Power-Sharing and Political Stability in Deeply Divided Societies

Allison McCulloch

• Publisher: Taylor & Francis
• Year: 2014
• eBook ISBN: 9781317682189
• Edition: 1
• Topic: Politique et relations internationales
• Subtopic: Politique

Nearly all the peace accords signed in the last two decades have included power-sharing in one form or another. The notion of both majority and minority segments co-operating for the purposes of political stability has informed both international policy prescriptions for post-conflict zones and home-grown power-sharing pacts across the globe.

This book examines the effect of power-sharing forms of governance in bringing about political stability amid deep divisions. It is the first major comparison of two power-sharing designs – consociationalism and centripetalism - and it assesses a number of cases central to the debate, including Nigeria, Sri Lanka, Fiji, Bosnia and Herzegovina, Burundi and Northern Ireland. Drawing on information from a variety of sources, such as political party manifestoes and websites, media coverage, think tank reports, and election results, the author reaches significant conclusions about power-sharing as an invaluable conflict-management device.

This text will be of key interest to students and scholars of ethnic conflict management, power-sharing, ethnic politics, democracy and democratization, comparative constitutional design, comparative politics, intervention and peace-building.

1 Power-sharing, political stability, and deep divisions

DOI: 10.4324/9781315773698-1

Nearly all the peace accords signed in the last two decades have included power-sharing in one form or another. The notion of both majority and minority segments cooperating for the purposes of political stability has informed much of the international policy prescriptions for post-conflict zones, as in Afghanistan, Bosnia and Herzegovina, Kosovo, and Iraq. Home-grown power-sharing pacts have been negotiated and implemented in places like Belgium, Burundi, Northern Ireland and South Tyrol. Power-sharing has also informed, or been proposed for, institutional settlements in a variety of divided places, including Cyprus, Fiji, Kenya, Malaysia, Nigeria, Rwanda, South Africa, Sri Lanka, and Zimbabwe. How power-sharing pacts are adopted, how they are implemented, and how (and if) they bring about political stability represent crucial questions in the study of ethnic conflict regulation.

It is especially critical because there are different ways by which to implement the power-sharing principle and the question of how power-sharing is institutionalized has important consequences for the pursuit of political stability. Consider the divergent paths ethnic conflict regulation has taken in Sri Lanka and Northern Ireland. In the former, the power-sharing principle is evident only in the electoral rules for the office of the president. Voters are allowed to rank the candidates in order of their preference and the winning candidate requires an absolute majority in order to be elected. This type of arrangement is thought to strengthen the influence of minority Tamil and Muslim voters; to reach the majority threshold, campaigning politicians have to reach out to the minority groups. Meanwhile, in the latter case, power-sharing has permitted representatives from both the British unionist and Irish nationalist groups to participate in the executive and to seek proportional representation of their respective communities in key political institutions, from the legislature to the police service.

On 26 January 2010, Sri Lankans went to the polls to elect a new president. It was the first national election since the government vanquished the Liberation Tigers of Tamil Eelam (LTTE), a violent and oppressive separatist group which claimed to represent the minority Tamil community and whose defeat marked the end of a protracted 26 year civil war. The election was announced in November 2009, with two years remaining on the mandate of incumbent president, Mahinda Rajapaksa. Rajapaksa, in seeking re-election, was hoping to take advantage of popular support among the majority Sinhalese community crediting him with the LTTE defeat. The gamble paid off: he was handily re-elected with nearly 58 per cent of the vote on the first count (Sri Lanka Department of Elections 2010). Though the electoral rules are meant to facilitate minority influence, it is not clear that this occurred. Rajapaksa polled only a quarter of the vote in the Tamil-majority areas of Jaffna and Vanni. Turnout in these areas was also limited – only 25 per cent in Jaffna, for instance. Many remained in refugee camps amid harrowing conditions; those that have gone home have returned to areas suffering years of government neglect. Rajapaksa’s closest opponent in the race, Sarath Fonseka, the military commander under whose watch the victory against the LTTE was achieved, also disputed the result, though he was quickly quieted. Within days of the election, he was arrested on allegations of plotting a coup and other charges. Neither Rajapaksa nor Fonseka represented legitimate choices for the Tamils. Rajapaksa’s political fortunes stem from his willingness to align with extremist Sinhalese parties, while Fonseka has previously stated that he believed the country belonged to the Sinhalese and that the Tamils “must not try to, under the pretext of being a minority, demand undue things” (Bell 2008). The fighting has stopped but for many the conflict is far from resolved and stability far from achieved.

While Sri Lankans were casting their ballots, Northern Ireland was once again in the spotlight. The region has enjoyed comparatively stable politics since the signing of the Good Friday Agreement in 1998, which implemented a widely inclusive form of power-sharing. To be sure, there have been several speed bumps along the way and early 2010 represented another critical moment in which the region stood “between stability and collapse” (Wolff 2005). The representatives of Sinn Féin and the Democratic Unionist Party (DUP), which together were jointly governing the region, were negotiating the possibility of devolving power for justice and policing to the Northern Ireland Assembly. These are deeply contentious issues in Northern Ireland and many feared an inability to reach agreement on these issues would push the polity to the brink of collapse. A series of negotiations, including at one point the involvement of the British Prime Minister and the Irish Taoiseach, eventually culminated in the Hillsborough Agreement. The Agreement set out a timetable for devolution of justice and policing, the rules for the selection of the Justice Minister, the establishment of a commission to study the parade issue, and the establishment of a working group to consider means for improving the function of the Executive (Northern Ireland Office 2010). Crisis, it would seem, averted.

Two deeply divided societies with two different interpretations of power-sharing and two different outcomes – one characterized by widespread instability and the other in the process of consolidating its stability. While the outcomes in Sri Lanka and Northern Ireland differ, the deep ethnic divisions that characterized their respective politics are not unique. Ethnic conflict is seemingly ubiquitous and manifestly disruptive. Many lives are lost in the name of such conflict. Sri Lanka has lost more than 80,000 lives over the course of its civil war. In Northern Ireland, 3,700 deaths are attributable to the Troubles. In Burundi, more than 300,000 were killed in a single decade; in Bosnia and Herzegovina, more than 100,000 perished between 1992 and 1995 while an estimated one million died during Nigeria’s civil war. These casualties are accompanied by widespread displacement, both within the country in question and across international borders, by economic disruptions, by destruction of property, infrastructure, and social bonds.

Polities may be divided in numerous ways, by ethnicity, religion, region, language, and ideology. A common conceptualization views a society as ethnically divided when it is “ethnically diverse and where ethnicity is a politically salient cleavage around which interests are organised for political purposes, such as elections” (Reilly 2001: 4). Under such conditions, perceived ethnic differences serve as the basis for the creation and perpetuation of segments – collections of individuals under ethnically prescribed banners – counter-posed to other similarly designated segments. When membership in an ethnic segment becomes overtly politicized and relations between segments continue in antagonistic and potentially violent fashion, a polity can be considered as deeply divided. This may entail, as Adrian Guelke (2012: 32) asserts, “a lack of consensus on the framework for the making of decisions and a contested political process [that] is commonly challenged by political representatives of one of the segments.”

Deep division can be empirically assessed by examining the political salience such divisions display and, in particular, by the form of the polity’s party system. The extent to which voters and their representatives perceive ethnic divisions as politically salient will be manifested in the sort of parties that receive the most electoral support. That is, it is possible to speak of the party system in so-called “ethnic” terms, where the primary organizational principle for party formation is some conception of ethnicity and the promotion of an ethnic group. This sort of party system configuration suggests the presence of deep ethnic division for two reasons. First, where such parties dominate the party system, voters consider ethnicity or ethno-nationalism a politically salient variable. Even the presence of moderate ethnic parties suggests that voters perceive something politically relevant about ethnicity and the need for group protection. Second, where the system includes both moderates and extremists, there exists an increased likelihood of instability and conflict. This is because such scenarios frequently lead to outbidding and the rejection of cross-community cooperation.

