¹ The phrase “peace-preserving” federalism is coined by Bermeo, Nancy, “The Import of Institutions,” Journal of Democracy 13 (April 2002)CrossRefGoogle Scholar.

² Recent research that has begun to address this question focuses on the political-economic conditions for self-determination movements and the degree to which a country's minority population is concentrated in one province. See Sambanis, Nicholas and Milanovic, Branco, “Explaining the Demand for Sovereignty” (Manuscript, Department of Political Science, Yale University, New Haven, May 2004)Google Scholar; and Hale, Henry, “Divided We Stand: Institutional Sources of Ethnofederal State Survival and Collapse,” World Politics 56 (January 2004)CrossRefGoogle Scholar.

³ In this regard, our study is consistent with “state in society” approaches. See Migdal, Joel S., State in Society: Studying How States and Societies Transform and Constitute One Another (Cambridge: Cambridge University Press, 2001)CrossRefGoogle Scholar.

⁴ Riker, William, Federalism: Origin, Operation, Significance (Boston: Little, Brown and Company, 1964), 11Google Scholar.

⁵ Botana, Natalio, El orden conservador: La politico Argentina entre 1880lyl916 (Buenos Aires: Editorial Sudamericana, 1998)Google Scholar; Rock, David, Argentina, 1516–1987: From Spanish Colonization to Alfonsin (Berkeley: University of California Press, 1987)Google Scholar.

⁶ Castro, Jaime, Descentralizar para pacificar (Bogotá: Editorial Ariel, 1998)Google Scholar.

⁷ Foner, Eric, Free Soil, Free Labor, Free Men: The Ideology of the Republican Party before the Civil War (New York: Oxford University Press, 1995)Google Scholar; Weingast, Barry R., “Political Stability and Civil War: Institutions, Commitment, and American Democracy,” in Bates, Robert H., Greif, Avner, Levi, Margaret, Rosenthal, Jean-Laurent, and Weingast, Barry R., eds., Analytic Narratives (Princeton: Princeton University Press, 1998)Google Scholar.

⁸ Simeon, Richard, “Canada: Federalism, Language, and Regional Conflict,” in Amoretti, Ugo M. and Bermeo, Nancy, eds., Federalism and Territorial Cleavages (Baltimore: Johns Hopkins University Press, 2004)Google Scholar.

⁹ Martin, Terry, The Affirmative Action Empire. Nations and Nationalism in the Soviet Union, 1923–1939 (Ithaca, N.Y.: Cornell University Press, 2001)Google Scholar.

¹⁰ Shakhrai, Sergei, “The Treaties on Differentiation of Responsibility between the Powers of the Russian Federation and Its Subjects,” Kazan Federalist 4 (Autumn 2003)Google Scholar.

¹¹ André Bächtiger and Jürg Steiner, “Switzerland: Territorial Cleavage Management as Paragon and Paradox,” in Amoretti and Bermeo (fn. 8), 34–35.

¹² Brass, Paul, Language, Religion and Politics in North India (New York: Cambridge University Press, 1974)Google Scholar; Atul Kohli, “India: Federalism and the Accommodation of Ethnic Nationalism,” in Amoretti and Bermeo (fn. 8).

¹³ Bermeo (fn. 1), 98–100.

¹⁴ Gurr, Ted Robert, Peoples versus States (Washington, D.C.: United States Institute of Peace Press, 2000)Google Scholar; Hechter, Michael, Containing Nationalism (New York: Oxford University Press, 2000)Google Scholar; Stepan, Alfred, Arguing Comparative Politics (New York: Oxford University Press, 2001), chap. 9Google Scholar; Saideman, Stephen M., Lanoue, David J., Campenni, Michael, and Stanton, Samuel, “Democratization, Political Institutions, and Ethnic Conflict: A Pooled Time-Series Analysis, 1985–1998,” Comparative Political Studies 35 (February 2002)Google Scholar.

¹⁵ Lijphart, Arend, “The Power-Sharing Approach,” in Montville, Joseph V., ed., Conflict and Peacemaking in Multiethnic Societies (Lexington, Mass.: Lexington Books, 1990)Google Scholar.

