1 WORLD DEVELOPMENT REPORT 2011 BACKGROUND PAPER HORIZONTAL INEQUALITIES AS A CAUSE OF CONFLICT A Review of CRISE Findings1 Frances Stewart Director, Centre for Research on Inequality, Human Security and Ethnicity (CRISE), Oxford University August 20, 2010 The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the World Development Report 2011 team, the World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. 1 I am very grateful for helpful comments on a previous draft from Andreas Wimmer. For a much fuller discussion of the issues covered here see Stewart (2008). 2 Violent conflict in multi-ethnic and multi-religious countries remains a major problem in the world today. Fortunately, there is plentiful evidence to show that such conflict is not an unavoidable ramification of ethnic difference, an outcome of ‘age-old ethnic hatreds’ as is popularly suggested, nor of an unavoidable ‘clash of civilisations’ (Huntington, 1993). As Fearon and Laitin (2003) have shown most multi-ethnic societies are peaceful. The critical question, then, is why ethnic or religious conflict breaks out in some circumstances and not in others. CRISE (the Centre for Research on Inequality, Human Security and Ethnicity) was established in 2003 at Oxford University, funded by the UK Department for International Development (DFID), with the basic aim understanding why many multi-ethnic and multi-religious societies countries are peaceful while others experience violent conflict, working in partnership with local scholars in three regions: Latin America (Bolivia, Guatemala and Peru), Southeast Asia (Indonesia and Malaysia) and West Africa (Côte d’Ivoire, Ghana and Nigeria). In each region, one of the countries studied has avoided serious national conflict while the remainder have experienced severe violent conflict in the recent, or relatively recent, past. In addition, the work has drawn on the experience of other countries and on econometric investigations covering a broader range of countries. CRISE research is multidisciplinary, and conducted jointly with partners in the countries in which we have worked. CRISE research has been primarily devoted to investigating the hypothesis that the presence of major ‘horizontal inequalities’ (HIs) or inequalities among culturally defined groups significantly raises the risk of conflict (Stewart 2000).2 This theory is based on the notion that ‘when cultural differences coincide with economic and political differences between groups, this can cause deep resentment that may lead to violent struggles’ (Stewart and Brown, 2007, p. 222)3. The hypothesis thus straddles two important strands in the literature on conflict: on the one hand, that ethnic or religious differences are at the heart of problem (the cultural hypothesis) (e.g. Huntington 19964); and on the other, that cultural differences are superficial and instrumentalised, and it is economic factors (or political), in which the fundamental roots are to be found (e.g. Collier and Hoeffler 2004; Fearon and Laitin 1996). According to the HI hypothesis it is a combination of cultural differences and political and economic inequalities running along cultural lines that, in part at least, explain contemporary violent conflict. If this is correct, it suggests important policy implications, for development policy generally as well as for policy in conflict-affected countries. HIs are inequalities among groups of people that share a common identity. Such inequalities have economic, social, political and cultural status dimensions, each of which contains a number of elements – some which may matter to people in some societies, but not in others.5 Economic HIs include inequalities in access to and ownership of assets—financial, human, natural resource-based and social, and also inequalities in income levels and employment opportunities, which depend on such assets and the general conditions of the economy. Social HIs include inequalities in access to a range of services, such as education, health care and housing, as well as to the benefits of educational and health care outcomes. Political HIs include inequalities in the distribution of political opportunities and power among groups, including control over the army, the cabinet, local and regional governments, parliamentary assemblies, the police and the presidency. They also encompass inequalities in people’s capabilities to participate politically and to express their needs. Cultural status HIs include disparities in the recognition and standing of different groups’ language, religion, customs, norms and practices. 2 Horizontal inequality differs from ‘vertical’ inequality (VI) in that the latter is a measure of inequality among individuals or households, not groups—furthermore, measurement of VI often is confined to income or consumption. 