifts result in cultural threat and conflict? We argue that cultural conflict erupts most frequently in those places where old social contracts permitted ethnic and religious criteria to guide the allocation of political and economic resources. That mode of resource allocation permitted the logic of identity politics to characterize and sometimes even dominate political competition. Where identity politics once prevailed and when institutions upholding the old social contract are weakened, the odds of cultural conflict and even violence increase.

Why would the practice of identity politics increase the odds of cultural violence? The logic of identity politics suggests that claims on resources based on ascriptive criteria like ethnicity and religion are often incompatible. Conflicting claims on resources based on cultural criteria are more prone to intense conflict than disputes between interest groups. This is because while interests are malleable and multiple, making compromises and log-rolling possible, cultural identity is fixed and non-negotiable. Disputes over resources among “identity groups” will thus prove to be particularly difficult to negotiate, raising the odds of violence. ₈ But even when the logic of identity politics dominates the political game, opposing cultural groups can commit themselves to pacts that ensure social order, and the institutions of central authority can help enforce those pacts. Even when one cultural group dominates others, repressive institutions can provide for “order,” although it is discriminatory and experienced by less powerful groups as unjust. Cultural conflict escalates into violence when these institutions are weakened, disrupted, or transformed in ways that undermine the commitment to uphold these contracts or repress dissent. Although there are many forces which undermine that commitment, globalization and economic liberalization are two of the most important current culprits. When the forces of globalization and the impersonal market have usurped control from domestic institutions and when those institutions can no longer credibly enforce agreements that ensured peaceful competition among politically relevant cultural groups, or when they can no longer enforce culturally discriminating policies, violence may be the only alternative course for political entrepreneurs making non-negotiable resource demands on behalf of distinct cultural groups.

The essays in this volume elaborate on, illustrate, and sometimes even “test” this argument. In this introduction, I begin with a brief discussion of our puzzle: why cultural violence appears to be everywhere on the rise and why there are significant variations in kind and levels of conflict despite historical, regional, and demographic similarities. I then turn to an examination of three alternative explanations for cultural conflict. I argue that these explanations are incomplete and suggest how an institutional approach can account for more variation in cultural conflict and violence. In the third section, I lay out the institutional argument in three steps. I begin by describing the process by which cultural identity does or does not become politically relevant and the importance of institutions in shaping and legitimating political identity. I then discuss the role of institutional disruption, weakening, collapse, and transformation in undermining and even breaking social contracts in ways that permit the rise of political entrepreneurs who wish to mobilize political support with cultural appeals; I further discuss the conditions under which those mobilization efforts can be successful and can lead to violence. Finally, I lay out the causes for institutional change and broken social contracts located in the processes of globalization and economic liberalization.

Purpose: Explaining the Rise and Extent of Cultural Conflict

Our purpose in this study is twofold: to generate hypotheses that will explain why ethnic and sectarian violence appears to be everywhere on the rise (particularly in the current period), and why some states experience high levels of it while others in similar circumstances experience little or none. We have taken an intellectual risk in our effort to engage in large comparisons and come to general conclusions. As a rule, those who believe in the uniqueness of each case of conflict enjoy a certain intellectual advantage over those who seek to argue that patterns of variation appear across time and space. The case for uniqueness can be made simply by enriching each story with so much historical detail that generalizations and comparisons seem artificial, oversimplified, and silly. Indeed as the stories in the coming chapters will reveal, no two cases of cultural conflict are really alike. What we attempt here, however, is to abstract from the historical and descriptive complexity of each case and challenge ideographic accounts by proposing a conceptualization of the causes of cultural conflict embodied in institutional arrangements. ₉

We base our conclusions on twelve cases that vary according to region; we compare conflicts in Western Europe, the former Communist world (the former Soviet Union and the Balkans), and the developing world (Muslim-majority states and India). While ethnic and sectarian conflict is lower in the West than in post-Communist countries, and conflict is higher in post-Communist regions than in Muslim-majority states or in less developed but relatively consolidated democracies, levels of conflict also vary significantly within developing, post-Communist, and Western industrial states. Algeria and Egypt have experienced high levels of sectarian conflict while Malaysia has not; the former Yugoslavia and Abkhazia experienced protracted and bloody separatist conflicts, while Bulgaria and Ajaria remained at peace. In contrast to England, Germany has experienced relatively high levels of cultural violence. Because the role and the strength of political institutions vary widely across regions and because there are large variations in conflict levels within regions, our primary comparisons are intraregional rather than cross-regional.

Table 1

Measures of Conflict by Country/Region

Country/Region	Level of Violence	Organization	Protraction
Yugoslavia	High: war of succession	Organized military units	1991–95; prolonged
Abkhazia	High: war of succession	Organized military units	1991–93; prolonged
Algeria	High: state-threatening civil war	Organized military units	1991–present; prolonged
Punjab	Medium: nonviolent political competition, violent protests, military repression; state reduces violence, separatism	Political parties, militias, Indian military	Chronic since 1975
Kashmir	Medium: secessionist violence, government repression; state reduces violence	Political parties, secessionist movement, Indian army	Chronic since 1974
Egypt	Medium: sporadic violence and state repression	Organized Islamic groups/state police	Chronic since 1974
Germany	Low: rising right-wing violence in 1980s, increased sporadic violence 1991–93, state repression; decline in incidents of violence since 1993	Relative lack of organized violence	Sporadic dramatic increases 1991–92
United States	Low: state repression of racial violence; recent violence directed against central government	Relative lack of organized violence	Sporadic
England	Low and declining: riots, 1958, 1980–81, 1985, mischaracterized as race riots	Unorganized violent attacks	Sporadic
Bulgaria	No ethnic or sectarian violence	———	None
Malaysia	No ethnic or sectarian violence	———	None since 1969
Ajaria	No sectarian violence	———	None

Sources: The data for this table were drawn from several sources. These are cited here, along with numbers of fatalities resulting from ethnic and sectarian violence, where possible. The scoring of high, medium, and low is a judgment of relative intensity of violence in comparison with the other cases. By 21 November 1995, 250,000 people had been reported dead or missing in the former Yugoslavia. Other figures for the former Yugoslavia are listed in note 4. For Abkhazia, see note 5. For Algeria there are various estimates for deaths ranging from 15,000 (John P Entelis, "Political Islam in Algeria: The Nonviolent Dimension," Current History 94, 588 [January 1995]: 13) to 40,000 (Economist, 8 July 1995). For the Punjab estimates of fatalities related to sectarian conflict range from 18,000 (Gurharpal Singh, "Punjab Elections 1992: Breakthrough or Breakdown?" Asian Survey 32, 11 [November 1992]: 988) to 20,000 (Hamish Telford, "The Political Economy of the Punjab: Creating Space for Sikh Militancy," Asian Survey 32, 11 [November 1992]: 969). For Kashmir, see Binder. For Egypt there have been an estimated 1,000 deaths from violent attacks by extremist Islamic groups (Sarah Gauch, "Terror on the Nile," Africa Report 38, 3 [May-June 1993]: 32). In Germany from 1991 to 1993, there were a total of 5,881 attacks against "foreigners" (see Jens Alber, "Towards Explaining Anti-Foreign Violence in Germany" [Cambridge, Mass.: Center for European Studies, Harvard University], Working Paper Series No. 53, table 1; cited in John Leslie's article in this volume). England has experienced sporadic social violence in recent years, which has been mischaracterized as "ethnic conflict." Elaine Thomas describes here both the violent incidents and their misrepresentation. For Bulgaria the only recorded deaths related to ethnic conflict occurred in riots in May 1989 in response to Todor Zhivkov's attempt to "Bulgarize" the country's ethnic Turks. This riot resulted in 600 deaths (Roland Flamini, "A Modern Balkan Exodus," Time 134, 7 [14 August 1989]: 39). There has been no cultural violence involving Pomak groups. In Malaysia the only recorded deaths from ethnic conflict occurred in a riot in May 1969, resulting in 200 deaths (see Leon Comber, 13 May 1969: A Historical Survey of Sino-Malay Relations [Kuala Lumpur: Heinemann Asia, 1983], p. 88, and Karl von Vorys, Democracy without Consensus [Princeton: Princeton University Press, 1975]). Georgi Derluguian in this volume argues that Ajaria experienced very little cultural conflict in the wake of Soviet collapse. The United States has experienced sporadic and intense periods of cultural violence, primarily directed against African Americans. Recent violence, discussed in this volume by Ronnie Lipschutz, has been directed against the federal government.

Cross-regional comparisons can be suggestive, however. Our cases vary according to ethnic and sectarian demographic composition. While the former Yugoslavia and the former Soviet Union, Bulgaria, Malaysia, the United States, and India can be considered multiethnic or multicultural states, England and Germany are relatively homogeneous; Egypt and Algeria are Muslim-majority states with small minority populations. Despite these demographic variations, however, ethnic and sectarian violence appears to have erupted in all of these in recent times; nonetheless, there are significant variations in intensity among these conflicts that do not correlate with levels of cultural diversity. And some social conflicts have been widely characterized as “ethnic and sectarian conflict” when they were not.

