Minorities and mistrust

Abstract

An enduring debate in the conflict management literature concerns the wisdom of recognizing versus avoiding reference to ethnic identities in institutions to manage ethnic conflict. Understanding why ethnic recognition occurs is crucial for informing this debate. We develop a theory based on functional and political mobilization effects of recognizing ethnic groups. Contrary to reasoning that minority leaders would be most interested in recognition, the theory suggests that recognition consistently favors the interests of leaders from larger, plurality groups, whereas minority leaders face a ‘dilemma of recognition’ between functional gains and mobilization threats. We use mixed methods to test our theory. For our quantitative analysis, we draw on an original coding of recognition in constitutions and comprehensive political settlements from 1990 to 2012. We find that for cases with leaders from plurality groups, recognition is adopted 60% of the time. With leaders from minority groups, the rate is about 40 percentage points lower, even after accounting for many background factors. Additional quantitative tests and a qualitative analysis present more detailed evidence to show that the processes correspond to the logic of our theory. Answering these questions about when and why recognition is adopted is a crucial step in evaluating its effects on conflict.

Keywords

ethnic conflict ethnic identity identity politics minority rights peace agreements

Introduction

Past violent conflict is a robust predictor of future conflict (Collier & Sambanis, 2002; Walter, 2010). Moreover, conflicts with an ethnic component are nearly twice as likely to recur (Mattes & Savun, 2009: 754). This raises important questions about institutional choices in the aftermath of ethnic violence. Current literature focuses intensely on institutions as the basis of societal harmony (Acemoglu & Robinson, 2012), including peace after ethnic conflict (Kuperman, 2015; Reilly, 2001; Reynolds, 2011). In such instances, a fundamental institutional choice is whether or not ethnic groups should be recognized explicitly. Indeed, this choice defines an ‘enduring debate’ in the literature and policy discussions (McGarry, O’Leary & Simeon, 2008).

An example illustrates the possibilities. With similarly troubled pasts and nearly identical ethnic and material structural conditions, Rwanda and Burundi have answered the institutional question differently, both ostensibly in the aim of peace. Burundi’s 2005 constitution entrenched public institution quotas for Hutus and Tutsis after the 1993–2004 civil war. In contrast, Rwanda’s 2003 constitution resolves to ‘eradicate ethnic, regional and any other form of divisions’, and mere reference to Hutu or Tutsi identity can be reason for prosecution under anti-genocide laws. What informs these diametrically opposed choices under such similar circumstances?

We study why some countries adopt ethnic recognition to manage conflicts while others do not. Understanding this institutional choice is a crucial first step toward assessing its impact on peace. By ethnic recognition, we mean the formal identification of ethnic groups by name in constitutions or political settlements. While the literature discusses philosophical merits of recognition and commonality of recognition-based policies (Krook & O’Brien, 2010; Reynolds, 2005), to our knowledge ours is the first systematic mapping of the adoption of recognition in the context of violent conflict. Analyses of recognition strategies tend to say little about the motivations for adopting them (Sisk, 1996: 77). Yet, it is important to study the origins of institutions, focusing on critical junctures that follow social upheaval (Acemoglu & Robinson, 2012).

Our theoretical analysis below points to a key factor that may inform leaders’ choices about recognition: their status as minority or non-minority group members. One may reason that minority leaders would be most interested in recognition to institutionalize respect for ethnic identities and minority groups (Taylor, 1992; Young, 1990). Recognition allows for ‘group-differentiated rights’ such as ethnic quotas or autonomy arrangements (Horowitz, 2000; Kymlicka, 1995; Kymlicka & Shapiro, 1997) that address risks of ‘tyranny of the majority’ (McGarry, O’Leary & Simeon, 2008).

However, we argue that such an assessment of the functional advantages of recognition overlooks the political mobilization effects. Our theoretical prediction is that non-minority leaders are likelier to prefer recognition. This may seem counter-intuitive, but the logic is simple. Amid identity-based conflict, leaders face intergroup mistrust and potential for identity-based remobilization (McGarry, O’Leary & Simeon, 2008). Recognition facilitates interethnic comparisons and ethnic mobilization and, thus, may entrench ethnicity as a political cleavage (Horowitz, 2000). This is to plurality groups’ advantage, but for minority groups it presents a risk.¹ The potential gains, paired with the risks, produce the ‘dilemma of recognition’ (De Zwart, 2005). This dilemma has political bite for minority group leaders, but not for plurality group leaders.

We use mixed methods to assess this theoretical prediction. We draw on a dataset that incorporates an original coding of ethnic recognition in constitutions and comprehensive settlements adopted amid violent political conflict from 1990 to 2012.² Using quantitative cross-national analysis, we find that for cases with leaders from plurality groups, recognition is adopted 60% of the time. With leaders from minority groups, the rate is about 40 percentage points lower, a pattern that holds even after accounting for many potential confounders. The quantitative analysis structures a qualitative investigation (Lieberman, 2005). We find credible illustrations in Rwanda, Burundi, and Ethiopia that the structural conditions underlying our theory hold in practice and that the dilemma of recognition plays out through mechanisms that conform to our theory.

Minority leaders and recognition

The goal of our analysis is to understand leaders’ strategies for managing ethnic conflict.³ We focus upon the choice of whether or not to adopt ethnic recognition. We operationalize ethnic recognition as the explicit naming of ethnic groups in constitutions or political settlements. Ethnic recognition is distinct from non-discrimination clauses that bar differential treatment based on ethnicity. Recognition also stands in contrast to bans on references to ethnicity (Basedau & Moroff, 2011; Ishiyama, 2009).

There are strong arguments both in favor and against recognition as a conflict management strategy. On one side, a growing group of scholars contend that accommodative institutions, guaranteeing rights to ethnic groups, build peace (Cederman, Gleditsch & Wucherpfennig, 2014; Lijphart, 1977, 1985; McGarry & O’Leary, 2006). Subordination along ethnic lines has historically provided a cause for violent intergroup conflict (Cederman, Gleditsch & Buhaug, 2013; Gurr, 1993; Wimmer, Cederman & Min, 2009). Recognition strategies have the potential to redress grievances by granting groups status or allowing for more precise targeting of resources along ethnic lines (Cunningham, Loury & Skrentny, 2002; Horowitz, 2000: 657–659). Other scholars make the case that integrative institutions that reduce the political salience of ethnic groups best prevent conflict (Horowitz, 1991, 2000; De Zwart, 2005). This could avoid ‘freezing’ divisions that emerged in the heat of conflict (Simonsen, 2005) and open more space for conflict transformation (Taylor, 2001). Lieberman & Singh (2012) argue that historically, institutionalizing ethnicity contributes to ethnic war. This debate between accommodative and integrative strategies – or recognition and non-recognition – as well as a number of intermediary strategies (Kuperman, 2015; Roeder, 2005; Sisk, 1996) continues. Our premise is that we need to understand conditions leading to the adoption of recognition to avoid confusing cause with effect.⁴

To understand patterns in the adoption of recognition, we must consider how leaders might view it. We analyze a very stylized setting that brings strategic dynamics into sharp relief. (A formal model is in the supporting information.) We demonstrate that the ‘dilemma of recognition’ is especially pronounced for minority leaders.

Our setting is one of recent or ongoing violent conflict where ethnicity is an important basis of political mobilization. Our analysis rests on a number of key observations. First, we observe that leaders who preside over a country’s political regime are typically concerned with political survival. Our analysis is based on the idea that such leaders have the power to decide how ethnicity will be treated under the regime and the ability to adopt recognition should they wish to do so. This assumption is relevant in circumstances of violent mobilization, when leaders often achieve a degree of power beyond ordinary politics (King, 2007), but may be less convincing in cases of usual legislative politics.