Under such conditions, ethnic conflicts often present themselves as intractable. Not only are such conflicts very difficult to resolve, many of the strategies utilized to address them are not only ineffective but also normatively unsustainable. As John McGarry and Brendan O’Leary note, countries are faced with two broad methods for the regulation of ethnic conflict: the elimination of differences, which includes genocide, forced mass-population transfers, partition or secession, and assimilation; or the management of differences, including hegemonic control, arbitration, federalization and power-sharing (McGarry and O’Leary 1993: 4). Assimilation is unlikely to be agreed upon by both majorities and minorities and many examples of third-party arbitration are temporary. Even a strategy like hegemonic control, in which one group monopolizes the state order but which brings about short-term stability, is difficult to justify. This type of minority repression will only add to the list of grievances of excluded groups and is thus unlikely to realize long-term stability or normative legitimacy (McGarry and O’Leary 1993: 23–26). To suggest that genocide and forced population transfers are normatively problematic is, of course, an understatement.

Democracy matters. Its commitment to treating citizens as political equals, its ability to provide authentic access to decision-making channels, and its service as an institutional mechanism for the non-violent regulation of conflict makes it a superior choice in addressing grievances between divided groups. Democratic institutions “allow conflicts to formulate, find expression and be managed in a sustainable way, via institutional outlets such as political parties and representative parliaments, rather than being suppressed or ignored” (Reilly 2001: 5). While democracy appears as a normative solution, not just any kind of democracy will suffice in deeply divided areas. Majority rule seems particularly problematic in such circumstances, where voting falls along ethnic lines, and is likely to entail hegemonic control (Lijphart 1977; McGarry and O’Leary 1993). Though it represents a movement away from traditional understandings of democracy, power-sharing governance may be what is required when society is deeply divided over questions of identity.

Democratic power-sharing is premised on the notion that some form of joint decision-making between the major social segments in a divided society represents the best prospect for achieving political stability. Rather obviously, as Florian Bieber (2005) suggests, power-sharing encompasses two distinction components – power and sharing. This means that the power exercised by the state is effective and not just symbolic and that that power is to be collectively managed. Power-sharing occurs in a political system where “the authority of the state is administered jointly and not only by one narrow constituency” (Bieber 2005: 85). By promoting a collectively managed conception of political power, power-sharing provides authentic access to decision-making channels for groups that, on the basis of simple demography, could be marginalized under a majoritarian system. To be clear, I do not consider power-sharing as a categorical means of successful conflict resolution. That is, I do not promote power-sharing as a panacea. As Adrian Guelke (2012: 156) reminds us, “it is rarely, if ever, possible for people in deeply divided societies to take for granted success in putting into place a political dispensation that removes any possibility of renewed conflict.” What I do argue is that consociationalism represents a powerful strategy by which to avert conflict and to foster conciliatory practices in divided places but with the proviso that it can be prone to failure. Critical to the success of power-sharing is the way in which it is institutionalized. First, power-sharing can be direct or indirect and this often entails a choice between consociationalism and centripetalism. Secondly, within consociationalism, there are two broad strategies to facilitate power-sharing, one premised on the pre-determination of the groups involved – what is often called corporate consociationalism – and another, liberal consociationalism, that allows groups to determine the extent of their involvement (discussed further in Chapter 2).

Consociationalism, which is closely associated with the works of Arend Lijphart (e.g. 1969, 1977, 1985, 1995, 2004, 2008) and John McGarry and Brendan O’Leary (2004, 2006a, 2006b, 2009a), advocates broadly inclusive institutions, entailing the representation and participation of all major social segments in the process of governing. Such institutions include grand coalitions or cross-community representation in the executive, proportionality, group autonomy, and veto rights. A consociation represents a direct form of power-sharing in that it explicitly entails the inclusion and representation of all major groups in government. Centripetalism, associated primarily with the works of Donald L. Horowitz (e.g. 1990, 1991, 2000c, 2002a, 2002b, 2008) and Ben Reilly (2001, 2002, 2011, 2012), advocates the adoption of political and electoral rules that require politicians to appeal to voters across the ethnic divide. This is thought to benefit politicians who present moderate platforms and to limit the rewards of extremism. Centripetal institutions include majoritarian-preferential voting rules, regional distribution requirements in presidential elections, centrist coalitions, and federal arrangements intended to cut across group lines. Centripetalism can be understood as both direct power-sharing, when it involves a coalition of moderate parties from different ethnic communities, and as an indirect form of power-sharing, when it involves presidential executives elected by centripetal electoral mechanisms (the alternative vote or majoritarianism plus regional distribution requirements). Indirect power-sharing is intended to make political decision-making broadly-based, even if sometimes the minority’s share of power is found in the guise of political influence on decision-making rather than in political representation. Some scholars, notably Lijphart (2000), appear keen on equating only consociation with power-sharing. This represents an overly restrictive definition given that other institutional strategies can also be understood as forms of broad-based decision-making. As Timothy Sisk (1996) suggests, the two models ought to be considered as “conceptual poles in a spectrum of specific conflict-regulating institutions and practices that promote power-sharing.” power-sharing, and the debates over its use, are thus best understood as the general collection of institutions that facilitate broad-based decision-making between communities and the process of selection among them.

There is little consensus as to which of these approaches is best in cases of deep division. This book attempts to resolve this impasse. It does so through a comparison of six cases of power-sharing amid deep division: Fiji, Sri Lanka and Nigeria’s Second Republic (which ran from 1979 to 1983) as examples of centripetalism and Burundi, Bosnia and Herzegovina and Northern Ireland as examples of consociationalism. Drawing a comparison between cases of consociationalism and centripetalism allows us to see which institutions promote stability, why they do so, and, most importantly, where they do so. This achieves two things: a clearer understanding of how power-sharing has worked (or not worked) in the specific cases under consideration and a stronger basis from which to make informed policy prescriptions for other divided societies.

On the basis of this comparison, two central arguments are advanced. The first is that centripetalism often proves very difficult to adopt and, even when adopted, its track record is one that tends to reinforce instability rather than to realize its aim of moderation. Second, consociationalism appears more likely to promote political stability in deeply divided areas but only on condition that it be implemented in a revised and expanded form – what O’Leary (2005) has called consociation plus – to incorporate issues such as refugee returns, security sector reform, and citizenship rights, which are the sort of issues most likely to promote instability if left unresolved. A consociational framework, in which both majority and minority groups participate actively, provides insecure groups with the security and self-sufficiency to collectively address contentious issues. Where trust is absent between groups, and where only the possibility of minority office-holding, rather than its guarantee, is assured appears insufficient to stabilize politics.

In assessing these six cases, my concern is with their level of political stability. Yet the concept of political stability remains an elusive one in political science. It has broadly been defined as “the regularity of the flow of political exchanges” and has variously been characterized by the absence of violence, the longevity of government and cabinet duration, the legitimacy of the constitutional regime, and the absence of structural change (Ake 1975; Hurwitz 1973; Lijphart 1977). Stability is best conceptualized as a spectrum. At one end are stable political order...

Table of contents

• Cover Page
• Half Title Page
• Series Page
• Title Page
• Copyright Page
• Table of Contents
• List of figures
• List of tables
• Acknowledgements
• List of abbreviations
• 1 Power-sharing, political stability, and deep divisions
• 2 Consociationalism, centripetalism, and the intellectual conflict
• 3 Consociationalism
• 4 Centripetalism
• 5 Context matters
• Bibliography
• Index