In measuring conflict levels, our scoring is largely judgmental and our rankings nominal rather than ordinal. We distinguish violent conflict from the legitimate and rule-based ethnic and sectarian conflict of identity politics. Where violent cultural conflict does occur, we measure its intensity by looking at whether it is prolonged, chronic, or sporadic, and by assessing whether it is organized or spontaneous. Higher violence is associated with separatist or civil wars threatening a central state; lower violence is associated with outcomes in which state elites repress uprisings and negotiate bargains with ethnic and sectarian groups. Variations in the conflicts are summarized in Table 1.

Alternative Explanations

Primordialism

How can these variations in cultural conflict best be explained? The literature on ethnic and sectarian conflict and its absence has offered three explanatory approaches. The first invokes the centuries of “accumulated hatreds” between “nations” with primordial origins. ₁₀ This essentialist or primordialist approach asserts that “the urge to define and reject the other goes back to our remotest human ancestors, and indeed beyond them to our animal predecessors,” ₁₁ and suggests that when the grip of central control is relaxed, “people reflexively grasp at ethnic or national identifications or what passes for them.” ₁₂ The argument suggests that the illiberal politics of identity, with its claims of collective exclusivity, and tendencies toward xenophobia and intolerance are more natural to human societies than liberal politics of interest, which hold that individuals hold multiple and cross-cutting identities and interests that can be aggregated and represented in the political arena.

But current peaceful relations, for example, between the Germans and French in Alsace-Lorraine; between whites, Africans, and African ethnic groups in South Africa; between Pomaks, Turks, and Bulgarians in Bulgaria; or between the Chinese and the Malay in Malaysia suggest that even if accumulated hatreds once fanned the flames of violent conflict, they can be attenuated by alternative memories, more current experience, and institutional incentives. Primordial explanations that call on “centuries of accumulated hatreds” cannot account for situations in which ethnic groups coexist peacefully. They cannot account for differences in the political expressions of cultural difference—i.e., separatist movements, efforts to control the state, or attempts at power-sharing with other cultural groups. They are thus guilty of selection bias.

Primordial explanations are flawed in two further respects. First, they fail to make the distinction between cultural identity and politically relevant cultural identity. They assume that cultural differences, such as language, religion, cultural traditions, and ethnicity, automatically lead to conflict because they assume that culturally defined groups are by nature exclusionary and are dominated by parochial values that outweigh universalistic norms. According to primordial accounts, parochial norms attributed to cultural groups are believed to isolate them and lead to extremism. ₁₃ Extremism raises the odds of violence.

But cultural differences do not necessarily trigger the extremism of cultural conflict. Bulgarian Muslims possess different cultural characteristics from Bulgarian Christians, and Ajaris see themselves as culturally different from Georgians. But Pomaks and Bulgarians, Ajaris and Georgians have managed to live peacefully together. Indeed there is usually more conflict within a culturally defined group than between different groups. ₁₄ What primordialists neglect is that cultural identities lead to conflict only when they have become politically charged. And cultural identity is politicized only when it becomes a criterion for discrimination and privilege in struggles over the distribution of political and economic resources, rights, and protection. Our findings thus support the claim that the political relevance of cultural identity is socially constructed. ₁₅

Second, primordial explanations ignore the role that the institutions of the state play in easing, perpetuating, or triggering cultural conflict by structuring incentives in ways that either exacerbate or attenuate the political relevance of cultural identity. We find that these institutional incentives, embedded in citizenship laws, rules of accountability, participation, and distribution, can structure political struggle in ways that either mute or encourage ethnic and sectarian political conflict. In Malaysia, for example, cultural identities have become politically relevant, but the institutions of state and economy have channeled identity politics into peaceful political competition. In Bulgaria, Pomak cultural identity was never politicized by state institutions and never became politically relevant. In England universal membership in the political community combined with supportive liberal institutions and the structure of party politics to depoliticize ethnic identity, thus blocking the foothold needed by ethnic entrepreneurs for significant participation in British politics. These examples suggest that the state’s institutional structure and strength should be placed at the center of any explanation of cultural conflict.

By tracing the ways in which cultural identity is transformed into political identity and showing how politically relevant ethnic and sectarian identities were legitimated and in some cases created by political institutions of recent origin, we will show that accumulated hatreds do not have primordial origins and they do not necessarily lead to violent conflict. Indeed we would be naive not to recognize that hatreds do accumulate and collective memories of victimization can lead populations to respond to politicians who draw on the reserve of those memories to foment hatred anew in their efforts to mobilize culturally defined groups for political action and even violence. But we will show that in all cases, politically charged divisions fueled by collective memory are more likely to become violent when institutions that would shape peaceful resolutions to that struggle are weak. ₁₆ Weak institutions permit political entrepreneurs to exploit those divisions in ways that lead to violence. We thus challenge the assumption that political identities are “given” or primordial. Indeed such a characterization of those identities simply perpetuates the myths that fan the flames of ethnic conflict and make social integration seem impossible.

Security Dilemmas

An alternative explanation for violent cultural conflict, drawing on the concept of the security dilemma in international relations theory, addresses this second deficiency in that it recognizes the importance of institutions in creating social order. ₁₇ This approach implicitly suggests that where institutions of central authority do not exist and where anarchy prevails, groups tend to seek security above all other goals. Their quest for security leads them to take measures that render other groups insecure, and those groups in turn take measures that threaten others. Thus a vicious cycle of escalating threats takes hold in the absence of a central authority that could reduce those threats. Institutional collapse or decline leads groups to seek an offensive advantage—that is, to grasp the resources that assure their security before these resources are seized by an opposing group. Thus the odds of violence are high when the security dilemma prevails.

Although this approach may indeed be useful in explaining the proximate causes of violence as a result of the actors’ strategic interaction, it is incomplete in four important ways. First, in equating communal conflict with interstate conflict, it takes the formation of politically relevant cultural groups as given; preferences are assumed and not explained. Like primordial explanations, accounts of ethnic conflict from this approach assume that cultural groups, like states, have conflicting and incompatible political interests, and thus they will automatically clash. These accounts are primarily concerned with how that clash might become violent. They ignore the evidence that different cultural groups do not necessarily have conflicting political interests, and they do not explore the ways in which the logic of identity politics diverges from the logic of political competition based on conflicts of interest under international anarchy.

Indeed the metaphor of the security dilemma in international politics is misplaced. The political interests of culturally defined groups, unlike the interests of states in the international system, are not necessarily divergent. A complete account of cultural conflict and its escalation must assess the political transformation of these cultural groups. It must explain why politicized social divisions did not take the form of interest group competition, ideological movements, or class-based organizations. It must also trace the process by which cultural differences led to identity politics and explain why some cultural divisions never became politically charged. Only then can it explain why cultural interests might clash in the political arena.

Second, the security dilemma metaphor makes the unwarranted assumption that the function of central authority is simply to mitigate and prevent conflict. Indeed the institutions of central authority can provide the necessary transparency and information to relieve the insecurity of competing groups. ₁₈ They also work to create a political community that cements the expectation that all groups will interact indefinitely, thus enhancing the preference for cooperation among them. And they can repress social violence. But this rather narrow focus on the function of institutional strength in mitigating violence disguises the role that the institutions of central authority play in creating and legitimating cultural conflicts or in attenuating their intensity. They play this role by defining not only the constraints, but also the incentives facing social and political elites. By providing elites with both constraints and incentives, institutions shape their interests and objectives and determine which of many social divisions will become politically relevant. As Levy et al. argue, institutions “shape the logic of the political game.” ₁₉ The institutions of central authority determine whether the logic of interest-based politics, ideology-driven politics, class conflict, or identity politics will dominate political competition. Institutions do not treat all forms of conflict impartially because their function is to channel conflict, not simply to mitigate and prevent it. ₂₀ Because of this channeling function, institutions not only constrain behavior, but also they shape political preferences and identities. In short, institutions can construct policies of discrimination and privilege that politicize ethnicity and religion. Or they can construct rules that prohibit cultural discrimination but provide for ideological, class, or interest-based competition.

Third, the metaphor of the security dilemma fails to explain cultural conflict and the outbreak of cultural violence in industrial societies where central authority is relatively strong and social contracts are largely considered legitimate. Accounts that use the security dilemma metaphor to explain ethnic and sectarian conflict fail to explain the outbreak of violence where security dilemmas do not exist. Finally, the approach fails to explain why institutions weaken and sometimes collapse. A complete account of cultural conflict must take the causal arrow one step back; it must account for the causes of institutional change and the process by which change triggers cultural violence as opposed to other kinds of conflict.

