Power Sharing in Deeply Divided Places

3. The Track Record of Centripetalism in Deeply Divided Places

CHAPTER 3 The Track Record of Centripetalism in Deeply Divided Places Allison McCulloch In ethnically divided places, exclusion from government tends to be associated with exclusion in the wider polity (Horowitz 1993). Centripetalism, as developed and defended by Donald L. Horowitz, is seen as a novel strategy for the design of political institutions intended to mitigate this sense of exclusion . Centripetalism, sometimes called the integrative approach, is thought to encourage minority influence on majority decision making through the use of electoral rules that necessitate cross-ethnic appeals on the part of political leaders. The underlying assumption is that leaders must be given incentives to “make moderation pay” (Horowitz 1990). Calling the electoral system “the most powerful lever of constitutional engineering for accommodation ,” Horowitz advocates the adoption of preferential electoral systems , specifically the alternative vote (AV) (1991, 163, 189). AV is a majoritarian-preferential system typically used in single-member districts: voters rank the candidates, and the winning candidate must receive at least 50 percent plus 1 in order to be elected. Centripetalists claim that AV sets in motion a cycle of moderation and cross-community cooperation that reduces the potential for ethnic conflict in deeply divided places. Because it is both majoritarian and preferential, political leaders will have to campaign (in heterogeneous constituencies) outside their own ethnic group in order to receive enough votes to get elected. To capture the lower-ranked votes of members of other groups, they will have to make explicit cross-ethnic appeals and in so doing will have to moderate their own claims. This gives rise to “coalitions of commitment” composed of the ethnic parties willing to pool Centripetalism 95 votes with other parties across the ethnic divide (Horowitz 2000, 379). It creates a scenario in which minorities help get majorities elected. Indeed, Horowitz’s argument hinges on the argument that minorities fare better when they can influence decision making even if they are not guaranteed their own representation: “in divided societies, officeholding is decidedly inferior to influence” (1997, 34). At first glance, the model appears to be compelling. It rests on an intuitive logic that suggests that political parties closer to the middle will find it easier to engage in cross-group cooperation, thereby enhancing political stability. The idea that moderate behavior and reciprocal relationships are the means by which to facilitate political stability and intergroup peace is both inherently plausible and normatively desirable. However, the model thus far has a poor track record. Its practice has shown that in cases of deep division, AV is unlikely to promote the sort of incentives imagined. Rather than promoting a politics of moderation, that logic is in many cases reversed , and a politics of extremism results. The disjuncture between the logic of moderation and the institutional outcome of extremism stems from two sources. First, deeply divided societies generally do not have the specific demographic configuration mandated by the model. Majoritarian-preferential voting arrangements are heavily dependent on both the spatial distribution of groups and on the number of groups engaged in conflict. Where one group constitutes an overwhelming majority and can be elected without minority support, it may find that there is little need to pool votes. Second, several factors related to how deeply divided the groups consider themselves to be make the implementation of the proposed recommendations difficult. The most important is the fact that cross-group voting is rare in situations of deep ethnic division where the stability of political institutions is precariously maintained in the face of high degrees of animosity between groups and where ethnic violence, on a small or large scale, is probable or already occurring . My intent in this chapter is first to demonstrate the problematic incentive structure at work in divided places that have adopted centripetal institutions , focusing on the cases of the Republika Srpska (RS) Entity of Bosnia and Herzegovina, Sri Lanka, and Fiji. The second goal is to sketch some contextual imperatives that help explain why the model is not appropriate for deeply divided places. I conclude with the possibility that Horowitz’s recommendations may still be of use to places that are mildly rather than deeply 96 Allison McCulloch divided—ironically what Horowitz has always claimed is true of consociation applies in greater force to his own prescriptions. The Practice of Centripetalism in Deeply Divided Places AV is intended to provide a “centripetal spin” to politics in divided places as it is thought to...