In settings of ethnic conflict, the leader is a member of an ethnic group that faces opposition ethnic groups. The political survival of the leader might be threatened by a critical mass of opposition group members deciding to contest his/her authority. In such situations, the leader would evaluate strategies for dealing with this threat. As Wimmer, Cederman & Min (2009) suggest, members of the opposition group would support contestation if they believe opportunities available to their group are below some critical threshold. We also observe that ongoing conflicts generate situations of interethnic mistrust (Collier et al., 2003; Posen, 1993; Snyder & Jervis, 1999). It is up to the leader to decide how to manage this mistrust, including how to deal with ethnic identities.

Two effects of recognition imply that the preference for recognition increases in the size of the regime leader’s ethnic group relative to that of opposition groups. These effects are that (i) recognition may allow leaders to overcome the mistrust of opposition members more efficiently than non-recognition, and (ii) that ethnic recognition confers a relative advantage to larger ethnic groups in their ability to mobilize in the future.

Effect (i), the functional effect of recognition in managing opposition mistrust, is based on three mechanisms that have been examined in the literature. First, the symbolic value of being recognized confers a direct benefit to opposition members (Young, 1990; Taylor, 1992). Second, by allowing for transparency about the allocation of resources along ethnic lines, opposition members are able to more precisely evaluate how well the regime is treating them, which in turn makes it easier for the regime to win opposition members’ trust (Cederman & Girardin, 2007; Weisskopf, 2004). Third, recognition may be a step toward institutionalizing opposition group rights, which reduces their uncertainty about how they will fare in the future (McGarry, O’Leary & Simeon, 2008). These functional effects make recognition attractive to leaders interested in ensuring stability regardless of whether they are from a minority or plurality group.

However, recognition effectively entrenches ethnic divisions, yielding effect (ii), the mobilization effect of recognition: should they wish to do so, groups can take advantage of the opportunity to use ethnic appeals to enhance their political position. The current literature is clear in identifying substantial coethnic advantages in political mobilization, and recognition lowers the barriers to such mobilization (Bates, 1983; Habyarimana et al., 2007; Varshney, 2007). Institutions that facilitate ethnic mobilization should be especially advantageous to larger ethnic groups, who will then be in a better position to bargain over the distribution of spoils (Posner, 2005). Social identity theory suggests that mere mention of groups can cause people to think about themselves differently (Tajfel, 1982), meaning that recognition provides a nudge toward mobilization along ethnic lines. In contrast, by avoiding recognition, minority groups may hope to ‘transcend’ ethnicity as the primary line of political competition, thereby overcoming a structural, demographic disadvantage (Smith, 1986: 214–217).

Plurality group leaders face no conflicting pressures when it comes to recognition: they benefit from both the functional and mobilization effects. But for a minority leader, the losses due to the mobilization effect may outweigh the gains from the functional effect. Holding all other considerations fixed, recognition should be a dominant strategy for plurality group leaders. For minority group leaders, the situation depends on the relative strength of the functional versus the political mobilization effects – a dilemma of recognition. This yields the following hypothesis:

H1: All else equal, minority group regime leaders should be less likely to adopt ethnic recognition than plurality group regime leaders.

The force of this logic depends on the starkness of the difference in the strategic positions of minority versus plurality groups. This starkness is affected by the level of ethnic fractionalization. Ethnic fractionalization measures the probability that two randomly selected individuals from a country belong to different ethnic groups (Alesina et al., 2003). Minority status matters much more in situations of lower ethnic fractionalization. It is in those cases that the minority stands distinct relative to a clear majority group (Collier & Hoeffler, 2004). Under high ethnic fractionalization, minority status may be less strategically meaningful because no one group demographically dominates society.

This theory relegates to the background some complicating factors. We presume that the regime leader is not concerned with how recognition affects potential dissent from his/her own group. We do not focus upon more perverted applications of recognition that seek to reinforce inequality rather than redress it (Cinalli, 2005). Our analysis also ignores the possibility that majority leaders may have an easier time enacting new provisions. Given these complexities, the relationship between minority/non-minority status and recognition is unlikely to be perfect. But if the logic developed above characterizes the essence of the dilemma of recognition, we expect general patterns to conform to our hypothesis.

Methods

We use a mixed-methods approach to test our theory. Using first a cross-national quantitative analysis, we then apply research design principles from Lieberman to select ‘well predicted cases’ to qualitatively trace whether the assumptions and mechanisms of our theory characterize what actually occurred (Lieberman, 2005: 444). Because of the impossibility of experimental variation in our primary explanatory factor of interest (minority status of leaders), the qualitative analysis allows us to evaluate threats to the validity of our interpretations of the ‘effect’ of minority status.

Data

We coded the adoption of recognition in constitutions or comprehensive settlements promulgated in the context of violent intergroup conflicts from 1990 to 2012.⁵ Constitutions and comprehensive settlements represent ‘constitutional moments’ that fundamentally define the terms of a regime⁶ and are more entrenched and difficult to revise than other legal measures (Lerner, 2011: 210–211). We acknowledge that recognition may be articulated in other ways (e.g. in legislation) and that a few of the non-recognition decisions that we coded may have little to do with the conflict (e.g. arguably the 1997 amendments to the Pakistan constitutions). This makes ours a conservative estimate of the rate at which recognition is adopted. We also note the possibility of informal recognition, for example through cabinet appointments. Our view is that recognition in a constitution or settlement more clearly signals leaders’ preferences than informal arrangements, which may be used for more tactical purposes and may be more ‘fragile’ (Spears, 2013). Recent research on the relationship between formal and informal recognition comes to different conclusions. In Africa, for example, some have found high rates of informal ethnic accommodation despite infrequent formal recognition (Francois, Rainer & Trebbi, 2015) while others have found informal arrangements to be less accommodative than formal institutions (Kuperman, 2015).⁷

Our main analyses below pool constitutions and settlements together. One could argue that they should be analyzed separately, because settlements may be more malleable and they only arise amid conflicts that are not resolved by military victory. We view this as an empirical question and so we include a robustness check (reported in the Online appendix) to evaluate whether adoption patterns are different for constitutions versus settlements.

We identified cases of violent conflict from the internal conflict datasets constructed by the Uppsala Conflict Data Program.⁸ These cases involve deadly violence reaching at least 25 deaths. We did not select cases on the basis of whether the conflicts are commonly labeled as ‘ethnic wars’ since the relevance of recognition may extend beyond such cases. We do, however, examine whether results change when we consider only cases commonly regarded as ethnic wars. We operationalized ‘constitutional moments’ as comprehensive peace agreements, constitutions, or constitutional amendments adopted since 1990 and that take place amid or immediately following (that is, within a year of) violent conflict.

We defined a ‘recognition’ variable that took a value of 1 if the constitution or settlement explicitly identified multiple ethnic groups, by name, as constituting the population of the country, and took the value of 0 if no such explicit mention is made. The coding does not consider the precise configuration of who is being recognized. Further, it is possible that a document both recognizes ethnic groups and bans particular uses of ethnic references. In such cases, we nonetheless coded the outcome as recognition. We consulted various primary, news, and academic sources, including direct communication with dozens of country experts, to finalize the set of cases as well as our coding. A list of our cases and coding is provided in Table A1 of the Online appendix.

Our main explanatory variable of interest is the minority status of the regime leader. We used the dataset on ‘ethnic minority rule’ produced by Fearon, Kasara & Laitin (2007). Their dataset covers up until 1999 and we applied their coding rules to our cases to complete the coding to 2012.⁹ This variable is based on the ethnicity of the head of state at the time that the settlement or constitution was put forward. It takes a value of 1 for cases where this head of state is from a group that is not the plurality group in the country, and 0 otherwise. An assumption in using this measure is that the head of state is pivotal in determining the terms of the political regime. This will not always be the case, for example in situations where insurgent forces have more influence over regime outcomes. For testing our hypothesis, this would represent measurement error that attenuates the relationship between our minority leader variable and recognition (thereby producing a bias that works against us).