The Panacea of Markets and “Democracy”

Despite our refutation of essentialist explanations and our empirical support for institutional arguments, our evidence challenges the currently fashionable claim that the rapid and simultaneous construction of liberal economic and democratic political institutions can mitigate ethnic and sectarian conflict. ₂₁ Free markets create wealth for all, the argument runs, erasing the need for violent struggle over resources. And democracy permits political aggregation and representation of all social interests, elevating conflicts of interest that can be adjudicated in the political arena over conflicts of identity that are more difficult to negotiate. The logic of liberal democracy suggests that the construction of democratic institutions makes the individual rather than collectivities the subject of legal protection and political participation. Democratic theory claims that if ethnic and religious conflicts do exist, they can be peacefully resolved if the organizing principles of the political system elevate tolerance and national unity above ethnic and religious domination and privilege. Furthermore, it claims that federalism, confederalism, and other forms of territorial decentralization that devolve political power to the local level create local and responsive government that will maximize individual freedom and satisfy the claims of some groups for autonomy and self-determination. ₂₂ In short, the classical liberal argument claims that the construction of markets and democracy and the decentralization of political and economic power ensures that individuals receive equal protection under the law and that economic and political competition need no longer be violent.

Despite widespread acceptance of these claims, however, the evidence suggests that perceived economic inequities, particularly those arising from current policies of economic liberalization and the longer-term effects of globalization, can undermine liberal political practices and lead to the illiberal politics that characterize ethnic and sectarian conflict. ₂₃ Indeed, the case of Germany shows that a well-established liberal capitalist democracy is not immune to ethnic violence and that identity politics are not always easily dissolved within pluralist polities.

We will show that where communal differences had become politically relevant in the past, the ethnic or religious card may be the easiest one to play in the effort to mobilize political support in the face of the uncertainties of economic decline, in the shift from welfare to market economies, and in the move from centralized to decentralized polities. This is particularly evident where both political and economic decentralization threaten to break down established community and the liberal focus on individual self-reliance threatens historical bonds and leads to deep insecurities. Secular economic decline and policies of economic liberalization require the dismantling of institutions of state resource allocation; weakened states are unable to provide equal protection for all who live within their territory.

Finally, the establishment of democracy is not a panacea. Liberal democracies can indeed mute cultural conflict with institutions of inclusiveness, universal representation, and electoral systems designed to encourage elite compromise. Indeed a robust liberal democracy may be one of the strongest defenses against cultural conflict. However, democracies are not all liberal. Illiberal democracies may possess many of the attributes of polyarchy, like free elections, freedom of speech, freedom of movement, freedom of association, and freedom of religion. But they pay only lip service to the rule of law, minority and citizen rights, and independent judicial review. ₂₄ Such systems can actually exacerbate cultural conflict. In periods of economic uncertainty and political transition, when states that once provided entitlements pull back or are dismantled, when illiberal democracies are so constructed that they fail to protect rights, and when the introduction of markets leads to deep insecurities, the rich symbolic resources of ethnicity and religion offer hope in their promise of collective power to those populations who feel powerless under these conditions.

The Argument

If primordial explanations are flawed, security dilemma metaphors limited, and liberalization panaceas utopian, then what explains variation in conflict and violence? As noted above, our explanation focuses on political institutions. We follow Douglas North’s rather broad definition:

Institutions are a set of rules, compliance procedures, and moral and ethical behavioral norms embedded in those rules and compliance procedures designed to constrain the behavior of individuals in the interests of maximizing wealth, social order, and the well-being of a society. Institutions establish the cooperative and competitive incentives in society by virtue of their norms, rules, and procedures. ₂₅

In short, institutions both constrain behavior and provide incentives for cooperation and compliance in norms, rules, and procedures for allocation, participation, representation, and accountability. As such, institutions embody the social contract between state and society. We will suggest that the various incentives for cooperation and competion that political institutions establish will hold the key to an explanation of cultural conflict and cooperation.

I begin with a discussion of how cultural identities are transformed into political identities and how politicized cultural identities are legitimated or attenuated by state institutions. I suspect that there will be higher levels of cultural violence in those areas where culture was historically politicized than in those areas where culture did not become politically relevant and other social divisions gave rise to political competition. This is not, however, an argument that all cultural conflict can be blamed on history. The institutions of the modern state are crucial in either cementing, creating, or attenuating cultural or identity politics that were created in historical power struggles. I then turn to the consequences of institutional transformation and propose that the odds of violence increase in those countries where ethnicity and religion were historically most politicized and where old institutions that perpetuated the political relevance of cultural identity collapsed.

Finally, I discuss the causes of those institutional transformations that can increase or decrease cultural conflict in the current period: globalization and economic liberalization. To the extent that economic liberalization and global integration bring economic growth and to the extent that old institutions remain strong and new institutions are resilient in creating incentives for cultural conflict in the face of economic and political pressures that might weaken or transform them, violence is attenuated and political conflict is constrained by the rules of peaceful political competition or the state’s ability to repress dissent. Violent communal conflict will be more intense in those areas where ethnicity and religion became politically relevant in the past and where the institutions of central authority are now weakened, are under pressure to change, or have simply reconstituted their constitutive and allocative rules and procedures under the pressures of economic liberalization and global integration. The structure of state institutions and their ability to establish and maintain a legitimate social contract is thus key to the prevention and attenuation of violent ethnic conflict; economic liberalization and globalization can threaten to change the terms of the social contract within these institutions, thus creating political “space” for the appeals of ethnic and sectarian political entrepreneurs and their offers of new social contracts.

Transformation and Political Legitimation of Cultural Identity

There are many ways to define social interests and identify divisions among people; not all of them become politically relevant. In most Latin American countries, for example, institutional transformation that accompanied industrialization resulted in the political dominance of divisions between capital and labor over other social divisions that could potentially be politicized—e.g., ethnic, religious, or urban-rural splits. ₂₆ Why and how do cultural divisions become more politically relevant than other social divisions?

One of the most important causes of cultural identity transformation can be found in the historical policies of colonial divide-and-rule that separated subjugated populations along ethnic and sectarian lines. The constraints and incentives offered by these policies created the opportunity for political entrepreneurs among colonized groups to draw on cultural identities to mobilize resistance to imperial control, gain access to political power and territory, and exercise power in the construction of new national institutions when colonial power collapsed: opportunities for drawing on other identities or interests were reduced by previous colonial policies. Although they differ substantially in most respects, the millet system of Ottoman rule and British colonial policies in India discussed below in this volume provide examples.

Not only external colonial domination but internal political domination and discrimination codified in political institutions—like apartheid in South Africa and the institution of slavery and Jim Crow laws in the United States—created similar cultural cleavages and led to similar historical struggles. Indeed with the founding of the United States came institutional biases in favor of people of Caucasian stock; the Naturalization Act of 1790 permitted any white immigrant to become a citizen after two years’ residence. The political struggles in the United States throughout the nineteenth century between the federal government and the states reflect struggles between those who would politicize cultural identity and those liberalizing forces who wanted to make other social divisions politically relevant. That debate politicized race in the United States.These and similar historical struggles in other regions that transformed cultural identity into political identity always required political entrepreneurs, individual leaders, and elites to interpret discrimination, oppression, and privilege in ways that made cultural identity politically relevant to their targeted constituencies. Their interpretations were often shaped by specific institutional incentives. The codification of racial segregation in the infamous Plessy v. Ferguson decision of 1896 led to systematic discrimination against blacks in voting, housing, transportation, and education. Black nationalist leaders pointed to that discrimination in their efforts to mobilize blacks in support for nonassimilation, arguing that systematic discrimination had denied them full citizenship rights. In 1954, however, the institutional basis for segregation was abolished, and the U.S. Supreme Court ordered integration “with all deliberate speed.” In the 1960s, however, a new generation of black political entrepreneurs drew on historical grievances similar to those of their predecessors’ and used evidence of discrimination to argue not against assimilation, but rather for the dismantling of discriminatory practices that prevented assimilation. Indeed political entrepreneurs will interpret cultural grievances and shape their particular political agenda in a number of ways, according to the institutional constraints and incentives that face them.

In particular, when arguing for a “state of their own,” political entrepreneurs have also attempted to mobilize support by instilling culturally defined groups with a mythical and heroic past, a sense of mission and messianism, or a belief that the group had intrinsic and unique rights to territory by virtue of its ethnic or religious identity. For example, as Ernst Haas has noted, the southern slaveholding elite in the United States, whose members controlled virtually all political offices in the South, believed that the white master race was a divinely sanctioned oligarchy. ₂₇

With arguments similar to these, many political entrepreneurs throughout the world have mobilized culturally defined groups for political action. Some of these arguments, like the one above, are ludicrous; other interpretations of discrimination and privilege are justified. The point here is not the truth of the interpretation, but rather that the most widely accepted interpretation of discrimination has almost always depended on the political acumen of the chief interpreters—the successful political entrepreneurs; in turn, their interpretations are shaped by institutional incentives. And these interpretations are the catalysts, transforming cultural identity into political identity.

Finally, the institutions of the central state, constructed in the critical historical period of nation-building, determined whether or not politicized cultural identity would be cemented in social and political practice and whether culturally defined groups would seek autonomy, separatism, or the right to participate with others in the political arena. Sometimes, as in parts of the former Soviet Union, these institutions created new cultural divisions that were not previously in place. At other times, as in postwar federal Yugoslavia, new institutions reinforced cultural divisions created by the historical interpretations of successful political entrepreneurs. Always, however, the institutions of the state set the terms of the social contract. The institutions that embodied the social contract structured the terms of membership in the political community, the rules of political participation and accountability, the relationship between a state’s central government and its various regions, and criteria for the production and distribution of material resources. ₂₈ In some cases, the social contract allocated resources according to previously established cultural criteria; in other cases, it eschewed those criteria in favor of merit, class, regional or territorial distinctions, or other criteria.