3 There are clear synergies between the concept of HIs and other approaches to understanding inequalities and the dynamics of mobilisation. For instance, Charles Tilly’s (1998) concept of ‘categorical inequalities’ describes similar group inequalities. Ted Gurr’s (1993) concept of ‘relative deprivation’ as a cause of minority rebellion represents a similar view about mobilisation. However, the HI hypothesis differs from relative deprivation in its view that the relatively rich, as well as the relatively poor, may initiate conflict. In Burundi, for example, the Tutsis have attacked the poorer Hutus; and the relatively rich area of Biafra initiated the Nigerian Civil War of the late 1960s. Such incidents seem to be motivated by fear that an existing situation is not sustainable without force and that the relative prosperity of the group is, or may be, subject to attack. 4 Huntington, of course, was primarily concerned with international rather than intra-national relations. 5 The relevance of any element depends on whether it is an important source of income or well-being in a particular society. For instance, the distribution of housing (a key source of discord between Catholics and Protestants in Northern Ireland in the 1970s) is likely to be more relevant in an industrialised country than in a country where people still build their own homes. 3 Any type of horizontal inequality can provide an incentive for political mobilisation, but political inequalities (that is, political exclusion of particular groups) are most likely to motivate group leaders to instigate a rebellion, as can be seen, for example, in the recent case of Kenya (Stewart 2010). By contrast, economic and social inequalities and those of cultural status are more likely to motivate the mass of the population. Moreover, cultural status inequalities bind groups together and thereby increase the salience of identity differences. Of course, political inequalities themselves may be partly responsible for other inequalities, for example as a result of clientilism and patronage, and decisions about the distribution of public services, as well determining cultural status inequalities. Some evidence. Evidence on the significance of HIs for conflict is drawn from both case studies6 and econometric investigations. The Côte d’Ivoire situation is one example that powerfully illustrates the connection between political and socioeconomic inequalities. Severe socioeconomic inequalities have long existed between the north and the south of the country, but an inclusive government under President Félix Houphouët-Boigny managed to keep the peace for several decades. After his death in December 1993, a number of important factors, including the political exclusion of northerners—the debarring of a northern politician from standing for president and the exclusion of many others from the rights to citizenship and voting— eventually led to civil war. The rebels’ Charte du Nord clearly expressed the economic grievances of northerners as well as their resentment at insufficient state recognition of the Muslim religion (Langer 2005).7 Econometric evidence has been produced both across and within countries, and concerns both socio- economic and political HIs, but data limitations constrain the range of countries covered. Socio-economic inequalities. Østby’s (2008) analysis across 55 countries for 1986–20038, reveals a significant rise in the probability of conflict in countries with severe economic and social HIs. In her models, she defines groups alternatively by ethnicity, religion and region, and finds a significant relation between HIs and the onset of violent conflict for each definition. Economic HIs are measured by average household assets and social HIs by average years of education. The effect of HIs is quite high: the probability of conflict increases threefold when comparing the expected conflict onset when all variables have average values, compared to a situation where the extent of horizontal inequality of assets among ethnic groups is at the 95th percentile.9 Cederman et al., 2010, again using cross-country evidence and the G-Econ data set for1991 to 2005, show that ‘groups with wealth levels far from the country average are indeed more likely to experience civil war’ (24). This is found whether the group is wealthier or poorer than the average. Brown (2010a) has come to a similar finding exploring the determinants of separatist conflicts. He covers 31 countries, from East and Western Europe, North and South America and South and East Asia. Again, the likelihood of a separatist conflicts increases the richer or poorer a region is in terms of GDP per capita, compared with the national average. In addition, intra-country studies demonstrate a positive relationship between the level of HIs and the incidence (or intensity) of conflict. Mancini (2008) uses district-level data to examine the connection between HIs and the incidence of conflict in districts of Indonesia. After controlling for a number of intervening factors, including economic development, ethnic diversity and population size, he finds that horizontal inequality in child mortality rates and its change over time are positively (and significantly) associated with the occurrence of deadly ethno-communal violence. Other measures of HI, in civil service employment, education, landless agricultural labour and unemployment, were also related to incidence of conflict, but the effects were less pronounced than those of child mortality. The Indonesian results suggest, too, that violent