Dealing with endogeneity

We also consider alternative explanations and confounders. Any relationship between the adoption of recognition and the minority status of leaders could be spurious to confounding factors. In identifying such confounders, we consider mechanisms that give rise to minority leadership. The literature suggests two mechanisms in particular. First, countries that have lower levels of exclusion and horizontal inequality may be more likely to have minority leaders and, presumably, also less likely to adopt recognition policies (Wimmer, Cederman & Min, 2009). Second, regionally concentrated minorities are unlikely to win broad appeal and thus have a disadvantage relative to larger groups in becoming heads of state (Posner, 2005). Presumably, regional concentration of groups also makes recognition more likely, to manage the concentrated interests of such groups. Confounding may also arise due to levels of development, political institutions, intensity and military outcomes of the conflict, the engagement of the international community, or interregional differences, variables conventionally incorporated into analyses of conflict resolution dynamics. We drew on various existing data sources for these variables. Finally, adoption patterns may depend on cultural values that vary across regions. As such, we include region dummy variables to account for such heterogeneity. (Variable sources and summary statistics are included in the Online appendix, Tables A2 and A3.) We use logistic regression that controls for these factors. The supporting information in the Online appendix also contains a robustness check using a non-parametric matching estimator.

Regression control strategies cannot fully overcome the fact that the emergence of minority leadership is deeply endogenous. Experimental variation is impossible and we could not identify a source of quasi-experimental variation. To attribute the outcomes that we observe to the variation in minority leadership that we measure, we seek four types of evidence. First, the effects should be of high magnitude, in which case their sensitivity to confounding from unobserved variables will be low (Imbens, 2003; Rosenbaum, 2002: Ch. 4). Second, the size of the effects should be robust to quantitative control strategies as described above. Third, we should find evidence for more elaborate implications of the theory (Rosenbaum, 2002: 5–6). For this, we test two interaction effects propositions: first is the proposition that minority leader effects are smaller when ethnic fractionalization is higher, and second is the proposition that minority leader effects are stronger in cases that are classified as ethnic conflict. Fourth, the qualitative analysis should demonstrate that the decisionmaking process follows the logic of our theory and does not present more convincing alternative accounts (Van Evera, 1997: 55–67).

Figure 1.

Proportion over time of new constitutions, constitutional amendments, and comprehensive political settlements in conflict-affected countries with recognition, 1990–2012

Table I.

Regional patterns in adopting recognition in constitutions, constitutional amendments, and comprehensive settlements in conflict-affected countries, 1990–2012

Region	Percentage with recognition	N
Sub-Saharan Africa	17%	40
Americas	67%	6
East/Southeast Asia	60%	10
Europe	100%	11
Middle East/N. Africa	56%	9
South/Central Asia	40%	10
World	43%	86

Trends

Figure 1 and Table I show temporal and regional trends in the adoption of recognition in constitutions, amendments, and settlements in conflict-affected countries from 1990 to 2012. There are 86 cases in our dataset. Cases of recognition represent 43% (37 cases), indicative of this period having been an ‘era of identity politics’ (Eisenberg & Kymlicka, 2011). The rate of adoption holds steady over the years that we cover.

We see substantial interregional differences in the rate of adoption. European cases include the Balkans, Northern Ireland, constitutional revisions in the Russian Federation, and the Caucasus; all 11 of these cases adopted recognition. We qualitatively explore the European cases in the supporting information and

Table II.

Patterns in adopting recognition by the minority status of the leader’s ethnic group for constitutions, constitutional amendments, and comprehensive political settlements in conflict affected countries, 1990–2012

		Recognition
Leader type		No	Yes	Total
Non-minority	N	18	27	45
	%	40%	60%
Minority	N	31	10	41
	%	76%	24%
Total	N	49	37	86
	%	57%	43%
Pearson χ2 = 11.10, p < .001

note that recognition in Europe may be affected by regional norms. But because all European leaders in our dataset represent pluralities, we cannot say definitively whether the patterns in Europe are driven more by our logic of recognition or by regional norms and external intervention. The European pattern is in stark contrast to sub-Saharan Africa, where only 17% (7) of the 40 cases we identified adopt recognition. Across the Middle East, North Africa, East and Southeast Asia, and the Americas, recognition occurs in a majority of cases, whereas in South and Central Asia recognition occurs in a minority.

Quantitative analysis

Our quantitative analysis tests our hypothesis that recognition will be substantially less likely in cases where the regime leader comes from a minority ethnic group. The outcome is our binary coding of recognition adoption. Table II shows the basic pattern. These basic results conform to the expectations of our theory. In cases where regime leaders come from non-minority (that is, plurality or majority) ethnic groups, recognition is adopted 60% of the time. But when the leader comes from a minority group, recognition is adopted only 24% of the time.

Table III provides results of logistic regressions that control for potential confounders. (The tables report marginal effects, which measure the estimated change in the probability of recognition given a unit change in the variable, holding all other regressors to their means.) This set of models adds the various controls described above. The results suggest that the basic relationship captured in Table II is robust to controlling for these factors.

Table III.

Minority ethnic group leader as a correlate of ethnic recognition, logistic regression estimates (marginal effects reported)

Outcome is ethnic recognition in constitution or settlement (0/1)
Model	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
Minority leader^d	–0.36**	–0.32*	–0.31*	–0.44**	–0.44**	–0.55*	–0.20	–0.26	–0.46	–0.55*	–0.70*
	(0.11)	(0.14)	(0.16)	(0.17)	(0.17)	(0.23)	(0.29)	(0.16)	(0.42)	(0.25)	(0.28)

Pre-violence minority leader^d		–0.06	–0.06	0.06	0.03	–0.03	0.02	0.05	–0.01	0.03	0.01
Pre-violence minority leader^d		(0.16)	(0.18)	(0.22)	(0.21)	(0.34)	(0.21)	(0.16)	(0.18)	(0.15)	(0.22)

Ethnic fractionalization			–0.16	–0.52	–0.58	–0.57	–0.52	0.74	0.92	0.79	–0.90
			(0.37)	(0.48)	(0.48)	(0.57)	(0.45)	(0.52)	(0.74)	(0.49)	(0.66)

Ethnic conflict^d			0.25†	0.12	0.16		0.28	0.10		–0.01	0.16
			(0.15)	(0.17)	(0.19)		(0.20)	(0.16)		(0.22)	(0.20)

Excluded proportion			0.02	0.35	0.42	0.55	0.38	0.24	0.42	0.26	0.41
			(0.26)	(0.26)	(0.26)	(0.39)	(0.28)	(0.44)	(0.31)	(0.45)	(0.29)

Regionally concentrated groups^d			0.14	0.31†	0.37**	0.45**	0.36**	–0.15	0.04	–0.22	0.37**
Regionally concentrated groups^d			(0.29)	(0.16)	(0.12)	(0.16)	(0.13)	(0.46)	(0.34)	(0.54)	(0.14)

log(GDP/capita)				0.02	0.01	–0.09	0.01	–0.12	–0.12	–0.13	0.00
				(0.08)	(0.08)	(0.11)	(0.08)	(0.13)	(0.08)	(0.11)	(0.07)