Where state institutions structure political membership and resource distribution according to ascriptive criteria, rewarding and punishing particular ethnic or religious groups, politicized cultural divisions become legitimate in the political arena, thus intensifying their political relevance. This means that the preferential political institutions themselves can have the effect of intensifying and even actively creating political groups that legitimate identity-based political struggles and the allocation of benefits. Preferential policies can then generate a backlash on the part of those groups excluded from benefits, intensifying the militancy of the beneficiaries and reinforcing the importance of ascription as the principle of choice in allocating benefits. ₂₉ Alternatively politically relevant cultural identity may be attenuated by state institutions whose rules of allocation, participation, and membership do not recognize cultural difference as politically relevant. These rules are considered to be one of the central hallmarks of the liberal state. Secular states, for example, weaken the political relevance of religious differences. Universal suffrage and citizenship rights weaken the political relevance of race and gender divisions.

The example of citizenship rights illustrates how membership rules can attenuate or exaggerate the political relevance of cultural identity. Citizenship rules can be either inclusive or exclusive. Inclusive citizenship rules make individual civic behavior the criteria for membership, regardless of the individual’s cultural attributes. Exclusive membership rules, in contrast, restrict membership in the political community to those people who are of a specified cultural origin, speak a specific language or practice a certain religion. Inclusive citizenship criteria weaken the political relevance of cultural difference; exclusive membership criteria strengthen it.

Membership in the political community, however, is not the only determinant or “cementer” of the content of political identity. Countries with inclusive citizenship laws need supportive institutions of participation and allocation to ensure that cultural identity will not become politically relevant or that its relevance will be weakened. They need a police force to protect some cultural groups against others who would perpetrate hate crimes. They need a political system that represents the interests of all citizens. They need a system of justice that bolsters equal citizenship rights. And they need allocative institutions to address problems of discrimination and to ensure that equality of citizens is a fact and not simply an empty right. Without a set of supportive institutions, formal laws granting inclusive membership in the political community cannot prevent the politicization of cultural identity by those who practice racist and discriminatory acts. As the British case described here will show, it is not specific or isolated institutions, but rather critical institutional “clusters” that determine whether or not cultural identity will be relevant in the political arena and how identity politics is practiced.

Countries whose institutions make cultural identity politically relevant also need supportive institutions to cement that culturally defined political identity and structure rules in which political competition among cultural groups is considered fair by all competitors if they wish to maintain social stability. Systems of proportional representation, for example, are often made more legitimate by rules for executive power-sharing. In Malaysia political parties are ethnically based with a moderate Islamic catch-all party for the Malays, ensuring representation for all cultural groups. The New Economic Policy (NEP) provides a system of resource allocation that privileges Malays who have experienced economic discrimination. The allocative system is intended to bring fairness to the system of distribution and thus bolster the legitimacy of the system of political representation.

Only through a set of linked supportive institutions can particular political identities be cemented and social practices consistent with institutional rules result. When institutions send conflicting signals about the character of political identity, their legitimacy decreases, and with decreased legitimacy, they can no longer effectively shape the political identity of those in their jurisdiction. This last point requires elaboration.

The Importance of Institutional Legitimacy and Strength. When political institutions make ascription—that is, cultural distinctions— a criterion for membership, participation, and allocation, identity politics is played out in the political arena. When the institutions of central authority are strong and perceived as legitimate, and when resource allocation is considered fair, political conflicts are less likely to become violent. Indeed perceptions of fair resource allocation are a key pillar of institutional legitimacy. Strong and legitimate institutions provide broadly accepted channels of political competition within which political actors operate in normal times. They allow central authorities to make credible commitments to distribute benefits and structure bargaining among various groups in ways that will be perceived as mutually advantageous. Institutional legitimacy enhances institutional capacity, thus reducing the threat of communal conflict by increasing the benefits of peaceful dispute resolution and reducing the benefits of violence. Although these institutions may privilege some groups over others, they can counter the threat of backlash with offers of side payments and compensation to those who see themselves as harmed by the preferential practices.

It would be wrong to assert that perfect social harmony is the result. These institutions often foster resentment because of these practices of privilege and compensation, but where they are considered essentially legitimate, their behavioral rules are echoed in other organizations and in the society at large. Thus these institutions can create “sticky” norms that shape social practice even in periods of institutional disruption. These norms, reflected in dominant public attitudes, act as a firebreak against ethnic and sectarian violence in that they provide the basis for a legitimate contract between state and society that ensures a degree of domestic order. In Germany, for example, a relatively strong set of institutions protecting individual rights weakened the discriminatory power of exclusive citizenship laws and created a large constituency that opposed discrimination and violence against immigrants. But the German case also suggests that social norms are ultimately dependent on institutional rules and procedures. Cultural violence in Germany was attenuated when those institutions were strong and violence was at its height when those institutions were weakened.

The opposite is true when state institutions are considered unfair, illegitimate, and oppressive. Often privilege is granted to one group, and others are excluded from the privileged resource allocation. Resentment is likely to build but will be repressed as long as the state is strong enough to exert coercive power to maintain social order. As we shall see in the pages that follow, for example, both Punjabi Sikhs and Georgian peasants in Abkhazia were excluded from privileged resource allocation. Thus both sought to secede from the governing state that they perceived as oppressive. As long as that state remained strong enough to repress dissent and as long as these two groups continued to be deprived of resources for mobilization, their grievances festered, but they did not resort to violence until the institutions of the central state weakened.

Our first set of propositions, then, is the following:

• The political relevance of cultural cleavages will be sustained and in some cases even created if they are legitimized by the dominant institutions of state and economy.
• Conversely, the political relevance of cultural identity can be weakened if it is not legitimated by those institutions.
• To the extent that those institutions are strong and legitimate, they can attenuate that political relevance or channel it in the direction of nonviolent political competition.
• To the extent that the institutions are strong enough to coerce but illegitimate in the eyes of the groups excluded from preferential access to resources, resentments build but will be repressed and violence will be attenuated.
• To the extent that institutions charged with maintaining order are weak, and to the extent that previous institutions encouraged identity politics and separatist solutions to redress grievances of particular cultural groups, the odds of violence increase.

Institutional Change, Political Entrepreneurs, Bandwagoning, and Ethnic Alliances

In those societies whose institutions are under pressure and weakened and whose institutional legacy perpetuated and either formally or informally politicized ethnic or sectarian cleavages, political entrepreneurs emerge who have both the incentive and the opportunity to exploit cultural cleavages and perceived inequities in an effort to mobilize popular support. In part, institutional legacies determine whether populations will respond to these mobilization efforts. If political institutions encourage identity politics and if prohibitions against the practice of extreme identity politics are weak, bandwagoning effects are likely to take hold. This means that when one individual sees others responding to an ethnic entrepreneur or engaing in ethnic protest, the costs of joining decrease; as the costs of joining are reduced, others are encouraged to join; indeed the costs of not joining might go up. Bandwagoning effects can escalate ethnic conflict to violence; the odds of violence increase when ethnic alliances are formed across borders.

When weakened states and resources are scarce, political entrepreneurs—whether in government or in opposition—have little to offer in exchange for support. Limited resources increase their incentive to distribute particular benefits to important supporters rather than to espouse general welfare policies whose benefits to individuals are diffuse. Under conditions of resource scarcity and institutional uncertainty and weakness, in societies where an entrenched tradition of cultural privilege and discrimination prevailed, politicians are tempted to privilege—or promise to privilege—the members of one ethnic or religious community over others. This is because their cheif constituency may be a particular cultural group and because cultural patronage networks—as allocative mechanisms—require few transaction costs. In Yugoslavia, for example, the weaker the central government became, the more allocative authority fell into the hands of regional party elites. The deepening economic crisis and the collapse of the social welfare system made their role and their patronage networks increasingly important because their aid became indispensable in keeping both enterprises and individuals afloat; they made significant allocative decisions in the economy, as well as political and administrative appointments based on ethnic and cultural bonds created in their local communities.

With regard to the institutions of representation in new or fragile democracies where resources are scarce and the legacies of ethnic machines still linger, the requirement for electoral support may provide more of an incentive for political entrepreneurs to make extremist appeals that promise more benefits to the targeted ethnic group than for them to make moderate appeals to a wider population. ₃₀ In explaining why he formed a nationalist Muslim political party in Bosnia, the Party of Democratic Action (SDA), Alija Izetbogović illustrated this logic: “Perhaps in four or five years we shall have passed through the minefield to the horizon of civil society. For now, unfortunately, our party must be sectional. . . . The parties that try to represent everyone are small and weak.” ₃₁ Before the elections of 1996, the SDA, following this logic, escalated its rhetoric of identity politics and actively and brutally oppressed the nonsectarian opposition.