conflict is more likely to occur in areas with relatively low levels of economic development and greater religious polarisation. In contrast, standard measures of (vertical) income inequality as well as other purely demographic indicators of ethnic diversity were found to have no significant impact on the likelihood of communal violence. Studies in other conflict-affected countries have shown a relation between HIs and intensity of conflict. In an examination of the Moro rebellion in the southern Philippines, Magdalena (1977) records a strong link between the relative deprivation of Muslims, measured in terms of 6 E,g, Rwanda; Sudan; Kenya; N.Ireland; Peru; Guatemala; Nigeria. Many more case studies are to be found in Stewart 2008; and CRISE Working Papers (www.crise.ox.ac.uk) 7 A large amount of evidence has also been collected to test the ‘relative deprivation’ hypothesis of Gurr using the Minorities at Risk data set (e.g. Gurr 1993; Walter 2006; Gurr and Moore 1997). However, the data in these studies are confined to minorities judged to have been at risk. For this reason, we do use their extensive investigations here. 8 The number of countries is limited by the availability of group data. She uses the Armed Confict data base (ACD) for conflict incidence and Demographic Health Surveys (DHS) for socio-economic data. 9 In the case of inter-regional HIs, the probability of conflict increases 2.5 times as HIs rise from the mean value to the 95th percentile value. See also Østby (2003). 4 differential returns to education, and conflict intensity. Murshed and Gates (2005), using a ‘gap’ measure of human development, note strong econometric support for a relationship between regional deprivation and the intensity of the Maoist rebellion across districts of Nepal. A later study by Do and Iyer (2007) replicates the finding that conflict intensity is related to regional deprivation, although in this case it is measured by the regional poverty rate and the literacy rate. They also indicate that caste polarisation affects conflict intensity. It is of course, perceptions which motivate people to action. Kirwin and Cho investigating perceptions in 17 African countries covered by the Afrobarometer found that , among other factors, ‘group grievances are strongly associated with both popular acceptability of political violence and higher levels of participation in demonstrations’ (Kirwin and Cho 2009: 1), where group grievances are defined as how often a respondent’s ethnic group is seen to be treated unfairly by the government. Political HIs. The data limitations on political HIs are even more severe than for socio-economic. However, Wimmer, Cederman and Min (2009) compiled a global data base for 1946–2005 and show that countries with high degrees of political exclusion are more likely to experience armed conflict. Cederman et al, 2010, confirm this finding, showing that political HIs (defined as exclusion from political power of major groups) adds to economic HIs in raising the risk of conflict. Østby, however, does not find an independent influence of her measure of political exclusion10 and conflict onset, but she finds a strong interaction effect between HIs in regional assets and political exclusion with a rise in the likelihood of conflict, for a given level of HIs, if there is political exclusion (Østby 2008) Other evidence on how far institutions which appear likely to be more inclusive politically, such as proportional representation (PR), federalism and decentralisation, produce ambiguous results, depending on the country situation (particularly the size and number of groups) and the design of the institutions. For example, , Reynal-Querol (2002; 2005) finds that PR has a positive influence in reducing conflict propensity and Schineider and Wiesehomeier (2008) finds that ‘controlling for the presence of the ethnic structure of a country, more inclusive arrangements pacufy intrastate relations’ ((p198), but this is not supported by Gleditsch et al., 2009. Federalism can reduce conflict propensity, but much depends on its design (Bakke and Wibbels (2006); Brown, 2010b), Bakke and Wibbels conclude that ‘co-partisanship’ between central and sub-national governments, which implies shared political power (regionally) and consequently lower political HIs, significantly reduces the chance of conflict. 11 Decentralisation of power to local levels within a unitary state has been argued to reduce national level conflicts, but often at the expense of increasing local level ones (Brancati 2006; Green 2008; Diprose 2009; Tranchant 2009). Case study evidence, as in the Cote d’Ivoire case above, suggests that it is where there are both socio- economic and political HIs, in the same direction that conflict is most likely. Conversely, where one group has political power and another is economically privileged (as in Malaysia and for much of the time Nigeria), or governments are broadly inclusive, conflict seems to be less likely. While this has not been tested systematically econometrically, for data reasons, some limited indirect evidence is available: for example, Barrows’ (1976) investigation of Sub-Saharan African countries in the 1960s finds