¹⁶ Weingast, Barry R., “Constructing Trust: The Political and Economic Roots of Ethnic and Regional Conflict,” in Soltan, Karol, Uslaner, Eric M., and Haufler, Virginia, eds., Institutions and Social Order (Ann Arbor: University of Michigan Press, 1998)Google Scholar.

¹⁷ Lake, David A. and Rothchild, Donald, “Containing Fear: The Origins and Management of Ethnic Conflict,” International Security 21 (Fall 1996)CrossRefGoogle Scholar.

¹⁸ Amoretti and Bermeo (fn. 8).

¹⁹ For example, Cain, Michael and Dougherty, Keith, “Suppressing Shays’ Rebellion: Collective Action and Constitutional Design under the Articles ofConfederation,” Journal of’Theoretical Politics 11 (April 1999)Google Scholar.

²⁰ Spiller, Pablo and Tommasi, Mariano, “The Institutional Foundations of Public Policy: A Transactions Theory and an Application to Argentina” (Manuscript, Department of Economics, University of San Andres, Buenos Aires, Argentina, 2006)Google Scholar.

²¹ Eaton, Kent, “Armed Clientelism: How Decentralization Complicated Colombia's Civil War” (Manuscript, Politics Department, University of California-Santa Cruz, 2005)Google Scholar.

²² Ugo M. Amoretti, “Italy: Political Institutions and the Mobilization of Territorial Differences,” in Amoretti and Bermeo (fn. 8).

²³ Roeder, Philip G., “Soviet Federalism and Ethnic Mobilization,” World Politics 43 (January 1991)CrossRefGoogle Scholar; Suny, Ronald Grigor, The Revenge of the Past: Nationalism, Revolution, and the Collapse of the Soviet Union (Stanford, Calif.: Stanford University Press, 1993)Google Scholar; Brubaker, Rogers, Nationalism Reframed: Nationhood and the National Question in the New Europe (Cambridge: Cambridge University Press, 1996)CrossRefGoogle Scholar; Bunce, Valerie, Subversive Institutions: The Design and the Destruction of Socialism and the State (Cambridge: Cambridge University Press, 1999)CrossRefGoogle Scholar.

²⁴ Suny (fn. 23), 87.

²⁵ Bunce (fn. 23), 138. While Bunce writes about the breakdown of communist countries, others have found a similar trend in the postcommunist era. Both Daniel Treisman and Henry Hale find that a high degree of regional autonomy has been positively correlated with separatism in Russia in the 1990s. See Treisman, Daniel, “Russia's ‘Ethnic Revival’: The Separatist Activism of Regional Leaders in a Postcommunist Order,” World Politics 49 (January 1997)CrossRefGoogle Scholar; Hale, Henry, “The Parade of Sovereignties: Testing Theories of Secession in the Soviet Setting,” British fournal of Political Science 30 (January 2000)Google Scholar. It is along similar lines that Jack Snyder argues that federalism (or power sharing) does not represent a means to contain nationalist conflict, as federal institutions lock in elite-driven hostile ethnic identities. See Snyder, , From Voting to Violence: Democratization and Nationalist Conflict (New York: Norton, 2000)Google Scholar.

²⁶ These provisions, the authors argue, send a costly and credible signal that all the parties to the peace agreement recognize “the need of the other side to be secure from coercion at the hands of the government during the initiation phase.” See Hoddie, Matthew and Hartzell, Caroline, “Power Sharing in Peace Settlements: Initiating the Transition from Civil War,” in Roeder, Philip G. and Rothchild, Donald, eds., Sustainable Peace: Power and Democracy after Civil Wars (Ithaca, N.Y.: Cornell University Press, 2005), 102Google Scholar.

²⁷ David A. Lake and Donald Rothchild, “Territorial Decentralization and Civil War Settlements,” in Roeder and Rothchild (fn. 26).