Freedom House ‘partly free’^d				0.08	0.18	0.33	0.11	0.13	0.36	0.20	0.16
Freedom House ‘partly free’^d				(0.17)	(0.18)	(0.23)	(0.19)	(0.15)	(0.22)	(0.15)	(0.19)

log(mountain percent)				0.22*	0.26*	0.20	0.25*	0.17†	0.21	0.17†	0.27*
				(0.10)	(0.11)	(0.13)	(0.10)	(0.09)	(0.14)	(0.09)	(0.12)

log(fatalities + 1)				0.01	0.01	0.01	0.01	0.01	0.00	0.01	–0.00
				(0.02)	(0.02)	(0.03)	(0.02)	(0.02)	(0.02)	(0.02)	(0.02)

PITF atrocities historical max				–0.08	–0.09†	–0.10	–0.09	–0.09*	–0.14*	–0.10*	–0.10
PITF atrocities historical max				(0.05)	(0.06)	(0.08)	(0.06)	(0.05)	(0.07)	(0.05)	(0.06)

Military victory^d				–0.26†	–0.32*	–0.34*	–0.31**	–0.23**	–0.18*	–0.22*	–0.32*
				(0.14)	(0.13)	(0.15)	(0.12)	(0.09)	(0.09)	(0.09)	(0.14)

Previous powersharing^d					–0.01	0.16	–0.01	–0.15	–0.07	–0.14	–0.03
					(0.20)	(0.23)	(0.21)	(0.22)	(0.18)	(0.21)	(0.22)

International engagement^d					0.25†	0.36†	0.22	0.25	0.52	0.33	0.28†
International engagement^d					(0.15)	(0.22)	(0.15)	(0.17)	(0.37)	(0.21)	(0.16)

Ethnic conflict X minority leader^d							–0.32			0.40
Ethnic conflict X minority leader^d							(0.25)			(0.28)

Ethnic fractionalization X minority leader^d											0.73
Ethnic fractionalization X minority leader^d											(0.88)

Minority leader + ethnic							–0.52**			–0.14
Conflict X minority leaderⁱ							(0.22)			(0.16)

Region dummies	N	N	N	N	N	N	N	Y	Y	Y	N
Year trend (linear + quadratic)	N	N	N	Y	Y	Y	Y	Y	Y	Y	Y
Observations	86	86	81	75	75	54	75	65	42	65	75
Omitted observations	None	None	No excl. prop.	No excl. prop. FH ‘free’	No excl. prop. FH ‘free’	No excl. prop. FH ‘free’ Non-eth. conf.	No excl. prop. FH ‘free’	No excl. prop. FH ‘free’ Europe	No excl. prop. FH ‘free’ Non-eth. conf. Europe	No excl. prop. FH ‘free’ Europe	No excl. prop. FH ‘free’

Marginal effects; standard errors in parentheses. Standard errors account for clustering by country. † p < 0.10, * p < 0.05, ** p < 0.01.

^d for discrete change of dummy variable from 0 to 1. ⁱ sum of main effect (Minority leader) and interaction effect (Ethnic conflict X Minority leader) coefficients.

‘No excl. prop.’ refers to cases omitted due to missing ‘Excluded proportion’ data: Djibouti 1994, 2001; Pakistan 2011; Sierra Leone 1999; and Sudan 2011. FH ‘free’ refers to cases coded as ‘free’ by Freedom House. These predict recognition perfectly and so these cases are dropped from models 4–10. The dummy variable for Europe predicts recognition perfectly and so these cases are dropped from models 8–10. ‘Non-eth. conf.’ refers to cases that are not coded as ‘ethnic conflicts’ in the Wimmer et al. (2009) dataset.

Figure 2.

Predicted probability of ethnic recognition under plurality (solid line) and minority (dashed line) leader regimes, over values of ethnic fractionalization

As Model 6 shows, the negative relationship between minority leaders and recognition adoption is considerably stronger when we limit ourselves to cases coded as ethnic conflict in the Ethnic Power Relations dataset (Cederman, Wimmer & Min, 2010). Rather than the 30–40 percentage point difference in rates of adoption that we see in the overall set of cases, in the subset of ‘ethnic conflicts’ the difference is 55 percentage points. The interaction term is not statistically significant, however, given the modest sample size (Model 7).

Models 8–10 use dummy variables for the regions shown in Table I to account for unmeasured interregional heterogeneity. This forces us to drop the European cases given that all 11 of them were cases with recognition. Nonetheless, when we do so, we find that the relationship between minority leaders and recognition remains strong in terms of the point estimates. Thus, the effect is not driven solely by the European cases nor is it spurious to interregional heterogeneity.

Model 11 considers how the effect of minority leadership is modified by the extent of ethnic fractionalization. Above, we discussed that an implication of our theory is that the effect of minority status should be stronger in situations with lower ethnic fractionalization (fewer groups), which makes minority status more relevant strategically. Model 11 includes the interaction of the minority leader variable and ethnic fractionalization. The nature of the effect is as we expect, although given the relatively small number of cases, the interaction term itself is not statistically significant. Nonetheless, as Figure 2 displays, in cases of very low ethnic fractionalization, the predicted rate at which ethnic recognition is adopted is much lower under minority leaders (about a 70 percentage point difference). Where ethnic fractionalization is very high, minority leaders do not affect the likelihood of adoption.

The supporting information contains other robustness checks. First, we evaluate robustness to model specification by using the non-parametric ‘bias-adjusted’ matching estimator of Abadie & Imbens (2011). The results are consistent with the regression estimates. Second, a possible objection is that the 86 cases that we study represent a ‘selected’ sample. That is, they represent cases in conflict-affected countries where a political agreement was reached. It could be that plurality leaders try to avoid agreements altogether. This would undermine our story if such avoidance were related to the possibility of recognition occurring should an agreement be reached. To evaluate this possibility, we extended our dataset to include all years of conflict in all countries from 1990 to 2012 and conducted an event history analysis of whether agreements were reached. The results suggest that at best, minority leaders are associated with a small and statistically insignificant increase in the likelihood that an agreement is struck in any given year. Next, we studied whether results differ for constitutions versus settlements, finding no indication that they do. After that, we studied interaction effects between minority leadership and levels of bloodshed, finding no significant evidence of such an interaction. Finally, we conducted a test of whether informal recognition tends to substitute for formal recognition (see supporting information section 10 in the Online appendix). We find that this is not the case – indeed, even after controlling for many background characteristics, formal recognition is associated with a pronounced reduction in future levels of ethnic exclusion. The finding that minority leaders are much less likely to adopt recognition is quite robust and appears to be meaningful.

Nonetheless, without experimental or quasi-experimental variation in minority leadership, such a quantitative analysis does not, on its own, seal the case for the interpretation provided by our theory. Rather, we turn to richer, qualitative evidence.

Qualitative analysis

We follow Lieberman’s (2005: 444) proposal to scrutinize good-fit cases to assess whether our interpretation of the quantitative results is valid. We include Rwanda (2003), a minority-led government that did not adopt recognition, and Burundi (2005), a plurality-led government that adopted recognition. These two countries share nearly identical structural conditions and similar histories of conflict, and yet have leaders who have pursued diametrically opposed strategies. We also include Ethiopia (1994), a minority-led government that adopted recognition in a country with very high ethnic fractionalization. Ethiopia is often said to have similar leadership to Rwanda (Matfess, 2015), allowing us some control for leadership style. If our theoretical explanation is valid, we should find two patterns. First, we should find evidence of the relevant structural conditions for our theory: that there is a credible regime leader, a basis of ethnic rivalry and mistrust, and the potential for remobilization. Second, we should find that the decisionmaking process follows the causal logic of our theory and that other confounding factors were not driving recognition decisions. For example, we should find that leaders considered minority or plurality status, ethnic fractionalization, and the functional and mobilization effects in making decisions about recognition. If our interpretation of quantitative results were wrong, we would expect not to observe these patterns and, moreover, find more convincing alternate explanations (Van Evera, 1997: 55–67).