consistently positive correlation between HIs and political instability across 32 Sub-Saharan African countries in the 1960s, with measures of inequality including share of political power and socioeconomic variables; and Østby (2008) reports that while political exclusion on its own (as an independent variable) does not affect the likelihood of conflict, it has a strong effect when calculated as an interactive variable with inter-regional asset inequality. One important finding that highlights the difficulties in tackling HIs is that they can endure over very long periods. Analysis of the histories of a range of countries (including Bolivia, Brazil, Côte d’Ivoire, Ghana, Guatemala, India, Nigeria and Peru) revealed such long-term persistence of inequality, with some groups remaining relatively deprived for centuries. Many HIs originated in colonial policies that privileged particular groups, but on-going interlocking elements sustain them. Different types of inequality reinforce one another, with inequality in one sphere, such as access to various types of capital—including education, finance, land and social assets—making inequalities in other realms more likely. Discrimination and political exclusion compound these inequalities. Such interlocking elements have made it almost impossible for some groups to escape deprivation. Comprehensive policies are needed to tackle these complex problems. Policy Implications 10 Defined according to the nature of the political system, not by evidence on actual exclusion as in Wimmer et al. 2009; and Cederman et al. 2010. 11 Suberu (2009) shows how federalism Nigeria has contributed to solutions to difficult and potentially conflict- provoking issues. 5 These findings have important implications for development policy. They suggest that policies to correct economic, social and political HIs should be prioritised in multi-ethnic societies—as part of general development policies—especially in post-conflict environments. However, the international policy community too often is blind to the issue of horizontal inequality. None of the issues on the leading agenda—notably, poverty reduction, promotion of economic growth and structural adjustment— incorporates consideration of HIs. However, the issue is increasingly acknowledged in post-conflict environments, although policies tend to be confined to a narrow range of issues and have limited impact, in most cases, in practice (DFID 2005; Fukuda-Parr forthcoming; Langer et. al., forthcoming). There is a much higher level of consciousness of the importance of HIs in the national policies of some heterogeneous countries, such as Malaysia, India or Brazil, and these states have adopted a range of different policy approaches. Similarly, as far as political systems and initiatives are concerned, Westerners generally advocate multiparty democracy and governance reforms, such as improved accountability and transparency. In practice, though, multiparty democracy can lead to exclusionary politics in heterogeneous societies and consequently elections can provoke violence. Again, the need for power sharing is acknowledged more frequently in post-conflict societies, such as Bosnia-Herzegovina, Iraq and Lebanon. Wider awareness of the need to reassess the design of democratic systems in multi-ethnic settings is, however, rare. Yet from a preventative perspective, there is a need to tackle socio-economic, political and cultural status HIs in all multiethnic societies, not only those that have already suffered violent conflict. A dearth of international statistics on the issue reflects (as well as causes) this lack of focus. Neither the United Nations Development Programme (UNDP) nor the World Bank includes statistics on ethnic, regional or religious HIs in their well-known data sets, although some national-level Human Development Reports provide ethnically or religiously disaggregated data. A notable exception is the Demographic and Health Surveys (supported by USAID and other donors) —55 of 77 countries covered to date (September 2009) have included an ‘ethnic’ variable. Østby’s analysis, cited above, uses this source. These surveys, though, do not cover all countries, do not include many relevant variables, and they are not conducted at regular intervals. The evidence supporting the view that HIs increase the risk of violent conflict, especially in cases where they are consistent across the, economic, social, political and cultural status realms, suggests that correction of such inequalities should be a major policy aim in any multi-ethnic society where HIs are severe. Moreover, this is important from the perspective of efficiency, justice and well-being, as well as for reducing the risk of conflict. Policies are needed that address specific inequities, including exclusionary political systems and inequalities in assets, employment opportunities, income levels and cultural status . At the same time, though, policies also must remedy the deeper causes of horizontal inequalities, particularly long embedded factors such as discrimination and prejudice. In addition, they need to improve the capabilities of members of deprived groups and enhance their ability