If the targeted ethnic group is in the majority, the temptation to make ethnic appeals for electoral support is high if opponents are making appeals on alternative bases, especially when past institutional legacies provide the political entrepreneur with an “ethnic machine” that can deliver voter support in exchange for a credible promise of resources. Even if the targeted ethnic group is in the minority, the temptation exists to play the ethnic card by demanding ethnic autonomy. That temptation is heightened when the ethnic entrepreneur stands to gain autonomous power over a specific territory. ₃₂

Nonetheless, even if political entrepreneurs practice identity politics in electoral campaigns and make appeals based on exclusive promises to specific cultural groups, there is no guarantee that those appeals will result in an enthusiastic response from the targeted population; indeed the efforts of ethnic entrepreneurs do not automatically result in successful political mobilization. We can expect that populations jolted out of previous roles and identities by institutional reform and collapse and soured by perceived inequities may be particularly open to political appeals that emphasize cultural oppression, discrimination, and privilege. But in the recent elections in post-apartheid South Africa, for example—a country undergoing a dramatic institutional and social upheaval—nationalist, separatist leaders in both majority and minority populations were unable to gain a significant political foothold. This was the result despite decades of apartheid and what some might call ancient hatreds. The ethnic appeals of Muslim Bulgarian political entrepreneurs also fell on deaf ears after the fall of communism there, despite institutional collapse. But ethnic appeals in the republics of Serbia, Croatia, and Bosnia after 1989, nativist appeals of German political entrepreneurs in the 1980s, “skinhead” rhetoric there shortly after German unification, and the separatist appeals of Abkhazian and Sikh radicals resonated with significant elements of the populations in these regions. Why?

As noted, a critical part of the explanation has to do with institutional legacies and current institutional practices—the extent to which identity politics was and continues to be cememted in political institutions and the extent to which the population expects that identity politics will dominate current political competition. The odds are higher that ethnic or sectarian political entrepreneurs fomenting cultural conflict will succeed in their mobilization efforts in those places where institutional legacies created the expectation that identity politics would dominate the political game, even if old institutions supporting that political logic weakened and collapsed. Note the “success” of Vladislav Ardzinba of Abkhazia, Slobodan Milošević of Serbia, and Franjo Tudjman of Croatia. When new institutions are able to even partially erase those legacies, their chances of success diminish. The South African case provides a good example.

Conversely, the odds are higher that political entrepreneurs who wish to mitigate cultural conflict and mobilize populations around ideological appeals or regional (as opposed to cultural) autonomy will be more successful when the institutional legacy supports their efforts—that is, when the institutional carriers of identity politics have not been strong and when new institutions do not encourage divisive identity politics. Kiro Gligorov’s relative success in Macadonia and Aslan Abashidze’s success in Ajaria illustrate the point. And where the institutional carriers of identity politics have been weak, political entrepreneurs fomenting cultural conflict have been less than successful. Note the relative lack of success of the VMRO in Macadonia, the radical Muslim party in Malaysia, and Kamen Burov in Bulgaria. In short, the resources available for the realization of the goals of political entrepreneurs vary according to the institutional legacies and degree of institutional strength in their country or region, and those resources will be good predictors of their success.In Britain, for example, where liberal institutions are relatively strong, the resources available to nativist political entrepreneurs who would foment violence against immigrants are scarce. In Germany, where institutions are predominantly liberal but citizenship is defined in ethnocultural terms and the political party system was undergoing change and experiencing gridlock, ethnic entrepreneurs had a wider range of available resources for political mobilization using ethnic appeals against immigrant populations. At the other extreme, in Croatia, Serbia, and Bosnia, where Titoist institutions had deeply politicized ethnic and religious cleavages and where old institutions quickly disintegrated, political entrepreneurs practicing identity politics had a vast array of resources to exchange for support before institutional transformation could begin.

Indeed success and failure in political mobilization efforts based on the rhetoric of identity politics is highly correlated with institutional legacies. But what is the process or mechanism by which successful mobilization occurs? Strategic interaction theories explain how support for an ethnic entrepreneur can spread beyond the support of those who receive direct and tangible benefits in exchange. ₃₃ They suggest that bandwagon effects influence each individual’s choice to engage in identity politics. Both subsequent reduced costs of joining and increased social pressure to join exacerbate bandwagon effects. When one ethnic group jumps on the ethnic bandwagon, other groups are motivated to take countermeasures and jump on bandwagons of their own in order to balance against the first group’s strength. Thus the importance of timing to the odds of escalating cultural conflict. The initial success of one ethnic entrepreneur lowers the costs and raises the incentives for other politicians from other ethnic groups to pursue exclusive ethnic or religious nationalist strategies. That is, when one ethnic bandwagon is filled, those who are excluded because they belong to a different ethnic group feel threatened, and they respond by creating bandwagons of their own. As bandwagons from opposing groups fill, holdouts feel increased pressure to build and jump on their own bandwagons lest they be left standing alone. Timur Kuran explains bandwagoning and balancing this way:

When members of one ethnic group start engaging in more ethnic activity, attention is drawn to society’s ethnic divisions. Members of other groups are thus reminded of their outsider status vis-à-vis the group that initiated the process. Motivated to ensure that some group accepts them, they feel pressured to make more public displays of their identity. ₃₄

But how do identity politics escalate and become violent? Just because ethnic entrepreneurs may practice identity politics and just because they can mobilize support for that practice does not automatically mean that violence will be the result. Kuran argues that to get the ethnic bandwagon rolling, political entrepreneurs must initially arouse emotions and evoke images through vivid public acts. Particularly effective are acts of violence or civil disobedience. These acts are sometimes small and sporadic, but they can create the initial social pressure required to join the ethnic bandwagon. Extremists are the ones likely to engage in these initial acts; they climb on the bandwagon first to set it in motion, while moderates join later on and thus move toward extremist positions, particularly when constraints on extremism are weak. The bandwagon can also start to roll when violence is initiated against a particular group by the state. In this case, moderates are likely to be rapidly discredited and the value of extremism is heightened. The Indian army’s storming of the Sikhs’ Golden Temple provides a good example.

This escalation to violence when institutions weaken or collapse is not unique to the practice of identity politics. Class conflict and ideological cleavages escalate to violence in a similar way. The difference is this: the odds that bandwagoning effects under the institutional conditions described here will escalate to violence are higher because resource claims based on exclusive cultural criteria are incompatible. Compromise is always difficult when extremists dominate the political arena; it is more diffucult because cultural identity is fixed and non-negotiable; exclusive claims on resources for distinct cultural groups mobilized by myths of superiority and intrinsic rights are particularly difficult to compromise.

We will witness this process in many of the stories told here. In the Yugoslav case, the escalation of ethnic nationalism that led to the 1970–71 crisis in Croatia began with banal incidents that were discounted and therefore tolerated, encouraging escalation. ₃₅ The central state, then still strong enough to do so, was able to quell the violence. The 1981 Albanian riots in Kosovo began in a similar fashion, but the central state was weaker and violence began to spread. Shortly before the war broke out, mass demonstrations to protest Albanian discrimination against Serbs in Kosovo began with small groups of protesters but expanded to crowds numbering as many as one million. In Croatia, Croat gangs, often aided by the police, firebombed Serb homes, smashed storefronts, and arrested Serb leaders, fueling Croat nationalist sentiments and thus strengthening the political appeals of Croat ethnic entrepreneurs. In the German case, the process of escalation to large-scale violence was precipitated by the attack of eight young skinheads on Vietnamese street merchants in Hoyerswerda. That initial act of violence was conveyed by electronic media, triggering similar incidents which spread not only throughout the east, but to the west as well. Once Hoyerswerda demonstrated the “success” of such action—that is, the action was not penalized by either state authority or public opinion—potential aggressors elsewhere risked little but expected to gain prestige among their peers by engaging in similar acts.

As the examples above suggest, political institutions can encourage or inhibit these bandwagoning effects. In Abkhazia, for example, institutions encouraged bandwagoning effects. Institutions providing preferential policies for the Abkhazian population spurred demands from the Georgian population to provide preferential policies for Georgians. In response, Abkhazian elites escalated their demands for more preferential policies. Georgians resisted; ethnic tensions heightened. The Yugoslav case tells a similar story.

Conversely, institutions can squelch the temptation to bandwagon. ₃₆ In post-Wall Germany ethnic bandwagoning escalated to violence when an institutional vacuum opened in the east; it halted when liberal institutions regained strength. Indeed the lesson of the German case is that it was the absence of institutions to prevent ethnic bandwagoning that permitted the violent escalation of ethnic tensions and the reinstatement of institutional constraints that squelched the bandwagoning effect. But in a more subtle sense, when institutions are too weak to meet the political demands of culturally defined groups, when they can no longer make concessions to those groups and adjudicate disputes, the benefits of nonviolent political action are reduced. When all nonviolent means are exhausted, these groups may feel that they have nothing to lose through violent acts. Indeed, as we shall see, this was the case in both Punjab and Kashmir.

Finally, bandwagoning accounts draw our attention to the importance of ethnic alliances, which raise the odds of the escalation of identity politics to cultural conflict. As the bandwagons begin to roll faster, costs of participation are further lowered if ethnic alliances form across borders, bringing material and symbolic resources to new ethnic movements and parties from the more established ones. These resources affect the available incentive structure of political elites and push them to jump on the ethnic bandwagon. ₃₇ In Bosnia “sister” Serb and Croat nationalist parties had preponderant resources, crowding out non-nationalist alternatives. Abkhazian separatists called on former KGB members, elements of the Soviet army, and the Confederation of the Mountain Peoples of the Caucasus for material support in their war of succession. Radical Islamic groups formed transnational coalitions to facilitate acts of violence across state borders. As both the Abkhazian and the Bulgarian cases demonstrate, Western human rights organizations and aid agencies have unwittingly abetted the agendas of ethnic and religious entrepreneurs in post-Communist regions and helped to swell the ranks of their supporters by providing or promising to provide material or symbolic support to targeted cultural groups. In those cases where powerful transnational patrons have encouraged a violent grasp for power on the part of their clients, in transforming societies with weakened state institutions, the odds of violence increase.

Our second set of propositions is thus the following:

• When state institutions that sustain the social contract or repress excluded groups are weakened and placed under pressure by internal or external forces in societies where cultural identity had become politically relevant, ethnic and sectarian violence can erupt.
• The eruptions will be more intense in states where political identity was most politicized and where institutions were most severely weakened. In those states, political entrepreneurs will be able to more easily mobilize support based on ethnic or sectarian appeals.
• The odds of violence increase when institutions either encourage bandwagoning effects or are too weak to stop them.
• The odds of violence increase when cultural alliances are formed across borders and the alliance partner encourages violence.

The Political Economy of Cultural Conflict

Economic factors operate in four ways to raise the odds that communal conflict will escalate to violence. First, economic discrimination and privilege on the basis of ascriptive criteria cause cultural identity to become politically relevant and intensify cultural identities that have already become politically charged. Economic discrimination and privilege can thus politicize cultural identity, even if the rules of political institutions encourage the political relevance of other social divisions. In many regions where cultural conflict has erupted, an ethnic or sectarian “division of labor” historically provided for the preferential allocation of resources to particular cultural groups. Over time, preferential systems of resource allocation created both collective resentments and claims on resources based on past entitlements and perceptions of an inherent right to those resources. These resentments and resource claims were woven into the fabric of the collective political identity of particular cultural groups. In Abkhazia, for example, the Abkhazes controlled the most lucrative crops, and their control over local administration gave them an edge in their economic competition with Georgian peasants. Georgian collective farms experienced tighter central control and fewer subsidies than Abkhazian farms. In India the Green Revolution made Punjab the breadbasket of the country, and thus the central government starved Sikhs in the region of industrial development, leading them to believe that they were being exploited for food supplies without adequate compensation. The resentments of both Georgian peasants in Abkhazia and Sikhs in Punjab provided them with strong incentives to free themselves from the oppressive yoke of central governments practicing economic discrimination. Georgians wanted Abkhazia to become part of Georgia, and Sikhs pushed for Punjab’s independence.

Second, economic factors can weaken the political institutions that uphold the social contracts that provide social stability. Economic factors can also weaken the institutions which repressed those cultural groups with political grievances. There are, of course, many reasons for institutional weakness and transformation—internal corruption, bureaucratic rigidity, rent-seeking behavior on the part of government elites, a failure to deliver on democratic promises, or simply constitutional changes, to name just a few. Even weakened institutions, however, can maintain a credible commitment to uphold the social contract if they have adequate material resources to exchange for social support or to repress dissent. When the resources of institutions that provided accepted distributive channels are diminished, however, those institutions withdraw from their distributive role in society. When they withdraw by dismantling their social welfare systems and privatizing state enterprises, those who are disadvantaged by that withdrawal often blame the state for their hardships. These populations become increasingly cynical about their political membership in those states; a gradual loss in government legitimacy is the result. ₃₈ Further, states that use discrimination and repression to maintain social order need resources to reinforce coercive policies targeted against specific cultural groups. When those resources are diminished, coercion can fail, and political entrepreneurs are likely to seize the opportunity to capture the state or secede from it. We must therefore look to the factors that diminish those resources if we are to understand the conditions under which institutional transformation raises the odds of cultural violence.

Our cases point to long-term globalization trends and short-term policy responses to those trends—i.e., forces that reduce the state’s role in the economy and reduce its sovereignty over political membership—as the key domestic causes of broken social contracts and failed coercive policies. Global immigration patterns undermine the contract for political membership. National economic growth and decline and the level of external debt affect the level of resources that the state can allocate, and short-term policies of economic liberalization yield up the state’s distributive powers to the market. Indeed when states make the decision to allow the market to pick economic winners and losers, they can break the social contract that once permitted them to soften some of the disadvantages suffered by particular cultural groups. The state’s ability to soften such disadvantages permitted the integration of those groups into the political community. All of these forces of globalization and liberalization thus have an important impact on the state’s ability to support institutions that provide social order or repress dissent.

Similarly, policies of economic decentralization can break the social contract that once permitted central governments to transfer resources to minority ethnic or religious groups who would lose those resources if local majority authorities gained control. Policies of decentralization can also break the cycle of economic discrimination against distinct regional cultural groups. These policies heighten the motivation of political entrepreneurs to confront the center and strengthen their hand with populations in their regions with their promises of autonomy. Policies of decentralization also break the chain of central control over repressive institutions, creating new opportunities for political opposition. When opposition political entrepreneurs have tangible resources to provide, they are likely to gain support.

Globalization and liberalization can trigger conflict only when economic hardships result and when those hardships fall disproportionately on distinct cultural groups. Indeed in some important cases, economic liberalization has mitigated cultural conflict. For example, as Malaysian prosperity expands the economic pie, increasing resources are available for all cultural groups. Because the allocative institutions that support the social contract distribute these resources in ways widely perceived as fair, this prosperity solidifies the legitimacy of the social contract and denies sectarian political entrepreneurs the grievances that could become the fuel for political mobilization. In Punjab, where central state restrictions have been lifted, market-stimulated growth may indeed benefit disgruntled Sikh farmers, who were excluded from a system of ascriptive resource allocation. But a note of caution is in order: it is only when states “win” in market competition and when previously disadvantaged groups benefit that economic transformation can mute cultural conflicts.

Third, economic factors not only affect the strength of institutions at the top within society, but they also directly create conditions at the bottom that make social groups receptive to the appeals of political entrepreneurs. ₃₉ Economic hardship provides a concrete justification for political grievances that can be transformed into a resource for political mobilization.

Economic crisis and change always cause social disruption and radical dislocation of communities. When secular economic trends lead to low growth, debt crises, rising unemployment, and rising rates of immigration, and when the resulting hardships and benefits are disproportionately allocated among various cultural groups, existing political cleavages based on cultural difference are exacerbated and new ones are created. In Bulgaria, for example, the introduction of markets and the restitution of land created disproportionate unemployment among the Muslim population. In England, as industry declined in the early 1980s, the resulting unemployment was disproportionately allocated to minority populations. As Elaine Thomas describes here, in 1984, 95 percent of Handsworth’s black school leavers and 84 percent of their Asian counterparts remained jobless. It was in this area that a major riot broke out.

Economic hardships that lead cultural groups to distrust the state can make these groups available for reassignment to new political identities. The losers in economic transformation will attempt to use their political resources and position to resist changes that disadvantage them. Economic difficulties that fall disproportionately on culturally defined social groups thus create the demand for the goods that political entrepreneurs promise to deliver, particularly when those same factors that fuel cultural grievances also reduce government resources to uphold the social contract.

State withdrawal from its allocative role, the introduction of markets, and disproportionate economic hardships are grist for the mill of eager political entrepreneurs. This is exemplified in the case of Bulgaria. There the former Communist regime provided the Turkish minority with economic security. Ethnic Turks were concentrated in the tobacco industry; the state purchased tobacco, ensuring full lifetime employment. With the fall of communism, however, the inefficient and uncompetitive tobacco industry was privatized, and its failure in global markets left the majority of Turks unemployed and destitute. Now Turkish political entrepreneurs in Bulgaria label unemployment “ethnic genocide” in their effort to mobilize the Turkish population against the liberalizing policies of the new regime.

Finally, the political entrepreneurs who are able to offer tangible resources to disadvantaged populations are those most likely to gain support. For example, privatization deprives trade union leaders of the resources needed to mobilize support around class identities; local officials, however, who have a cache of resources or foreign support from a diaspora ethnic or religious community will be able to mobilize support around an ethnic or religious identity. If would-be ethnic or sectarian nationalists have no tangible resources to offer, they may be able to initiate random terrorist acts, but they will be unlikely to gain the necessary support to play a role in the political arena, capture the state, or secede from it.

The level of material resources available to competing political entrepreneurs is dependent upon two conditions: 1) the degree to which cultural criteria were historically used to allocate these resources; political systems in which resources were once allocated according to ethnic and sectarian criteria provide some modern ethnic and sectarian political entrepreneurs with a stash of material resources available for political mobilization, and 2) the level of resources provided by international alliances. In the Bulgarian case cited above, Turkish political entrepreneurs had few tangible resources to offer in exchange for support. Because cultural identity in Bulgaria was not politically charged and because the Turkish party—the MRF—was legally prohibited from constituting itself as an ethnic party, its influence with the Turkish population was minimized. Although the MRF plays an important political role in Bulgaria, it has not been able to mobilize enough support in order to introduce the logic of identity politics into the political arena.

In regions where ethnic and sectarian entrepreneurs had more tangible resources to offer, they were able to gain increasing political power. Particularly where fiscal crises were deep, space was opened for politically relevant cultural groups to usurp the state’s previous allocative functions and deliver alternative economic benefits to targeted cultural groups. By providing an alternative channel for economic resources, these groups either provide the first impetus for the politicization of cultural identity or they further mobilize political support against state institutions and intensify the political relevance of cultural identity. By becoming an alternative distributor of material goods in Egypt, for example, the moderate Muslim Brotherhood has been able to deepen and sometimes create the political relevance of Muslim identity. Indeed the Muslim Brotherhood and the FIS (Islamic Salvation Front) in Algeria have far more support that can be translated into political power than Gamma or Islamic Jihad, which can offer only an alternative identity but few tangible resources. These groups obtained tangible resources from transitional Islamic groups. In the former Soviet Union regional officials with access to material goods have been able to prevail over political competitors who can offer only symbolic resources. Because they were well positioned to convert the local state apparatus into a mechanism for the distribution of patronage, they have been able to create new political identities—often ethnic or sectarian—among their clients.

Where the fiscal crisis is deep, where cultural identities have been politicized but where political entrepreneurs have few tangible resources to offer, they may be able to organize for only sporadic violence against the regime in power or against groups who allegedly caused the economic hardship (often minority immigrant populations). In East Germany after 1990, for example, the increasing presence of immigrants provided a focus for dissatisfactions and resentments, as the native population saw public expenditures going to “foreigners” while they were faced with rising unemployment and housing shortages. Sporadic violence against foreigners was the result. When economic crisis was attenuated and political institutions were restored, violence receded.

In short, when economic decline, rising immigration, liberalization policies, and decentralization reduce central state resources required to uphold the social contract, that contract can break. If the costs of broken social contracts fall disproportionately on culturally defined groups, the door opens to political entrepreneurs bent on mobilizing populations along ethnic and sectarian lines. Cuts in resources to culturally defined populations ensure that their appeals will not fall on deaf ears. But only those political entrepreneurs who have tangible resources to offer in exchange for support are likely to succeed in meeting their goals.

Our third set of propositions is the following:

• Economic discrimination and privilege on the basis of ascriptive criteria cause cultural identity to become politically relevant.
• Economic globalization and domestic economic policy responses can weaken the political institutions that upheld the social contract that provided social stability or the institutions that repressed dissent. Broken social contracts and weakened repressive institutions open political space for ethnic and sectarian entrepreneurs to mobilize support for their efforts to capture the state or secede from it.
• Economic hardships that fall directly and disproportionately on culturally defined social groups create the demand for the goods that ethnic and sectarian political entrepreneurs promise to deliver.
• The political entrepreneurs who are able to offer tangible resources to disadvantaged populations are most likely to gain support. The resources available to them are directly related to legacies of institutional resource allocation—whether cultural criteria were used as allocative principles or not—and to the level of resources available from international alliances.
• The extent of institutional disruption brought on by economic transformation and disproportionate economic deprivation, combined with the extent to which cultural distinctions were previously politicized and the level of resources available to political entrepreneurs who promote violence, will largely determine whether identity politics escalates to communal violence and will largely determine the intensity of that violence.

Conclusions

In sum, the central argument of this volume can be distilled as follows: Cultural identities can be transformed into political identities when cultural groups are targeted for privilege or discrimination and when economic factors, no matter how “impersonal,” lead to disproportionate hardships among culturally defined populations. The institutions of the central state can either legitimate the political relevance of cultural identity and channel cultural conflict in ways that achieve social harmony, or they can mute the political relevance of cultural identity and construct channels of allocation without regard to cultural differences. Economic strength contributes to the institutional strength that mutes the political identity of culturally defined groups or channels identity politics to peaceful conflict resolution. When these institutions are weakened, political entrepreneurs in divided societies can mobilize support around cultural identities where they had been previously legitimized, or they can act to transform cultural into political identity by pointing to disproportionate hardships where politicized cultural identities did not previously exist. Economic factors can thus impact both identity transformation and institutional strength. If institutional legacies encourage identity politics, if economic grievances are defined in cultural terms, if political entrepreneurs are initially successful in their cultural appeals for political support, bandwagoning effects and ethnic alliances increase the odds that identity politics will escalate to violence.

The chapters in this introductory section elaborate on the central concepts inroduced here, and the case studies follow. In the concluding section, I draw upon the wealth of material in the stories told in those case studies in order to defend each step of the argument.

---

Endnotes

Note 1. David Binder with Barbara Crosette, “As Ethnic Wars Multiply, U.S. Strives for a Policy,” New York Times, 7 February 1993, pp. A1, A14.

Note 2. Cited in David C. Rapoport, “The Importance of Space in Violent Ethno-Religious Strife” (San Diego: Institute on Global Conflict and Cooperation, University of California, January 1996); IGCC Policy Paper No. 21.

Note 3. See Carol Bellamy, “State of the World’s Children in 1996”; UNICEF report; cited in Barbara Crosette, ”Unicef Report Sees Children as Major Victims of Wars,” New York Times, 11 December 1995, p. A7.

Note 4. By August 1995 UNHCR reported that more people had been uprooted in Bosnia than anywhere else in the former Yugoslavia, both absolutely and as a share of the population. Of the prewar population of 4.4 million, 1.3 million were displaced within Bosnia. Another 500,000 sought shelter in Serbia and Croatia, and over 500,000 live outside the former Yugoslavia. Before the war, there were 600,000 Serbs in Croatia of a total population of 4.8 million. With the Croatian offensive in the Krajina, 150,000 left. Before that, perhaps 200,000 had already left Croatia (Economist, 19 August 1995, p. 42).

Note 5. See Komsomolskaia pravda, 28 June 1994, and Moskovskie novosti, no. 23, 1995.

Note 6. See, for example, Ethan Kapstein, “Workers and the World Economy,” Foreign Affairs 75, 1 (May–June 1996): esp. pp. 17, 22, and 26.

Note 7. Political entrepreneurs resemble their economic counterparts in that they seek to maximize their individual interests and in doing so, have an effect on aggregate interests. The political entrepreneur seeks to maximize political power, while the economic entrepreneur seeks to maximize wealth. But like their economic counterparts, political entrepreneurs engage in risk-taking behavior to maximize their returns. For additional comments on political entrepreneurs, see David Laitin, “Hegemony and Religious Conflict: British Imperial Control and Political Cleavages in Yorubaland,” in Bringing the State Back In, ed. Peter B. Evans, Dietrich Rueschemeyer, and Theda Skocpol (New York: Cambridge University Press, 1985), pp. 285–316; Paul R. Brass, “Ethnicity and Nationality Formation,” Ethnicity 3, 3 (September 1976): 225–39.

Note 8. This assumption has not been systematically tested. A good test would compare the intensity of conflict and level of violence of “identity group” conflicts with “interest group” conflicts, ideological conflicts, class conflicts, and interstate conflicts.

Note 9. For important but more specific institutional approaches with general implications, see Donald Horowitz, A Democratic South Africa? Constitutional Engineering in a Divided Society (Berkeley: University of California Press, 1991), and Arend Lijphart, Democracy in Plural Societies (New Haven: Yale University Press, 1977).

Note 10. Clinton’s inaugural address; cited in New York Times, 21 January 1993, p. A15.

Note 11. Cited in Bernard Lewis, “Muslims, Christians, and Jews: The Dream of Coexistence,” New York Review of Books, 26 March 1992, p. 48. Literature in this essentialist school includes James G. Kellas, The Politics of Nationalism and Ethnicity (London: Macmillan, 1991); R. A. Schermerhorn, Comparative Ethnic Relations (New York: Random House, 1970); Uri Ra’anan, “Nation and State: Order out of Chaos,” in State and Nation in Multi-Ethnic Societies: The Breakup of Multinational States, ed. Uri Ra’anan, Maria Mesner, Keith Armes, and Kate Martin (Manchester: Manchester University Press, 1991), pp. 13–14; and Clifford Geertz, “The Integrative Revolution,” in Old Societies and New States: The Quest for Modernity in Asia and Africa, ed. Clifford Geertz (New York: Free Press, 1963). For a more nuanced version, often called the “ethnic-continuationist” version of nationalism, see Anthony D. Smith: The Ethnic Origins of Nations (Oxford: Basil Blackwell, 1986), and The Ethnic Revival (Cambridge: Cambridge University Press, 1981).

Note 12. See, for example, James B. Rule, “Tribalism and the State,” Dissent, Fall 1992, p. 519.

Note 13. See Russell Hardin, One for All: The Logic of Group Conflict (Princeton: Princeton University Press, 1995).