Rwanda

Rwanda’s 2003 constitution fits our theory. Promulgated under President Paul Kagame, a minority Tutsi, not only are ethnic groups not recognized, the constitution sets to ‘eradicate […] ethnic, regional and other divisions and promot[e] national unity’ (Article 9).

Evidence of structural conditions

The structural conditions that underlie our theory are evident for Rwanda. The constitution was adopted after the civil war (1990–93) and 1994 genocide that left, according to government estimates, nearly one million people dead. While many Hutus were killed during the genocide, Tutsis are considered to have been the primary targets. The genocide ended through a military victory by the Tutsi-dominated Rwandan Patriotic Front (RPF), putting an end to more than 40 years of majority-Hutu-led government since independence. According to common wisdom, but no longer measured as per the current non-recognition policy, Tutsis represent 14% of the Rwandan population in contrast to Hutus, who represent roughly 85% and Twas comprising just 1%. Based on a long history, as well as the ongoing presence of Hutu extremists in the Congo, ethnic mistrust and the potential for opposition group remobilization remained high (Rafti, 2006). Finally, Kagame’s role as the regime agenda-setter was well established by 2003 (Reyntjens, 2004).

Evidence of dilemma of recognition in the decisionmaking process

Consistent with our theory, the Rwandan leadership was very conscious of the majority/minority dynamics in determining a post-genocide identity strategy. In theory, after a military victory, the RPF would have been in a good position to enshrine recognition for the Tutsi minority it most represented, and/or the Hutu majority to help assuage mistrust, had it wished to do so. Yet, according to numerous scholars, for a Tutsi minority who wishes to maintain and ‘mask [its] consolidation’ (Reyntjens, 2004: 178) of disproportionate power, ‘ethnic amnesia’ is a good strategy to detract attention from their dominance and therefore prevent mobilization against them (Lemarchand, 1996; see also Bradol & Guibert, 1997; King, 2014; Pottier, 2002). To support their ban on ethnicity, the government often equates ethnic ‘division’ with ‘categorization’ and has vague divisionism and ethnic ideology laws upon which people can be jailed. The government has ‘created a phobia of talking about ethnicity’ (King, 2014: 141).

Additional considerations

Rwandan history provides additional counterfactual support for our theory: had a majority been in power after the genocide, the leadership would likely have supported recognition. Indeed, after the 1959 ‘social revolution’ and violence against Tutsis surrounding independence, the Parmehutu party (French acronym for Party of the Hutu Emancipation Movement), whose leader ultimately became the first President of independent Rwanda (1962–73), advocated for recognition. The ‘Hutu Manifesto’ became government policy, emphasizing Hutus’ historical marginalization by Belgians and Tutsi leaders and stated that ‘we are strongly opposed, at least for the time being, to removing the labels “Mututsi”, “Muhutu” and “Mutwa” from identity papers. Their suppression would create a risk of preventing the statistical law from establishing the reality of facts’ (cited in Prunier, 1997: 46). Two successive Hutu governments ‘emphasised sharp ethnic contours’ (Pottier, 2002: 62), for instance, decreeing that national educational needed to indicate the ‘racial’ proportions of Hutus, Tutsis, and Twas in schools. The governments also introduced ethnic quotas for promotion past primary school and public employment (King, 2014). In contrast, the Union Nationale Rwandaise (UNAR), a pro-monarchy Tutsi-dominated party, called on the ‘Children of Rwanda’ to ‘unite our strengths’ and insisted that ‘There are no Tutsi, Hutu, Twa. We are all brothers!’ (cited in King, 2014: 45). Reyntjens (2004: 187) explains that the RPF’s denial of ethnicity today is ‘an essential element of the hegemonic strategies of small Tutsi elites, such as the powerful in Rwanda during the 1950s and in Burundi between 1965 and 1988’.

Burundi

In contrast to Rwanda today and to Burundi historically, Burundi’s 2005 constitution includes explicit quotas for members of the Hutu and Tutsi ethnic groups. This accords with our theory, in that the constitution was adopted under the incumbency of majority-Hutu president Domitien Ndayizeye, although below we discuss some nuances associated with the strong bargaining position of the largest rebel faction, the Conseil national pour la défense de la démocratie – Forces pour la défense de la démocratie (CNDD-FDD).

Evidence of structural conditions

Like its neighbor Rwanda to the north, Burundi’s ethnic structure is conventionally understood as being constituted as 85% Hutus, 14% Tutsis, and 1% Twas. Also like Rwanda, Burundi had endured a post-independence history marked by interethnic mistrust, rivalry, and violence (Lemarchand, 1996, 2009; Ngaruko & Nkurunziza, 2000). Key events include the purging of military officers and coup in 1966 and a Hutu uprising in 1972 that triggered a genocidal crackdown by the military regime, resulting in some 150,000–200,000 deaths, mostly Hutus (United Nations, 1996). Decades of repression gave way to a tumultuous attempt at democratization in 1993, when elections brought into power Hutu leader Melchior Ndadaye (Reyntjens, 1993). Ndadaye was assassinated four months after being elected in a bungled coup attempt by officers thought to be associated with a hardline Tutsi faction (United Nations, 1996). As southern Tutsi elites initiated a restoration of their authority, Hutu elites mobilized for insurgency. The CNDD-FDD became the largest insurgent group. Major fighting ended with the 2003 Pretoria agreement, with the CNDD-FDD having taken control over large swathes of territory, forcing the southern-Tutsi dominated army to concede a radical redistribution of military power that, in turn, provided the basis for the CNDD-FDD’s political rise (Samii, 2014).

The recognition outcome under Ndayizeye, a majority leader, conforms to the expectations of our theory. At the same time, attention to Ndayizeye distracts from the fact that constitutional negotiations were driven by the interaction between the ascendant CNDD-FDD and Tutsi elites (Lemarchand, 2009). Hutu leaders, both Ndayizeye and the CNDD-FDD leadership, would have had many reasons to be sensitive to the potential for Tutsi remobilization. This includes the memory of the coup of 1993, the restoration of Tutsi authority in neighboring Rwanda, and a pugnacious movement of Tutsi intellectuals (Lemarchand, 2009: 159–162).

Evidence of dilemma of recognition in the decisionmaking process

Given such interethnic mistrust, our theory proposes that majority leaders would view recognition as an efficient method for securing the acquiescence of the ethnic opposition. That recognition also favors the majority group in its ability to mobilize politically is an added advantage. The leadership’s enthusiasm for recognition-based strategies is evident in the extraordinary extent to which the constitution uses them in defining quotas. Article 124 requires that vice-presidents be from different ethnic groups, while Article 129 mandates that Hutus constitute no more than 60% of ministers and vice ministers, while for Tutsis the ceiling is 40%. Article 143 applies the same formula for quotas in the public administration, while Article 164 mandates a 60–40 distribution of Hutu and Tutsi deputies in the national assembly. Article 255 calls for reforms to the security forces that ensure ethnic balance. Article 266 requires ethnic balance in the electoral commission and local administrative units.

Superficially, one could view the quotas in the 2005 constitution as merely a mechanism for consolidating Hutu gains. But the nature of the quotas is indicative of an intention to manage the mistrust of the Tutsi opposition. The quotas are far more generous to Tutsis than what one would expect under either fair division by ethnic distribution or division on the basis of mass political support.

Additional considerations

While we acknowledge important differences, the structural similarities between Burundi and Rwanda allow us to view them comparatively. The two countries differ enormously in the ways that they have addressed ethnicity recently in a manner that corresponds with the propositions of our theory. We can also look at change over time in Burundi. The 1992 constitution, under which the ill-fated 1993 elections took place, was promulgated under the rule of Tutsi president Pierre Buyoya. It contains no quotas or other recognition provisions and emphasizes a need for ethnic unity.