²⁸ For work with a similar argument but focused largely on political parties, see Brancati, Dawn, “Decentralization: Fueling the Fire or Dampening the Flames of Ethnic Conflict or Secessionism,” International Organization 60 (July 2006)CrossRefGoogle Scholar. In general, this is a research agenda that is promoted by Amoretti and Bermeo (fn. 8), but they do not systematically examine the conditions under which peace-preserving federalism works.

²⁹ Nor does a cure versus curse debate allow us to investigate the track record of conflict within federal states. We know, for example, that the conflict score for India is due to conflicts primarily in Assam, Jammu and Kashmir, Mizoram, Nagaland, Punjab, and Tripura—just a few of India's thirtyfive states and union territories. Likewise, Russia's conflict score is due to conflicts and protests mainly in Buryatia, Chechnya, Dagestan, Ingushetia, Karachay-Cherkessia, Sakha, Tatarstan, and Tuva—a small share of Russian's eighty-nine regions.

³⁰ Wheare, K. C., Federal Government (London: Oxford University Press, 1953)Google Scholar.

³¹ Weingast, Barry R., “The Economic Role of Political Institutions: Market-Preserving Federalism and Economic Growth” Journal of Law, Economics, and Organization 11 (Spring 1995)Google Scholar.

³² The authors recognize that implementation of any political institutions requires careful consideration of how those institutions may interact with other institutions and their economic/cultural environment. However, the authors focus on institutions, maintaining that even though the “supergame” of norms, conventions, and culture matters, it “lies outside the realm of conscious design so that we can focus on formal rules and the question of whether choices exist that encourage federal stability regardless of culture.” Filippov, Mikhail, Ordeshook, Peter C., and Shvetsova, Olga, Designing Federalism: A Theory of Self-Sustainable Federal Institutions (Cambridge: Cambridge University Press, 2004), 161CrossRefGoogle Scholar.

³³ Sharma, Brij Mohan, Federalism in Theory and Practice (Chandausi, India: Bhargava, 1953)Google Scholar.

³⁴ Horowitz, Donald, Ethnic Groups in Conflict (Berkeley: University of California Press, 1985), 613–28Google Scholar.

³⁵ Kaplan, Robert D., Balkan Ghosts: A Journey through History (New York: Vintage Books, 1993)Google Scholar.

³⁶ Petersen, Roger D., Understanding Ethnic Violence: Fear, Hatred, and Resentment in Twentieth- Century Eastern Europe (Cambridge: Cambridge University Press, 2002)CrossRefGoogle Scholar.

³⁷ Posen, Barry, “The Security Dilemma and Ethnic Conflict,” Survival 35 (Spring 1993)CrossRefGoogle Scholar; Lake and Rothchild (fn.17).

³⁸ Gagnon, V. P., “Ethnic Nationalism and International Conflict: The Case of Serbia,” International Security 19 (Winter 1994–95)CrossRefGoogle Scholar.

³⁹ Kaufman, Stuart J., Modern Hatreds: The Symbolic Politics of Ethnic War (Ithaca, N.Y.: Cornell University Press, 2001)Google Scholar.

⁴⁰ Hewstone, Miles and Greenland, Katy, “Intergroup Conflict,” International Journal of Psychology 35 (April 2000)CrossRefGoogle Scholar.

⁴¹ Toft, Monica Duffy, The Geography of Ethnic Violence: Identity, Interests, and the Indivisibility of Territory (Princeton: Princeton University Press, 2003)Google Scholar.

⁴² Bermeo (fn. 1).

⁴³ Bunce (fn. 23); Valerie Bunce and Stephen Watts, “Managing Diversity and Sustaining Democracy: Ethnofederal versus Unitary States in the Postcommunist World,” in Roeder and Rothchild (fn. 26); Philip Roeder, “Power Dividing as an Alternative to Ethnic Power Sharing,” in Roeder and Rothchild (fn. 26).

⁴⁴ Hale (fn. 2).

⁴⁵ Muller, Edward N. and Seligson, Mitchell A., “Inequality and Insurgency,” American Political Science Review 81 (June 1987)CrossRefGoogle Scholar.

⁴⁶ Gurr, Ted Robert, Why Men Rebel (Princeton: Princeton University Press, 1970)Google Scholar.

⁴⁷ Gurr (fn. 14).

⁴⁸ Alesina, Alberto, Spolare, Enrico, and Wacziarg, Romain, “Economic Integration and Political Disintegration,” American Economic Review 90 (December 2000)CrossRefGoogle Scholar; Sambanis and Milanovic (fn. 2). In two recent large-N studies of civil war, Fearon and Laitin and Collier and Hoeffler find no support for the proposition that inequality contributes to conflict. See Fearon, James D. and Laitin, David A., “Ethnicity, Insurgency, and Civil War,” American Political Science Review 97 (February 2003)CrossRefGoogle Scholar; and Collier, Paul and Hoeffler, Anke, “Greed and Grievance in Civil War,” Oxford Economic Papers 56 (October 2004)CrossRefGoogle Scholar. The problem with these findings, however, is that the empirical measure for inequality is national Gini coefficients, which measure inequality at an individual level, while the theoretical arguments concern group-level inequality.

⁴⁹ Horowitz (fn. 34), chap. 4.

⁵⁰ Gurr (fn. 14).

⁵¹ Stewart, Frances, “Horizontal Inequalities: A Neglected Dimension of Development,” CRISE Working Paper, no. 1 (Oxford: Oxford University, October 2003), 38Google Scholar; “Policies towards Inequality in Post-Conflict Reconstruction,” CRISE Working Paper, no. 7 (Oxford: Oxford University, March 2005), 42Google Scholar.

⁵² Mancini, Luca, “Horizontal Inequality and Communal Violence: Evidence from Indonesian Districts,” CRISE Working Paper, no. 22 (Oxford: Oxford University, November 2005), 41Google Scholar; Østby, Gudrun, “Horizontal Inequalities, Political Environment, and Civil Conflict: Evidence from 55 Developing Countries,” CRISE Working Paper, no. 28 (Oxford: Oxford University, August 2006), 30Google Scholar.

⁵³ Boix, Carles, Democracy and Redistribution (Cambridge: Cambridge University Press, 2003)CrossRefGoogle Scholar; Acemoglu, Daron and Robinson, James, “Why Did the West Extend the Franchise? Democracy, Inequality, and Growth in Historical Perspective,” Quarterly Journal of Economics 115 (November 2000)CrossRefGoogle Scholar.

⁵⁴ Toft (fn. 41); Walter, Barbara, “Building Reputation: Why Governments Fight Some Separatists but Not Others,” American Journal of ‘Political Science 50 (April 2006)Google Scholar. According to both authors, government's willingness to accommodate ethnic challengers depends on the number of potential ethnic challenges in a country. The higher the number of potential challengers, the less likely governments are to accommodate.

⁵⁵ Sambanis, Nicholas, “The Demand for Sovereignty: From Decentralization to Secession” (Manuscript, Department of Political Science, Yale University, New Haven, April 2006)Google Scholar.

⁵⁶ Krugman, Paul, Geography and Trade (Cambridge, Mass.: MIT University Press, 1992)Google Scholar.

⁵⁷ These demands resulted in a nine-year-long violent conflict with Delhi that began in 1984, after the Indian army attacked the Sikhs’ most important shrine, the Golden Temple in Amritsar.

⁵⁸ See, for example, Judge, Paramjit Singh, Religion, Identity, and Nationhood (Jaipur, India: Rawat Publications, 2005)Google Scholar.

⁵⁹ From the party's Anandpur Sahib Resolution of 1977, reprinted in Grover, Verinder, ed., The Story of Punjab Yesterday and Today, 2nd ed. (New Delhi: Deep and Deep Publications, 1999), 3:307.21.Google Scholar

⁶⁰ As stated by G. S. Dhillon, a Sikh historian who was in charge of writing the Sikh version of the events leading up the Golden Temple massacre in 1984: “(T)he facts indicate that there has been a calculated plan to denude Punjab and its people of its natural wealth and thereby to seriously jeopardize the economic, industrial, and agricultural destiny of the State.” See Dhillon, G. S., India Commits Suicide, 3rd ed. (Chandigarh, India: Singh and Singh Publishers, 2004), 63.Google Scholar

⁶¹ Inman, Robert and Rubinfeld, Daniel, “The Political Economy of Federalism,” in Mueller, Dennis, ed., Perspectives on Public Choice: A Handbook (Cambridge: Cambridge University Press, 1997)Google Scholar; Buchanan, James M., “Federalism as an Ideal Political Order and an Objective for Constitutional Reform,” Publius 25 (Spring 1995)Google Scholar.

⁶² Hechter (fn. 14); Simeon (fn. 8); Liesbet Hooghe, “Belgium: Hollowing the Center,” in Amoretti and Bermeo, (fn. 8). Likewise, Rotimi T. Suberu, who argues that “(f)ederalism remains the lifeblood of Nigeria's survival as a multiethnic country,” considers the country's fiscal overcentralization as a source of conflict. See Suberu, “Nigeria: Dilemmas of Federalism,” also in Amoretti and Bermeo (p. 346).

⁶³ Tiebout, Charles, “A Pure Theory of Local Expenditures,” Journal of Political Economy 64 (October 1956)CrossRefGoogle Scholar; Oates, Wallace, Fiscal Federalism (New York: Harcour Brace Jovanovich, 1972)Google Scholar.

⁶⁴ Hayek, F. A., Individualism and Economic Order (Chicago: University of Chicago Press, 1957)Google Scholar.

⁶⁵ Hechter (fn. 14), 153.

⁶⁶ Linz, Juan J. and Stepan, Alfred, “Inequality Inducing and Inequality Reducing Federalism” (Paper presented at the World Congress of the International Political Science Association, Quebec City, Quebec, Canada, August 1–5, 2000)Google Scholar; Leibfried, Stephan and Pierson, Paul, eds., European Social Policy: Between Fragmentation and Integration (Washington, D.C.: Brookings Institution, 1995)Google Scholar; Swank, Duane, Global Capital, Political Institutions, and Policy Change in Developed Welfare States (New York: Cambridge University Press, 2002)CrossRefGoogle Scholar.

⁶⁷ Prud’homme, Remy, “The Dangers of Decentralization,” World Bank Research Observer 10 (August 1995)Google Scholar.

⁶⁸ Hooghe (fn. 62), 71–73.

⁶⁹ Cai, Hongbin and Treisman, Daniel, “Does Competition for Capital Discipline Governments? Decentralization, Globalization, and Public Policy,” American Economic Review 95 (June 2005)CrossRefGoogle Scholar.

⁷⁰ Peterson, Paul, The Price of Federalism (Washington, D.C.: Brookings Institution, 1995)Google Scholar.

⁷¹ Castles, Francis, Comparative Public Policy: Patterns of Post-war Transformation (Cheltenham: Edward Elgar, 1999)Google Scholar.

⁷² Swank (fn. 66).

⁷³ Oates (fn. 63).

⁷⁴ For example, Rodden, Jonathan, “The Dilemma of Fiscal Federalism: Grants and Fiscal Performance around the World,” American Journal of’ Political Science 46 (July 2002)Google Scholar.

⁷⁵ Wibbels, Erik and Rodden, Jonathan, “Business Cycles and the Political Economy of Decentralised Finance: Lessons for Fiscal Federalism in the EU,” in Wierts, Peter, Deroose, Servaas, Flores, Elena, and Turrini, Alessandro, eds., Fiscal Policy Surveillance in Europe (New York: Palgrave Macmillan, 2006)Google Scholar.

⁷⁶ Suberu (fn. 62).

⁷⁷ Of Russia's thirty-two ethnically defined regions, twenty-five declared sovereignty during the 1990–91 period, and many followed up by either adopting their own constitution, asserting the right to control natural resources, or even—as in the case of Chechnya and Tatarstan—declaring outright independence.

⁷⁸ Treisman, Daniel, After the Deluge: Regional Crisis and Political Consolidation in Russia (Ann Arbor: University of Michigan Press, 2001)Google Scholar. In a small-N study, Aleman and Treisman find limited support for this theory across federations. See Eduardo Aleman and Daniel Treisman, “Fiscal Politics in ‘Ethnically-Mined,’ Developing, Federal States: Central Strategies and Secessionist Violence,” in Roeder and Rothchild (fn. 12).

⁷⁹ Filippov, Ordeshook, and Shvetsova (fn. 32).

⁸⁰ Ibid.; Riker (fn. 4); Stepan (fn. 14).

⁸¹ Filippov, Ordeshook, and Shvetsova (fn. 32).

⁸² Brancati (fn. 28).

⁸³ Khemani, Stuti, “Partisan Politics and Sub-national Fiscal Deficits in India—What Does It Imply for the National Budget Constraint?” (Manuscript, World Bank Development Research Council, Washington, D.C., 2001)Google Scholar.

⁸⁴ Ansolabehere, Stephen and Snyder, James M., “Reapportionment and Party Realignment in the American States,” University of Pennsylvania Law Review 153 (November 2004)CrossRefGoogle Scholar.

⁸⁵ Jones, Mark, Sanguinetti, Pablo, and Tommasi, Mariano, “Politics, Institutions, and Public Sector Spending in the Argentine Provinces,” Journal ofDevelopment Economics 61 (April 2000)Google Scholar.

⁸⁶ Lake and Rothchild (fn. 27).

⁸⁷ Filippov, Ordeshook, and Shvetsova (fn. 32), 189.

⁸⁸ Wilkinson, Steven, Votes and Violence: Electoral Competition and Ethnic Riots in India (Cambridge: Cambridge University Press, 2004)CrossRefGoogle Scholar.

⁸⁹ Bunce (fn. 23).

⁹⁰ See Elazar, Daniel, Federal Systems of the World: A Handbook of Federal, Confederal, and Autonomy Arrangements (Essex, U.K.: Longman Group, 1994)Google Scholar; Watts, Ronald, Comparing Federal Systems (Kingston, Canada.: Institute of Intergovernmental Relations, 1996)Google Scholar; Treisman, Daniel, “Defining and Measuring Decentralization: A Global Perspective” (Manuscript, Department of Political Science, University of California, Los Angeles, 2002)Google Scholar.

⁹¹ The data set builds considerably on that collected by Jonathan Rodden and Erik Wibbels by including region-specific measures of inequality, identifying ethnic-majority regions and the share of those regions governed by the party governing nationally, and adding several cases: Belgium, Ethiopia, Russia, South Africa, and the three communist countries that disintegrated in 1991 and 1992. See Rodden, and Wibbels, , “Beyond the Fiction of Federalism: Macroeconomic Management in Multitiered Systems,” World Politics 54 (July 2002)CrossRefGoogle Scholar.

⁹² Gurr (fn. 14); Sarkees, Meredith Raid, Wayman, Frank Whelon, and Singer, David J., “Inter- State, Intra-State, and Extra-State Wars: A Comprehensive Look at Their Distribution over Time, 1816–1997,” International Studies Quarterly 47 (March 2003)CrossRefGoogle Scholar.

⁹³ Aléman and Treisman (fn. 78). Several countries have regions that are commonly known as ethnic regions even though the ethnic groups in question actually make up a fairly small percentage of the region's population. For example, the Russian Federation has thirty-two regions that are designated as ethnic regions, but in some of these only a small percentage of the population belongs to the ethnic group that the region is named after. In order to capture more of these regions, we construct a second measure for ethnic minority regions where we include regions in which at least one-fourth of the population belongs to an ethnic group that is a national minority. Using this latter measure does not affect our findings. See Appendix 1 for data sources.

⁹⁴ For states where provincial GDP data are not available, we use either provincial income data (Switzerland) or provincial primary school enrollment data (Ethiopia and Nigeria).

⁹⁵ For each country, as an alternative measure for the overlap of ethnic concentration and regional income we calculate the average ethnic region's GDP per capita as a share of the entire country's average GDP per capita. See discussion in fn. 122 below.

⁹⁶ Most important, the IMF counts automatic transfers from national taxes to regional governments as revenue raised by the regions themselves for several of our cases, including Germany, Mexico, and Argentina, and as revenue to the central government. In other words, those revenues are incorrectly double-counted. Given the importance of distinguishing between spending financed by regional taxes and that financed by national transfers for our third hypothesis, we have relied on national sources for these cases.

⁹⁷ Because of Switzerland's unusual executive power, where neither national nor provincial executives are embodied in one person but are instead part of a collegial body of seven persons at the national level and collegial bodies of five to nine members at the provincial level, we code it as follows: In each province, we considered the number of executive members who were from the four parties that have been represented in the national executive since 1959. If all members of a province's executive body were from one, two, three, or four of the parties in the national executive, it was coded as 1. If one of the members of the canton's executive body was from any of the parties represented in the national executive, the measure for copartisanship was “one over total number of seats in the provincial executive” body, and so on. The yearly number in the data set is the average for all the ethnic provinces in that year. For country-years with authoritarian rule, the variable was coded a

⁹⁸ Data until 1999 from Fearon and Laitin (fn. 48), available at http://www.stanford.edu/~jfearon/.

⁹⁹ Bermeo (fn. 1).

¹⁰⁰ Fearon and Laitin (fn. 48).

¹⁰¹ Collier and Hoeffler (fn. 48) see the same indicator as measuring how poor or rich a certain state is and find that richer states are less likely to experience conflict than poor ones. Again, we would think that wealth may be of particular import in federations, at least if the central government is to engage in significant redistribution among the regions of the state.

¹⁰² Ibid.

¹⁰³ See Suberu (fn. 62) on Nigeria.

¹⁰⁴ Treisman (fn. 25).

¹⁰⁵ Gurr (fn. 14).

¹⁰⁶ Data as in fn. 98.

¹⁰⁷ Brancati (fn. 28) and Bermeo (fn. 1).

¹⁰⁸ Clarke, Kevin A., “The Phantom Menace: Omitted Variable Bias in Econometric Research,” Conflict Management and Peace Science 22 (Winter 2005)CrossRefGoogle Scholar.

¹⁰⁹ Clarke's point is that including too many control variables can make the bias on our coefficients of interest equally bad or worse than potentially omitted variable bias.

¹¹⁰ Fearon and Laitin (fn. 48). Relying on the area under the Receiver Operating Characteristic (ROC) curve as a measure for accuracy, we compared our models to that of Fearon and Laitin's model on our sample and violent conflict variables. The area under the ROC curve ranges from 0.5 to 1. Numbers closer to 1 are preferred because 1 indicates that the diagnostic test for the model achieves both 100 percent sensitivity and 100 percent specificity. See King, Gary and Zeng, Langche, “Improving Forecasts of State Failure,” World Politics 53 (July 2001)CrossRefGoogle Scholar. Both in terms of armed conflict and ethnic conflict, we found that our model produced a slightly higher number (for the logit model of armed conflict, 0.9868 compared with 0.9744, and for the logit model of ethnic conflict, 0.9611 compared with 0.9395). Fearon and Laitin did not design their model to explain nonviolent conflict, such as our measure for major ethnic protest.

¹¹¹ Gleditsch, Nils Petter, Wallensteen, Peter, Eriksson, Mikael, Sollenberg, Margareta, and Strand, Havard, “Armed Conflict 1946–2001: A New Dataset,” Journal of Peace Research 39 (September 2002)CrossRefGoogle Scholar; Harbom, Lotta and Wallensteen, Peter, “Armed Conflict and Its International Dimensions, 1946–2004,” Journal of Peace Research 42 (September 2005)CrossRefGoogle Scholar. We use version 3 of the data set.

¹¹² More precisely, 0 is no conflict; 1 is minor conflict, which means more than twenty-five battlerelated deaths per year every year over the course of the conflict; 2 is intermediate conflict, which means more than twenty-five battle-related deaths per year and a total conflict history of more than 1,000 battle-related deaths; 3 is war, which means more than 1,000 battle-related deaths for every year over the course of the conflict. The unit of the analysis in the armed conflict data set is conflict-year. In cases where there were more than one conflict per country-year in our data set, we marked the highest level of conflict.

¹¹³ The data set is maintained by the Center for International Development and Conflict Management at the University of Maryland and is available via the center's Web site, www.cidcm.umd. edu/inscr/mar/ (accessed October 13, 2006). The Minorities at Risk (MAR) data is group-level data, and we use the MARGene software to transform the group conflict scores into country-year indicators. In the event of there being more than one group in conflict in a country in any given year, we used the highest level of conflict. Prior to 1985, the MAR indicators are available on a five-year basis, so we use MARGene's interpolation function for the 1978–1985 period. (We also ran our analysis for ethnic rebellion and protest on the sample including only country-years after 1985, and it did not affect our main findings.)

¹¹⁴ See, for example, Hechter (fn. 14).

¹¹⁵ For each of them, we chose a cut-off point approximately one standard deviation above the average in our sample. For armed conflict, this means that the cut-off point is 1, i.e., minor armed conflict. For ethnic rebellion, we focus on whether in any given year a country has experienced a conflict categorized as local rebellion or a higher level of conflict, such as guerilla activities or protracted civil war. For ethnic protest, we focus on whether in any given year a country has experienced a protest categorized as medium or large demonstration.

¹¹⁶ Franzese, Robert, Kam, Cindy, and Jamal, Amaney, “Modeling and Interpreting Interactive Hypotheses in Regression Analysis,” (Manuscript, Department of Political Science, University of Michigan, Ann Arbor, 2001)Google Scholar; Clark, William, Gilligan, Michael, and Golder, Matt, “A Simple Multivariate Test for Asymmetric Hypotheses,” Political Analysis 14 (Summer 2006)CrossRefGoogle Scholar; Braumoeller, Bear, “Hypothesis Testing and Multiplicative Interaction Terms,” International Organization 58 (Fall 2004)CrossRefGoogle Scholar.

¹¹⁷ Franzese, Kam, and Jamal (fn. 116).

¹¹⁸ An alternative would be to plot conditional coefficients, for example, Clark, Gilligan, and Golder (fn. 116), but we find the conditional predictions more substantively interesting.

¹¹⁹ We use King, Tomz, and Wittenberg's Clarify software to generate these predicted values. See King, Gary, Tomz, Michael, and Wittenberg, Jason, “Making the Most of Statistical Analyses: Improving Interpretation and Presentation,” American Journal of Political Science 44 (April 2000)CrossRefGoogle Scholar.

¹²⁰ Lijphart (fn. IS).

¹²¹ For example, Alesina, Spolare, and Wacziarg (fn. 48); and Sambanis and Milanovic (fn. 2).

¹²² For example, Horowitz (fn. 34), Gurr (fn. 14), and Stewart (fn. 51). In order to try to distinguish among these two hypotheses, we exchanged the interaction term Ethnic Concentrationinterregional Inequality with an indicator measuring the relative wealth of a country's ethnic regions (calculated as the average of the ethnic regions’ wealth over the country's average level of wealth) and found the indicator to be negative but not significant, suggesting that more research is needed to determine whether it is relatively wealthy or poor regions that may spur ethnic rebellions.

¹²³ King and Zeng (fn. 110).

¹²⁴ The area under the ROC curve ranges from 0.5 to 1, and numbers closer to 1 are preferred because 1 indicates that the diagnostic test for the model achieves both 100 percent sensitivity and 100 percent specificity. In our case, for the logit analysis of armed conflict, the area under the ROC curve of the full model was 0.9868, while the corresponding number for the analysis excluding the interaction terms was 0.9845. For the logit analysis of ethnic conflict, the area under the ROC curve for the full model is 0.9611, while the number for the limited model is 0.9477 and the corresponding numbers for ethnic protest are 0.9386 and 0.9282.

¹²⁵ Saideman et al. (fn. 14).

¹²⁶ Filippov, Ordeshook, and Shvetsova (fn. 32), 5.