Note 14. See Michael Ignatieff, Blood and Belonging: Essays in the New Nationalism (New York: Farrar, Straus and Giroux, 1994).

Note 15. See Smith, The Ethnic Origins of Nations, for a discussion of essential and constructed identities.

Note 16. An important discussion of collective memory can be found in Andrei Markovits and Simon Reich, “The Contemporary Power of Memory: The Dilemmas for German Foreign Policy”; paper presented at the conference on the Postwar Transformation of Germany, Center for German and European Studies, University of California, Berkeley, December 1995.

Note 17. See Barry R. Posen, “The Security Dilemma and Ethnic Conflict,” in Ethnic Conflict and International Security, ed. Michael E. Brown (Princeton: Princeton University Press, 1993), pp. 103–24; James Fearon, “Commitment Problems and the Spread of Ethnic Conflict”; IGCC Project on the International Spread and Management of Ethnic Conflict, 30 October 1995; and David A. Lake and Donald Rothchild, “Containing Fear: The International Spread and Management of Ethnic Conflict”; conference draft for the IGCC Project on the International Spread and Management of Ethnic Conflict, 30 October 1995. Stephen Saideman makes a similar argument in “Is Pandora’s Box Half-Empty or Half-Full? The Limited Virulence of Secessionism and the Domestic Sources of Disintegration” (San Diego: Institute of Global Conflict and Cooperation, University of California, 1996); IGCC Policy Paper No. 18. A similar discussion involving the role of population growth and migration can be found in Lawrence R. Robertson, “Ethnic Minorities, Migration, and Conflict Contagion in Eastern Europe and the Former Soviet Union”; conference draft for the IGCC Project on the International Spread and Management of Ethnic Conflict, 30 October 1995.

Note 18. Charles A. Kupchan and Clifford A. Kupchan, “Concerts, Collective Security, and the Future of Europe,” International Security 16 (Summer 1991): 130–33.

Note 19. See Jonah Levy, Robert Kagan, and John Zysman, “The Twin Restorations: The Political Economy of the Reagan and Thatcher ‘Revolutions’”; (Berkeley, October, 1995); unpublished manuscript.

Note 20. See E. E. Schattschneider, The Semisovereign People: A Realist’s View of Democracy in America (New York: Holt, Rinehart, and Winston, 1960), p. 72.

Note 21. See, for example, Surjit S. Bhalla, “Freedom and Economic Growth: A Virtuous Cycle?”; paper presented at the South Asia Seminar, Center for International Affairs, Harvard University, February 1994; Laurence Whitehead, “Political Democratization and Economic Liberalization: Prospects for Their Entrenchment in Eastern Europe and Latin America”; paper presented at the Southern California Seminar on Political and Economic Liberalization, University of Southern California, Los Angeles, 30 January 1995. Barbara Geddes makes a more narrowly economic argument when she claims that authoritarian regimes are not required for development (see “Challenging the Conventional Wisdom,” Journal of Democracy 5, 4 [October 1994]: 104–18).

Note 22. Lake and Rothchild, pp. 36–37.

Note 23. Benjamin Barber makes similar connections, although his logic of explanation diverges from ours. He argues that economic globalization also globalizes politics by creating new sources of dominance, surveillance, and manipulation, thereby weakening the nation-state. The state is thus increasingly less important as a focus of political life. Global processes do not require democracy to expand. With the decline of the nation-state as the locus of political life and the increasingly undemocratic globalization of political and economic life, subnational communities governed by fanatical hierarchies attempt to localize politics. These groups are also undemocratic in that they demand loyalty to the group above loyalty to the individual, and rights are only real for the dominant group. The result is the decline of democracy and democratic, integrating nation-states. Barber does not explain why local politics would take a nondemocratic form. See Jihad vs. McWorld (New York: Random House/Times, 1995). The main themes of this book were published in an earlier article by the same name in the Atlantic Monthly 209, 3 (March 1992): 53.

Note 24. On the distinction between liberal and illiberal democracies, see Robert Dahl, Polyarchy: Participation and Opposition (New Haven: Yale University Press, 1971); Guillermo O’Donnell and Philippe Schmitter, Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies (Baltimore: Johns Hopkins University Press, 1986); Terry Lynn Karl: “Dilemmas of Democratization in Latin America,” Comparative Politics, October 1990, and “Imposing Consent? Electoralism vs. Democratization in El Salvador,” in Elections and Democratization in Latin America, 1980–1985, ed. Paul Drake and Eduardo Silva (La Jolla: Center for Iberian and Latin American Studies, UCSD, 1986); Adam Przeworski, Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America (New York: Cambridge University Press, 1991); Guillermo O’Donnell, “On the State, Various Crises, and Problematic Democratization,” World Development 21, 8 (August 1993); David Collier and Steven Levitsky, “Democracy ‘with Adjectives’: Conceptual Innovation in Comparative Research” (South Bend, Ind.: Kellogg Institute, University of Notre Dame), working paper. For other arguments about the potential negative impact of democracy on ethnic harmony, see the essays in Brown, ed.

Note 25. Douglas C. North, in Structure and Change in Economic History, pp. 201–2.

Note 26. Barbara Geddes, “A Comparative Perspective on the Leninist Legacy in Eastern Europe,” Comparative Political Studies, 28, 2 (July 1995): 239–75.

Note 27. See Ernst B. Haas, “Nationalism in the United States” (Berkeley, 1995); unpublished manuscript.

Note 28. By constructing the rules for membership, distribution, accountability, and participation, institutions shape actor preferences by rewarding some choices and punishing others. They not only constrain some behaviors and encourage others, but also they provide power resources to some social actors and deny them to others. They thus shape the strategies of political elites, both inside and outside government, and structure the relations of power among social groups. See Mark Granovetter and Richard Swedberg, The Sociology of Economic Life (Boulder: Westview Press, 1992); Walter W. Powell and Paul J. DiMaggio, eds., The New Institutionalism in Organizational Analysis (Chicago: University of Chicago Press, 1991); James G. March and Johan P. Olsen, “The New Institutionalism: Organizational Factors in Political Life,” American Political Science Review 78 (1984): 734–49; Sven Steinmo, Katherine Thelen, and Frank Longstreth, eds., Structuring Politics: Historical Institutionalism in Comparative Analysis (Cambridge: Cambridge University Press, 1992).

Note 29. Myron Weiner, “The Political Consequences of Preferential Policies: A Comparative Perspective,” Comparative Politics 16, 1 (October 1983): 35–53.

Note 30. See Alvin Rabushka and Kenneth Shepsle, Politics in Plural Societies: A Theory of Democratic Instability (Columbus, Ohio: Merrill Press, 1972).

Note 31. Cited in Noel Malcolm, Bosnia: A Short History (New York: New York University Press, 1994), p. 219.

Note 32. Studies of electoral processes suggest that specific promises to members of an identifiable group are more appealing to that group than moderate promises from which a diffuse population stands to gain. The logic suggests that when a political entrepreneur makes exclusive promises to a targeted ethnic group, he is more likely to win the votes of that group than if appeals are targeted broadly to a multiethnic population.

Note 33. Important examples include Mark Granovetter, “Threshold Models of Collective Behavior,” American Journal of Sociology 83 (May 1978); James De Nardo, Power in Numbers (Princeton: Princeton University Press, 1985); Timur Kuran, “Sparks and Prairie Fires: A Theory of Unanticipated Revolution,” Public Choice 61 (April 1989); Dennis Chong, Collective Action and the Civil Rights Movement (Chicago: University of Chicago Press, 1991); Susanne Lohmann, “The Dynamics of Informational Cascades: The Monday Demonstrations in Leipzig, East Germany, 1989–91,” World Politics 47 (October 1994): 42–101.

Note 34. Timur Kuran, “Ethnic Dissimilation and Its Global Transmission” (Los Angeles, 1995); unpublished manuscript, p. 15.

Note 35. See Dennison Rusinow, “Nationalities Policy and the ‘National Question,’” in Yugoslavia in the 1980s, ed. Pedro Ramet (Boulder: Westview, 1985), p. 134.

Note 36. For a game-theoretic approach to institutional “squelching” of ethnic conflict, see James D. Fearon and David D. Laitin, “Explaining Interethnic Cooperation”; paper prepared for the annual meeting of the American Political Science Association, Chicago, 31 August–3 September 1995.

Note 37. On ethnic alliances between states, see Will H. Moore and David R. Davis, “Does Ethnicity Matter? Ethnic Alliances and International Interactions”; paper presented at meeting of the International Studies Association, Vancouver, Canada, March 1995.

Note 38. Guillermo O’Donnell makes this argument in “On the State, Various Crises, and Problematic Democratization.”

Note 39. Our argument about the role of economic factors in the current period builds on work that demonstrates a linkage between economic liberalization, and globalization, on the one hand, and political mobilization along ethnic or religious lines on the other. See, for example, Gail W. Lapidus, Victor Zaslavsky, and Philip Goldman, eds., From Union to Commonwealth: Nationalism and Separatism in the Soviet Union (Cambridge: Cambridge University Press, 1992).