Ethiopia

Like Rwanda, Ethiopia’s 1994 constitution was adopted under a minority government, leading us to expect non-recognition. In contrast, the constitution states, ‘every Nation, Nationality and People in Ethiopia has an unconditional right to self-determination, including the right to secession’ (Article 39.1). It also lays out rights that each ‘Nation, Nationality and People’ has specific rights to language, culture, the preservation of history and self-government (Article 39.2). This recognition outcome is unexpected yet not entirely inconsistent with our theory: in contrast to Rwanda and Burundi, with ethnic fractionalization scores of .18 and .29, respectively, Ethiopia’s level of ethnic fractionalization is much higher at .76. Our theory suggests that minorities will be more likely to adopt recognition if ethnic fractionalization is high, because the political mobilization effects are less threatening.

Evidence of structural conditions

Meles Zenawi, the leader of the Tigrayan People’s Liberation Front (TPLF), was President at the time of the 1994 Constitution. Tigrayans represent roughly 6% of the Ethiopian population, constituting the third largest ethnic group, and are territorially concentrated in the northernmost region of Ethiopia, Tigray. There are over 80 ethnic groups in Ethiopia, with Oromo (35%) and Amhara (27%) being the two largest groups. According to Zenawi himself, ‘the key cause of the war all over the country was the issue of nationalities’ (quoted in Spears, 2010: 78). The situation was arguably one of high mistrust and potential for remobilization. Ethiopia had endured civil war since 1974 leaving over 1.4 million dead. The TPLF was a founding member of the Ethiopian Peoples’ Revolutionary Democratic Front (EPRDF). The EPRDF brought into alliance several ethnically based groups who ultimately defeated the Dergue in 1991.

Evidence of dilemma of recognition decisionmaking process

The dilemma of recognition predicts that concern over potential ethnic mobilization against them may have prevented the minority-Tigray leadership from recognizing ethnicity, although mobilization is less of a concern with high ethnic fractionalization. Indeed, in this case, it appears that the functional benefits of recognition outweighed the mobilization concerns. The Tigrayan leadership well understood the implications of stemming from an ethnic minority and were ‘shrewd’ in their consequent calculations. As Spears (2010: 83) argues, ‘[t]he ethnic politics and the constitution […] were not arrived at with the overall interests of Ethiopia in mind, with the belief that this was the best way to maintain Ethiopian unity. It was essentially a form of self-preservation. That was the agenda.’¹⁰

The Ethiopia case highlights the differences in strategic calculations when ethnic fractionalization is high. The TPLF allied with other groups in order to win a military victory and needed to take their concerns into account in the constitution. Zenawi explained, ‘without guaranteeing these rights [for which people had been fighting, it] was not possible to stop the war, or prevent another one from coming up’ (quoted in Spears, 2010: 78). Such reasoning highlights the functional gains of recognition. At the same time, by establishing nine ethnically based states, the government directed political competition to these areas and promoted Tigrayan interests in the center through control of the army, security forces, and economy. Critics suggest that they hoped to divide the opposition, in other words, an effort at stemming mobilization against them. Given their geographic concentration, Tigrayan leaders may have seen self-determination and secession as a fall back in the event the TPLF were unable to dominate Ethiopia.

Conclusion

The peacebuilding value of ethnic recognition is debated. Some point to its utility in addressing grievances and imbalances between groups, while others point to the dangers of entrenching ethnicity as a basis of mobilization. In this article we propose that such considerations, which define the ‘dilemma of recognition’, also play into the strategic calculations of leaders deciding on recognition policies. For leaders from plurality groups, recognition is a win–win: it provides functional benefits in managing the mistrust of opposition ethnic groups and allows leaders to take advantage of their groups’ numerical superiority in the event they wish to do so. For leaders of minority groups, the functional and mobilization effects of recognition work in opposite directions, presenting a dilemma. Examining constitutions and comprehensive political settlements from 1990 to 2012, regimes with leaders from plurality groups adopt recognition about 60% of the time, while for regimes with minority leaders, the rate of adoption is about 40 percentage points lower, even after accounting for many background factors. The difference is even larger when we restrict ourselves to conflicts conventionally understood as ethnic wars. When ethnic fractionalization is low, in which case minority–majority differences are starker, the pattern is yet more pronounced. We pursue further evidence of the plausibility of the interpretation through qualitative analyses. The results provide additional evidence in favor of our theory. We show that the strength of the association between minority leaders and recognition means that even if it is not the only reason for variation in recognition, it is likely a crucial one.

Our analysis draws attention to the enduring importance of ethnic dynamics in political decisionmaking. It is crucial to understand what drives the institutionalization of identity, as this itself may have consequences (Lieberman & Singh, 2012). At the same time, non-institutionalization of ethnic identity does not necessarily imply an ‘escape’ from ethnic politics. In cases with histories of ethnic mobilization, it may be the continued salience and fear associated with ethnic mobilization that drives such non-institutionalization. The implication is that one cannot neglect the politics that give rise to the adoption or non-adoption of recognition when considering their effects.

A limitation of this study is due to challenges of measurement. Coding the outcome variable required judgment calls. Also, it is not always clear who is the agenda-setting ‘leader’ during the negotiation of a political settlement. Nor is the assessment of the minority status of leaders always straightforward. Table A1 in the Online appendix suggests that among our cases, South Africa’s 1993 constitution is the most pronounced anomaly of a minority-led government (led by then-president FW de Klerk) adopting recognition (recognizing languages of ethnic groups and calling for judicial institutions that are representative in terms of race). A reasonable interpretation of this case is that Nelson Mandela and the African National Congress were the true agenda setters (Waldmeir, 1998). At the other end of the spectrum, Sudan’s 1998 constitution is the most pronounced anomaly of a plurality leader (president Omar al-Bashir of the Arab plurality) avoiding recognition. Some, however, have argued that Bashir’s membership in the northern Ja’aliyyin tribe constitutes the more significant identity, given that such minority northern tribes have dominated the national government since independence (El Tom & Salih, 2003; Musa, 2010). Of course, other cases of measurement error may not fall in line so neatly with our theoretical expectations.

Our analysis is a necessary first step toward estimating the effects of recognition. Understanding when and why recognition is adopted helps us to (i) distinguish the effects of such policies from the conditions that promoted their adoption in the first place (an internal validity concern), (ii) identify good cases to investigate and compare (another internal validity concern), and (iii) define scope conditions for the generality of findings that we derive from particular cases (an external validity concern). We hope that future research might take the results that we have developed here to study the effects of recognition – a crucial question for conflict management.

Footnotes

Replication data

The dataset, codebook, and do-files for the empirical analysis in this article, as well as the Online appendix, can be found at .

Acknowledgements

The authors thank Maria Carreri, Lars-Erik Cederman, Kanchan Chandra, Michael Gilligan, Macartan Humphreys, and Brendan O’Leary, as well as participants at the Queen’s University Ethnicity and Democratic Governance group, CAPERS, the International Political Science Association meeting (2014) and Folke Bernadotte Academy working groups. We also thank Adriana Castro Gonzalez and Grace Pai for outstanding research assistance.

Funding

We gratefully acknowledge funding from the Folke Bernadotte Academy, Sweden.

Notes

References

Abadie

Alberto

Imbens

Guido W

(2011) Bias-corrected matching estimators for average treatment effects. Journal of Business and Economic Statistics 29(1): 1–11.

Acemoglu

Daron

Robinson

James

(2012) Why Nations Fail: The Origins of Power, Prosperity and Poverty. New York: Crown.