and confidence to exercise their rights. There are three distinct approaches to managing HIs: Direct approaches—groups are targeted directly through quotas for the allocation of jobs, distribution of assets or educational access, for example. These can be effective, but they risk increasing the salience of identity difference and antagonising those who do not benefit from the policy initiative. Generally, direct approaches should be of limited duration to avoid mounting opposition and the corruption that is often associated with their execution. Indirect approaches—these are general policies that have the effect of reducing group disparities, such as anti-discrimination policies, policies to decentralise power, progressive taxation or regional expenditure policies. Such measures are less likely to increase the significance of identity, but they may be less effective in reducing HIs. ‘Integrationist’ approaches—these aim to diminish the salience of group boundaries by, for instance, promoting national identity and shared activities across groups. These policies decrease the significance of group differences, but they can conceal rather than reduce inequalities. These are illustrated in the matrix below. 6 Policy approach Direct HI-reducing Indirect HI-reducing Integrationist D im en si o n Political Group quotas; seat reservations; consociational constitution; ‘list’ PR Voting system designed to require power sharing across groups (for instance, two-thirds voting requirements in an assembly); specification of boundaries and seat numbers to ensure adequate representation of all groups; human rights legislation and enforcement Geographical voting spread requirements; ban on ethnic/religious political parties (national party stipulations) Socioeconomic Quotas for employment or education; special investment or credit programmes for particular groups Anti-discrimination legislation; progressive taxation; regional development programmes; sectoral support programmes (for example, Stabex) Incentives for cross- group economic activities; requirement that schools are multicultural; promotion of multicultural civic institutions Cultural status Minority language recognition and education; symbolic recognition (for example, public holidays and attendance at state functions) Freedom of religious observance; no state religion Civic citizenship education; promotion of an overarching national identity The cases of Malaysia and Northern Ireland illustrate how a comprehensive effort can improve economic and social HIs and contribute to reduced conflict propensity. Both employed a combination of direct and indirect approaches, although arguably neither put enough emphasis on integrationist policies. In the case of Malaysia, the New Economic Policy adopted in 1970, following serious violence against the Chinese systematically favoured the Malays in economic activities and in education and the gaps have narrowed. During the 1997 financial crisis, neighbouring countries suffered violence directed at the Chinese but this did not occur in Malaysia. In N.Ireland, after centuries of discrimination and inequalities of all kinds, from the late 1970s anti- discrimination policies began to take effect, closing the gaps markedly over the subsequent twenty years, which was probably an important factor in permitting peace effective peace negotiations. Chart 1a and b illustrates. Yet relations between the communities remain poor according to surveys, and this may be due to the policies themselves and to the absence of active (integrationist) policies to improve relations. 7 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1970 1979 1990 1999 Malaysia: Mean incomes relative to average Bumipu tera Chinese Indian Finally, one should note some caveats concerning policies to correct HIs: These are not the only policies needed. Wherever possible, they should complement other development policies focused on employment expansion, economic growth and poverty reduction; where there are trade-offs, priorities will need to be determined. There is no one-size-fits-all approach to HIs. It is essential to understand the nature and extent of HIs in a particular context in order to design appropriate and effective policies, requiring the gathering and analysis of data. It is important that decision-makers are conscious of and sensitive to the tensions and controversies that might arise following the implementation of policies aimed at redistributing resources among groups. In summary, policies to correct HIs are desirable to maintain peace and security and to establish a just and inclusive society, but one needs to introduce them with care and sensitivity. References Bakke, K. M. and Wibbels, E. (2006) Diversity, Disparity, and Civil Conflict in Federal States. World Politics 59:1 , pp. 1-50 Brancati, D. (2006) ‘Decentralisation: fueling the fire or dampening the flames of ethnic conflict’, International Organisation, 60, 651-685. Booth, D. and Z. Curran (2005) Aid instruments and exclusion. London: Overseas Development Institute. Brown, G.K. and F. Stewart (2006) ‘The Implications of Horizontal Inequality for Aid’. CRISE Working Paper, No. 36. Oxford: CRISE, University of Oxford. 8 Brown, G.K. (2010a), ‘Inequality, ethnicity and violent secessionism: A mixed-level and survival analysis approach to disaggregating the study of civil war’ CRISE Working Paper 80. Brown, G.K. (2010b), ‘Federalism, Decentralisation and Horizontal Inequalities’, CRISE Briefing Paper 3; www.crise.ox.ac.uk Cederman, L.A.. Gledditsch, K.S. and Weidmann, N.B., 2010 ‘ Horizontal inequalities and Ethno-Nationalist Civil War: a Global Comparison’, paper prepared for presentation at Yale University April 2010. Collier, P., and Hoeffler, A. 2004. Greed and grievance in civil war. Oxford Economic Papers 56(4): 563-595. DFID (Department for International Development) (2005) Fighting poverty to build a safer world: A strategy for security and development. London: DFID. Diprose, R. (2009) Decentralisation, horizontal inequalities and conflict management in Indonesia. Ethnopolitics 8:1 Do, Q.T. and L. Iyer (2007) ‘Poverty, Social Divisions and Conflict in Nepal’. World Bank Policy Research Working Paper, No. 4228. Washington, DC: World Fearon, J.D. and Latin, D.D. (2003)’Ethniciity, insurgency and civilwar’, American Political Science Review, 97,1, 7590. Fukuda-Parr, S. forthcoming. Poverty Reduction Strategy papers (PRSPs) and Horizontal Inequality’ in Langer et l. (eds.) Post-Conflict Reconstruction and Horizontal Inequalities. Basingstoke: Palgrave Macmillan. Gleditsch, Nils Petter, Håvard Hegre, and Håvard Strand. (2009). "Democracy and civil war." Pp. 155-192 in Handbook of War Studies III: The Intrastate Dimension, edited by Manus I. Midlarsky. Ann Arbor: University of Michigan Press. Green, E.D. (2008) ‘Decentralisation and conflict in Uganda’ Conflict, Security & Development, 1478-1174, 8,4, 427 – 450 . Gurr, T.R. (1993) Minorities at Risk: A Global View of Ethnopolitical Conflicts. Washington, DC: United States Institute of Peace Press. Gurr, T.R. and Moore, W.H. (1997), ‘Ethnopolitical rebellion: a cross-sectional analysis of the 1980s with risk assessments for the 1990s’ American Journal of Political Science, 41, 4, 1079-1103. Huntington, S.P. (1993) ‘The Clash of Civilisations?’ Foreign Affairs, 72(3), pp. 22–49. Kirwin, M. and W. Cho (2009) ‘Weak States and Political Violence in Sub-Saharan Africa’ , AfroBarometer Working Papers 111. Langer, A. (2005) ‘Horizontal Inequalities and Violent Group Mobilisation in Côte d'Ivoire’. Oxford Development Studies, 33(1), pp. 25–45. Langer, A., F. Stewart and R. Venugopal (eds.) (forthcoming) Post-Conflict Reconstruction and Horizontal Inequalities. Basingstoke: Palgrave Macmillan. Magdalena, F.V. (1977) ‘Intergroup Conflict in the Southern Philippines: An Empirical Analysis’. Journal of Peace Research, 14(4), pp. 299–313. Mancini, L. (2008) ‘Horizontal Inequality and Communal Violence: Evidence from Indonesian Districts’. In F. Stewart (ed.) Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan. Pp 106-135. Murshed, S. and S. Gates (2005) ‘Spatial-Horizontal Inequality and the Maoist Insurgency in Nepal’. Review of Development Economics, 9(1), pp. 121–134. OECD (Organisation for Economic Co-operation and Development) (2001) The DAC Guidelines: Helping Prevent Violent Conflict. Paris: OECD. Østby, G. (2003) Horizontal Inequalities and Civil War: Do Ethnic Group Inequalities Influence the Risk of Domestic Armed Conflict? Cand. Polit. Thesis. Trondheim: Department of Sociology and Political Science, Norwegian University of Science and Technology. Østby, G. (2008) ‘Inequalities, the Political Environment and Civil Conflict: Evidence from 55 Developing Countries’. In F. Stewart (ed.) Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan. Pp136-159. Reynal-Querol, M. (2002) ‘Ethnicity, Political Systems, and Civil Wars’. Journal of Conflict Resolution, 46(1), pp. 29–54. Schneider,G. and Wiesehomeir, N. (2008), ‘Rules that matter: political institutions and the diversity-conflict enxus’ Journal of Peace Research, 45, 2, 183-203. Stewart, F. (2008c) Note for Discussion: Kenya, Horizontal Inequalities and the Political Disturbances of 2008. Oxford: CRISE, University of Oxford. Stewart, F. (ed.) (2008) Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan. Stewart, F. 2000. The root causes of humanitarian emergencies. In Nafziger, W.E., Stewart, F., and Väyrynen, R., eds, War, Hunger and Displacement: The Origins of Humanitarian Emergencies, Vol. 1. Oxford: Oxford University Press. Stewart, F. 2010, ‘Horizontal Inequalities in Kenya, and the Political Disturbances of 2008: some implications for aid policy, Conflict, Security and Development. Suberu, R. (2009) ‘Religion and Institutions: Federalism and the Management of Conflicts over Sharia in Nigeria’. Journal of International Development, 21(4), pp. 547–560. Tilly, C. (1998) Durable Inequality. Berkeley, CA: University of California Press. Tranchant, J.-P. (2009) ‘Fiscal decentralisation, institutional quality, and ethnic conflict: a panel http://www.crise.ox.ac.uk/ 9 data analysis’, 1985-2001. Conflict, Security and Development 8:4, 491-514. Wimmer, A., L.-E. Cederman and B. Min (2009) ‘Ethnic Politics and Armed Conflict: A Configurational Analysis of a New Global Data Set’. American Sociological Review, 74 (April), pp. 316–337.