Ethnicity 1/31 Ethnicity, Cleavages and Conflict H˚avard Mokleiv Nyg˚ard Peace Research Institute Oslo March 10, 2015 Ethnicity Introduction 2/31 Summary of lecture Main question: Do ethnic cleavages matter for conflict, and how? Main point: Ethnic cleavages matter in combination with relative inequality between groups Overview of lecture: 1. Context and conventional wisdom 2. Concepts 3. Research questions and popular theories 4. What do the empirics say?: Country-level studies (the ‘sceptics’) Group-level studies (against the ‘sceptics’) Micro-level research 5. Summary and discussion Ethnicity Introduction 3/31 Context A very popular claim that ethnic cleavages generate internal armed conflict Very popular in the 1990s (Bosnia, Rwanda etc.) Robert D. Kaplan‘s ‘The Coming Anarchy’ article, 1993 Samuel Huntington‘s ‘Clash of Civilizations’ Still extremely widespread (e.g right-wing immigration sceptics in Norway) Main ideas: a) Many conflicts are driven by ‘ancient hatreds’ between groups → b) ethnic cleavages generate inter-group antagonisms → c) ethnic heterogeneity will lead to conflict This lecture will show that these claims must either be: Rejected, or (more likely)... qualified: Ethnic cleavages matter in combination with political and economic context Ethnicity What is an ethnic group 4/31 Concepts: Ethnic group What is ethnicity? What is an ethnic group? First, an ethnic group needs a common marker Second, a group needs a common identity Ethnicity What is an ethnic group 4/31 Concepts: Ethnic group What is ethnicity? What is an ethnic group? First, an ethnic group needs a common marker Religion (i.e Alawites in Syria) Language (i.e French-Quebequouis in Canada) Common origin or nationality (i.e Indians in South Africa) Second, a group needs a common identity The members of the group must actively identify themselves as members Shared sense of belonging to that group Often also common narratives relating to history etc. Ethnic groups are “imagined communities” (e.g Benedict Anderson) (BUT, not epistemically subjective! Ethnic groups exist) External attribution is often also mentioned (others must also identify the group as a group) Ethnicity What is an ethnic group 5/31 Concepts: Ethnic conflict What constitutes ethnic conflict? The definition most of the studies stick to: ‘Internal conflict between a government and an organized internal challenger (Uppsala definition!) in which the challenger is defined (and defines itself) along ethnic lines.’ Another common definition (from the MAR project) ‘episodes of violent conflict between governments and national, ethnic, religious, or other communal minorities (ethnic challengers) in which the challengers seek major changes in their status’ Ethnicity What is an ethnic group 6/31 Why ethnic conflict: ancient hatred and the security dilemma Why do groups fight? (‘onset’ of conflict) 1 Ancient hatreds (Kaplan) Historical rivalries, Retribution and revenge Almost tautological! 2 Security dilemmas (Posen) Since groups (where the state is weak or predatory) cannot be sure that other groups are benign, and vice versa This leads to a security dilemma where groups attack each other pre-emptively (offense is the best defense) Ethnicity Classics 7/31 Why ethnic conflict: the state and relative deprivation 3 State institutions The Lijphart-Horowitz debate Lijphart: Majoritarian ‘winner-take-all’ institutions lead to ethnic conflict Horowitz: Lijphartian power sharing institutions ‘reify’ and harden ethnic cleavages, leading to more conflict Note: Przeworski’s argument about the fundamental un-democratic nature of power sharing 4 Relative deprivation (Gurr) ‘Why Men Rebel’ Frustrations relating to relative deprivation generate conflict Relative deprivation: a discrepancy between ‘the goods and conditions of life to which people believe they are rightfully entitled’ and the ‘goods and conditions they think they are capable of attaining or maintaining, given the social means available to them’. This is the most developed theoretical framework Ethnicity Classics 8/31 Why ethnic conflict: opportunity and feasibility Opposing framework: Opportunity theory (‘ethnic-grievance skepticism’) Tilly (and Skocpol): Grievances are ubiquitous – what is needed is a minimum level of resources and organization Ethnic (and other) antagonisms are endemic, opportunities for conflict are not Ethnic grievances not important Focus on economic opportunities for conflict (e.g Collier and Hoeffler 2004) or feasibility (e.g Fearon and Laitin 2003) Ethnicity Three waves 9/31 Three ‘waves’ of research 1 The skeptics: Seminal studies (on armed conflict) are Collier and Hoeffler (2004) and Fearon and Laitin (2003) Country-level studies, looking at civil conflict Using Ethnolinguistic Fractionalization (ELF) as a measure of country-level ethnic heterogeneity General finding: Ethnic heterogeneity does not increase risk of civil war 2 The horizontal-inequality wave: Østby (2008), and Cederman and colleagues Finds that substantive inequalities (political and economic) between groups lead to more conflict Most research here is at the group level, using the Ethnic Power Relations (EPR) dataset 3 Micro-research For example the contributions of Lyall on the syllabus Less interested in whether ethnic grievances matter, and more in how they matter Ethnicity Three waves 10/31 Empirical Research: The skeptics Figure: A prominent representative of “grievance-skepticism”: Professor Paul Collier Ethnicity Three waves 11/31 Empirical research: The skeptics The skeptics Use the ELF, which measures the: ‘probability that two randomly drawn individuals in a country are from different ethnolinguistic groups’ (Fearon and Laitin, 2003, p.78) FL and CH find no statistically significant effects of ELF on civil war But, CH find an effect of ‘Ethnic dominance’ (i.e one group being in a majority) Some studies using alternative fractionalization measures are less skeptical (e.g Montalvo and Reynal-Querol 2005) The ‘take home point’ in the wake of FL and CH is that ethnic cleavages matter less than what is commonly believed Ethnicity Three waves 12/31 Empirical Research: The skeptics Figure: New York Times coverage of Fearon and Laitin Ethnicity Three waves 13/31 Empirical research: The horizontal inequality wave The horizontal inequality wave Argues against looking at ELF at the country-level Not enough to count number of ethnic groups! We should look at substantive (economic and political) inequalities between groups The book on the syllabus sums up a large body of this research Note: two crucial recent papers: Cederman, Wimmer and Min 2010: ‘Why Do Ethnic Groups Rebel?’ Cederman, Weidmann, Gleditsch 2011: ‘Horizontal Inequalities and Ethno-Nationalist Civil War: A Global Comparison’ Ethnicity Three waves 14/31 Empirical research: The horizontal inequality wave Why do Ethnic Groups Rebel? Core question: Does political exclusion/inclusion increase the risk of ethnic conflict Introduces the EPR dataset Codes all ‘politically relevant’ ethnic groups in the world, 1946-2005. Rely on theories of relative deprivation The ‘polity model’: A government versus several contenders Excluded versus included groups Ethnicity Three waves 15/31 Empirical research: Why do Ethnic groups Rebel? Figure: Cederman, Wimmer and Min‘s Polity model Ethnicity Three waves 16/31 Empirical research: Why do Ethnic groups Rebel Independent variables: Political exclusion (not being included in government, or regional autonomy) Political ‘downgrading’ Dependent variable: Ethnic Armed Conflict onset: UCDP internal conflicts, where challenger “pursued ethnonationalist aim” Ethnicity Three waves 17/31 Empirical Research: Why do Ethnic Groups Rebel? Figure: Findings in Cederman and collegues Ethnicity Three waves 18/31 Empirical Research: Why do Ethnic Groups Rebel? Figure: Findings in Cederman and colleguesn Ethnicity Countries and groups 19/31 From countries to groups Cederman and colleagues show that politically excluded groups are more prone to conflict And that groups who lose power are also more likely to rebel Discussion: The ‘politically relevant ethnic group’ criterion The dependent ‘ethnic conflict’ variable Coding ‘power status’ The endogeneity of inclusion/exclusion (this is affected by conflict-potential!) Ethnicity Countries and groups 20/31 From countries to groups Core question: Does economic inequality between groups lead to conflict? They look at horizontal