Alesina

Alberto

Devleeschauwer

Arnaud

Easterly

William

Kurlat

Sergio

Wacziarg

Romain

(2003) Fractionalization. Journal of Economic Growth 8(2): 155–194.

Basedau

Matthias

Moroff

Anika

(2011) Parties in chains: Do ethnic party bans in Africa promote peace? Party Politics 17(2): 205–222.

Bates

Robert

(1983) Modernization, ethnic competition, and the rationality of politics in contemporary Africa. In: Rothchild

Donald

Olorunsola

Victor A

(eds) State Versus Ethnic Claims: African Policy Dilemmas. Boulder, CO: Westview, 152–171.

Bradol

Jean-Hervé

Guibert

Anne

(1997) Le temps des assassins et l’espace humanitaire, Rwanda, Kivu, 1994–1997 [The time of assassins and the humanitarian space, Rwanda, Kivu, 1994–1997]. Hérodote 86–87: 116–149.

Brancati

Dawn

Snyder

Jack L

(2011) Rushing to the polls: The causes of early post-conflict elections. Journal of Conflict Resolution 55(3): 469–492.

Brancati

Dawn

Snyder

Jack L

(2013) Time to kill: The impact of election timing on post-conflict stability. Journal of Conflict Resolution 57(5): 822–853.

Cederman

Lars-Erik

Girardin

Luc

(2007) Beyond fractionalization: Mapping ethnicity onto nationalist insurgencies. American Political Science Review 101(1): 173–185.

10.

Cederman, Lars-Erik; Kristian Skrede Gleditsch & Halvard Buhaug (2013) Inequality, Grievances, and Civil War. New York: Cambridge University Press.

11.

Cederman

Lars-Erik

Gleditsch

Kristian Skrede

Wucherpfennig

Julian

(2014) Explaining the decline of ethnic conflict: Was Gurr right and for the right reasons? Paper presented at the European Political Science Association Annual Conference, 28-31 August, Washington, DC (http://www.observatoriodetierras.org/wp-content/uploads/2014/11/EXPLAINING-THE-DECLINE-OF-ETHNIC-CONFLICT-CEDERMAN-Y-OTORS.pdf).

12.

Cederman

Lars-Erik

Wimmer

Andreas

Min

Brian

(2010) Why do ethnic groups rebel? New data and analysis. World Politics 62(1): 87–119.

13.

Chandra

Kanchan

(2006) What is ethnic identity and does it matter? Annual Review of Political Science 9: 397–424.

14.

Cinalli

Manlio

(2005) Below and beyond power sharing: Relational structures across institutions and civil society. In: Flynn

Ian O

Russell

David

(eds) Power Sharing: New Challenges for Divided Societies. London: Pluto, 172–187.

15.

Collier

Paul

Hoeffler

Anke

(2004) Greed and grievance in civil war. Oxford Economic Papers 56(4): 563–595.

16.

Collier

Paul

Sambanis

Nicholas

(2002) Understanding civil war. Journal of Conflict Resolution 46(1): 3–12.

17.

Collier

Paul

Elliott

Hegre

Håvard

Hoeffler

Anke

Reynal-Querol

Marta

Sambanis

Nicholas

(2003) Breaking the Conflict Trap: Civil War and Development Policy. Washington, DC & Oxford: World Bank & Oxford University Press.

18.

Cunningham

Clark D

Loury

Glenn C

Skrentny

John David

(2002) Passing strict scrutiny: Using social science to design affirmative action programs. Georgetown Law Journal 90(4): 835–882.

19.

De Zwart

Frank

(2005) Targeted policy in multicultural societies: Accommodation, denial, and replacement. International Social Science Journal 57(183): 153–164.

20.

Eck

Kristine

Hultman

Lisa

(2007) One-sided violence against civilians in war: Insights from new fatality data. Journal of Peace Research 44(2): 233–246.

21.

Eisenberg

Avigail

Kymlicka

Will

(2011) Bringing institutions back in: How public institutions assess identity. In: Eisenberg

Avigail

Kymlicka

Will

(eds) Identity Politics in the Public Realm: Bringing Institutions Back In. Vancouver: UBC, 1–30.

22.

Tom

Osma

Abdalla

Mohamed Salih

(2003) Review of the black book of Sudan. Review of African Political Economy 30(97): 511–513.

23.

Fearon

James D

(2003) Ethnic and cultural diversity by country. Journal of Economic Growth 8(2): 195–222.

24.

Fearon

James D

Kasara

Kimuli

Laitin

David D

(2007) Ethnic minority rule and civil war onset. American Political Science Review 101(1): 187–193.

25.

Francois

Patrick

Rainer

Ilia

Trebbi

Francesco

(2015) How is power shared in Africa? Econometrica 83(2): 465–503.

26.

Gleditsch

Nils Petter

Wallensteen

Peter

Eriksson

Mikael

Sollenberg

Margareta

Strand

Håvard

(2002) Armed conflict 1946–2001: A new dataset. Journal of Peace Research 39(5): 615–637.

27.

Goemans

Henk E

Gleditsch

Kristian Skrede

Chiozza

Giacomo

(2009) Introducing Archigos: A dataset of political leaders. Journal of Peace Research 46(2): 269–283.

28.

Gurr

Ted Robert

(1993) Why minorities rebel: A global analysis of communal mobilization and conflict since 1945. International Political Science Review 14(2): 161–201.

29.

Habyarimana

James

Humphreys

Macartan

Posner

Daniel N

Weinstein

Jeremy M

(2007) Why does ethnic diversity undermine public goods provision? American Political Science Review 101(4): 709–725.

30.

Horowitz

Donald L

(1991) A Democratic South Africa? Constitutional Engineering in a Divided Society. Berkeley, CA: University of California Press.

31.

Horowitz

Donald L

(2000) Ethnic Group in Conflict, 2nd edition. Berkeley, CA: University of California Press.

32.

Imbens

Guido W

(2003) Sensitivity of exogeneity assumptions in program evaluation. American Economic Review, Papers and Proceedings 93(2): 126–132.

33.

Ishiyama

John

(2009) Do ethnic parties promote minority ethnic conflict? Nationalism and Ethnic Politics 15(1): 56–83.

34.

King

Charles

(2007) Power, social violence, and civil wars. In: Crocker

Chester

Hampson

Fen

Aall

Pamela

(eds) Leashing the Dogs of War. Washington, DC: United States Institute of Peace Press, 115–130.

35.

King

Elisabeth

(2014) From Classrooms to Conflict in Rwanda. New York: Cambridge University Press.

36.

Krook

Mona Lena

O’Brien

Diana Z

(2010) The politics of group representation: Quotas for women and minorities worldwide. Comparative Politics 42(3): 253–272.

37.

Kuperman

Alan J

(2015) Designing constitutions to reduce domestic conflict. In: Kuperman

Alan J

(ed.) Constitutions and Conflict Management in Africa: Preventing Civil War Through Institutional Design. Philadelphia, PA: University of Pennsylvania Press.

38.

Kymlicka

Will

(1995) Multicultural Citizenship: A Liberal Theory of Minority Rights. Oxford: Oxford University Press.

39.

Kymlicka

Will

Shapiro

Ian

(1997) Introduction. In: Shapiro

Ian

Kymlicka

Will

(eds) Ethnicity and Group Rights. New York: New York University Press, 3–21.

40.

Lemarchand

René

(1996) Burundi: Ethnic Conflict and Genocide. New York: Cambridge University Press.

41.

Lemarchand

René

(2009) The Dynamics of Violence in Central Africa. Philadelphia, PA: University of Pennsylvania Press.

42.

Lerner

Hanna

(2011) Making Constitutions in Deeply Divided Societies. Cambridge: Cambridge University Press.

43.