inequality (overlapping cleavages in Rokkan‘s terminology) Test a ‘grievance’ argument: Relative inequality between groups lead to conflict Ethnicity Countries and groups 21/31 From countries to groups How do they measure group-level wealth? Use GIS Combine information about where groups reside, with geographical estimates of local economic activity Dependent variable: Ethnic Armed conflict between a group and the state Independent variables: The ration between the GDPpc of the given group from the average GDPpc in the country One variable measuring this ratio when it is lower than the average Another capturing this variable when it is higher Ethnicity Countries and groups 22/31 From countries to groups Figure: Combining Spatial Wealth Estimates with The Location of Ethnic Groups Ethnicity Countries and groups 23/31 From countries to groups Figure: What Cederman et.al Find: Horizontal inequality Ethnicity Countries and groups 24/31 From countries to groups Figure: What Cederman et.al Find: Inequality and exclusion Ethnicity Countries and groups 25/31 From countries to groups Cederman and colleagues find that groups that are poorer or richer than country average are more prone to rebel Discussion: This shows strong support for ‘grievance explanations’ Potential measurement problems? Causal mechanisms Ethnicity Ethnicity in war 26/31 Empirical research: Micro-studies In addition to country- or group-level studies, there has also been a surge of micro-studies These look at mechanisms, and ask how ethnicity matters, not just whether it matters at the macro level Research questions in micro-literature: Where does ethnic violence occur? Mobilization/coordination within ethnic groups Ethnic collaboration/defection Two examples: Lyall 2010, “Are Coethnics More Effective Counterinsurgents? Evidence from the Second Chechen War.’ See also: Weidmann 2011, ‘Violence ‘from above’ or ‘from below’? The Role of Ethnicity in Bosnia’s Civil War.’ Ethnicity Ethnicity in war 27/31 Empirical research: Micro-studies, Lyall 2010 Lyall, 2010 Core question: Are ethnic insurgents more effectively fought by members of the same ethnic group Case: The Russian war in Checnya, 2000-2005 Points: Finds that Pro-Russian chechens are better at conducting ‘sweep’ operations in insurgent villages in Chechnya (2000-5). Argues that members of same ethnic groups have information advantage when it comes to identifying insurgents Ethnicity Ethnicity in war 28/31 Empirical research: Micro-studies, Lyall 2010 Figure: Findings from Lyall, 2010 Ethnicity Ethnicity in war 29/31 Empirical research: Micro-studies, Lyall 2010 Why should co-ethnics be more effective counterinsurgents? Lyall suggests a number of mechanisms: 1 Because Chechens are more likely to cooperate with co-ethnics? Unlikely, since these are seen as traitors 2 Chechen tactics more effective? Chechen units for example use kidnapping more often But, problematic since these tactics have no effect in the data 3 Being of the same ethnicity as the insurgents reduces uncertainty Co-ethnics can access existing social networks more easily, to obtain information Co-ethnics can find who they are looking for more easily 4 ‘To catch a thief’ mechanism: Prior rebel experience Pro-Russian Chechens more likely to have previously been insurgents They know the insurgents better A comparison of defector versus non-defector units speaks against this mechanism Ethnicity Ethnicity in war 30/31 Summary: Where are we now? Summary: What do we know about ethnic cleavages and conflict? Ethnic heterogeneity might not matter much in itself Ethnic cleavages matter in combination with horizontal inequalities Groups that are somehow aggrieved are more likely to rebel This supports the “grievance explanation” for conflict Denny and Walter present arguments for why cleavages that are ethnic are so potent The emerging micro-literature can tell us more about mechanisms Ethnicity Ethnicity in war 31/31 What next? We know more about correlations, less about causality Huge endogeneity problems, almost never dealt with Ethnic groups are treated as ‘black boxes’ Ethnicity treated as ‘static’ More work needed to establish mechanisms We know little about what drives horizontal inequality