Lieberman

Evan S

(2005) Nested analysis as a mixed-method strategy for comparative research. American Political Science Review 99(3): 435–452.

44.

Lieberman

Evan S

Singh

Prerna

(2012) The institutional origins of ethnic violence. Comparative Politics 45(1): 1–24.

45.

Lijphart

Arendt

(1977) Democracy in Plural Societies. New Haven, CT: Yale University Press.

46.

Lijphart

Arendt

(1985) Power-Sharing in South Africa. Berkeley, CA: Institute of International Studies, University of California.

47.

Matfess

Hilary

(2015) Rwanda and Ethiopia: Developmental authoritarianism and the new politics of African strong men. African Studies Review 58(2): 181–204.

48.

Mattes

Michaela

Savun

Burcu

(2009) Fostering peace after civil war: Commitment problems and agreement design. International Studies Quarterly 53(3): 737–759.

49.

McGarry

John

O’Leary

Brendan

(2006) Consociational theory, Northern Ireland’s conflict, and its Agreement 2: What critics of consociation can learn from Northern Ireland. Government and Opposition 41(2): 249–277.

50.

McGarry

John

O’Leary

Brendan

Simeon

Richard

(2008) Integration or accommodation? The enduring debate in conflict regulation. In: Choudhry

Sujit

(ed.) Constitutional Design for Divided Societies. Oxford: Oxford University Press, 41–90.

51.

Melander

Erik

Pettersson

Therése

Themnér

Lotta

(2016) Organized violence, 1989–2015. Journal of Peace Research 53(5): 727–742.

52.

Musa

Abdu Mukhtar

(2010) Marginalization and ethnicization in the Sudan: How the elite failed to stabilize a diverse country. Contemporary Arab Affairs 3(4): 551–562.

53.

Ngaruko

Floribert

Nkurunziza

Janvier D

(2000) An economic interpretation of conflict in Burundi. Journal of African Economies 9(3): 370–409.

54.

Posen

Barry R

(1993) The security dilemma and ethnic conflict. Survival 35(1): 27–47.

55.

Posner

Daniel N

(2005) Institutions and Ethnic Politics in Africa. Cambridge: Cambridge University Press.

56.

Pottier

Johan

(2002) Re-Imagining Rwanda: Conflict, Survival and Disinformation in the Late Twentieth Century. New York: Cambridge University Press.

57.

Prunier

Gérard

(1997) The Rwanda Crisis: History of a Genocide 1959–1994. London: Hurst.

58.

Rafti

Marina

(2006) Rwandan Hutu rebels in Congo/Zaire, 1994–2006: An extra-territorial civil war in a weak state? In: Reyntjens

Filip

Marysse

Stefaan

(eds) L’Afrique des Grands Lacs: Annuaire 2005–2006 [Great Lakes Africa: Directory 2005–2006]. Paris: L’Harmattan, 55–83.

59.

Reilly

Benjamin

(2001) Democracy in Divided Societies: Electoral Engineering for Conflict Management. New York: Cambridge University Press.

60.

Reynolds

Andrew

(2005) Reserved seats in national legislatures: A research note. Legislative Studies Quarterly 30(2): 301–310.

61.

Reynolds

Andrew

(2011) Designing Democracy in a Dangerous World. Oxford: Oxford University Press.

62.

Reyntjens

Filip

(1993) The proof of the pudding is in the eating: The June 1993 elections in Burundi. Journal of Modern African Studies 31(4): 563–583.

63.

Reyntjens

Filip

(2004) Rwanda, ten years on: From genocide to dictatorship. African Affairs 103(411): 177–210.

64.

Roeder

Philip G

(2005) Power dividing as an alternative to ethnic power-sharing. In: Roeder

Philip G

Rothchild

Donald

(eds) Sustainable Peace: Democracy and Power After Civil Wars. Ithaca, NY: Cornell University Press, 51–82.

65.

Rosenbaum

Paul R

(2002) Observational Studies, 2nd edition. New York: Springer.

66.

Samii

Cyrus

(2014) Military integration in Burundi, 2000–2006. In: Licklider

Roy

(ed.) New Armies from Old: Military Integration After Civil War. Washington, DC: Georgetown University Press, 214–228.

67.

Simonsen

Sven Gunnar

(2005) Addressing ethnic divisions in post-conflict institution-building: Lessons from recent cases. Security Dialogue 36(3): 297–318.

68.

Sisk

Timothy D

(1996) Power Sharing and International Mediation in Ethnic Conflicts. Washington, DC: United States Institute of Peace Press.

69.

Smith

Anthony D

(1986) The Ethnic Origins of Nations. Oxford: Blackwell.

70.

Snyder

Jack L

Jervis

Robert

(1999) Civil war and the security dilemma. In: Walter

Barbara F

Snyder

Jack L

(eds) Civil Wars, Insecurity, and Intervention. New York: Columbia University Press, 15–37.

71.

Spears

Ian S

(2010) Civil War in African States: The Search for Security. Boulder, CO: First Forum.

72.

Spears

Ian S

(2013) Africa’s informal power-sharing and the prospects for peace. Civil Wars 15(1): 37–53.

73.

Sundberg

Ralph

Eck

Kristine

Kreutz

Joakim

(2012) Introducing the UCDP non-state conflict dataset. Journal of Peace Research 49(2): 351–362.

74.

Tajfel

Henri

(1982) Social Identity and Intergroup Relations. New York: Cambridge University Press.

75.

Taylor

Charles

(1992) Multiculturalism and the Politics of Recognition: An Essay with Commentary. Princeton, NJ: Princeton University Press.

76.

Taylor

Rupert

(2001) Northern Ireland: Consociation or social transformation. In: McGarry

John

(ed.) Northern Ireland and the Divided World: Post-Agreement Northern Ireland in Comparative Perspective. Oxford: Oxford University Press, 36–52.

77.

United Nations (1996) Final Report of the International Committee of Inquiry for Burundi. New York: United Nations Documentation Center.

78.

Van Evera

Stephen

(1997) Guide to Methods for Students of Political Science. Ithaca, NY: Cornell University Press.

79.

Varshney

Ashutosh

(2007) Ethnicity and ethnic conflict. In: Boix

Carles

Stokes

Susan

(eds) The Oxford Handbook of Comparative Politics. Oxford: Oxford University Press, 274–294.

80.

Vaughan

Sarah

(1994) The Addis Ababa Transitional Conference of July 1991: Its Origins, History and Significance. Edinburgh: Centre of African Studies.

81.

Waldmeir

Patti

(1998) Anatomy of a Miracle: The End of Apartheid and the Birth of a New South Africa. New Brunswick, NJ: Rutgers University Press.

82.

Walter

Barbara

(2010) Conflict relapse and the sustainability of post-conflict peace. World Development Report 2011 background paper. Washington, DC: World Bank (http://web.worldbank.org/archive/website01306/web/pdf/wdr%20background%20paper_walter_0.pdf).

83.

Weisskopf

Thomas E

(2004) Affirmative Action in the United States and India. London: Routledge.

84.

Wimmer

Andreas

Cederman

Lars-Erik

Min

Brian

(2009) Ethnic politics and armed conflict: A configurational analysis of a new global data set. American Sociological Review 74(2): 316–337.

85.

Young

Iris Marion

(1990) Justice and the Politics of Difference. Princeton, NJ: Princeton University Press.

86.

Young

John

(1997) Peasant Revolution in Ethiopia: The Tigray Peoples Liberation Front 1975–1991. Cambridge: Cambridge University Press.

87.

Young

John

(2004) Post-civil war transitions in Ethiopia. In: Ali

Taisier M

Matthews

Robert O

(eds) Durable Peace: Challenges for Peacebuilding in Africa. Toronto: University of Toronto Press, 19–59.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB