The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions?

The Paradox of Federalism Does Self-Rule Accommodate or Exacerbate Ethnic Divisions?

Edited by Jan Erk and Lawrence M. Anderson

I~ ~~o~~!;n~~~up LONDON AND NEW YORK

First published 2010 by Routledge y

2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN

Contents

Simultaneously published in the USA and Canada by Routledge 270 Madison Avenue, New York, NY 10016

Notes on Contributors

ROlltledge is all imprim of the Taylor & Francis Group, all illforma business

Abstracts © 20lO Taylor & Francis

The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions? .Jail Erk & Lawrellce Alldersoll 2

Typeset in Times by Value Chain, India Printed and bound in Great Britain by TJI Digital, Padstow, Cornwall All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers.

British Library Catalogllillg ill Publicatioll Data A catalogue record for this book is available from the British Library

ISBN1O: 0-415-56494-S ISBNl3: 97S-0-415-56494-6

Ethnofederalism and the Mismanagement of Conflicting Nationalisms Philip G. Roeder

vi Vlll

1

13

3

The Political Dynamics of Secession and Institutional Accommodation Hudsoll Meadwell

30

4

Federalism in a Unitary State: a Paradox too Far? Stephell Tierney

45

5

The Partisan Logic of Decentralization in Europe Jasoll Sorells

62

6

The Paradox of Ethnic Partition: Lessons from de facto Partition in Bosnia and Kosovo Erin K. Jellne

80 97

7

State, Society and Separatism in Punjab Kristill M. Bakke

8

The Paradox of Federalism: Some Practical Reflections David Cameron

115

Index

126

Notes on Contributors

NOTES ON CONTRIBUTORS

vii

called "The Dynamics of Secession: Spatial Models of Secession-proofness and Equilibrium Size", due out in the journal, Quality and Quantity, in 2010.

Philip G. Roeder is Professor of Political Science at the University of California, San Diego, USA. His current research examines institutional arrangements to accommodate ethnic and national diversity and gives special attention to the Soviet successor states. Lawl'ence Anderson teaches at the University of Wisconsin-Whitewater. His current research interest focuses on the link between federalism and secessionism. He is currently completing a book manuscript on federalism and secessionism in the antebellum American South. His work has appeared in the following journals: Regional and Federal Studies, Publius: the Journal of Federalism, Nationalism and Ethnic Politics, TheO/y and Society, and others. Kl"istin M. Bakke is a Lecturer in Politics and International Relations at University College London. Her research focuses on decentralization and intrastate struggles, dynamics of political violence, and post-conflict societies. She received her PhD in Political Science from the University of Washington, Seattle (2007). David Cameron is interested in Canadian government and politics, particularly questions of federalism and Quebec nationalism, ethnocultural relations and the politics and constitution-making of emerging federal countries, particularly Sri Lanka and Iraq. He is the author of Nationalism, Selj~Determinatio/l and the Quebec Question; The Social

Thought ofRousseau and Burke: A Comparative Study; Taking Stock: Canadian Studies in the 90's (ed.); The Referendum Papers: Essays on Secession and National Unity; and Cycling Into Saigon: The Conservative Transition in Ontario (with Graham White); Disability and Federalism: Comparing Dijferent Approaches to Full Participation (ed. with Fraser Valentine).

Jan EI'k teaches at the University of Leiden. He has research interests in various areas of comparative politics, including federalism. His work has appeared in the journals

Comparative Politics, Comparative Political Studies, Publius: the Journal ofFederalism, Journal of Common Market Studies, Nations and Nationalism, West European Politics, Regional and Federal Studies, Canadian Journal of Political Science, Journal ofPublic Policy, Nationalism and Ethnic Politics, among others. Erin K. .Jenne has a PhD in Political Science from Stanford University, specializing in conflict management and ethnic politics. She is currently an associate professor at the International Relations and European Studies Department at the Central European University in Budapest, Hungary. Her most recent book is Ethnic Bargaining: The Paradox of Minority Empowerment (Cornell University Press), which won the 2007 Edgar S. Furniss Book Award. Hudson Melldwell teaches in the Department of Political Science at McGill University. He has a general research interest in the reconstruction of political choices when there are strong political commitments in play. He has a methodological interest in the role of rationality in explanation in the social sciences. He is the author of a forthcoming book

Jason Sorens is Assistant Professor of Political Science at the University at Buffalo, SUNY. He received his PhD from Yale University in 2003. He works on secessionism, comparative subnational politics and federalism. Stephen Tierney is Professor of Constitutional Theory, School of Law at the University of Edinburgh. He works in the areas of comparative constitutional law and international law.

Abstracts

Abstracts

ix

the Scotland Act 1998, while in many ways a coherent attempt to meet the demands of national diversity, may also, paradoxically, contain elements that in the long run have the potential to destabilize the UK. We address the non-federal model that has been used to manage the plurinational UK, highlighting certain elements of this ad hoc arrangement which seem useful to the management of pluralism and others which seem to exacerbate the risk of secessionism.

The Paradox of Federalism: Does Self-Rule Accommodate 01' Exacerbate Ethnic Divisions'!

The PIII·tisan Logic of Decentnllization in Europe

JAN ERK & LAWRENCE ANDERSON

JASON SORENS

The paradox of federalism is about whether self-rule accommodates or exacerbates ethnic divisions. A federal arrangement that formally recognizes ethno-linguistic diversity to help manage divisions can also pave the way for eventual disintegration. In this introductory piece, the editors of this book highlight a number of common reference points for the study of the secession-inducing and secession-preventing features of federalism: First, the political will of the secessionists and their capacity to mobilize to this end; secondly, the characteristics of federal institutional/constitutional design; and, thirdly, economic and sociological uncodified factors that have a bearing upon these questions.

Since the 1970s, a decentralizing trend has gathered pace in several Western European countries. Governments in Spain, Italy, Belgium and the United Kingdom have moved to bestow significant powers on certain regions, while France and Portugal have made more limited reforms. The fact that countries facing nationalist challenges in the periphery have been more likely to decentralize poses a puzzle, because research shows that greater autonomy does not necessarily decrease secessionist sentiment and may even increase some forms of nationalist agitation. Why then do governments decentralize? This paper argues that the explanation lies in partisan political calculations, which can also explain the timing and character of devolution.

Ethnofederalism and the Mismanagement of Conflicting Nationalisms

The Paradox of Ethnic Partition: Lessons from de facto Partition in Bosnia and Kosovo

PHILIP G. ROEDER Recent discussions of federal solutions to ethnic conflict have focused on ethnofederal arrangements; in these the constituent units are homelands for ethnic minorities. Like autonomy arrangements in non-federal states, these institutional arrangements structure subsequent politics in ways that increase the likelihood of escalating conflict that results in nation-state crises. Tinkering with the institutional details of these arrangements is unlikely to exorcise these problems.

The Political Dynamics of Secession and Institutional Accommodation I-JUDSON MEADWELL Although not at the core of the history of ideas, federalism has a distinguished pedigree in political theory. This paper does not turn directly to federalism, however, and to the question of whether its institutional arrangements can be fine-tuned so as to reconcile territorial integrity and cultural heterogeneity. I propose instead a focus on the political dynamics of secession. This focus reveals the sensitivity of institutional accommodation to degrees of heterogeneity, showing that stable accommodation may depend on imposition rather than self-limiting behaviour or mutual enforcement.

Federalism in a Unitary State: a Paradox too Fm"? STEPHEN TIERNEY This paper takes the devolution settlements in the UK as a model of accommodation of territorial diversity, with a focus mainly upon devolution to Scotland. It is argued that

ERIN K. JENNE This article argues that ethnic partition, rather than resolving ethnic security dilemmas endemic to ethnic civil wars, has the paradoxical effect of reproducing wartime ethnic cleavages in the post-war period. This is because segregating combatant groups into militarily defensible self-governing territories tends to undermine the central government, ensures successive electoral victories of ultra-nationalists, and puts state resources in the hands of ethnic militia leaders who have incentives to perpetuate the conflict. This argument is illustrated in the cases of post-war Bosnia and Kosovo, which show that the unwillingness of the international community to implement the integrationist elements of the peace arrangements has amplified the challenge of rebuilding peaceful state societies today.

State, Society and Sepm'atism in Punjab KRISTIN M. BAKKE Why do decentralized states differ in their capacity to preserve peace within their borders? This is the question motivating this study, which maintains that an understanding of decentralization's divergent effect on intrastate conflicts calls for a consideration of how these institutions are embedded in the societies they govern. In particular, this article suggests that the impacts of policy and fiscal decentralization are conditioned by any given region's ethnic make-up and wealth. The argument is anchored in a case study of separatism in Punjab in India.

x Abstracts The Paradox of Fedenllism: Some Pnlctical Reflections DAVID CAMERON This paper explores the promise and paradox of federalism in Iraq, Sri Lanka and Quebec and Canada. The author has doubts as to whether the paradox can be effectively resolved with institutional fine-tuning. Rather, for him, questions of political justice prevail when exploring whether federalism leads to or calms secessionism. The challenge then is not institutional but pre-institutional-things that must be agreed upon before normal politics can operate. While it is comparativeiy easy to adjust institutions, it is more difficult to adjust-let alone bring about-these pre-institutional features. Despite the risks inherent in the institutional set-up of federalism, there might be little else on the table to keep divided societies together in a liberal democratic system that respects the basic demands of justice.

The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions? JAN ERK* & LAWRENCE ANDERSON** 'Department of Political Science. University or Lelden. Leiden. The Netherlands. "Department of Political Science. University of Wisconsin-Whitewater. Whitewater. USA

Introduction In the last few years, the study of federalism has come to enjoy a new-found prominence (Erk, 2006, 20()7). From the European integration process to the World Bank policies in the industrializing world. the boom in the study of federalism is accompanied by growth in its applied side. One particular area where federalism is increasingly prescribed is in the accommodation of territorial divisions and the management of ethno-linguistic conflict. It is especially marketed as a palliative to secessionist conflict. That is, federalism has come to be seen as a way to accommodate territorially based ethnic, cultural and linguistic differences in divided societies, while maintaining the territorial integrity of existing states. Here. however. we have a paradox that puzzles students of federalism. Territorial recognition of minorities through the adoption (or strengthening) of federalism may intuitively seem to be the best way to manage ethno-linguistic conflict

2

1. Erk & L. Anderson

but, in the long run, such recogl1ltlon perpetuates and strengthens the differences between groups and provides minority nationalists with the institutional tools for eventual secession. Further, federalism provides opportunities for conflict between regions and centres that might otherwise not exist. The fundamental question, then, is whether federalism provides a stable, long-lasting solution to the management of conflict in divided societies or is, instead, a temporary stop on a continuum leading to secession and independence. A federal arrangement that formally recognizes ethno-linguistic diversity to help manage the political system can also set this newly-or increasingly-federal state on a path to eventual disintegration. Here, in a nutshell, is the paradox: federalism has features that are both secession inducing and secession preventing. While forms of collective representation are generally seen to be a positive measure for stability in divided societies, there are also signilicant risks. The paradox is, in many ways, part of the broader question of recognition of diversity: Institutions, policies and practices that are designed to manage (ethnic, racial, social, linguistic, religious and economic) divisions may also ensure the perpetuation of these very divisions. Self-rule tends to reinforce and strengthen the divisions by institutionally 'freezing' them in various forms. Measures designed to guarantee minority representation and thereby bring inclusion can also act as a base for further separation-both in physical form and in mentality. This "dilemma of recognition" is inherent in all forms of group rights (de Zwart, 2(05). Group recognition ensures the perpetuation of the differences and provides minority elites with a vested interest in the continuation of the divided system. Recognition also means that collective groups will have the institutional tools to strengthen their internal cohesion, heightening the 'us vs. them' mindset. The paradox of collective representation is that it perpetuates the very divisions it aims to manage. Furthermore, it provides the tools that reduce the costs of secession, thereby making it a realistic option. Ethnic conflicts are often rooted in a desire for increased autonomy from the central state (Gurr, 2000: 195). Group demands may range from a minor devolution of political authority to complete formal independence. These demands are often rooted in the belief that the group's social, economic or cultural survi val is threatened by the actions or inactions of the central state, or the group may simply chafe at the perceived efforts of the central state to interfere with issues that are considered exclusively regional concerns. Given the region's desire for increased independence and the presence of international law that privileges the ambiguous norm of national self-determination-not to mention the norm of maintaining the territorial integrity of the state-it should come as no surprise that one mechanism of conflict reduction explored by social scientists includes the creation (or strengthening) of regional political structures of self-rule. Federalism is one of the most important tools of collective representation, providing autonomy to the constituent regional political structures. Of course, self-rule for constituent groups co-exists with federal shared-rule (Elazar, 1987). A defining feature of federalism is that self-rule and shared-rule are constitutionally (or otherwise) enshrined (Riker, 1964). Decentralization, ethnic partition and devolution are other forms of self-rule designed to give groups collective representation. While they are marketed as mechanisms of conflict management, tools of collective representation have features that might exacerbate divisions under certain

The Paradox of Federalism: Self-Rille alld Ethnic Divisions circumstances. The very same institutions that appear able to calm secessionism, reduce or eliminate the possibility of conflict and manage diversity might actually work in the opposite intended direction. These institutions might freeze identities that are meant to be fluid, provide incentives to mobilize in favour of separation and, most alarmingly, provide institutions that can be used to overcome the collective action problem and accomplish secession. These institutions hold over into independence, thereby reducing the fairly significant costs of secession. Self-rule, then, might actually promote secessionism rather than resolve it.

Ethnic Conflict and Federalism Students of ethnic conflict and federalism often acknowledge the paradoxical characteristics inherent in self-rule and have tried to find ways to reconcile the secessioninducing and secession-preventing features inherent in federalism, yet quests to resolve the paradox have so far fallen short of a clear consensus. In her analysis of federalism and unitarism in divided societies, Nancy Bermeo (2002) stated that she expected to find that federalism exacerbated ethnic conflict. Instead, Bermeo (2002: 97) found that "federal institutions promote successful accommodation". According to her analysis, this conclusion is borne out both in advanced democracies in which "federalism has helped to keep states unified and democratic in the face of possible secession by territorially based minorities" and in less developed countries, which "have all evinced the positive effects of federal structures" (Bermeo, 2002: 98). Bermeo (2002: 1(8) claimed that "no violent separatist movement has ever succeeded in a federal democracy", painting federalism as an unmitigated success as a method of ethnic conflict resolution. Other advocates of self-rule tend to offer more nuanced endorsements of federalism. In their analysis of ethnic conflict regulation, John McGarry and Brendan O'Leary (1993: 4) identified federalization (or cantonization) as a "macro-method" of "managing differences". Federalization "can be used to manage ethnic differences in ways which are fully compatible with liberal democratic norms" (McGarry and O'Leary, 1993: 30). Federalism, however, is unlikely to satisfy groups that have not achieved a critical mass of demographic dominance within the constituent political unit in question. Although confident in employing federalism as a method of conflict regulation, McGarry and O'Leary noted that "democratic federations have broken down throughout Asia and Africa", but they still consider "genuine democratic federalism" an "attractive way to regulate ethnic conflict" (McGarry and O'Leary, 1993: 34, 35). Donald Horowitz's Ethnic Groups in COll/liet (1985) continues to be a central text in the study of ethnic and regional conflict. In it, Horowitz (1985: 6(2) noted that the "skillful division of authority between regions or states and a centre has the potential to reduce conflict", but he did not exhibit the overconfidence that is common among more recent advocates of self-rule. He warned that federalism may be little more than a resting point on the road to secession. His case study of Nigeria showed that "federalism can either exacerbate or mitigate ethnic conflict" (Horowitz, 1985: 6(3). He wrote: "the most potent way to assure that federalism or regional autonomy will not become just a step to secession is to reinforce those specific interests that groups have in the undivided state" (Horowitz, 1985: 628). In other words, would-be secessionists need

4

1. Erk & L. Anderson

to be convinced of the continued benefit of remaining within the extant state. Among the benefits of membership that can be stressed or strengthened are the security umbrella provided by the state and central state-supplied social policies that benefit the group (Bartkus, 1999). Regardless of the potential for problems, Horowitz (1985: 619) was confident that "federalism or at least some devolution has conflictreducing possibilities for many more countries than have so far contemplated it". Ted Robert Gurr (2000: 195) asserted that most "of the ethnic wars of the last half century have been fought over issues of group autonomy and independence". "Negotiated autonomy", he stated, "has proved to be an effective antidote for ethnonational wars of secession in Western and Third World states" (GUlT, 2000: 366). Gurr advocated "preventive diplomacy", which may include supporting negotiations for the pre-emptive granting of autonomy for territorially concentrated substate groups whose goal is independent statehood. Like other supporters of self-rule, Gurr recognized that this method of conflict resolution has its drawbacks: States may not be willing to devolve power to the regional unit. For those concerned with resolving ethno-linguistic conflict, Yash Ghai (2000: 483) advocated exploring "the potential of autonomy". Like Horowitz, however, Ghai's confidence in self-rule is tempered by the concern that federalism may freeze and entrench what would otherwise be a fluidly forming and reforming of group identity (Ghai, 2000:499). Ghai (2000: 50 I) also warned that federalism may serve as a "springboard to secession". Despite these concems, he is confident that autonomy "can play an important constructive role in mediating relations between different communities in multiethnic states". It is, he wrote, a "valuable option, notwithstanding its own diffkulties" (Ghai, 2000: 524). For Ghai, self-rule is a tool of conflict reduction because it promotes integration, not disintegration: it provides a basis for interaction between the region and the centre that is satisfactory to both. He concluded (Ghai, 2000: 525), "Autonomy should be chosen not because of some notion of preserving sovereignty but in order to enable different groups to live together, to define a common public space". This is the essence of the "shared rule" side of federalism, of course. Others have been less enthusiastic about the palliative potential of federalism. While a great deal of recent political science literature sings the praises of self-rule as a method of conflict resolution, some social scientists have recently begun to question the enthusiasm with which it has traditionally been put forward as a solution. In his examination of minority ethnic mobilization in the Russian Federation, Dmitry Gorenburg (2003: 25) found that "ethnic mobilization is most likely to occur in countries that combine an ethnically based federal state structure with efforts to assimilate minority groups". Philip Roeder (1991: 199) made a similar claim about the antecedent Soviet federalism: "Autonomous homelands provide essential resources for the collective mobilization of ethnic communities". Others have found a similar dynamic operating in other former communist systems. Jack Snyder (2000) wrote, "While ethnofederalism does not always produce ethnic violence in late-developing, transitional societies, it does create strong incentives for their elites to mobilize mass support around ethnic themes. When other factors are favorable for intense nationalist mobilization, the legacy of ethnofederalism heightens the likelihood of conflict" (Snyder, 2000: 2(2). Snyder argued that ethnofederalism in Yugoslavia helped to weaken the central state and fuel nationalism (Snyder, 2000: 210). Those who have expressed concerns about

The Paradox (~l Federalism: SellRule and Ethnic Divisions

5

the effectiveness of self-rule as a method of conflict resolution tend to be those who study the consequences of autonomy in the context of former communist states (Brubaker, 1994: Dorff. 1994: Treisman, 1997; Bunce, 1999; LefT. 1999; Cornell, 2(02). In between the advocates and opponents of self-rule as a way to accommodate ethnic divisions are those who have ventured to propose ways to determine under what conditions federalism's potential can be realized. Henry Hale, for example, attempted to resolve the paradox by focusing exclusively on demographic institutional structures in ethno-federal states, arguing that states with core regions (defined as a "single ethnic federal region that enjoys dramatic superiority in population") are more likely to be vulnerable to secessionist pressures than states without core regions (Hale, 2004: 166; see also Levy, 2(07). Michael Hechter (2001: 146) has also tackled the paradox "Whereas [federalism I may provide cultural minorities with greater resources to engage in collective action, leading to a rise in protest events, at the same time it may erode the demand for sovereignty". This reduction in the demand for sovereignty ought to reduce the incidence of secessionism. Thus, while decentralization enhances protest events, it does so in a way that cllltaiis secessionism. However, Hechter also argued that the relationship between federalism and secession is highly dependent upon the specific context in question. A decentralized environment that is able to contain secessionist conflict may, thanks to exogenous forces, end up facilitating secessionism. For Hechter, resolving the paradox of federalism requires taking exogenous factors into account. Lustick et al. (2004: 223) explored the impact of power sharing on secessionism and found that such institutions "seem to inhibit secessionism". They accounted for the paradox by suggesting that power-sharing institutions, such as federalism, may decrease the chances of secession, but that they increase the likelihood of mobilization along ethnic lines; that is, analysts of federalism and secessionism who see groups mobilizing along ethnic lines have mistakenly identified mere ethnic mobilization as secessionism. From this point of view, the paradox is simply a case of mistaken identity (see also Snyder, 2(00). Dawn Brancati (2006) looked at regional political parties as an intervening variable that resolves that paradox. While decentralization might reduce the chance of secessionism, it can increase the chances that regional parties will develop. Thus, the federal bulwark against secessionism obtains when regional political parties are absent; it does not obtain when parties are present (for more on the importance of political parties for federal stability, see Filippov et al., 2(04). Allen Buchanan's (1995: 55) solution to the paradox is legalistic one: "if intemational law unambiguously rejects the principle that an existing federal unit may secede if there is a plebiscite in that unit in favor of secession". However, to what extent intemationallaw could have averted the Yugoslav civil war is open to different interpretations. So why are some federations beset by strong secessionist pressures, while others are virtually free of such forces? When and under what circumstances is federalism secession inducing or secession calming? In short, can the paradox of federalism be resolved? The literature on ethnic conflict and federalism does not seem to provide us with an unambiguous verdict on the paradox. In some cases, federalism does seem to work as advertised: satisfying groups that are or might be in conflict with the centre or with one another and managing diversity within a single state, all the while keeping international boundaries intact. In other cases, federalism works as feared: freezing

6

1. Erk & L. Allderson

identities, creating incentives and opportunities to pursue secession rather than other strategies and creating institutions through which secession can be pursued-with fewer costs than if federal institutions were not in place.

Secession-inducing or Secession-preventing Factors While no magic formula for resolving the paradox is yet unearthed, there are nevertheless useful paths to explore. Certain factors might tip the federal balance in the secession-inducing or secession-preventing direction. In a quest to put the spotlight on these factors, we propose to highlight three dimensions: (I) the political will of the subunits and their institutional and societal capacity: (2) federal institutional design codilied in the constitution: (3) uncodified economic and social factors. 'The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions?' aims to unpack the conditions under which self-rule induces or prevents secession. From devolution in the UK to federalism in Iraq (Anderson, 2007), the expected returns from trying to resolve the paradox are not only academic. Ranging from Sri Lanka to Bosnia-Hercegovina there are ongoing federal experiments seeking to create secure and stable democracies in deeply divided societies. But the paradox confounds the study of federalism and real-world efforts to design stable institutions: the same institutions that seem to be able to resolve differences, acknowledge diversity and prevent states from breaking apart along various fault-lines seem to encourage contlict, harden divisions, facilitate the break-up of states. A central question is whether it is possible to design federal solutions in ethnically divided societies, which are stable over time. The applied side of federalism thus has immediate contemporary relevance. We believe that focusing on three dimensions could lead toward a fuller understanding of the paradox Will and Capacity

The twin factors of will and capacity together form a dimension that focuses on the internal politics of the subunits. There is a complex relationship between the two. While the will to secede might be high in a unitary setting where minority groups have no option for self-government other than separation, they often lack the capacity to bring this about. While the capacity of the subunits to secede increases in a federal system, there is often less of a political will to do so. However, things are unfortunately not this simple. The will to secede can increase if self-rule proves to be a success. Or the opposite can happen, i.e. the unacceptably high costs of secession in a unitary system can dampen the will to secede. The two are, therefore, closely interlinked. Secession is unlikely without it being sought by at least some subset of a popUlation. Where does this will to secede come from? How does it develop? What is its link to the background conditions in which the state was formed and the present conditions of the state itself? In the existing literature, the presence of the will to secede is grounded (typically) in some dissatisfaction or grievance with the status quo. This dissatisfaction might take the form of retrospective displeasure with policy governed by the centre, or it might be prospective hope for better policies and more economic growth.

The Paradox of FederalislIl: SelFRule alld Ethllic Divisiolls

7

Self-rule brings with it a number of institutional and societal tools for the management of territorial diversity and even contlict reduction, but these very tools of self-rule can then be the bases that make secession possible. What role is played by the institutions of self-government and autonomy in the development of secessionism? How much govemment is there at the regional level? What role does more autonomy (or less) play in secessionism? The will to secede can be expressed through these self-governing institutions. Capacity, then, is a critical element of the paradox. Capacity is precisely what is sought by groups seeking increased autonomy from the centre-including groups seeking outright independence. Depending upon other factors (e.g. a will to secede), capacity can contribute to secessionism, but increased capacity might also satisfy an aggrieved group short of independence. While will and capacity are factors internal to the subunit where secessionist tendencies exist, the overall institutional structure of the federal system in question has immediate consequences for the paradox. Illstitl/tiollal Design

The federal institutional design codified in the constitution is often the first dimension that attracts scholarly attention. This dimension includes questions such as drawing subunit boundaries, the number of subunits, constitutional division of powers, representation in central institutions, integration of markets and legal systems, constitutional amending formulae, shared vs. separated jurisdictions, and secession clauses in constitutions. Some of these questions are primarily about tinding a proper distribution of authority between the centre and subunits, but others have an indirect impact on the secession-reducing and secession-inducing features of federalism. At the end of the day, the core concern is whether there are elements of federal institutional design that make one federation more prone to secession (or secessionism) than another. The number (and size) of constituent subunits plays an important role in reducing or exacerbating conflict between the subunits and the central government as well as between the subunits themselves. The general observation seems to be that federalism tends to be more stable with multiple constitutional units rather than two or three large units or a single dominant one. A federal system defined by mUltiple units produces more room for shifting alliances and reduces an 'us vs. them' mindset. At the same time, when only a handful or one of the subunits is ethno-culturally distinct (Spain, Canada), subunits of the minority culture(s) might feel overpowered by the rest of the subunits representing the majority culture. Although not formally part of the federal institutional design, electoral systems have indirect influence on the workings of the federal system and are thus part of the overall institutional structure. The number and nature of political parties are immediately linked to the electoral system in place; and parties, in turn, playa critical role in how a federal system functions. Proportional electoral systems provide incentives for political actors to reinforce their bonds with core homogeneous groups of supporters, while majoritarian electoral systems reward parties that bridge appeals to heterogeneous groups (Norris, 2004: 4). While majoritarian electoral systems can help weaken the divisiveness of group identities, they provide little protection to distinct minorities that resist co-optation into majoritarian politics.

8

J. Erk & L. Anderson

Another key institutional element influencing the relationship federalism has with secessionism is the constitutional role of the jUdiciary. Whether the judiciary has the linal authority to rule on conllicts between levels of government has direct implications on the choices and constraints secessionist movements face. While the courts can provide a critical role as the neutral umpire on the relationship between subunits and the central government, often at times the issues at play are more political than legal. Uncod(fied Factors

The third dimension we seek to highlight is the economic and social factors that normally would not be codified in the federal constitution. The main emphasis here is to explore whether there are extra-constitutional features that make some federations more prone to secession than others. Economic and social factors that lack formal institutional recognition can often playa key role in exacerbating secessionist tendencies or mollifying them. Economic disparity across regions-especially if it corresponds to ethno-cultural distinctiveness-can increase secessionism in a federal system. Economic disparity can be based on the level of economic development or the distribution of natural resources. Sociological factors-in particular, the social cleavage structure-is part of the uncodified dimension. Whether cleavages overlap and reinforce one another or whether they cross-cut and thereby decrease the political saliency of particular cleavages is of critical importance here. While a reinforcing cleavage structure often leads ethnicity, language and religion to converge and consolidate the minority cultural identity, this makes federalism an ideal territorial solution to demands of the minority. In a cross-cutting cleavage structure, it is often harder to reach territorial solutions that would satisfy all groups involved. An additional important social factor is whether kin-groups exist in neighbouring territories. The existence of such kin-groups often fuels secessionist irredentism. Of course, at the end of the day none of the three dimensions exists in a vacuum. They interact with one another leading to complex combinations and difficulty in separating out which factor is responsible for what outcome. The existing literature on the factors that inlluence secessionism focuses, by and large, on these uncodifled factors. Grievance at economic (or political or cultural) injustices creates a will to secede that is then expressed through available institutions and is designed to right the perceived wrong. Clearly, however, this dominant story misses too much: it does not explain secessionism and it does not explore the interaction between will and capacity, institutional design and the uncodified, but critical, features of the federation. The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions?', thus, seeks to bring a degree of order to factors that would otherwise be explained away as contingency.

Different Perspectives on the Paradox The paradox has been the object of study that crosses over various subfleld boundaries with the aid of a wide-collection of theoretical and conceptual tools. One of the aims of

The Paradox (~f Federalism: Se(f-RlIle and Ethnic Divisions

9

this collection is to link up the literatures on ethnic conllict, comparative federalism, nationalism, political philosophy and group rights and, finally, constitutional law and institutional design in divided societies. In addition to the different theoretical repertoires of these various literatures, the paradox is studied through different methodological perspectives ranging from the rationalist to nonnative. The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions?' includes contributions from various methodological schools focusing on the same puzzle. And, finally, the substantive context of the paradox ranges from the constitutional debates around devolution in the UK to the violent secessionist struggle in Kosovo. The following articles explore the paradox in various guises of self-rule, ranging from de facto partition to devolution. As editors, it was our aim to pull together different theoretical, methodological and substantive perspectives on the paradox. The end product suggests that neither perspective has the claim to have unearthed the magic formula to resolve the paradox. As the following chapters show, the approaches to the puzzle are diverse, but the overall verdict seems to point in a common direction: on its own, federalism is not a panacea for conllict. Nor, however, is it simply the next step on the road to secessionist conllict. It is, at once, a problem and a solution. Whether it is one or the other seems to depend a great deal on factors that are extemal to the design of federalism itself-the uncodifled factors identified earlier. The first empirical contribution to The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions?' is by Philip Roeder. In a piece entitled 'Ethnofederalism and the Mismanagement of Conllicting Nationalisms', Roeder offers a tOllr de force exploration of the problematic career of 'ethnofederalism' in the world. Here, he pulls together his significant body of work in this area, examining the consequences of ethnofederalism and autonomy arrangements. The core lesson that seems to emerge is that institutions shape subsequent politics. The adoption of ethnofederal institutions gives groups in conllict with political centres what they want without stopping them from wanting more. In short, ethnofederal arrangements institutionalize competing national projects, increasing the chances that secessionist conflict will develop. This fate is in contrast to the prospect of simple (or non-ethnic) federalism, which has the potential to calm and divide these altemate loyalties. In the following piece, Hudson Meadwell approaches the same analytical puzzle through a different angle. Based on a game-theoretical orientation, 'The Political Dynamics of Secession and Institutional Accommodation' challenges the conventional interpretation of the paradox of federalism. It is not possible, Meadwell argues, to fine-tune the institutions of federalism to make them secession-proof. Instead, secession-proofness and secession-proneness are functions of the underlying strategic dynamic of the situation. In this view, the hope that tinkering with institutions might prevent, reduce or eliminate the prospects of secession is simply misplaced. Thus, by definition, there can be no solution to the paradox. The question then becomes: to what extent does federalism impact the underlying logic? Meadwell concludes that, to be successful, a solution to the problem of secessionism will have to be imposed; in the absence of such imposition, there is no way to guarantee that any solution-federal, or otherwisewill succeed. The fourth chapter explores the paradox in the context of British constitutional politics. Stephen Tierney's goal is to see how the paradox of self-rule

10

J. Erk & L. Anderson

reveals itself in a unitary state. What makes this a particularly interesting case is that self-rule to Scotland (and Wales and Northem Ireland) has been devolved from the centre without a federal redesign of the constitution or. indeed. without a structured system of inter-govemmental relations. That is. the regime guiding self-rule in the UK is an ad hoc one that exists outside the traditional parameters of a formal written constitution. According to Tiemey. the result of this is that the tensions defining some of the other case are partly relieved and partly exacerbated by this ad /zoe regime. One point that Tierney's paper deals with is. of course, the role party politics played in the Scottish devolution process. In the following piece entitled 'The Partisan Logic of Decentralization in Europe' Jason Sorens expands on this angle. The analytical puzzle guiding the paper approaches the paradox of self-rule through the angle of the host state: Given the possibility that federalism will not work as planned. why do some states risk it and make the decision to decentralize? What explains the timing and character of this move? Sorens finds that the decision to decentralize is related to the structure and character of political party competition. particularly the electoral prospects of a given party in the central (or regional) government. Parties that expect to perform well in newly decentralized regions will support decentralization. Thus, for Sorens, resolving the paradox of federalism means explaining why central governments ever agree to decentralize, given the dangers that seem to attend this choice. The answer lies in the structure of party competition. Whether or not self-rule helps accommodate ethnic divisions or whether it exacerbates them is a question not limited to the context of electoral party politics, of course. Erin Jenne tackles the paradox in two cases emerging from ethnic conllict and political violence. Her chapter focuses on the unintended consequences of ethnoterri torial partition in Kosovo and Bosnia. Here. the paradox revolves around the implementation of post-conflict institutions that are designed to prevent future conIlict. Rather than resolving and preventing connict, separating warring parties into ethnic enclaves manages to harden these identities and create opportunities for connict. Integration with measures taken to ensure ethnic security, not partition. Jenne argues. has a better chance of creating a lasting peace. The seventh paper of this collection examines the paradox in the context of the world's largest democracy, which also happens to be one of the most ethnically diverse-India. Many observers tend to see federalism as the key to India's success in holding together, in spite of ethnic tensions and separatist tendencies. However, Punjab and ongoing conllicts in India's North East and Kashmir are notable exceptions. It is within this substantive context that Kristin Bakke explores the paradox of self-rule. In order to explain under what conditions federalism induces or prevents secession, Bakke focuses on state-society interactions. Federal institutions respond to societal traits. in particular ethnicity and wealth. and innuence whether-and which part of-the state becomes a target of political mobilization. In her in-depth analysis of the paradox in the Punjabi case, Bakke analyses how state-society relations interact to affect political mobilization. how political mobilization in turn affects the institutional organization of the state, as well as the consequence of such institutional changes. The key theoretical point which emerges from this article is that understanding the ways in which federal institutions can help preserve peace requires that we theorize and examine how those institutions interact with societal traits. Thus. just as there

The Paradox

(~lFederalislll:

Se(rRlIle and Ethnic Divisions

II

seems to be no single response to the question of whether federalism calms or encourages secessionism. there is-and, by definition. cannot be-no "one size tits all" federal solution to connicts in divided states that is independent of the societies in which the institutions are embedded. As a scholar of federalism who has been involved in the applied side of devising federal solutions to divided societies, David Cameron offers the final word on the topic. His contribution explores the promise and paradox of federalism in Iraq, Sri Lanka and Quebec and Canada. The paradox we have identified exists to various degrees in the environments he explores. but he has his doubts as to whether the paradox can be effectively resolved with institutional fine-tuning. Rather. for him, questions of political justice prevail when exploring whether federalism leads to or calms secessionism. It is here that Cameron recommends that students of ethnic connict/federalism pay more attention to the centrality of the liberal democratic principle of justice in all this. The challenge then is not institutional but pre-institutional-things that must be agreed upon before normal politics can operate. While it is comparatively easy to adjust institutions. it is more difficult to adjust-let alone bring about-these pre-institutional features. Having said all of that, Cameron acknowledges that justice might require the adoption of federal institutions. Despite the risks inherent in the institutional set-up of federalism, there might be little else on the table to keep divided societies together in a liberal democratic system that respects the basic demands of justice. The aim of 'The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions?' was to bring together theoretically. methodologically and substantively diverse perspectives to examine the very same puzzle. While a clear resolution to the paradox did not materialize, one important lesson seems to emerge. It seems that whether federalism is a problem or a solution depends a great deal on factors that are external to-and, indeed, deeper than-the design of federalism itself. Most contributors display a fair amount of scepticism that political institutions can be constructed or finely-tuned to eliminate. reduce or prevent secessionist pressures. That being said. Philip Roeder cautions us that the guiding logic of federal institutional design, i.e. whether it is designed to asymmetrically empower ethnic groups or bring about a symmetrical division of powers between the centre and the regions, does playa role. But we must be equally sceptical of the conclusion that federalism should be avoided at all costs because of its propensity to encourage challenges to central states. David Cameron reminds us that, despite its imperfections, sometimes federalism could be the only tool available.

Aclmowledgement The editors would like to thank Charlie Jeffery, Wilfried Swenden. the anonymous referees and all the contributors to the original special issue.

References Anderson. L. M. (2007). Theorizing Federalism in Iraq. Regiollal lIlld Federal Sludies, Vo1.17. No.2. pp.159-171. Bartkus, V. O. (1999). The DYllalllic (J(Secessioll. Camhridge: Cambridge UllIversity Press.

12

J. Erk & L. Anderson

Benneo. N. (2002), The Import of Institutions. JOllrnal (!(Democracy, Vo1.l3. No.2. pp.96-110. Brancati, D. (2006), Decentralization: Fueling the Fire or Dampening the Flames of Ethnic Conllict and Secessionism, Illternariollal Orgallizatioll, Vol.60, No.4, pp.65 1-685. Brubaker, R. (1994), Nationhood and the National Question in the Soviet Union and post-Soviet Eurasia: An Institutional Account. Theory alld Society, Vol.23, No.1. pp.47-78. Buchanan, A. (1995), Federalism, Secession, and the Morality of Inclusion, AriZOlla Law Review, Vol.37. No.1. pp.53-63. Bunce, V. (1999), Subversive Illstitutiolls: the Desigll alld the Destructioll of Socialism alld the State. Cambridge: Cambridge University Press. Cornell, S. E. (2002). Autonomy as a Source of Conflict: Caucasian Conllict in Theoretical Perspective, World Politics, Vol.54, No.4, pp.245-276. de Zwart, F (2005), The Dilemma of Recognition: Administrative Categories and Cultural Diversity, 77wO/)' alld Society, Vo1.34, No.2, pp.137-169. Dorff, R. H. (1994), Federalism III Eastern Europe: Part of the Solution or Part of the Problem? Pllblills: the JOllrnal (if Federalism, Vol.24, No.2, pp.99-114. Elazar, D. J. (1987), Explorillg Federalism. Tuscaloosa, AL: The UniverSity of Alabama Press. Erk, J. (2006), Does Federalism Really Maller? Comparative Politics. Vol.39, No.1, pp.I03-120. Erk, J. (2007), Federalism as a Growth Industry, Pllblills: the JOllrnal of Federalism, Vo1.37, No.2, pp.262-278. Filippov, M., Ordesi1ook, P. C. and Shvetsova, O. (2004), Desigllillg Federalism: A Theory ofSeilSlIstaillable Federal Illstitlltiolls. Cambridge: Cambridge University Press. Ghai, Y. (2000), Autonomy as a Strategy for Diffusing Conllict, in P.e. Stern and D. Druckman (eds), 11111'1'lIatiollai COlljlict Resoilltioll ajierthe Cold War. Washington, D.C: National Academy Press, pp.483-530. Gorenburg, D. (2003), Millority Ethllic Mobilizatioll ill the Russiall Federatioll. Cambridge: Cambridge Ulllversity Press. Gurr, T. R. (2000), Peoples versus States: Millorities at Risk ill the New CCIllllry. Washington. DC: United States Institute of Peace Press. Hale, H. E. (2004), Divided We Stand: Institutional Sources of Ethnofederal State Survival and Collapse, World Politics, Vol.56, No.2, pp.165-193. Hechter, M. (2()(JI), COlltaillillg Natiollalism. New York: Oxford University Press. Horowitz. D. L. (1985), Ethllic Groups ill COlljlict. Berkeley: University of Californm Press. LefT, C. S. (1999), Democratization and Disintegration: Federalism and the Break-up of the Communist Federal States, World Politics, Vol.51. No.2, pp.205-235. Levy, J. T. (2007), Federalism, Liberalism, and the Separation of Loyalties, Americall Political Sciellce Review, Vol.lOI. No.3, pp.459-477. Lustick, I. S., Miodownik, D. and Eidelson, R. J. (2004), SecessIOnism in Multicultural States: Does Sharing Power Prevent or Encourage It? Americall Political Sciellce Review, Vol.98, No.2, pp.209-229. McGarry, J. and O'Leary, B. (1993), Introduction: the Macro-political Regulation of Ethnic Conflict, in J. McGarry and B. O'Leary (eds), The Politics of Ethllic CO/!tlicr Regulatioll. London: Routledge, pp.I-47. Norris, P. (2004), Electoral Ellgilleerillg: Votillg Rilles alld Electoral Behavior. Cambridge: Cambridge University Press. Riker. W. (1964), Federalism: Origill. Operatioll. Sigllificallce. Boston: Lillie. Brown and Company. Roeder. P. G. (1991). Soviet Federalism and Ethnic Mobilization. World Politics. Vol.43. No.2. pp.196-232. Snyder. J. (2000). From Votillg to Violellce: Democratizatioll alld Nariollalist COlljlict. New York: Norton. Treisman. D. S. (1997). Russia' s 'Ethnic Revival': the Separatist Activism of Regional Leaders in Postcommunist Order. World Politics. Vol.49. No.2. pp.2 I 2-249.

Ethnofederalism and the Mismanagement of Conflicting Nationalisms PHILIP G. ROEDER Department of Political Science. University of California. San Diego. USA

Introduction Federal solutions to ethnic conflicts have received surprisingly favourable attention in the past decade. For example, on 26 September 2007 by a vote of 75 - 23 the US Senate adopted the Biden-Brownback amendment to the defence spending bill; this had 17 sponsors from both the Democratic and Republican parties. I The authors of the amendment explained that it should be seen as a call for the Iraq government to establish a federation of autonomous ethnic regions. The amendment's principal proponent, Delaware Senator Joseph Biden, writing in the New York Times, explained, "The idea, as in Bosnia, is to maintain a united Iraq by decentralizing it, giving each ethno-religious group-Kurd, Sunni Arab, and Shiite Arab-room to run its own affairs". He told reporters, "A federal Iraq is a united Iraq". In a parallel but not entirely independent effort, Edward Joseph of the Nitze School and Michael O'Hanlon of the Brookings Institution appeared on television and in print to explain their case for a federal Iraq that would comprise three ethnic states; they labelled their proposal "soft partition" (Joseph and O'Hanlon, 2007). And, in US newspapers and periodicals, there was a well-funded public relations campaign behind such proposals for a federal Iraq.

14

P. C. Roeder

But what is still more remarkable is the wide arTaY of conllicts around the world for which federal solutions have been proposed over the past decade? Some, like the English-Scottish conllict, seem nowhere near coming to blows-except in sports arenas. Others, like the so-called frozen conllicts in the Soviet successor states, have already escalated to the level of civil wars at least once (Coppieters, 200 I). In the Georgia-Abkhazia conllict, for example, in November 2001 the Group of Friends of the Secretary General of the United Nations presented the so-called Boden Paper, which proposed a federal agreement that would guarantee Abkhazia "broad powers" as "a sovereign entity ... within the State of Georgia" (quoted in International Crisis Group, 2007: 9). Non-governmental organizations have advanced still more imaginative federal solutions for this and the other frozen conllicts in the Caucasus, such as a Transcaucasus federation of Armenia, Azerbaijan, Georgia, Abkhazia, South Ossetia, Nagomyi Karabakh and, perhaps, still more constituent entities or a 'dual federation' that would link Nagomyi Karabakh to both Armenia and Azerbaijan but keep the latter two states separate (Abasov and Khachatrian, 20(6). Almost all of these proposals entail a specific type of federalism-ethnofederalismin which at least some, if not all, the constituent units of the federation are homelands controlled by their respective ethnic groups. This ethnofederalism is more akin to the autonomy <mangements found in non-federal states than simple (that is, non-ethnic) federalism: Although ethnofederalism and autonomy arrangements are like federalism in that they are state institutions designed to decentralize power among territorial jurisdictions, the first two types of arrangements also devolve powers to specinc minority communities within those jurisdictions (Roeder, 2007: 43-46). All devolved powers are concentrated in jurisdictions that can become proto-nation-states within the larger state (what I will label the common-state). Rather than capturing the benents of simple federalism in an ethnically plural society, ethnofederalism shapes the agenda of politics and distributes power in ways that increase the likelihood of escalating challenges to the common-state From alternative nation-state projects. This enthusiasm for ethnofederalism as a solution to conllicts in ethnically divided societies is a recent phenomenon-particularly in democracies. In the first half of the past century, ethnofederal and autonomy arrangements were typically attempts by dynastic or communist autocracies to Forestall the processes of national liberation; thus, the most commonly cited examples were the Austro-Hungarian, Ottoman and Russian empires and China, Czechoslovakia, the Soviet Union and Yugoslavia (Table I). Only one democracy-Canada-maintained an ethnofederation in which one province was based on a minority homeland; one other democracy-Switzerland-maintained a federal state within an ethnically plural society but no canton (with the possible exception of Ticino) constituted a homeland for one of the major ethnic groups. Following World War II, more ethnofederations were born in the process of decolonization, usually in an attempt to hold together previously autonomous jurisdictions, including Burma (1948), Indonesia (1949), India (1950), Ethiopia (1952), Pakistan (1956), Nigeria (1960), Malaysia (1963) and Tanzania (1964). It was only in the last decades of the twentieth century that European democracies borrowed this institutional innovation, with the introduction of ethnofederal arrangements in Spain (1978), Belgium (1980) and Bosnia (1995). The introduction of autonomous homelands for ethnic minorities within non-federal states also took off only after World War II.

Ethnofederalisl11 and the Mismanagement of COI!flicting Nationalisms

15

Table 1. Ethnofederations, non-federal states with autonomous ethnic regions. and federations, 1901Ethnofederations

Non-federal states with autonomous ethnic regions

Austro-Hungary

UK (1707-)

USA (1789-)

(1867-1918) Canada (1867 - ) USSR (!922-91) Yugoslavia (1945-92) Pakistan (! 947 -)

Russian Empire (1809-1917) Ottoman Empire (1832-1913) Finland (1922-) Denmark (1948- )

Switzerland (1848-)" Argentina (1853-) Venezuela (1864 - ) Germany (1867 - 1938.

Burma/Myanmar

Italy (1948-)

Brazil (1891-1934.

Indonesia (1949-50) Ethiopia (1952-62) India (1953-)

China (1951-) Uganda (1962-67) Sudan (1972-83. 2005-)

Australia (190 1- ) Mexico (1917 -)" Austria (1918-38,

Nigeria (1960-) Malaysia (1963-65)

South Africa (1972-94) France (1982-)

Libya (1951-63) Malaya (1957-63,

Tanzania (1964- ) Czechoslovakia (1969-92) Spain (1978-)

Philippines (! 990-) Azerbaijan (1991 - ) Georgia (1991 - )

Mali (1960-60)" Cameroon (1961 - 72)" United Arab Emirates

Belgium (1980-) Russia (1991 - ) Serbia-Montenegro

Tajikistan (1992-) Ukraine (1991-) Uzbekistan (1991-)

Comoros (1978 - )" SI. Kitts-Nevis (1983-) Micronesia (1986-)"

Federations

1949-) (1948-62)

1946-)

1945-)

!965-)

(1971-)

(1992-2006) Bosnia (1995-)

Moldova (1995 - ) Papua New Guinea (2000-)

This list includes only independent ethnofederations. states and federations. It excludes colonial federations. "The society is ethnically diverse (that is the indigenous minority popUlation constituted at least 25% of the total population in 2(00). but the constituents are not minority homelands. Source: Based on data in Elazar (1994). Lake and Rothchild (2005). Lemco (1991). Roeder (2007) and Walts (1999).

The enthusiasm with which ethnofederalism and autonomy arrangements have been embraced in the past decade is all the more remarkable because it appears to have run headlong into a substantial body of prior expert opinion warning against this. Eric Nordlinger (1972: 31) not only excluded federalism or ethnofederalism From his list of six conllict-regulating practices, but wamed that this "may actually contribute to a conllict's exacerbation and the Failure of conllict regulation". John McGarry and Brendan O'Leary (1993: 34-35) concluded that, although federalism has "obvious moral advantages over pure control", "unfortunately federalism has a poor track record as a conllict-regulating device in multi-ethnic states ... ". Ronald Watts (1999: 110) warned that when several divisions such as "language, religion, social structure, cultural traditions, and race" reinforce the territorial divisions of federalism (a correspondence that is inherent in ethnofederalism), stresses may grow within Federations to a breaking point (as in Malaysia, Pakistan, Nigeria, Yugoslavia

16

P. G. Roeder

and the USSR). Referring to this type of ethnofederal arrangement, Will Kymlicka (1998: 112-113, 138-139) concluded, "in general, it seems to be unlikely that federalism can provide an enduring solution to the challenge of ethnocultural pluralism". Moreover, "the very success of federalism in accommodating self-govell1ment may simply encourage national minorities to seek secession". Consistent with this academic scepticism, the policy community previously had been reluctant to embrace ethnofederalism or autonomy as measures to protect the rights of minorities. As Ruth Lapidoth (1996: I 1-19) noted, there is no endorsement (and at most only a passing mention) of either institutional arrangement in the 1966 Intell1ational Covenant on Civil and Political Rights, the 1990 Document of the Copenhagen Meeting of the Conference on the Human Dimension of the Commission on Security and Cooperation in Europe (CSCE), the 1992 UN Declaration on the Rights of Persons Belonging to National or Ethnic, Religious, and Linguistic Minorities, or the 1995 Council of Europe Framework Convention for the Protection of National Minorities. Only in the Lund Recommendations of 1999 did the Organization for Security and Cooperation in Europe (OSCE) give fuller attention to either institutional arrangement (Wolff and Weller, 2005: 3). This scepticism was supported by substantial evidence that ethnofederalism or autonomy arrangements seldom succeed when faced by serious nationalist divisions. Ruth Lapidoth (1996: 29) ended with the sobering assessment that the preponderance of her evidence shows that autonomy cannot heal a society that is already deeply divided: "So far. no arrangements of autonomy have succeeded in a hostile atmosphere ... If there is hatred and frustration, it is too late, and autonomy will not be able to soothe the strained atmosphere". Data presented by David Lake and Donald Rothchild (2005: 110-112) show that among the eleven ethnofederations or non-federal states with autonomy ammgements live subsequently split through de jure secession (Czechoslovakia, Malaysia and Pakistan) or de facto secession (Azerbaijan and Georgia). Two more (Nigeria and Uganda) became centralized states. Only four (Belgium. Canada, India and Papua New Guinea) survived. They concluded that their evidence shows that even in the absence of a previous civil war "territorial decentralization is a fragile political institution that ... is often quickly abandoned". Where conllict has already escalated to the level of a civil war, the prospects for ethnofederalism are grimmer still: they "find no evidence of successful institutionalization of these provisions in any postwar constitutional order" between 1945 and 1999. This initial scepticism was prudent. Ethnofederal and autonomy arrangements in ethnically divided societies structure politics inside ethnic communities, among ethnic communities and between those ethnic communities and the central govell1ment in ways that bring political instability. These institutions privilege some identities and interests and distribute coercive and defensive capabilities in a way that increases the likelihood of escalation of conllict into acute nation-state crises. Thus, these institutional arrangements are more likely than simple federal or unitary institutions to empower ethnic entrepreneurs who threaten the peace, the survival of democracy and the unity of the state. Particularly after civil wars, ethnofederalism and autonomy arrangements represent imprudent institutional choices. In this paper I will argue that these consequences inhere in the institutional arrangements of ethnofederalism and autonomy and, thus, distinguish these from simple

Etlll/{~t'ederalisl/I

and the Misl/lanagel/lent

(~l COI!tlicting

Nationalisl/ls

17

federalism or a unitary state. Tinkering with the institutional details of different forms of ethnofederalism or autonomy is unlikely to exorcise the demons, for the devil is to be found in ethnofederalism and autonomy aITangements themselves. The analysis in this paper begins with the assumption, which is at least as old as James Madison's Federalist No.5 I , that the structure of political institutions can shape the ambition of politicians in ways that disrupt or reinforce public order. Not only do these institutions check power holders, but they shape the configuration of interests likely to be pressed from society through political action. In this paper the key independent variable is the simple dichotomous distinction whether ethnofederal institutions are present or not. The outcome of interest (that is, the dependent variable) is stabilization of the political order through these institutions, where failure is measured by the dissolution of the state (often through escalating conllict or secession) or by dissolution of the institutions (often through centralization). The analysis begins with a 'model', which presents an abstract, logically derived simplification that draws our attention to the political dynamics that follow from different institutional constraints. The intervening variables are, first, the identities and claims that shape the agenda of politics and, secondly, the coercive and defensive capabilities that shape the balance of means in politics. The expectations generated by this model are tested against the cases shown in Table l. which I believe comprise the universe of cases from 1901 to mid-2008. Illustrations of key points are drawn disproportionately from Central Eurasia-that is, the region that was once the Soviet Union and now consists of the fifteen Soviet successor states. These cases provide a consistent body of evidence over an extended period of time-indeed, one of the few sets of cases that permit us to observe long-term processes in a variety of contexts.

Weaknesses in the Case for Ethnofederalism and Autonomy The argument for ethnofederalism and autonomy usually focuses on their purported immediate impact on agendas and capabilities. First, concell1ing agendas, ethnofederalism or autonomy is supposed to represent a compromise on the issue of statehood that will be acceptable to secessionists as well as central govell1ments. Both institutional arrangements purportedly give the secessionists some of what they demand in the form of a dependent but autonomous state, but permit the leaders of the central governmentto keep their common-state together (Lapidoth, 1996: 9; GUIT, 2000: 2(9). When conllict between a minority community and the central government has already taken a turn towards competing nation-state projects in which leaders of the minority community seek to secede, Michael Hechter (2000: 142-152) argued, ethnofederalism ought to reduce the demand for sovereignty by giving the local governments a voice in the provision of some collective goods, such as requiring a local language as the medium of instruction in schools or supporting one religion over others. Even when conllict has not yet taken a turn to competing nation-state projects, ethnofederalism is purportedly the best way to satisfy demands for political empowerment from ethnic entrepreneurs and to create a separate space for minority communities to flourish. Lustick et al. (2004) argued that regional autonomy permits extensive representation of ethnic groups in the local bureaucracy and that this, in turn, may prevent the spread of secessionist identities. Secondly, concerning capabilities,

18

P. C. Roeder

ethnofederalism or autonomy is supposed to represent the best means of balancing power so as to protect the rights of the secessionist community from predatory policies of the central government. Ethnofederalism in particular is supposed to provide minorities the defensive resources necessary to stay the hand of central government if it should try to diminish minority rights (Horowitz. 1985; Lijphart. 1996; Ourr. 2(00). Levy (2007) argued that ethnofederal provinces, when large and powerful enough. can protect individual liberties by acting as "bulwarks" that check the central state and by creating a "separation of loyalty" that checks any "unhealthy veneration" of the common-state. These arguments for ethnofederalism and autonomy typically suffer from five related weaknesses. First, they tend to focus on the short-term problem of finding a compromise to satisfy the demands of the parties to a conniet, but give less consideration to 'the days after' the new institutional arrangement is in place. This may be a smart approach for politicians seeking to muddle through a crisis and pass the problem on to their successors, but may leave the successors with a still worse problem. Secondly, these arguments tend to focus on finding an institutional compromise that will satisfy the greed and grievances on both sides in the current connict, but give less consideration to ways in which greed and grievances, in fact, are endogenous to the institutional arrangements. Ethnofederal and autonomy compromises may actually lead to escalation of demands in the next round of play. Thirdly, these arguments give little consideration to the ways in which these institutions structure not only greed and grievances, but also identities, capabilities and opportunities for escalating conniet over competing nation-state projects in the next round. Fourthly, to the extent they look to 'the days after', these arguments tend to focus on avoiding one type of coercive assault that may challenge post-settlement politics-predation from the commonstate government and recentralization. These arguments give less attention to predation by the homeland governments once they are empowered and the common-state government is weakened. Fifthly, these arguments fail to appreciate how ethnofederalism and autonomy arrangements trap politics between two perils-centralization and dissolution-with no stable equilibrium between these two extremes and no easy escape from this trap: Once ethnofederalism or autonomy is institutionalized, typically the only reforms on the bargaining table entail steering directly toward one peril in order to avoid the other. A Model of Ethnofederal Constraints on the Escalation of Nationalist ConrIict

Ethnofederal and autonomy institutions shape the agenda of politics and distribute power in ways that create or keep alive connicts among competing nation-state projects. 3 These institutional arrangements produce these effects by structuring politics on the periphery (that is, inside and among the ethnic communities) and between the periphery and the centre (that is, between those communities and the common-state government). Ethnofederal and autonomy institutions tend to create or to keep alive conllicts in which the issue is not simply about rights of different ethnic communities within a common-state, but whether the communities even belong in a common-state at all or should go their separate ways as distinct sovereignties. In these nation-state crises the possibility of secession by the homelands is on the bargaining table. In the absence

Etl/ll(~t"ederalislll

and the MislIlanagement ~t" COI!t"licting NationalislIls

19

of autonomous homelands, ethnic entrepreneurs are far less likely to have the support, resources and opportunities to press their claims successfully against a common-state government. Several students of simple federalism have documented the potential for recentralization in federations (Bednar et al., 1999; Lake and Rothchild, 20(5); ethnofederations also suffer from this more usual form of political instability, Since World War II, we have witnessed this recentralization take place in Burma. Ethiopia. India, Indonesia, Nigeria. Pakistan. Tanzania and. most recently. the Russian Federation (Lijphart. 1996; Roeder. 20(7). Ethnofederations and autonomy arrangements, unlike simple federations. also have a unique potential for the opposite form of breakdown. however-successful secession. Indeed. the designers of ethnofederations and autonomy arrangements and the politicians who subsequently conduct politics within these institutional frameworks are caught between the Scylla of over-centralization and the Charybdis of over-devolution; lying between the perils of recentralization that extinguishes autonomy and devolution that dissolves the common-state is a hard-to-identify knife-edge equilibrium of the properly designed and balanced constitution. Finding and maintaining that equilibrium is complicated because the parties at the bargaining table of constitutional design and post-settlement polities usually have short-term incentives to push the ship of state in opposite directions. And the state rocks perilously as a recurring crisis of politics focuses on the competing nation-state projects that pit homeland governments against the common-state govemment. Ethnofederalism. in particular, and autonomy arrangements, to a somewhat lesser extent. shape the subsequent agenda of politics so that the discussions come to focus almost exclusively on whether the state can avoid these perils by strengthening the central government or by further devolving powers to the homeland governments; thinking outside this box crafted by the constitutional designers and reinforced by the political actors empowered by the constitution becomes very difficult. Yet, it remains unclear how an altemative that avoids these perils might be found once non-federal solutions. like empowering cross-cutting interests. have been taken off the bargaining table or have simply been put out of mind (for an illustration. see Watts. 1999: 111-112). This institutional instability and the likelihood of nation-state crises, which inhere in ethnofederal and autonomy institutions, arise from four consequences that shape the agenda of politics and distribution of power (see Figure I). The political agenda is shaped by the tendency of these institutional arrangements to create and reinforce national-identity connicts and to focus centre-periphery bargaining on constitutional and sovereignty issues that express this national-identity connict. The distribution of power is shaped by the tendency of these institutional arrangements to strengthen politicians associated with secessionist claims by expanding their coercive capabilities and to weaken the common-state govemment by reducing its defensive capabilities. First, ethnofederal and autonomy institutions create or reinforce connict between identities associated with competing nation-state projects supporting the incompatible sovereignty claims of the homelands and the common-state. The creation of official homeland governments privileges one set of nation-state projects over the alternatives on the periphery and the creation of official homeland governments privileges the political elites associated with these nation-state projects (also see Rousseau and van del' Veen, 2005: 694.705- 7(6). On the periphery the introduction of official homeland

20

Etl/l/(~federalislll

P, G. Roeder

~

I

Institutional arrangement

r-

Centre-periphery agenda: National-identity conflict Allocation of decision rights

I--

K ~

Distribution of capabilities: Coercive capabilities Defensive capabilities

Institutional survival

I

-

Figure I. Schcmatic representation of an institutional model

governments is likely to reduce competition among multiple alternative nation-state projects (the cross-sectional effect) and to privilege over time one set of oflicial national identities (the temporal effect) associated with the homelands. In the absence of homeland governments it is more likely that none of the competing nation-state projects on the periphery will be strong and that individual nation-state projects, which might otherwise grow to challenge the common-state, will be ephemeral. In the absence of homeland governments there is likely to be a proliferation of nation-state projects as intellectuals compete with original new formulations, but each project is likely to garner only a limited following, to remain a pet project of its author and only a small circle of students, and to have a short half-life. The introduction of homeland administrations, based on one of these alternative ways of dividing the population and territory, changes this profoundly, creating conditions for one set of nation-state projects to become hegemonic over the alternatives and to survive. The very existence of the new states (such as union republics) within a common-state (such as the Union of Soviet Socialist Republics) focuses national identities, but the process is often helped by more tangible support li'OIn the homeland governments themselves. The leaders associated with these homelands are able to use their powers to establish the hegemony of the homeland's ofllcial nation-state project. Homeland leaders can offer inducements for other intellectuals to abandon their own separate nation-state projects and to join the new homeland. They can create institutes to invent, to elaborate and to perpetuate the cultures, traditions and histories of these peoples imagined by their nation-state projects. They can establish universities and pedagogical institutes to develop an official curriculum that propagates these cultures, traditions and histories among the next generation. No other nation-state projects on the periphery that challenge the common-state enjoy these advantages and the alternatives, for the most part, wither away. Ethnofederal and autonomy institutions typically create dual national identities; this is very different from dual or multiple ethnic identities found in most political systems because dual national identities constitute competing claims to membership in nations with rights to sovereign states of their own. Where dual national identities exist, the homeland leaders may come to playa critical role in mediating between the competing demands of the citizens' roles as members of two nation-states. This 'switchman role' is particularly important when the claims of the citizens' two roles come into conllict and the homeland leaders are the primary or sole point of contact

and tlte MislIlanagelllent

(~f COI({licting

NationalislIls

21

that many in the homeland population have with the common-state. In the late Soviet Union, for example, when the demands of their citizens' roles came into connict, the homeland leaders determined whether most would be mobilized into politics as citizens of the union republics or citizens of the Soviet Union. Even where loyalty to the Soviet Union continued to be particularly strong, as in Belarus, leaders such as Stanislau Shushkevich could use the powers of their oflices to shift popular participation in politics away from support of the Soviet Union and toward support of separate, independent nation-states. Secondly, ethnofederal and autonomy institutions risk focusing increasingly more interactions between the homeland and the common-state on constitutional and ultimately sovereignty issues that mirror this national-identity connict. The agenda of politics, which is established in bargaining between oflicials of the common-state and homeland governments, comes to focus on the allocation of decision rights between these governments. In the Soviet and post-Soviet experience, for example, the agendas of common-state politics structured by ethnofederalism or autonomy arrangements have tended to focus on three sets of demands by homeland govemments-limiting the power of the common-state government to write the rules of (or even to intervene in) the political life of the homelands, transferring more decision rights from the common-state to the homeland governments, and expanding the decision rights of homeland governments within the common-state government. These institutional arrangements shape the ways in which many other issues appear on the agenda of centre-periphery bargaining. Interests within the homelands seeking the most eflicient ways to get their issues on the policy agenda of the common-state have strong incentives to re-cast their demands in terms of the rights of the homeland. Cross-cutting claims that are not recast as homeland rights are often threats to the hegemony of homeland elites within their homelands and to their privileged position in centre-periphery bargaining, and so hegemonic elites within the homelands use their powers to co-opt or silence the voices of these competitors. Most clearly this is true when these entail alternative nation-state projects. Yet, this suppression is not limited to competing nation-state projects: In the last years of the Soviet Union, for example, the issues of environmental protection and economic reform were on the table of central decision-making organs in Moscow. The leaders of union republics worked to transform discussions of environmental protection and economic reform into debates over whether the common-state or homeland governments should exercise decision rights in these policy areas and over which government should exercise the sovereign power of the people to reallocate these decision rights (and possibly to dissolve the union). Ethnofederal and autonomy arrangements not only establish incentives for politicians to engage in these confrontations over decision rights, but also to link these confrontations to competing nation-state projects that can escalate to sovereignty claims. That is to say, escalation to nation-state crises is endogenous to these institutional arrangements. Homeland leaders demand a greater share of decision rights in order to shore up their positions within their homelands and extend their stay in ofllce. The autonomy to design the rules governing homeland elections is one of the most important tools used by homeland leaders to exclude their opponents from elections (for example, by language requirements for office holding) and for establishing

22

P. C. Roeder

a homeland political machine to extend the homeland leaders' political lives. Homeland leaders demand a greater share of decision rights controlling the allocation of economic resources in order to skim rents that homeland leaders can distribute to their followers. Homeland leaders demand a greater share of decision rights in order for their homelands to develop in divergent ways in politics and economics as well as culture and to isolate their societies from competitive and potentially homogenizing pressures from common political parties, a common economy, a common media, or inter-regional migration (Ghai, 2002: 158). As divergent development moves the homeland further along its separate path as a separate nation-state within the larger common-state, homeland leaders see less reason for common-state policy and may even see common-state policy as threatening to their continued separate development; so they demand a still larger share of decision rights. Most destabilizing of all are the incentives established by ethnofederalism for homeland leaders to play the sovereignty card-that is, to privilege their claims to greater powers by linking these to the claim that their minority community, which is the basis of their state within a state, constitutes a sovereign nation. In this way homeland leaders allempt to privilege their claims over the claims of other interests and against the counter-claims of common-state leaders. It is in the context of this agenda of politics that the distribution of means in ethnofederal and autonomy arrangements is likely to take on particular significance. First, ethnofederal and autonomy institutions are more likely than simple federal or unitary institutions to shift the balance of coercive capabilities to the advantage of secessionists. These arrangements expand the coercive leverage that leaders associated with secessionist nation-state projects exercise over the common-state government. The proponents of most nation-state projects lack access to the enormous coercive capabilities exercised by a state and so are unable to induce common-state governments to concede or even pay allention to them. Ethnofederalism and autonomy hand over some of the powers of the state to the proponents of a select set of such projects. The decision rights given to homeland governments can become institutional weapons-that is, decision rights that are then used to pry still more powers from the common-state government. In the Soviet and post-Soviet experience, for example, this leverage has included such institutional weapons as the ability to tip the outcome or even to disrupt the holding of common-state elections, to withhold federal taxes collected on the territory of the homeland, to embargo the flow of goods and energy within the common-state economy, and to force a showdown between police or troops controlled separately by homeland and common-state governments (Chapman and Roeder, 2007: 681; Roeder, 2007: 232). Secondly, alongside the shifting balance of coercive capabilities is often increasing vulnerability of the central government and a shift in the balance of defensive capabilities towards the periphery: The transfer of decision-making rights to homeland governments strengthens them against pressure from the common-state government, but weakens the common-state government so that it may no longer be able to defend itself (also see Manslleld and Snyder, 2002: 3(1). On one side of this balance, leaders of homeland governments who have used their autonomy rights to consolidate their own hegemony within their homelands are in much better positions to withstand counter-pressure from the common-state government. On the other side, where homeland governments exercise vetoes in common-state decision making, homelands

Etl/l[(~lederalism

and tfte Mismanagelllent of COI!llicting Nationalisms

23

can force deadlock in the common-state government. This, in turn, weakens the common-state government's ability to defend itself against the demands and even predation by homeland administrations. In the last months of the Soviet Union, for example, decision-making powers were transferred to union-republic governments from Mikhail Gorbachev's government; decision making within what remained of Gorbachev's government was held hostage to unanimity rules in the Council of the Federation (later the Council of State) that gathered the leaders of the union republics. As a consequence, the ability of the common-state government to defend itself from predatory homeland leaders withered away.

Consequences: Institutional Failure These processes appear to be common in ethnofederations, rather than the exception, and do not appear to be limited to communist or autocratic ethnofederations. The number of failed ethnofederations is most impressive. Of the eighteen ethnofederations that existed from 190 I to mid-2008 (the first column of Table I), eight broke up through secessions (Austro-Hungary, Czechoslovakia, Ethiopia, Malaysia, Pakistan, Serbia-Montenegro, the USSR and Yugoslavia). Another five saw recentralization or never actually implemented the devolution of powers due to autocratic rule (Burma, Indonesia, Nigeria, Tanzania and Russia). One remained under international administration (Bosnia). Only four remained whole, independent and democratic-the usual four of Belgium, Canada, India and Spain. Among the nineteen non-federal states with autonomies during the same 108 years, outcomes were only modestly better. Seven autonomy arrangements were terminated by imperial collapse (the Ottoman and Russian Empires), recentralization (South Africa, Sudan and Uganda), or de facto secession (Azerbaijan and Georgia). Another live were never implemented because power remained in the hands of the central government (China, Moldova, Tajikistan, Ukraine and Uzbekistan). Only seven survived (Denmark, Finland, France, Italy, Papua-New Guinea, the Philippines and the UK). The distinction between the consequences of ethnofederalism and autonomy arrangements, on the one hand, and federalism, on the other, is underscored by the fact that 33 ethnofederal or autonomy jurisdictions, but no constituents of simple federations, have seceded in the past 108 years. (One exception to the latter might be the break-up of the 61-day-old Mali Federation in 1960, but only if we stylize this as secession by Senegal from a simple federation.) Between 1901 and mid-2008, 41 new states were created by secession. Over 80% of these successful secessions (and 96% of the 26 secessions after 1945) resulted from the elevation of a pre-existing ethnofederal or autonomous jurisdiction to sovereign independence, such as the elevation of the Ukrainian SSR to independence in December 1991. AlLernati vely, none came about by elevating a simple federal jurisdiction that was not also a homeland for an ethnic group. Even in the four ethnofederations that have survived, the step-like escalation in identity conflict, privileging of the agenda of competing nation-states, expansion of the coercive capabilities of the homeland administrations and weakening of the common-state government can be seen. In India, Ramesh Dikshit (1975: 237) lamented after two decades of ethnofederalism,

24

P. C. Roeder despite a large number of apparent diversities, language has come to be recognized as the only cleavage-forming factor in the national life at the level of the common man. As internal political boundaries have now coincided with the linguistic ones, the cleavage lines between states have hardened, and language has begun to play divisive mischief.

In Canada the nation-state crisis exploded within less than five years of the granting of full sovereignty by the Constitution Act of 1982. Belgium survived the months-long deadlock over the formation of a common-state government in 2007 only because the common-state government had become so insignil1cant that the burden of governance was carried by the separate regional governments. With each new constitutional crisis and compromise reform the Belgian common-state appears to be withering away. In Spain the Basque Parliament on 30 December 2004 pressed its claims for greater autonomy in yet the next step of a slow progression of escalating claims on the Spanish common-state by playing the sovereignty card once again with a measure calling for a still looser relationship with Spain in a 'free association' and declaring the region's right to secede. In a parallel move the Catalonian Parliament pressed Madrid for extensive expansion of the region's decision rights. The socialist government of Jose Luis Rodriguez Zapatero, echoing the claim of Mikhail Gorbachev, has rationalized that further devolution is the best way to avoid violence, save ethnofederalism and keep the common-state whole. 4 In each of these cases observers have debated whether they are on a death watch for these ethnofederations; the inconclusive nature of this debate and the very uncertainty about the future of these countries are signs of the nation-state crisis in each of the purportedly successful ethnofederations.

Institutional Details that are Unlikely to Exorcise the Demons Since at least the time of Horowitz's (1985) pioneering speculations about the role of federalism in an ethnically divided society, many institutionalists advocating ethnofederalism or autonomy arrangements have insisted that constitutional designers need only "get the institutions right" in order to avoid the dangers described previously in 5 this paper. Yet, the accumulation of evidence in recent years suggests that institutional variations on ethnofederalism and autonomy arrangements do not profoundly change the consequences for post-settlement politics. Most of these variations simply entail adjusting the allocation of powers (decision rights), but this tinkering with the details highlights a variation on the central dilemma just discussed-ethnofederal and autonomy arrangements box subsequent politics into a narrowing range of reform options that limit subsequent reformers to a choice between the Scylla of recentralization and the Charybdis of secession. This also highlights that the devil is not in the minor details that distinguish one form of ethnofederalism or autonomy arrangement from others, but in the fundamental institutional difference between ethnofederal and autonomy institutions, on the one hand, and simple federal and unitary institutions, on the other. The former institutionalize and nest proto-nat ion-states within a larger nation-state, creating the conditions for a subsequent nation-state crisis. Four of the most commonly proposed tinkering reforms seek to guard against break-up by strengthening the common-state government and weakening the homeland

Etl/l/(~federalisll1

and the Mismanagement of Conflicting Nationalisms

25

governments, but by avoiding Charybdis they sail headlong into Scylla. First, constitutional designers can curb the decision rights of the homeland governments (Watts, 1999). They may limit the threat of deadlock in the common-state government by curbing the voting rights of homeland governments within common-state organs. Institutional arrangements can also limit the opportunities for regional leaders to free ride on the federation (also see Rodden and Wibbels, 2002). In many of the autonomy arrangements (compared to the ethnofederations) in Table I, the balance of power is relatively more favourable to the common-state government and this may account for the relatively lower break-up rate among the non-federal states with autonomy arrangements. Yet, this reallocation of decision rights simply increases the likelihood of recentralization. As Levy (2007) noted, in the logic of ethnofederalism the regional governments must have a plentitude of powers to become a bulwark against commonstate predation on minority rights. Secondly, constant change in ethnofederal jurisdictions can disrupt the processes by which regional leaders consolidate identity and political hegemony within their homelands and this may keep their homeland governments weaker relative to the common-state government. Several students of Nigerian federalism have noted that the multiplication of states has prevented new secessionism from gaining traction (Horowitz, 1985: 604). The very fluidity and impermanence of states may have weakened secessionism. Nigerian federalism has not permanently privileged one configuration of ethnic identities, but has progressively empowered more, smaller ethnic identities at the expense of larger, more inclusive groups. Nigeria began independence with three homelands but, in 1963, this became four, four years later this became 12, eight years after that this became 19, twelve more years passed and the number of homelands became 21. then four years after that it grew to 30, and after another five years it became 36. Yet, this is probably not a practice that could be implemented in too many ethnofederations; this seems a procedure that is hard to implement except under autocratic rule. The regional governments empowered in the initial ethnofederal settlement, if they are powerful enough to defend the interests of their communities against common-state predation, are likely to resist the threat to their power from further division. But when this resistance can be overcome, this second fonn of institutional tinkering simply creates an opportunity for recentralization, as it did in Nigeria. Thirdly, ethnofederalism can be used to disrupt the most divisive nation-state projects by empowering alternative ethnic communities that have not yet developed strong national identities. In India at the time of independence the most active nation-state projects that challenged unity were associated with religious communities and with the princely states, such as Hyderabad. The introduction of ethnofederalism with boundaries drawn along linguistic lines had the immediate effect of creating cross-cutting categories of empowerment that dissipated the immediate threat of secession. By contrast, Burma/Myanmar did not create cross-cutting jurisdictions to eliminate the princely states, but empowered these through ethnofederalism, and has faced continuing secessionism from these ever since. Yet, this strategy may only work for a while before the new ethnofederal jurisdictions become the source of secessionism. In India the linguistic nationalisms have grown in political strength. In the Soviet Union, cross-cutting ethnofederalism was the strategy that the Soviets used in Central Asia in the national delimitation of 1924. Moscow thought that the union republics would

26

P. C. Roeder

disrupt the most threatening nation-state projects of the day-notably the Turkestan, Bukhara and Khiva projects. With time, however, as the union republics consolidated their own nation-state projects, they became the major threat in Central Asia to the unity of the Soviet Union. Fourthly, various forms of asymmetrical federalism, in which only minority ethnic groups enjoy uniquely empowered homeland governments, may strengthen the common-state government. Henry Hale (2004: 166-167) argued that the threat of secession can be contained if the core nationality in an ethnofederation, such as the Russians in the Russian Federation, is divided among several federal units: In the Russian Federation, during its first decade of independence, the 89 constituent units of the federation included only 32 that were rormally homelands of some minority group; 57 of the constituent units were provinces or rederal cities with Russian majorities. The provinces did not enjoy the same rights to selr-government exercised by the republics based on ethnic-minority homelands. This, Hale argued, prevented the developments that he contended brought the break-up of the Soviet Union-fosterage of an alternative nation-state project ror the core nationality (Russians within the USSR), emergence of a revolutionary 'dual power' crisis between the governments or the common-state and the homeland administration of the core nationality (Gorbachev versus Yeltsin in Moscow), and reduction of the capacity of the common-state government to credibly commit to protect minority rights. Yet, asymmetry may be difficult to introduce and maintain and may generate its own conllicts. As Yash Pal Ghai (2002: 160-162) noted, asymmetry that grants minority homelands greater powers of selfgovernment creates resentment within the numerically larger core nationality. Pressure from the Basque Country, Catalonia, Corsica and Quebec for greater autonomy than enjoyed by other regions, departments or provinces has already brought a backlash from the majority populations and their many regional governments in each country (Kymlicka, 1998: 128-136). Any scheme to represent the minority homeland governments in central decision making is likely to become caught up in a bruising light in which the provincial govelllments of the core nationality demand equal votes for all regions-a scheme that typically guarantees the core nationality many more votes than anyone or the minorities. In response, the homeland governments of the minorities are likely to reject this equality among regions and demand that their votes be weighted so that the minorities (individually or collectively) are closer to parity with the core nationality in total votes in common-state decision making. In Russia this very issue brought the rederation to a nation-state crisis in 1993. The demand rrom republics to weight their votes above those or provinces in the Council of the Federation and to extend greater powers to the republics produced a backlash among the province leaders who threatened to proclaim their jurisdictions republics; several republic leaders threatened to separate their homelands further from the Federation ir the centre gave in to the demands of the province leaders. This was resolved only by a presidential coup in autumn 1993 that imposed a solution denying the republics' demands. Moreover, asymmetry does not appear to be a guarantee against secession: the de facto secessions of Abkhazia and South Ossetia from Georgia and of Nagornyi Karabakh from Azerbaijan, as well as the dejllre secessions of Singapore from Malaysia and of Eritrea from Ethiopia suggest that, even if asymmetrical ethnofederations do not break up, pieces of asymmetrical ethnofederations often break away.

Ethnofederalism and the Mismanagement of COI!f/icting Nationalisms

27

Restatement: Ethnofederalism is Not Simply Federalism This discussion underscores the fundamental dirferences between federalism and ethnofederalism-both in their structure and in their consequences for politics in ethnically divided societies. Indeed, simple federalism can be used to disrupt the processes described previously that lead to nation-state crises: Various forms.of ~anto-. nization or federalism that divide ethnic communities may prevent the consolidatIon 01 national identities on the periphery that compete with the roundations of common-state unity and may block the conllation of political agendas into a connict over alternative nation-state projects (McGarry and O'Leary, 1993).6 Ethnically homogeneous cantons that divide ethnic communities may encourage inter-regional competition within ethnic groups. Ethnically heterogeneous cantons that divide ethnic, communiti~s may also 'quarantine' inter-ethnic connicts and encourage cross-eth111c co-operatIOn on behalr of the region's interests in inter-regional conllicts. It may matter less whether the states within the common-state are homogeneous or heterogeneous, than that they divide potential nation-states within the common-state into smal.ler communi.ties and interests. This was the logic of the Vance-Owen project for Bos11la and part 01 the logic ror preserving the eighteen provinces or Iraq in a rederal arrangement rather than concentrating power in three ethnic regions. In Nigeria, after 1976, Hausas were divided among six states, the Yoruba among fjve states and the Ibo between two (Horowitz, 1985: 604-613). This discouraged consolidation of a pan-northern ethnicity uniting Hausa and Fulani, and it empowered still smaller communities and intel:ests, such as Kano Hausas, within the major ethnic groups. To the extent that It becomes harder to form coalitions among ethnic kin, because they now must reach across provincial lines, this weakens the ethnic groups against the common-state government and may forestall the consolidation of nation-state pro~ects that ch,allenge the unity or the common-state. To the extent that any new eth111c hegem0111es that emerge within the smaller states are proportionately weaker in any confrontation with the common-state government, the threat of break-up declines. Of course, cantonization brings the rami liar risk associated with weakening the challenge of alternative nation-state projects: The threat of recentralization vis-it-vis the ethnic minorities rises. This use of federalism to avoid the dangers of ethnofederalism highlights the key difference between the two institutional arrangements. Each shapes the agenda of politics difrerently. When the constituents of a federation are homelands, these constituents are more likely to become alternative nation-states within the common-state and are more likely to press successfully for independence. Homelands do not nece~saril~ have more powers than federal states but, in ethnofederalism, the concentratIOn 01 devolved powers in homeland governments increases the likelihood that a sel~ct set of national identities will become privileged over other identities on the penphery and come into connict with the national identity that holds the common-state together. Centre-periphery bargaining is more likely to rocus on a connict over competing nation-state projects. The coercive capabilities of the periphery are more likely to fall into the hands of secessionists who use their powers to weaken the commonstate government. Because institutions empower politicians whose careers depel~d on these arrangements, once ethnofederalism or autonomy arrangements are 111 place, homeland and common-state leaders are likely to limit constitutional options

28

P. C. Roeder

on the bargaining table to tinkering with the ethnorederation in ways that rock the state precariously between centralization and dissolution.

Notes I'Biden Plan for Iraq Passes Senate' on http://www.biden.senate.gov (accessed 17 October 2007); 'Biden Defends Senate Vote on Partitioning Iraq' on http://www.cqpolitics.com (accessed I October 20(7); Helene Cooper, '(Political Equations) Iraq Math: From One, Make Three', The Nell' York Times, 29 July 2007. Section 4. pp.l. 12. 2 For example. in the Israeli-Palestinian conflict. proposals have once again been advanced for a "bi-national state" linking Palestine and Jordan in a confederation; see Ian Bremmer. 'A Difficult Plan Whose Time I-las Come'. Illternatiollal Herald Triblllle. IS June 2007. 'This section draws on a model and evidence presented in Roeder (2007) and Chapman and Roeder (2007). 4 Renwick McLean, 'Spanish Basques Approve Secession Measure'. The Nell' York Times. 31 December 2004. p.A8; Renwick McLean. 'Voters in Catalonia Approve a Plan for Greater Autonomy'. The Nell' York Times, 19 June 2006. p.A8. 51 lahel Horowitz's discussion 'speculations' only hecause he had so few empirical cases to examine at that time, not to detract from the contribution of his propositions. 61n addition to federalism. however. there are alternative forms of territorial devolution that may protect the rights of minority communities without creating a knife-edge equilihrium (Roeder. 200S). The multiple-majorities or power-dividing strategy devolves power both horizontally and vertically to multiple. overlapping. functionally specific jUrISdictions rather than a single layer of general-purpose homeland administrations (also see Hooghe and Marks. 2(03).

References Ahasov. A. and Khachatrian. H. (2006). The Karabakh COlljlict: Variallts (!{ SelllemellI. COllcepts alld Reality. Baku and Yerevan. Available at http://www.ca-c.org (accessed 27 February 2(09). Bednar. J., Eskridge, W. N. Jr. and Ferejohn, J. (1999). A Political Theory of Fedemlism, in J. Ferejohn. J. N. Rakove and J. Riley (cds). COllstill/tiollal CllllUre alld Democratic Rille. New York: Camhridge University Press, pp.223-267. Chapman. T. and Roeder. p, G. (2007), PartitIOn as a Solution to Wars of Nationalism: The Importance of InstitutIOns, Americall Political Sciellce Review. Vol. 101. NoA. pp.677-691. Coppleters. B. (2001). Federalismalld COlljlict ill tile CauCllsus. London: Royal Institute of InternatIOnal Affairs. Dikshit, R. D. (I 97S). The Political Geography (!f Federalism: All IlIqlliry 11110 Origills alld Stability. New York: John Wiley and Sons. Elazar. D. J. (1994). Federal Systems (if the World: A Halldbook of Federal. COllfederal. alld Alllollomy Arrallgemellts. London: Longman. Glmi. Y. P. (2002), Constitutional Asymmetries: Communal Representation. Federalism. and Cultural Autonomy, in A. Reynolds (ed.), The Architectllre (if Democracy: COllstitutiollal Desigll. COIl/liet Mallagemellt, alld Democracy. New York: Oxford University Press, pp.141-170. GUff, T. R. (2000), Peoples Versus States: Millorities at Risk ill the Nell' COIlIII)'. Washington. DC: United States Institute of Peace Press. Hale. H. (2004), Divided We Stand: International Sources of Ethnofederal State Survival and Collapse. World Politics, VoI.S6, NoA, pp.16S-193. Hechter. M. (2000), COlltaillillg Natiollalism. New York: Oxford University Press. Hooghe. L. and Marks, G. (2003), Unraveling the Central State. but How? Types of Multi-level Governance, Americall Political Sciellce Review. Vo1.97, No.2, pp.233-243. Horowitz, D. L. (l98S), Ethllic GrollPS ill COlljlict. Berkeley: University of California Press. International Crisis Group (2007), Abkhazia: Ways Forward. Brussels, Europe Report No. 179. 18 January.

Ethl/(~federalism

and tlie Mismanagement of COI(fiicting Nationalisms

29

Joseph. E. P. and 0' Hanlon. M. E. (2007). The Case fill' Soft Partitioll ill Iraq. Washington. DC: Brookings Institution, The Saban Center for Middle East Policy Analysis Paper No.12. Kymlicka, W, (1998), Is Federalism a Viable Alternative to Secession? in P. B. Lehning (ed.). Tlleories of Secessioll. New York: Routledge. pp.1 I I-ISO. Lake. A. D. and Rothchild. D. (200S), Territorial Decentralization and Civil War Settlements, in P, G. Roeder and D. Rothchild (eds), SlIstaillable Peace: Power alld Democracy after Civil Wars. Ithaca: Cornell University Press. pp.1 09-132. Lapidoth. R. (1996). AutollOIllY: FIt'xible SOlllliolls to Ethllic COIl/lictS. Washington. DC: United States Institute of Peace. Lemco. J. (1991). Political Stability ill Federal Govel"llmellls. New York: Praeger. Levy, J. T. (20(J7), Federalism, Liberalism, and the Separation of Loyalties. Americall Political Scie/lce Review, VoI.IOI, No.3, ppAS9-477. Lijphart, A. (1996). The Puzzle of Indian Democracy: A Consociational InterpretatIOn. ilmericall Political Sciellce Review, Vo1.90, No.2, pp.258-268. Lustick.l. S .• Midodownik. D. and Eidelson, R. J. (2004), Secessionism in Multicultural States: Does Sharing Power Prevent or Encourage It? Americall Politica/ Sciellee Review. Vo1.98. No.2. pp.209-229. Mansfield, E. D. and Snyder. J. (2002). Democratic Transitions. Institutional Strength. and War. IllterlIatiollal Orgalliz.atioll. VoI.S6. No.2, pp.29 7-337. McGarry. J. and O'Leary. B. (1993). The Politics (~f Ethllic COIljlict Regll/atioll: Case Studies of Protracted Ethllic COIl/licts. New York: Routledge. Nordlinger, E. A. (1972). COIl/lict Regll/atioll ill Divided Societies. Camhridge. MA: Harvard UniverSity, Center for International Affairs Occasional Papers No.29. Rodden. J. and Wihhels. E. (2002). Beyond the Fiction of Federalism: Macroeconomic Management in Multitiered Systems. World Politics. VoI.S4. NoA. ppA94-S31. Roeder. P. G. (200S), Power DiViding as an Alternative to Ethnic Power Sharing. in P. G. Roeder and D. Rothchild (cds), SlIstaillable Peace: POlI'er alld Democracv l!{ter Civi/ Wars. Ithaca. NY; Cornell University Press. pp.SI-S2. Roeder. P. G. (2007). Where Natioll-States COllie From: 11Istitutiollal Clzallge ill tlze Age (!{ Natiollalism. Princeton: Princeton University Press. Rousseau, D. and van der Veen. A. M. (200S). The Emergence of a Shared Identity: An Agent-Based Computer Simulation of Idea Diffusion. jOlll"lla/ of COIl/lict Reso/lltioll. Vo1.49, No.S, pp.686-712. Watts. R. L. (1999), COlllparillg Federa/ Systems. 2nd edn. Montreal-Kingston: McGill-Queen's University Press. Wolff. S. and Weller. M. (200S). Self-determination and Autonomy: A Conceptual Introduction. In M. Weller and S. Wolff (eds). Autollom\" Se!f-Gol'e/"l/allce, alld COIl/liet Resollllioll: Ill/lovative Approaches to Illstitllliollal Deslgll ill Divided SOCieties. New York: Routledge. pp.I-25.

Secession and Institutional Accollll1lodation

The Political Dynamics of Secession and Institutional Accommodation HUDSON MEADWELL Department of Political Science, McGill University, Montreal, Quebec, Canada

31

of conflict in divided societies". Under certain fairly general conditions, I argue, accommodation up to and including federalism will not be stable without further manipulation of the costs of secession by the state or enforcement of the terms of accommodation which define the political equilibrium in question. I am sceptical that institutional design can induce stability in these situations, given the initial condition assumed in this discussion, namely deep division, without presupposing that stability is imposed, whether by one group over another or by the actions of a third party. The analysis here, moreover, takes up this problem of institutional design in a way that differs from the Introduction. Some of the strategic features of these situations are considered so as to take into account some of the 'uncodified factors' identified by the Co-Editors (including 'will and capacity'), which have consequences for the stability of political equilibrium. I begin, however, by situating my discussion in some of the literature on secessionism, both contemporary and historical. I

Introduction The literature on secession is largely a product of the last twenty or twenty-live years and is primarily a response to contemporary political history, although there is older literature on the legal aspects of secession and on individual cases. I Contemporary politics, however, has motivated sustained interest in the normative and empirical dimensions of secession. An important part of this work, both normative and empirical, addresses the desirability and feasibility of institutional arrangements which might take into account the kinds of heterogeneity associated with secessionism, without breaking apart existing states. Federalism lies within these institutional solutions to the general problem of divided societies. The leading edge of the argument developed below is not federalism per se, however, but the political dynamics of secession? The paper, thus, takes up the central issue posed in the Introduction to this Special Issue-whether institutional accommodation is 'secession preventing' or 'secession inducing'. It does so by explicitly posing the problem of institutional accommodation as a problem of political equilibrium. In this way, a related question posed in the Introduction is also taken up: whether " ... federalism provides a stable, long-lasting solution to the management

I want to draw attention away from one kind of situation that has influenced work on secession and related phenomena, particularly ethnic conflict. This is the analysis of situations in which ethnic groups or nations have incentives to pre-empt rivals and to move toward new state formation in the short term-to use force sooner rather than later in interactions with a state or other groups. These are often situations, moreover, in which states have either begun to fail or are well along the path of complete collapse and new political relationships are being forged. The paradigm case in one of the most influential treatments of the dynamics of these situations (Fearon, 1998) is the interaction between Croats and Serbs in Croatia after the Croatian parliament's declaration of independence in 1991. Attention is drawn instead to situations in which short-term incentives, associated with insecurity and fear of the future, are not necessarily present but in which an interest in secession continues to be present. Even when basic survival is not in question, secessionism does not always wither away if institutional accommodation is not self-enforcing. What would secessionism look like over the longer term, even without the kind of insecurity about basic survival which is emphasized in this literature?} There are some fairly clear signs that in some cases secessionists can have long time horizons, such that their basic calculation might be put: "Better not to try and secure more now, better to build on concessions won". The theoretical interest of this calculation is that, if taken up in a secessionist movement, it implies that activists do not believe that the new political arrangements after concessions are self-enforcing. Rather, these arrangements are considered to be vulnerable to further political mobilization. There are secessionist movements which span decades and generations. Since these movements often are related to earlier forms of political mobilization that are not strictly secessionist, they might arguably be said to have even longer histories. These movements, moreover, need not be located in zones of war and violence, nor must we assume a history of conflict and war from which actors have recently exited as a backdrop to the discussion.

32

H. Meadwell

Rather, the motivating condition here is one or disagreement about the distribution or territorial power, when there are no short-term incentives to fight one's way out, and the question is: under what conditions can this kind or situation be secession-proored? More directly, how can territorial decentralization, up to and including federal institutions, contribute to secession-proorness?4 II One of the central issues in political mobilization in general arises in secessionist mobilization as well and it bears on this question. The issue is timing: "[I It is a question of catching the right instant. Too early you break your neck. Too late you lose your time". These are the words of a Breton nationalist, writing in the late I 960s and rel1ecting on his political activity in the interwar period. s The political proposition embedded in these words is quite recognizable. Debates in secessionist movements between fundamentalists and gradualists or between purists and pragmatists are certainly in part about tactics and timing, about when (ir at all) to settle for less, without giving up the ultimate goal, of how concessions in the short term might enable (or disable) further mobilization in the future and thus further concessions in the future. In this light, it is unlikely that gradualists or pragmatists are going to act as enforcers, on an enduring basis, of an agreement that secured concessions. They do not see the agreement as something permanent, and it is this attitude that secures their position in the debate with rundamentalists. It may secure their position but, since these debates can be violent, it may not secure their lives, when fundamentalists use rorce to challenge agreements which fall shOit of their expectations At the same time, it is unlikely the case that those in the group who did not seek or support a demand for territorial concessions have the power to enforce such an agreement since, by hypothesis, if they were this powerful within the group, concessions might not have to have been orfered by the state agent in the I1rst place. These kinds or dynamics are common in secessionism, such as in the run up to sequential secession in the American South in the period from the nullification crisis in 1832 to 1860 (Meadwell and Anderson, 2(08). A striking reature of political mobilization in this case is the existence early on of what can only be called a secessionist plan, which focused on accomplishing secession through time. While care should be taken not to read too much forethought into any historical record, central actors in the radical core of the secessionist movement knew what they were planning. Secession was both end and means-a means by which to protect a slave order and an end against which political tactics would be evaluated. This is not to say that the radical core of the secessionist movement spent every waking moment mobilizing support or plotting the Southern revolution. However, it is very difficult to make sense of the politics or secession without taking into account the strategic thinking or secessionists and how through time they quite consciously built a political stronghold in South Carolina rrom which to mobilize outward. There were radical secessionists in South Carolina who preferred immediate secession, even if it meant that secession would be a unilateral act, unsupported by other states. The dominant tendency within the vanguard from Nullification onward rejected unilateral secession as a political option because of the political risks associated with secession in one state. Yet, the goal of

Secession and Institutional Accol/llllodation

33

secession was not abandoned or sacrificed; rather, the decision was to bide one's time until the moment was ripe and the likelihood or support from other states was higher. In Ireland after the signing of the Anglo-Irish Treaty in 1921, there was impassioned debate within the nationalist movement, indeed within most or Irish society, about the terms of the Treaty. Discussion centered, not primarily on the partitioning of Ireland which followed from the negotiations between the Irish plenipotentiaries and the British government, but rather on the symbolic issue of the Irish Republic and the oath of allegiance to the British monarch, which was tied to Dominion status within the British Empire. The problem from the point of view of the fundamentalists was not just that Ireland was now divided. The ratification debate in the Dail Eireann instead revealed their rejection of what was conceded by the British in the negotiations-Dominion status. When Michael Collins, one of the Irish negotiators, argued that the Treaty should be ratil1ed by the Dail, he presented a classic kind of defence or the value of concessions won in the short term for further mobilization in the future: "I do not recommend it [the Treaty] for more than it is. Equally, I do not recommend it for less than it is. In my opinion, it gives us freedom, not the ultimate rreedom that all nations desire and develop to, but the rreedom to achieve it".6 This was the 'stepping-stone' derence or the results or the negotiations. 7 That defence engaged an equally classic kind of response. To cite just one example: [Nlor do I intend to criticise the actions of those who support the Treaty honestly, on the grounds that it is a stepping stone to freedom. That may be so; time will tell. For my part I feel some day they will have a very rude awakening; to my mind, to get on that stepping stone you must drop rundamental principles. I cannot rollow them, never more so than when that involves the sovereign independence or my country".H In the contemporary sovereignist movement in Quebec, there have been longstanding differences about etapisl1le, a political strategy that seeks to progressively move toward independence in stages, by building on concessions achieved in the short term, especially concessions that strengthen the institutional capacity of the provincial state to shape identity-ronlliltion and to regulate culture and economy (Meadwell, 1993). The contemporary history or the Scottish National Party (SNP) (e.g. Levy, 1990; Miller, 1999) is marked by similar dynamics between gradualists and fundamentalists. There are differences or course since the legislature in Scotland is or recent provenance but now that the SNP shares power in Edinburgh the problems or how power is to be used, when a referendum is to be held and, if held, how it is to be worded, will ensure that these debates within the party and movement continue to be of consequenceY The politics of the Basque country (Lecours, 2(07) continue to exhibit these kinds of dynamics, such as during the debate about the Ibarretxe Plan, although made more complicated by the ract that, like Ireland, the movement has civil and military wings. 10 III Secessionists prerer independence to the status quo modified by concessions. With a plan that stretches in time into the future, they can accept concessions as a stepping

34

H. Meadwell

stone and continue to work toward the implementation of their plan. Sincere secessionists with a plan approach secession indirectly, neither immediately rejecting concessions because they leave them unsatisfied, nor adapting to them by fully accepting them. If concessions take the form of territorial decentralization, secessionists can use them as a baseline for further challenges; every concession won in terms of territorial decentralization brings them closer to their ideal point (Sorens, 2(04). They can act in these ways as long as their time horizon is not so short as to preclude this kind of planning. Further, if secessionists can do all of this, then the political status quo after concessions is not self-enforcing; rather it continues to be vulnerable to secessionist challenges, even if secessionists forego their preference for secession in the shott term. A decision to defer in time a direct move toward independence is a form of selfdenial. The political sacrifice which is implied when one's time horizon is long is, in fact, to acknowledge the possibility that independence will not be realized in one's lifetime, and to accept this fate, but to work to ensure that the next generation will be able to reap the full benefits of the concessions that you, and like-minded individuals, have won from the state. The political goal of the sincere secessionist with a plan, then, is to increase the odds of successful secession even if it does not occur in one's lifetime. In this light, it is worth noting that a state agent might have shorter time horizons than a sincere secessionist with a plan. This is particularly the case if the agent of the state is elected and elections are relatively free and competitive, such that governments are regularly replaced. By hypothesis, sincere secessionists are relatively insensitive to partisan governments-their time horizon and basic political commitments account for this insensitivity. A partisan politician, however, is sensitive to electoral incentives and, in general, these incentives encourage the development of a short time horizon tied to elections and electoral cycles. As a consequence, elected politicians may be willing for partisan reasons to make concessions in the short term if these concessions improve their competitive position in real-time partisan competition, even if these concessions produce arrangements which are closer to the ideal point of sincere secessionists." The kinds of concessions in mind here, moreover, are highly visible; they are publicly known rather than hidden. Political opportunism produces private political benelits. In effect, the politician is deferring the full cost of concessions until later; these are costs that will be borne by others. Thus, the elected politician reaps the partisan benellts of making concessions while passing on the potential costs to other politicians who come later. And, as a result, sincere secessionists with a plan move closer to their ideal point. Politicians in the future then may lind themselves boxed in by earlier partisan-driven decisions made by others. For example, it becomes dif(1cult to roll back concessions made by earlier politicians as long as secessionists playa sophisticated game which conforms to the basic rules of the political game. The secessionist, thus, has a strategic advantage but this advantage can be cashed in only later rather than sooner. These considerations might not be operative when (a) politicians are elected but are insensitive to electoral incentives, or (b) politicians are elected and are sensitive but there are no incentives to make concessions, or (c) when a state agent is not elected and thus not vulnerable to electorally driven pressure. The IIrst situation (a) is not

Secession and Institlltional Accom11/odation

35

an oxymoron but a politician who resists short-term electoral incentives may soon be out of a job. To avoid the latter outcome, a politician would have to respond to the incentives, either positively by following them or by resisting them, seizing the opportunity to demonstrate leadership and lead rather than follow, risking political defeat in order to refashion the political dynamics of the situation. The second situation (b) is imaginable, of course-there may be situations or relationships in which there is never a shott-term incentive to make concessions. The third (c) is also imaginable and illustrates the value of thinking about the time horizons of state agents and considering the consequences of varying incentives according to political competition. Electoral competition is just one sort of political competition; even unelected state agents may face competition, often of a nastier sort. While further variations on these themes might be pursued elsewhere, the general point relevant for this paper has now been made. A secessionist with a long time horizon may have a strategic advantage in pmticular situations. What a state agent seems to need, however, if concessions are to remove the possible strategic advantage of the sincere secessionist with a plan, is an enforceable contract that ensures concessions are exchanged for good behaviour-do not continue your secessionist ways. Given the partisan logic introduced above, it is dif(1cult to see how this could be accomplished. Sincere secessionists would not have to agree to this exchange, simply because the concessions will be made whether or not the secessionists agree to give up their secessionist ways. This is the essence of the sincere secessionist's strategic advantage. As a consequence, state agents today or their successors tomorrow are left free to use concessions if there are partisan advantages in making them, while passing the costs on to later decision makers, and secessionists with a plan move closer to their ideal point. This is not to say that, in general, the partisan logic cannot run the other way as well. If there can be partisan incentives to make concessions, there also can be incentives to roll back concessions or to hold fast in the face of demands for further concessions. When this is the case-when the partisan logic runs both ways-one can see why even sincere secessionists might have to seriously consider and perhaps accept some kind of agreement, ({they believe that the agreement fixes in place and protects a political arrangement closer to their ideal point than another political arrangement that the state agent for partisan reasons, might be interested in pursuing. However, it seems to be the case in some important instances that concessions made are effectively invulnerable. The likelihood that the Scottish parliament will be rolled back must be about zero. The odds that the Catalan parliament will have fewer powers in ten years than it has now are negligible. The likelihood that the institutional capacity of Flemish political institutions will be weakened to the advantage of Walloons is rather low. The probability that the provincial government of Quebec will lose its capacity to regulate immigrant entry is basically zero. Moreover, in these cases, and in a growing number of others, the politics of independence is a legitimate form of political activity. The goal of political independence does not have to be renounced or given up as long as certain norms are respected, particularly by renouncing the use of violence. It is a short step from the legitimacy of this political activity to the legitimacy of independence itself. The political implication is that the goal of that activity-independence-is, in principle, legitimate and feasible,

36

H. Meadwell

depending on when and how it is accomplished. Secessionists cannot be prosecuted under sedition laws nor persecuted as traitors. All of this changes the social and political imaginary signilkantly, which in turn modi lies the strategic structure of these situations. 12 In these kinds of situation, state agents cannot credibly threaten to roll back concessions in an attempt to induce sincere secessionists to agree to give up their secessionist ways. While those agents might still be able to manipulate the costs of secession, particularly the transition costs, state agents are thus in a weaker position. The prevention of secession then can come down to the comparative willingness of secessionists and the state or other groups to bear costs and to inllict costs on others and their ability to demonstrate that willingness. The particular features of the institutions of accommodation no longer bear directly on secession prevention. A question emerges here, however. Why wouldn't state agents move earlier to demonstrate their willingness to bear costs and to innict costs in order to prevent secession, since they were in a stronger position earlier rather than later? We answered this question by hypothesis earlier in this section: state agents have partisan motivations to make concessions early. Sincere secessionists with a plan still can be vulnerable to challenges from fundamentalists. The risk that sincere secessionists are willing to take is the nip side of the fear of the fundamentalists. They fear that independence will be endlessly deferred. There is the further motivation, which can underpin the fundamentalist point of view: the motivation to act sooner rather than later so that independence can be achieved and experienced in one's own lifetime. In de felTing a move to independence until later, sincere secessionists do risk not achieving independence at all. However, if the social and political imaginary is changed, as outlined above, fundamentalists are in a weaker position to challenge sincere secessionists with a plan within the secessionist movement. These changes are a measure of success. Sincere secessionists with a plan have been able to move the situation from one in which new state formation was illegitimate to one in which political activity in support of independence is legitimate, and in which, as a consequence, the possibility of independence becomes politically imaginable. The likelihood of independence has increased, even if fundamentalists are in a weaker position.

IV The concept of a contract has been introduced here to underline the separability of concessions from their enforceability. The existence of feasible concessions does not secession-proof these kinds of deeply divided societies. {{political arrangements are secession-proof it is because concessions are backed up by an enforceable contract. Enforceability must be separately specilled in order to fully grasp the potential problems of concession making. Concessions do not have natural limits; moreover, by hypothesis, concessions are being made to a secessionist who prefers independence to the status quo modi lied by concessions. Territorial decentralization, including federal arrangements, is often proposed as a way of accommodating certain forms of cultural heterogeneity which characterize

Secession and Institutional Accollllllodation

37

divided societies. Further, decentralization is sometimes considered a mechanism by which to diffuse conllict and violence in these kinds of societies (e.g. Brancati, 2006). The argument to this point shows why and how these proposals might be underspecilled. Decentralization on its own may conduce to stable accommodation only if the agreement underlying accommodation is enforceable. Decentralization also might lower the likelihood of violence, thus defusing a conIlict situation, but this does not remove the interest in secession. Rather, what it can induce is a change in tactics from violent to non-violent means among sincere secessionists. Giving up physical violence is not the same as giving up independence. And this is another place where sincere secessionists might act strategically in order to win concessions, including decentralization, which, if accomplished, leaves them closer to independence and, absent enforcement, leaves them able to pursue a longer-term agenda of independence. The argument also shows that there is something like a problem of "precedentsetting" in the relationship between a state agent and sincere secessionists. The state agent needs to take into account how concessions in this dyadic relation between state and secessionists affect their relationship in the future. The problem in this relationship is how or whether concessions can be enforced; whether accommodation can be instantiated in such a way as to discourage demands for further concessions in the future. Precedent-setting is typically posed as a problem in a multi-ethnic or multi-national setting and arises for states confronted by a demand for sovereignty from a territorially concentrated ethnic group or nation: "[Tlhe key question is whether the secession of the group will set a precedent for other groups, thus spurring subsequent secession" (Duffy-Toft, 2003: 26). This problem is related to the importance of reputation-building in relations between states and separatists: "If a government believes it could face multiple additional challenges over numerous pieces of ten'itory, it has greater incentives in building a reputation for toughness" (Walter, 2006: 313) in order to discourage challenges to its authority in the future. 13 However, reputation and precedent also matter in interactions between state agent and sincere secessionists in a single dyadic relationship, as long as the state agent is concerned that concessions at one point in time will be vulnerable to future exploitation at some later poinl. 14 Therefore, the importance of the existence of partisan reasons to make concessions (introduced in an earlier section) is further reinforced here. These reasons work against the incentive to establish a reputation for toughness, a reputation which, if established, would have the consequence of deterring further demand-making in the future. An incentive to establish a reputation for toughness works against concession-making; partisan reasons, in contrast, can encourage concession-making. And, if concessions are made, the issue that arises is whether accommodation is stable. Enforceability, even in the hypothetical form of a complete contract, is rarely, if ever, completely invulnerable, perhaps for some of the same reasons that institutional design cannot reduce to zero problems of implementation and compliance. Any contract is made in the (irst instance among individuals or, perhaps, more broadly a generation. Can future generations feasibly be bound? Does the hypothetical contract have a time limit'? If it does, the implication is that it is renegotiated on a regular basis.

38

H. Meadwell

But what is negotiated? The point of the contract then is not simply to enforce but in a more line-grained way to define those issues that are insulated from bargaining and those that are not. The contract comes to more resemble a constitution. The argument about enforceability and a contract should still extend to a constitution, however. The purpose of a contract cum constitution is to effectively put some issues outside the reach of both or several parties, so as to encourage compromise and exchange on other issues. But no human artefact can be immunized from change, no constitution written or unwriuen is immutable, and the separation between the two realms is no eternal firewall. In a general way, however, it is possible to describe the moves that would make the contract null and void: If the state agent rolls back concessions, the foregoing of independence by once-secessionists is off: if the secessionist plan is revived, the concessions made can be renounced by the state agent. Moreover, the contract might include clauses that stipulate what this kind of bad faith can trigger, such as a move toward independence if concessions are rolled back, and the initiation of bargaining between the rump state and the new state to distribute the transaction costs of new state-formation. Yet it is unlikely that the basic indeterminacy in this situation can be eliminated, whatever the content of these kinds of clauses, for clearly in this hypothetical case, for example, secessionists now have an interest in triggering the consequences of this clause by inducing the state to renounce concessions. But the best way to induce this rolling back of concessions is to revive the secessionist plan, which is exactly what the contract CUIlI constitution was designed to remove. 15

v As noted in detail by the Co-Editors, one immediate institutionally orientated response to this argument would look to the details of institutional design, posing such questions as: what is wriuen in the constitution with regard to the right to secede: how is the upper house or second chamber organized; do electoral rules encourage coalition formation in govemments: how are languages and other elements of identities protected and recognized: how much fiscal autonomy do substate units possess: do subunits have autonomous militaries: how many subunits are appropriate and how are the boundaries of subunits drawn? 16 No one can deny the intrinsic interest of these questions. Taken together, as a complex pattern of institutional checks and balances, which provide centripetal counterweights to centrifugal concessions, the requisite arrangements imply a kind of calculus. Here, I (ill in one way to think about this calculus in light of the analysis of the last section, since I believe that this perspective strengthens the discussion in the Introduction of the problem of institutional design. Together, these desiderata above imply that the arrangements which they instantiate are self-enforcing. These arrangements are a stable equilibrium because they induce mutually self-limiting behaviour. On the one hand, the very interest in institutional design, on its own, speaks to the difliculty of the problem at hand. Such a design has to be seriously elaborate, however simple its basic calculus. The design is complex, the calculus is simple. On the other hand, what reasons do we have to think that these situations can be stabilized by fine-tuning institutions?I7

Secession and Institutional Accollllllodation

39

VI Even federal arrangements, however fine-tuned they are institutionally, may depend for their stability on something more than a self-enforcing equilibrium, which is threat-free and violence-free. Instead of self-enforcement, it is the distribution of force between the parties-whether group to group or state to group-and the distribution of costs, and the ability of states or groups to inlluence the transaction and transition costs, which would be borne by members of the seceding group in seceding and establishing a new state and by members of the rump state, which sets limits. Much then would seem to come down to comparative advantage in capabilities and force and the comparative willingness of states and groups to inllict and bear costs. Can the state agent demonstrate more of a willingness to inllict costs in order to prevent secession than secessionists can demonstrate a willingness to bear costs in order to ,~ch~eve ind~pendence? This exchange of threats and signals may induce a stable equilibrium but It would be stretching the meaning of accommodation to call such an equilibrium 'institutional accommodation'. I would further suggest, then, that these arrangements of accommodation are not self-enforcing, given this sensitivity to degree of heterogeneity. A basic feature of much of the interest in institutional design, a feature which evades in a way the problem of deeply divided societies, is the presumption that there is enough commonality among groups in these situations to make secession unlikely, but there is still enough cultural heterogeneity to induce the search for a point of institutional accommodation, which modi lies the political status quo by decentralizing political power between the parties involved. Moreover, there has to be enough heterogeneity in these situations for a threat of secession to be credible enough to induce concessions from a state agent but there cannot be so much heterogeneity that independence is prefelTed to the status quo modified by concessions made to the seceding group by the state or by another group. Under these assumptions, there is always a point of accommodation from which neither party has an incentive to deviate. But this is an unlikely scenario, or so I have argued in this paper. The argument of this paper questions this scenario in a fUl1her way. An implicit assumption is that the preferences in question are completely endogenous to institutional arrangements, so that in the face of institutional accommodation, sincere secessionists will change their preferences and give up the goal of forming a state. However, as I have argued, sincere secessionists with any strategic sense at allactors who are more than passive registers of institutional change-do not have to modify their preferences in lock-step with changing institutions, ir they calculate that they can exploit these changes in the ruture. Sincere secessionists have already rormed preferences, perhaps this claim about the endogeneity or preferences is meant to be an argument about prererence-formation among those who are born under the institutions of accommodation and who come to political maturity within them. Perhaps the expectation is that these individuals will be less likely to form secessionist prererences, given institutional accommodation. ~ut this does not completely convince: Institutional accommodation almost always lI1creases the institutional capacity or the group in question and, in so doing, provides the inrrastructure to shape identity-rormation and thus, at least indirectly,

40

Secessioll alld lllstitlltiollal Accollll/lodatioll

J-/. Meadwell

preference-formation. On the contrary, then, institutional accommodation may allow secessionism to be reproduced, even in the face of generational replacement There is another way to understand the full logic underlying the scenario outlined above. Institutional accommodation, it can be argued, might have the consequence of inducing stability when accommodation isolates secessionists within the group, rather than inducing them to change their preferences. However, there are two implications of this position which point to problems. First, the claim that accommodation isolates secessionists denies or contradicts the more basic claim that motivates the problem of accommodation-namely that this is a deeply divided society. If the institutional arrangements of accommodation can be so easily insulated from secessionist challenges from within the group in question, then there is much less division than the basic claim implies. Moreover, secessionists cannot be isolated by depriving them of voice. If the stability of the political equilibrium depends on the silencing of secessionists, the equilibrium in question depends on some form of imposition or even violence. The second problem is this. For the claim that accommodation isolates secessionists to be made good, it should be the case that accommodation does not follow on the use of secession as a threat to induce a state agent to make concessions. If the threat of secession is used instrumentally in order to bargain for concessions short of independence, those making the threat have to be able to credibly promise not to use the threat again alld they must live up to that promise if the concessions in question are to be stable. Otherwise secessionists have not been isolated. This problem reproduces the basic issue of enforceability, which has been the focus of this paper. There are subtle ways to use secession, and the presence of secessionists within the group, in bargaining with a state agent The expectation that secession can be threatened in order to induce concessions at one point in time, and henceforth that secessionists will be isolated within the group without being deprived of voice, once concessions are put in place, is simply not credible. It implies that the threat of secession will be used once and then freely forsaken once and for all. Then, how are these institutional arrangements stabilized if they are not self-enforcing in the face of cultural heterogeneity? If these arrangements are not self-enforcing, they cannot be stabilized endogenously. The stability of institutional accommodation must be exogenous in this kind of situation. There must be an element of imposition in these arrangements.

41

the cultural composition of the society in question. This move questions some kinds of institutional arrangements but without the desire to have a state of one's own. Instead the motivation is the fear of the excessive centralization of power, a classical concern in the intellectual history of work on federalism. Even without cultural heterogeneity, there remains an issue of political power and its distribution. There is a difference, however, between seeking a state of your own on cultural grounds and seeking to hollow out a state by decentralizing public authority. The former is not anti-statist in the same way as the latter. The first kind of challenge is much more likely to be nationalist and sovereignist: the second challenge is much more sceptical in general about public authority centralized in states and thus potentially sceptical about both sovereignty and nationalism. IX Secession merely reproduces the state structure of international society, from this point of view. The issue with regard to the second challenge is whether this political position can achieve and reproduce political decentralization deep enough to produce stable institutional accommodation in a divided society. The relevance of cultural heterogeneity thus arises in a different way when the fundamental challenge to public authority is rooted in the fear of the centralization of power rather than in the politics of identity. The investments that persons make in local identities and practices might provide a way to protect a division of power which is required to decentralize power in light of the problem of the centralization of power. 19 This reverses the order of things: identities would function to protect a division of power and authority. However, if these local cultures are to function as acceptable limits on power and authority, they must be neither too strong nor too weak because of the difference mentioned earlier between 'state-seeking' and 'hollowing-out' These cultures must be strong enough to function as a barrier to political centralization yet, at the same time, they must be self-limiting-their members must be satisfied with 'hollowing out'. Either cultures might be too weak to do the work they are supposed to do with regard to the division of power or they do this work because they are themselves concentrated. This kind of justification then seems to be self-defeating. The tradeoff appears to be this: betlercultural homogenization locally than political concentration centrally. This move then will beg the question: given this tradeoff, how are these local cultures, whether nations or ethnic groups, held together institutionally in political equilibrium? >

VIII

VII We should consider, however, just how to interpret heterogeneity before we take up this last point In the first instance, this is not a problem. Heterogeneity is meant to refer to those phenomena that have the potential consequence of challenging the territorial status quo. Typically here we think of distinctive identities, practices, interests and loyalties, which are socially organized and territorially concentrated. Hence, the question: are there decentralized political institutions which are secession-proof in the presence of significant heterogeneity of this sort? One proviso now needs to be added. An interest in federal institutions can be motivated by the problem of the concentration of political power, independent of

These provisos are entered in light of the final issue to be noted here. Is this political equilibrium self-enforcing, or does it depend on imposition? If the stability of decentralized institutional arrangements is not self-enforcing but is enforced, then it depends on imposition. But this is exactly the problem that motivates the political fear of centralization. It is consequential who is doing the institutional nne-tuning. This point is of particular importance in those instances when groups are bargaining over an agreement that would lead them to share political institutions. Institutional design can be presented in an oddly idealized way-it is as if the persons doing the designing provide the solution without being a part of the problem-design from the outside, so to speak, as if institutional design was solely about limiting the action of others. But an institutional

42

Secession and jnstitlltional Acco/lll/lodation

H. Meadwell

43

design that limits others but not oneself is, at first blush at least, imposed on those others and is exogenous to the actions of those on whom it is imposed, unless the first moverthe institutional designer-is a principal or if really "outside", if she is 'correlated' with a particular principal (e.g. has been 'captured' by a principal), in which case the design is not strictly exogenous even if imposed. Yet without imposition it then is not immediately clear that the design can be instantiated because, if 'correlation' is present, for example, it is hard to believe that the design does not contain unjustified prejudicesthat is, biases-and, if it does, then why would one party agree to it?

presupposed. In some situations, basically disintegration of state institutions, the question might not be whether a particular set of institutions does or does not prevent secession but whether an institutional arrangement (a) works as a peace settlement after conflict and (b) is suffiCient to induce agreement to form a new state which incorporates both parties within shared political institutions. Failure to agree to form a new state under (b) IS not secession strictly speaking, according to the definition above, since a state has not been established from which to secede. JJenne (2009) discusses the weaknesses of the literature on ethnic secunty dilemmas in her paper on partition in this Special Issue. She argues that the application of the ethnic security dilemma to the analysis of these situations sanctions partition and partition institutionalizes ethnic tensions. In effect. I separate the problem of the security dilemma from the problem of accommodation and enforcement.
Conclusion

Meadwell (2008). 'Olier Mordrelto John Legonna. II November 1968. John Legonna Papers. NatlOlJaI Library of Wales. 6Michael Collins, Dail Eireann. 19 December 192!. 7The metaphor of the stepping slone was lirst introduced into the debates in the Dail by Eoin O'Duffy on 17 December 1921. "Count O'Byrne, Dail Eireann, 7 January 1922. 9For a fuller discussion of unresolved issues in Scottish politics following the devolution settlement of 1998, see the contribution of Stephen Tierney (2009) to this Special Issue. IOThis was a plan, first mooted in 2000-01 and presented to the Basque Parliament in 2003 by the President of the Basque government, which proposed an arrangement providing much more autonomy to the Basque Country without full sovereignty (and permitting referendums). It finessed the differences between moderate and radical Basque nationalists but without solving the problem of violence. liThe partisan logic of decentralization is discussed more fully in the contribution to this Special Issue by

An institutional design that is imposed is, strictly speaking, not self-enforcing but it may be secession-proof; a design that is not imposed must be self-enforcing to be secessionproof but self-enforcing arrangements are difficult to strategy-proof. The discussion in this paper points to three different kinds of political opportunism which are associated with this dilemma. (I) The opportunistic use of secession as a threat by individuals in the group who are not secessionists in order to extract concessions from the state agent. (2) The opportunism of politicians who make concessions in order to improve their competitive position in partisan politics. (3) The opportunism of sincere secessionists with a plan who accept concessions in the short term in order to continue to organize in support of independence in the longer term. There are limits to these kinds of opportunism. Too much partisan-driven opportunism might raise the question of political competence. The instrumental use of secession as a threat cannot be too blatant or used too often, otherwise it may become less credible. Independence cannot be postponed indelinitely without, in effect, giving up the goal of independence altogether. None the less, despite these limits, the dilemma identified above is real. There seems to be a deeper feature of these situations which produces these results. It is the existence of sincere secessionists within the group who (a) neither reject concessions because they do not immediately satisfy their desires nor respond to concessions by relinquishing the goal of independence and (b) who have renounced violence. In an increasing number of cases, these kinds of secessionists have significantly changed the political imaginary and strategic structure of these situations. The goal of independence has become a legitimate political goal. and political activity in support of secession has become legitimate political activity. 20 When viewed in this light, secessionism is of consequence, not simply because of its challenge to the identity of particular territorial states nor because secessionism reproduces the central feature of state-centred international society, but because it exposes the difficult balance between authority and liberty in liberal democracy.

Notes ISee, for example, Bucheit (1978), Young (1976: 460-5(4), Sisson and Rose (1990). 2My understanding of secession is conventional. Secession refers to "the formal withdrawal of a constituent unit from an established, internationally recognized state and the creation of a new sovereign state" (Bartkus. 1999: 9). This definition may well restrict the extension of the concept, depending on how we interpret the qualifier: "established", A significant degree of political rationalization is

Jason Sorens (2009). 12The notion of the social and politicaillnaginary here draws loosely on the work of Taylor (2004) and Anderson (1991) to suggest that the imagined range of institutional possibilities is enlarged when IIldependence is given tacit legitimacy. "Walter's argument is an application of the chain store paradox, first described in the economics literature. This literature addresses how a monopolist deters the entry of competitors by sacrificing shortterm profits for greater long-run profits (Kreps, 1990: 43). I"The essential point is the relevance of time. The literature on precedent-setting and reputation assumes a senes of potential challengers over time. But this is not to say that time is not relevant III the dyadic relationship between a state agent and a secessiolllst group. This relationship IS iterated. Hence, the importance of the enforceability of institutional accommodation over time. 15 T his pomt bears on the literature on the consequences of 'constltutionalizing' a right to secede. For different perspectives on thIS issue, see Sunstein (2001: 95-114) and Wellman (2(05). 16See , for example, Filippov el al. (2004), de Figueiredo el al. (2007). 17For discussion of the difficulties of institutional design in ethnofederations, see the contribution by Philip Roeder (2009). I"The literature on pI uri national democracies is explicitly 'post-sovereign', There is some discussion of this literature in the contribution by Tierney (2009) to this SpecHiI Issue. 19There is an excellent discussion of this kind of argument and of the contribution it makes to liberal theories of federalism in Levy (2007). foresight might suggest that state agents begin to prepare for the possibility of secession in such cases. Yet, to publicly announce the preparation of plans for territorial division, to be implemented should It come to that eventuality, might make more likely that very eventuality. Such

20 Po litical

are the dark arts of politics.

References Anderson. B. R. (l991), [magilled COl/llllwlilies (rev. edn). London: Verso, Bartkus, V, O. (1999), The DYllamic ofSecessioll. Cambridge: Cambridge UllIversity Press.

44

H. Meadwell

Brancati. D. (2006). DecentralizatIOn: Fueling the Fire or Dampening the Flames of Ethnic Conflict and Secessionism. /mematiollal Orgallizarioll. Vo1.60. No.3. pp.65 1-685. Bucheit. L. C. (1978). Secessioll: 71w Legitimllcy of Self-De term illati 011. New Haven: Yale University Press. Duffy-Toft. M. (2003). The Geography of Ethllic Violellce. Princeton. NJ: Princeton University Press. Fearon. J. (1998). Commitment Problems and the Spread of Ethnic Conflict. in D. Lake and D. Rothchild (cds). The /mernatiollal Spread of Ethllic COllflict. Princeton. NJ: Princeton University Press. pp.107-126. de Figueiredo. R. J. P. Jr .. McFaul. M. and Weingast. B. R. (2007). Constructing Self-enforcing Federalism in the Early United States and Modern Russia. Publius: the Journal of Federalism. Vol.37. No.2. pp. 160-1 89. Filippov. M .• Ordeshook. p, C. and Shvetsova. O. (2004). Desigllillg Federalism. Cambridge: Cambridge University Press. Haimanko. 0 .. LeBreton. M. and Weber. S. (2005). Transfers in a Polarized Country: Bridging the Gap between Efficiency and Stability. Journal of Public Ecollomics. Vo1.89. No.7. pp.1277-1303. Haimanko. 0 .. LeBreton. M. and Weber. S. (2007). The Stability Threshold and Two Facets of Polarization. Ecollomic Theory. Vo1.30. No.3. pp.415-430. Jenne. E. K. (2009). The Paradox of Ethnic Partition: Lessons from dej(lcto Partition in Bosnia and Kosovo. Regiollal lIlId Federal Studies. Vo1.l9. No.2. pp.273-289. Kreps. D. M. (1990). Game TheO/)' alld Ecollomic Modellillg. Oxford: Clarendon Press. LeBreton. M. and Weber. S. (2003). The Art of Making Everybody Happy: How to Prevent a Secession.IMF Staff Papers. V01.50. pp.403-435. Lecours. A. (2007). Basque Natiollalism alld the Spall ish State. Reno. NV: University of Nevada Press. Levy. J. (2007). Federalism. Liberalism and the Separation of Loyalties. Americall Political Sciellce Review. VoI.lOI. No.3. pp.459-477. Levy. R. (1990). Scottish Natiollalism at the Crossroads. Edinburgh: Scottish Academic Press. Meadwell. H. (1993). The Politics of Nationalism in Quebec. World Politics. Vo1.45. No.2. pp.203-241. Meadwell. 1-1. (2008). The Dynamics of Secession: Spatial Models of Secession-Proofness and Equilibrium Size. Unpublished manuscript. Meadwell. H. and Anderson. L. M. (2008). Sequence and Strategy in the Secession of the Amencan South. Theory alld Society. Vo1.37. No.3. pp.199-227. Miller. W, L. (1999). Modified Rapture All Round: The First Elections to the Scottish Parliament. GOI'emmemalldOppositioll.VoI.34.No.3. pp.299-322. Roeder. p, G. (2009). Ethnofederalism and the Mismanagement of Conflicting Nationalisms. Regiollal lIlId Federal Studies. Vol.J9. No.2. pp.203-219. Sisson. R. and Rose. L. R. (1990). War alld Secessioll: Pakistall. Illdia lIlId the Creatioll of Ballgladesh. Berkeley: University of California Press. Smens, J. (2004). Globalization. Secessionism and Autonomy. Electoral Studies. Vo1.23. No.4. pp.727-752. Sorens. J. (2009), The Partisan Logic of Decentralization in Europe. Regiollalalld Federal SII/dies. Vol. 19. No.2. pp.255-272. Sunstein. C. R. (2or)!). Desigllillg Democracy. What COllstill/liolls Do. Oxford: Oxford University Press. Taylor. C. (2004). Modem Sociallmagillaries. Durham: Duke Universlly Press. Tierney. S. (2009). Federalism in a Unitary State: a Paradox too Far? Regiollal alld Federal Studies. Vol. 19. No.2. pp.237-253. Walter. B. (2006). Building Reputation: Why Governments Fight some Separatists but Not Others. Americall Journal of Political Sciellce. Vo1.50. No.2. pp.313-330. Wellman. C. (2005). A Them)' (J{Secessioll. Cambridge: Cambridge University Press. Young. C. (1976). The Politics (!{ Cultural Pluralism. Madison: University of Wisconsin Press.

Federalism in a Unitary State: a Paradox too Far? STEPHEN TIERNEY School of Law. University of Edinburgh. Edinburgh. Scotland. UK

Introduction This paper addresses the devolution settlements in the UK as a model of accommodation of territorial diversity. with a focus mainly upon devolution to Scotland. It is argued that the Scotland Act 1998, while in many ways a coherent attempt to meet the demands of national diversity. may also. paradoxically, contain elements that in the long run have the potential to destabilize the Anglo-Scottish union. To begin. there are two important contextual issues which will be important points of reference throughout the paper-one is the fact that territorial diversity within the UK is plurinational in nature. and the second is the non-federal nature of the model of constitutional accommodation that has been used to manage it. As will be seen, this ad hoc constitutional solution to territorial plurinationalism contains some elements that seem useful to the management of pluralism and others that seem to exacerbate the risk of secessionism. By way of introduction we might distinguish two rationales for a federal state. One justification for a division of governmental power along territorial lines is that this is needed to accommodate strong ethnic or cultural distinctiveness; this is an important dynamic of Canadian federalism for example. Another rationale applies regardless of the existence of deep diversity within the state and considers that territorial

46

S. Tierney

decentralization is a value in itself, either because localized decision making can lead to more efficient government and/or because democracy. in keeping with the principle of subsidiarity, is enhanced if decisions are made at the most local level feasible for their effective execution (US and German federalism in differing ways represent this model). We are concerned with the former 'territorial diversity' rationale and, in approaching this question, a fUIther important point of distinction within this model is required: that between states which are territorially diverse on cultural or ethnic grounds but which remain uninational in orientation, where even those territorial sub-units that have strong cultural differences share one and only one national identity (e.g. Germany and, arguably, Switzerland), and those that are in fact plurinational. The paradox of federalism as explained in the introductory paper to this collection (Erk and Anderson, 2(09) is most apparent in plurinational states, as it is here that federalism is most often seen as, or feared to be, a stepping stone to secession or at least a pathology that can lead only to disruptive internal disputes and, consequently, constitutional stasis. Put simply, it is within plurinational states that we lind territorial units with a sufficient sense both of distinctiveness and normative entitlement (the latter encapsulated in the principle of self-determination) with which to generate claims to statehood. The fear is that institutional recognition of such a territory merely intensilles this sense of nationhood while also supplying the institutional infrastructure ('capacity' per Erk and Anderson, 2(09) through which secessionist nationalists can more easily bring about independent statehood or destabilize the state in the quest for independence. The idea of the plurinational state has now become a fairly well-settled category in accounts of federalism (Requejo, 2(05) and, in recent times. detailed studies have been published analysing and critiquing the political and constitutional nature of the plurinational state (Keating, 200 I a). It seems that one way in which this area of literature needs now to be advanced is in the careful study, case by case, of the differences as well as the similarities between plurinational states; the highly particular constitutional implications of plurinationality within developed industrial societies, which we have seen in recent constitutional studies, suggest that each case requires to be addressed with a high degree of specillcity. This is not to deny the feasibility or importance of comparative work, but suggests that in the process of drilling down within each plurinational state in examination of its constitutional structure, constitutional culture, political system(s) and demotic composition, we are IInding extensive constitutional and societal particularities which must inform the task of constitutional prescription in appropriately-tailored ways. In the second section, we will begin by developing this argument further, with particular reference to the specificities. or indeed idiosyncrasies, of the UK as one such state. Secondly. this paper, in addressing the UK-and Scottish devolution, specillcally-is, of course, concerned with the management of diversity not through a federal system at all, but through the unwritten constitution of what remains. in terms of constitutional formalism. a unitary state. Since the supreme law-making function in the UK lies indivisibly in the hands of the central legislature, I we are addressing a state which fails one of the classical tests of federalism. 2 Therefore, it is important to explore the advantages and disadvantages of such a system for the accommodation of territorial difference in comparison to a federal model. Having discussed the unique demotic composition of the UK in the next section. the argument will be made in

Federalism ill a Unit[//}' State: a Paradox too Far?

47

the section that follows that perhaps the most crucial issue for a plurinational constitution is not in fact whether it is federal or otherwise. but how carefully its model of decentralization is tailored to the societal conditions prevailing in the state and how effectively a nuanced approach can be taken to accommodate such conditions in all their complexity. In this light the non-federal model in the UK will be seen to have both beneficial and detrimental aspects as a constitutional structure through which to manage a Byzantine model of territorial diversity.

The UK as a Plurinational State: Specific Characteristics Let us begin then with a challenge or, perhaps less provocatively, an invitation to scholars of plurinational federalism. Our focus has understandably been concentrated hitherto upon drawing connections across plurinational states. However, what this wave of scholarship has revealed is not only the pertinent parallels but also the significant variations from case to case within the plurinational model (Tierney, 2004; Keating, 200 I b). An important point of difference for the purposes of this paper is the relative strength or weakness of a 'state' identity. In Canada and Spain, for example, existing alongside one or more minority national identity(ies), the dominant national identity within the rest of the state doubles also as a state identity (Canadian; Spanish). The UK by contrast is a more complex model. The dominant national identity combines, in a fairly inchoate way. that of a sub-state nation (Englishness), with the state identity (Britishness). In demographic terms England is overwhelmingly dominant, comprising approximately 84% of the population.} In consequence of this, in the era before devolution for Scotland, Wales and Northern Ireland, there was little need for English national identity to be distilled by English people from its commingling with that of Britishness; and, therefore, for many people the two badges were virtually synonymous. This has served to leave Britishness itself as a very hazy tag that applies in highly varying ways across the state (e.g. taking on a very different meaning for Northern Ireland unionists than it does for English people, leaving aside the strong particularities of meaning to be found within these two territories) and which since devolution has, consequently, fluctuated considerably in terms both of its meaning and strength as a component of various compound identities across the state (Kiely et al., 2(05). This is not to say that such compound or mixed identities do not exist in Canada, Spain and other plurinational states, it is simply to point out that identity compounds are very different in nature from state to state (Henderson, 20(7). And it is important to observe this level of specificity because these distinct identity patterns in turn inform very different constitutional aspirations among highly particular sub-state nations. We see this within states, as well as across them. Taking the UK as our example, what is notable is not only that the four discrete territories of the UK contain live very different national groups (Scottish, English, Welsh, Northem Ireland British and Northem Ireland Irish) but that within each territory in the late 1990s what emerged were very different sets of constitutional demands. And partly in response to this, the UK model of devolution, which is in turn highly distinctive, has been modelled largely in a bottom-up way (itself a paradox, given that devolution in formal constitutional terms involves a top-down grant of powers from the centre) to reflect this complex identity and aspirational matrix.

48

S. Tiemey

For example, enthusiasm for devolution in Wales has historically been far more muted than in Scotland. There was a very clear rejection of a model tabled in a referendum in 1979 and, in 1997, the settlement eventually to be implemented by way of the Government of Wales Act secured only bare majority endorsement in another referendum held in that year. The fact that the push for devolution within the Welsh political class and civil society prior to 1997 faced strong opposition is one of the main reasons that we nnd in Wales a very weak model of devolution. In contrast, the two territories where we nnd much stronger models of self-government (Scotland and Northel11 Ireland) are those where a distinctive national identity, or in Northel11 Ireland's case identities, are much more salient, bringing with them in the late 1990s a more intense impetus for meaningful change from political elites and civil society. In the case of Northern Ireland the demand for change was, of course, of a deeply contested nature. Containing two national groups, Northern Ireland is in many ways best classified as a divided society and, in this sense, constitutional change in 1998 was a half-way house between the irredentist aspirations of Irish nationalists and a general contentment with the status quo felt among British unionists. The resulting model of devolved government, therefore, owes as much to conflict resolution and the engagement of the Republic of Ireland as the kin-state of Irish national ists, as it does to decentralization for the traditional reasons we find in more amicable environments. Scotland has a more homogeneous sub-state national identity than either Wales or NOIthel11 Ireland. As such, it provides the best (and perhaps the only) appropriate comparator to set alongside the classic case studies of Catalonia, the Basque Country, Quebec and Wallonia in Belgium as a fully functioning and politically coherent sub-state national society that harbours genuine constitutional aspirations to reconstitute the state on a plurinational basis. The model of devolution for the UK has, therefore, been carefully shaped to meet this deep pluralism, resulting in a heavily asymmetrical model. In the nrst place, devolution does not extend to every sub-state nation. England, in contrast to the other territories, has shown little inclination for devolution either as a whole or by region (Bond and McCrone, 2(04). Therefore, it continues to be governed by the central institutions of the UK. This makes sense given England's size etc., but it can lead to imbalance in terms of the representation of other territories at the centre, as we shall see in the next section. Secondly, an interesting lesson emerges from the UK experience that may be relevant for other plurinational states. There are well-established arguments that levels of asymmetry are justifiable to create distinctive models of autonomy etc. for sub-state nations compared to those models appropriate for devolved/federalized territories that are part of the dominant national group (Kymlicka, 200 I: 91-120; Requejo, 2005). But the British experience takes this flllther in the way that devolution differs among sub-state nations themselves, rel1ecting the specific history, demography, identity patterns, institutional infrastructure and constitutional aspirations of each. Another interesting background feature of devolution is the informality of the UK constitution. This is often seen as a weakness by those who consider devolution to be an inadequate model for sub-state autonomy. The argument is that since the devolved settlements are created by ordinary Acts of the Westminster Parliament they can be amended or even repealed in the same simple way, a danger that is avoided within an entrenched federal system. From another perspective, however, the l1exibility of

FederalislII in (/ Unit(lf)' State: (/ Paradox too Far?

49

the UK model can be seen as a strength. This elasticity has offered the opportunity to create asymmetrical models of devolved government in an ad hoc way through the relative simplicity of ordinary legislation, leaving also adaptability for further easy change in the future. This raises the interesting question whether on an institutional level federalism is the only, or indeed the best, way to manage national diversity within the polity. Given the argument presented here that in a plurinational state the constitution needs to be crafted in a highly particular way, it may be that the very complex national mix in the UK would not lend itself well to the rigidities that we associate with a federal system, whereby the constitution can be very difflcult to change and may require broad territorial consensus in highly-charged metaconstitutional moments; the inter-provincial conferences which, in Canada, proved such a failure in 1987 and 1992 are a salutary reminder of the disadvantages as well as the democratic benellts of a highly formalized constitutional process. In contrast to the Canadian experience, the ease with which UK devolution was created; was suspended to meet various emergencies (Northern Ireland); and has been amended to extend further devolved powers (e.g., to Wales)4, suggests considerable benefits exist in the UK model. The experience of devolution has been very successful also in that hitherto no competence disputes have come before the courts (Hazell, 2(07). However, such a system depends upon a high degree of political consensus and good faith on the part of the central executive and legislature. And as this UK-wide consensus is presently weakening, we will see in the next section, in reference to Scotland, how the ad hoc model within the UK may be a double-edged sword. Building upon this, in the following section, aspects of the Scottish model will be identified which, it will be argued, makes more likely the prospect of, or at least more easy the opportunity for, secession. Some of these potential pathologies are common to any institutional attempt to accommodate diversity, whereas others seem particular to the way UK devolution has been effected and, hence, potentially avoidable.

The UK Constitution and the Accommodation of Scotland There are various features that show the development of Scottish devolution to be an innovative attempt to accommodate a sub-state nation's aspirations for constitutional recognition. We will consider four in turn that bring out the sense of ownership, felt within Scotland, over the devolution settlement and the extent to which it has been crafted to meet the specific aspirations of Scottish political elites in an asymmetrical way. In other words, ways in which devolution can be seen as a promising attempt to accommodate national sentiment and, in doing so, offering at least a prima facie disincentive to secessionist sentiment. (i) The Indigenous Origins (~l Scottish Devollllion

The initial feature that marks Scottish devolution as a suitable model to match the country's national aspirations is the extent to which the 'offlcial' process towards devolution, adopted by the Labour-led UK Parliament in 1997, built upon, rather than supplanted, the extra-parliamentary campaign for devolution within Scotland as orchestrated by the political opposition and certain important institutions of civil

50

S. Tierney

society through the 1980s and 1990s. This openness shown by Parliament to an extra-parliamentary initiative is not altogether surprising, given that this campaign was led in large measure by the Labour Party in Scotland. Indeed, we can identify the starting point of the process towards Scottish devolution in the failure of a devolution proposal put forward by the Labour government in 1979. Although a detailed model of devolution was enacted by Parliament in 1978, this failed to meet the requisite threshold of support in a referendum the following year. 5 However, with the election or a Conservative government in 1979 and its subsequent re-election in 1983, a push ror devolution was set in train by opposition parties who felt that the Conservative agenda was given to excessive centralization. This movement continued to o ITer constitutional proposals throughout the next two decades. The extra-parliamentary campaign ror constitutional change began with the Campaign ror a Scottish Assembly launched in 1985. This resulted in a document, A Claim ofRiglzt for Scotland, issued in 1988. What is notable about this paper and others that subsequently emerged through this period, is the extent to which they would draw upon the notion of 'union' as the fundamental constitutional principle of the UK. This had three implications. First, the Clailll l?f Right asserted the distinctive national identity and cultural and institutional speci(icity of Scotland, each of which had been recognized in the Acts of Union 1707, with the commitment that this plurinational fact should continue to inform British constitutional practice. Secondly, the document aired the grievance that the 'union state' pact stemming from 1707 had been undermined by subsequent UK constitutional practice (Edwards, 1989: 19). Thirdly, the Claim declared an entitlement to Scottish selfgovernment based upon the notion that a distinctive national identity carried with it a legitimate, indeed inherent, political right or self-determination. The extra-parliamentary process was largely the preserve of elite actors and suffered from the lack of participation of both the Scottish Conservative Party, which, favouring a centralized state, never took part, and the Scottish National Party which withdrew after early involvement, alleging a failure to address independence for Scotland as a serious constitutional option. None the less, in the referendum in 1997 there was very high level of support for devolution,6 which perhaps reflects the credibility of the proposed model as one generated within the Scottish body politic and civil society. (ii) Devolutioll Modelled by Territorial Actors

Important elites within Scotland were not only able to engage in the extra-parliamentary process, they were also heavily influential in shaping the substantive model eventually enacted in the Scotland Act. We can see the origins of this influence in the sophisticated task of constitutional prescription undertaken by these actors. The Claim of 1988 recommended that a cross-party Scottish Constitutional Convention (SCC) be established, which would have the task or drawing up a model of devolution that would generate popular support and hence "assert the right or the Scottish people to secure the implementation of that scheme". This SCC was inaugurated on 30 March 1989 and, over the next seven years, involved illter alia, not only the Labour and Liberal Democratic parties,7 but also local authorities, churches and the Scottish Trades Union Congress (McFadden, 1995). This resulted in a series or publications,

Federalism in a Unitmy State: a Paradox too Far?

51

the most important of which (Scotland's Claim. Scotland's RiRtzt, 1995) set out a detailed blueprint for devolution. The level of detail in this document meant that the process by which legislation was passed after Labour came to power in 1997 was very swift. The political climate in 1997 -98 was highly conducive to devolution; a Labour administration elected with a large majority in 1997 was committed to the principle, and so, heavily influenced by Scottish Labour MPs, the government drove the devolution settlement through in the (irst session of the new Parliament. Upon taking oflice, and relying heavily upon the detail in Scotland's Claim. Scotland's Right, the government issued a White Paper, Scotland's Pariiament, which was, therefore, able quickly to set f0l1h a comprehensive plan ror devolution. Although some minor changes were made by Parliament to the devolution proposal after the referendum in 1997, the model (inally enacted in the Scotland Act was in substantive terms the same as that voted for in the referendum, which in turn reflected heavily the pre-1997 deliberation. In this way elite actors within Scotland were essentially able to frame a model of devolution that suited their aspirations without the difficulties involved in bargaining a rull constitutional package across the state, with the compromises and the disparities in power that can attend such processes.

(iii) RecoRnition: the Constitutional POlVer (~f SymbolislII A third feature of the settlement marks it out as a powerrul model of accommodation, namely the symbolic recognition of Scottish nationhood that it either directly delivers or, by its process, implies. Devolution is, of course, often considered to be the poor cousin of rederalism ror sub-state nations seeking constitutional accommodation. And certainly its failure to entrench the powers that a devolved territory acquires seems to confirm this. And it might also be seen to be so in symbolic terms. The word 'devolution' suggests a cession or power from the centre (Requejo, 200 I: i 14), rather than the division of sovereignty which lies at the heart of federal systems. In this sense, it might be argued that devolution cannot possibly represent a symbolic recognition of a plurinational reality when substantively it does not contain a division of sovereign powers (Kymlicka, 1998: 135). This is a powerful point, but we should be careful to enter certain caveats in the case of Scotland. The IIrst is that insofar as federalism is a road not travelled, we should be wary to view this purely as an instance of central hegemony, whereby the centre decided to cede the fewest powers possible. As we have seen, the model of devolution for Scotland was organically generated and enjoyed widespread support, as it continues to do; indeed, federalism then and now was not a popular option in Scotland. A second point is that even in the absence of federalism, or perhaps on account or its absence, the symbolic recognition accorded to Scottish nationhood within the constitutional and political systems of the UK is striking. And, therefore, we should avoid an over-reliance upon the intuition that assumes a strong to weak level of plurinational realignment runs rrom federal to devolved states in a descending line. Devolution for Scotland contains elements that, despite a clearly hierarchical structure whereby the UK Parliament continues to insist upon its own legislative supremacy, offer some recognition of national distinctiveness. There is IIrst, what we might call

52

S. Tierney

'symbolic asymmetry'. Only sub-state national societies have acquired devolution (although the status of Northern Ireland as a sub-state national society, rather than as a territory containing two sub-state national societies, is, of course, heavily contested); and, related to this, the models of devolution are very different inter se. Thus Scotland's pivotal role in the union is recognized not only in the fact of devolution but also in its strength, when Scotland's extensive devolved powers are compared, for example, to those accorded to Wales. Another form of recognition is more explicit. This has a long provenance in the union nature of the state and is visible in the very name of the United Kingdom, in the design of its flag and in the common usage of the term 'nation' to describe Scotland, Wales and England. We see it also in the White Paper, which set out the UK government's plan for devolution. Here, Prime Minister Blair declared: "Scotland is a proud and historic nation in the United Kingdom" (Scotland's Par/iament, 1997). The use of the term 'Parliament' to name the devolved legislative assembly is another significant totem, applying the same name as that used both by the pre-union Scottish Parliament and, of course, Westminster itself. Here, we see a sense of asymmetry again amollg the devolved arrangements: Wales and Northern Ireland have 'assemblies'. a term thatintentionally or not-serves to emphasize that, of each territory, Scotland has the strongest set of devolved powers. The significance of this symbolism is brought home when one considers the tortuous debates other sub-state nations have engaged in as they try to achieve similar levels of recognition. Quebec has, of course. long sought a constitutional declaration of its national status. but even attempts to entrench the term 'distinct society' proved unpalatable to English Canada, a situation that changed to some extent only in 2006 with a House of Commons resolution that is some way short of full constitutional recognition of Quebec's nationhood. x In Spain. the Preamble to the Constitution recognizes the country's cultural plurality and the right to self-government of the nationalities and regions of which it is composed, but this falls short of any recognition of pi urinationality.'! Although the Catalan Statute of Autonomy of 2006. ratified by the central legislature. uses the word 'nation' for the first time, it does so by declaring this a Catalan self-perception. while making clear that the Spanish constitution conceives of Catalonia as a 'nationality' only.IO What is particularly notable is that the symbolic recognition of nationhood has been less forthcoming in these other plurinational states than in the UK, even though the models of autonomy in Canada and Spain are federal or "quasi-federal" (Moreno. 20(1), respectively. This suggests that the level of accommodation of symbolic nationhood is not necessarily better simply because the devolution of power is more formalized and that, in fact. within some manifestations of classical federalism there may be inherent incompatibility with pi uri national recognition. One reason for this is the in-built hostility towards asymmetry in certain approaches to federalism. An intuitive feature of federalism for Illany people is the equal distribution of powers among all sub-state territories (Dion. 1994) and this can make the accommodation of the particularities of sub-state nations especially difficult to achieve. Another difficulty in reaching plurinational recognition is the more practical one that where an original federal settlement is symmetrical in nature this can be. in practice, very hard to change. The embedded nature of federal constitutional settlements

Federalism in a Unitary State: a Paradox too Far?

53

can Illean that constitutional amendment requires the unanimolls consent of all sub-state territories; in such a process the political pressures not to concede privileges to only one or a few territories can be strong, as, once again, we have seen in Canada. And so it seems no small matter that the ad hoc process of using ordinary legislation in the UK, with separate statutes for each of Scotland, Wales and Northern Ireland, has made it relatively easy to tailor recognition on a case-by-case basis in bilateral processes between the state and each sub-state territory one by one, thereby avoiding the need to achieve a multi-party agreement across the whole.

(iv) Autollomy: Strong and Asymmetrical Devolution for Scotland also shows that a strong system of autonomy can be created within a unitary state. The Scotland Act contains a 'retaining model' of devolved power, whereby those matters that are reserved to the exclusive competence of Westminster are explicitly articulated within the Scotland Act. with all other matters devolved. There is an extensive range of devolved matters (Scotland Act, sections 29 and 30 and schedules 4 and 5) and, with few concurrent or shared powers. the Scottish Parliament effectively has exclusive competence in these areas. which extends as far as a power to repeal existing UK legislation in devolved areas of jurisdiction. Certainly there is a residual power for the UK Parliament to legislate in devolved areas (Scotland Act, section 28(7) (which is, in any case, arguably superfluous in light of the doctrine of parliamentary supremacy), but there is now a constitutional convention to the effect that the UK Parliament will not do so without the consent of the Scottish Parliament. II The asymmetry we see at the symbolic level extends to the allocation of powers. In terms of the range of devolved matters, the Scottish Parliament has a wider area of competence than its Welsh or Northern Irish counterparts. The UK constitution's llexibility has also been of use in extending further powers to the Scottish Parliament without the need even for primary legislation. far less constitutional amendment. 12 Also we have seen n1<~or changes put in place for Wales. which begin to move the model of devolution there more in the direction of the Scottish model. This has been effected simply through ordinary legislation (Government of Wales Act 2(06). And, in the case of Northern Ireland. powers have remained with both the UK Parliament and the UK government to suspend operation of devolution when agreement between republicans and unionists broke down at various points between 1998 and 2006. This adaptability has been very controversial. but arguably it has allowed breathing space until a linal agreement could be reached to revive the institutions on a, thus far. settled basis in 2006.

Tensions within the Devolved Model Section three has highlighted several successful aspects of UK devolution as a model for accommodating a complex level of national diversity within the state. However. in this part, and focusing again on Scotland, we will identify a fllliher four features that may. in the medium to longer term, tend further to unsettle the Anglo-Scottish union. possibly even making the prospect of secession for Scotland more rather than less likely. The first three of these features are paradoxical in that they seem to be

54

S. Tierney

Federalism ill a Vnitmy State: a Paradox too Far?

unavoidable consequences of any attempt to accommodate national identity by means of extending strong autonomy to sub-state territories. However, the fourth is a pathology specillc to UK institutional design that may well have been avoidable, and is perhaps still retrievable, if further constitutional change moving the UK towards a quasi-federal model is effected in the future. (i) Institutional Accollllllodation of Difference can Accentuate Difference

There are two ways in which attempts to accommodate difference can, in fact, serve to accentuate it, thereby intensifying conditions that might further unsettle a sub-state nation's place within the state. One involves patterns of identity: devolution may help create a greater sense of distinctiveness within the devolved territory. which might in turn strengthen secessionist sentiment. Secondly, the establishment of autonomous institutions of government makes the local more pertinent to people's lives. and the central state less meaningful or relevant. It is at least a persuasive hypothesis that this again might slowly erode attachments to the state within the sub-state territory. Addressing the IIrst issue. we need to ask what impact devolution has had upon national identity within the UK. From work conducted in 200 I, it is clear that British citizens continue to possess multiple and mixed national identities, but at this time it also seemed that Scottishness and Englishness were growing stronger in contrast to Britishness. The so-called 'Moreno' test takes into account the possibility of dual identity by asking not simply whether people feel. for example, Scottish or British. but by offering a range of live identity variables: Scottish, not British; more Scottish than British; equally Scottish and British; more British than Scottish; or British. not Scottish (Moreno, 1988). As Table I shows, citizens throughout Great Britain in 200 I continued to profess mUltiple identities, such as English and British. Scottish and British etc. Indeed a majority in each territory felt some sense of Britishness: Scotland 60%, Wales 73% and England 75%. Authors of one major research project addressing national identity since devolution. therefore concluded in 2004: "while devolution may not have strengthened adherence to Britishness. it has certainly not eroded it sufficiently to threaten the cohesion of the UK Union" (ESRC, 2004: 8). However, while these results were far from cataclysmic for the UK, they also seemed to suggest that adherence to 'Britishness' has weakened following devolution. "with a growing proportion in each country preferring to identify themselves as English (up from 34% to 43%) or Scottish (up from 72% to 77%) than British since 1997" (ESRC, 2004: 8). This led the authors to the conclusion that "devolution has not strengthened adherence to Britishness and seems to have sharpened national identities in England. Scotland and Wales" (ESRC, 2(04). Of course, such changing patterns of identity do not necessarily lead inexorably to secession. In the IIrst place, it is not clear that increasing attachment to sub-state national identities at the expense of the state national identity translates in any direct way into constitutional aspirations for independent statehood (recent figures suggest only 23% of Scots want independence (Leverhulme/ESRC, 2(07». However, it is also at least plausible that such patterns may in the longer term be a factor in a slow weakening of bonds to the union (Park et aI., 2(05). On the one hand. Britishness in 200 I seemed to be less important to Scots than it was previously. It might be anticipated that this

55

Table l. 'Moreno' national identity in Great Britain, 2001 Percentage

Scotland Wales England

x not British

More x than British

Equally x and British

More British than x

British not x

36 23 17

30 22 13

24 29 42

3 11 9

3 11 11

SOllree: Scottish Social Attitudes survey 200 I.

may in time dilute opposition to independence. A greater sense of English nationalism can also be a factor in making the unity of the state, over time, a less salient concern. 13 Whether or not there is a growth in reality of English resentment with the powers Scotland has, this has become a political issue. stoked up by the media and focused towards the privileges that Scots are perceived to enjoy.14 This could have a vicious circle effect of heightening tensions which may. in turn, [lIliher weaken Scots' identillcation with the union, whilst also diffusing any strong sense within England that it is in England's interests to dissuade Scotland from seceding. Another factor stemming from the attempt to accommodate difference may also be a danger to the long-term future of the Anglo-Scottish union, namely the simple fact that the devolved branch of government comes to be seen as more and more relevant to peoples' lives, while the UK government becomes more distant. There are trends in this direction. As has been noted, powers have continued to be devolved since 1998 and there are ongoing demands by a number of parties within Scotland for further devolution. including that of fiscal powers.15 The Scottish Executive has taken to calling itself the Scottish Government, a term that has also generally been adopted by the media. Although symbolic, this is another factor that may enhance the sense that the Scottish Executive is the 'real' government of Scotland. Although, of course. the increasing relevance of territorial government and the concomitant decrease in relevance of the central government is pa!i and parcel of any policy of giving autonomy to a sub-state nation, this is exacerbated in the UK case by the fact that autonomy for sub-state nations has not been accompanied by a reworking of their representational role within the central organs of the state. Arguably this is the greatest failure of the UK model, as we will explore in section (iv). Finally, the expanding role of the European Union as an external site of governance that now rivals the central UK institutions may well act to weaken further the relevance of the state in the eyes of citizens (Tierney, 2(05).

(ii) Nationalists have the Chance to Prove Thelllselves Another paradoxical feature of giving autonomy to a sub-state nation is that the new institutions can demonstrate to citizens that the territory is able to function effectively with a high degree of independent power. Since fear that the territory will not have the capacity and competence to be self-governing is one factor that might discourage people from supporting secession, it is ironic that this attempt to defuse secessionist pressure might in the longer run help legitimize and consolidate secessionism.

56

S. Tierney

This can be a particular issue where a nationalist party comes to power, as it then has a platfonn to show that it has the gravitas and maturity to govern, offering another string to its bow in the campaign for independence. These two goals: of showing Scotland's capacity for a high level of autonomy, and the competence of the Scottish National Party (SNP) to govern, are clear planks of the SNP's current strategy. 16 Furthermore, when a secessionist party comes to power in a devolved administration, it has the opportunity to demonstrate tensions in the existing model. For example, it can make political capital from the fact that powers which would be attractive are, in fact, retained by the centre, and can attempt to show how this retention works to the detriment of the devolved telTitory. This carries the flJlther message that the settlement overall is unjust or at least unduly restrictive. There is also an incentive for nationalist parties to provoke competence disputes or, in any event, little political incentive to try to avoid such tensions. By at least one analytical scenario, either these disputes will be settled in a way favourable to the devolved administration or, when the devolved administration fails in its objectives, it can use this failure to highlight again the inequities in the system and how the state uses its hegemonic power to the disadvantage of the devolved telTitory. An analogy with the 'unachievable demands' strategy used by revolutionary socialists to highlight the inherent deflciencies of a capitalist state seems apposite. It would be wrong, however, to ignore the fact that the coming to power of a secessionist party can, of course, work both ways. In Quebec the Parti Quebecois, despite spells in government, has so far not been successful in leading the province to sovereignty, a fact that suggests Canada's 'containment strategy', embodied in the high levels of autonomy and central representation which Quebec enjoys under federalism, may well be a sufficient incentive to dissuade voters from secession. However, in Scotland the electoral system is an important factor differentiating the situation there from that in Quebec. Quebec operates a flrst-past-the-post electoral system for provincial elections, which means that the winning party in an election invariably gains an overall m.\iority. Scottish parliamentary elections in contrast operate with an additional member system of proportional representation. This has resulted in the election in 2007 of an SNP Executive that does not enjoy an overall majority, operating now as a minority government. This makes the passage of its legislative programme more difficult, a fact that at first sight will clearly frustrate the SNP's objectives. In other words, the goals of showing Scotland's capacity for increased autonomy and the competence of the SNP to govern are, it seems, harder to achieve than they would be if the SNP enjoyed an overall majority. But, by another reading, it might be considered that the SNP has the opportunity to paint itself as co-operative in its dealings with other parties, while, if it operates skilfully, it can lay the blame on a 'unionist popular front' for the failure to pass 'nationalist' policies which enjoy public support. By the same token, not having an overall majority means that it does not have the same power to make unilateral mistakes which the other parties could then exploit for political gain. Paradoxically, therefore, an electoral system, devised by the Labour Party, which seemed to promise either to keep Labour in power, or in any event to deny any real power to the SNP were it ever to become the strongest party, may well backfire during the course of this or future administrations. Instead, it may allow the SNP to portray itself as a moderate party capable of working with unionists and, hence, as a prototype of an attractive, progressive government for an

Federalism ill a UnitalY State: a Paradox too Far?

57

independent Scotland. And so again we see how, ironically, a model designed to draw the sting from nationalism may have the opposite effect. (iii) Devoilltion and the Mechanics of Secession

Another quandary of the devolution settlement is that it has also provided the institutional infrastructure through which an indigenous push towards constitutional independence might be effected. With the establishment of a Scottish Parliament and Scottish Executive, for the first time the legislative and administrative mechanisms are in place with which, respectively, to authorize and conduct a referendum on independence. Indeed, the SNP Executive has moved quickly to begin planning such a poll (Scottish Executive, 2007). It seems clear that under the Scotland Act the Scottish Parliament is competent to pass legislation authorizing and managing a consultative referendum on independence. The Scotland Act provides that the Union is reserved (Schedule 5(1)( I )(b) and 5(Il)B6). This clearly makes unilateral secession illegal. However, there is no provision that proscribes the holding of a referendum that seeks to canvass the views of the Scottish people or asks them to authorize the Scottish Executive to enter negotiations with the UK government with a view to agreeing to Scottish independence; and so, by the logic 17 of the retaining model, such a power is devolved. The political importance of this power is potentially considerable. Neil MacCOImick pointed out that the power of the Scottish Executive or Scottish Parliament to hold an advisory referendum would be "absolutely critical" given the political significance which a Yes vote would carry. Even if the result had no legal weight, MacConnick (2000: 726) argued that a request for negotiations following such a result would be binding in political terms, in the same way as previous devolution and Common Market referendums forced Westminster's hand. In addition, with the power to hold an advisory referendum comes the power to control that process. By analogy it is interesting how control of voting rules, question wording and timing were used to considerable effect by the Parti Quebecois in 1994-95 in preparing a referendum that so nearly led to Quebec's secession. However, in this respect the electoral system seems clearly to be of aid to unionists in Scotland. An Act authorizing a referendum cannot be passed without the support of unionist MSPs (Member of the Scottish Parliament), and these same people would have considerable control over the process rules, thereby preventing the referendum process being controlled in a pmtisan way by the SNP, should the latter seek to do so. (iv) Weaknesses in Alltonomy and Representation IInder a Devolved System The model of Scottish devolution is attacked by nationalists as unduly restrictive. This critique has two main targets: a lack of autonomy and a failure to create suitable pathways for representation in central decision making. Turning Ilrst to autonomy, despite the devolution of extensive powers, there remain notable areas of substance that are not devolved. In particular, critics often target the fact that the Scottish Parliament lacks tax-varying powers, with the only exception to central Ilscal control being the power of the Scottish Parliament to vary the basic rate of income tax by three pence in the pound (Scotland Act, section 73).

58

FederalislIl in a Unitary State: a Paradox too Far?

S. Tierney

Tuming to representation at the centre, there have been few political disputes so far conceming the nature of intra-state relations, but it is widely perceived that the devolution settlement for Scotland and, indeed, for Wales and Northern Ireland, is unsatisfactory in this regard, and that amicable relations have owed a great deal to the fact that from 1999-2007 the Labour Party was in power in Edinburgh as the dominant partner in a coalition government, thus helping to maintain smooth relations with the Labour government in London. The potential for enduring problems based upon the institutional design of the system stem from what is essentially the ad hoc, or some might say incomplete, nature of the settlement. In this respect, we will focus on the features of the devolved model that are weaker than a federal system, noting their possible propensity to unsettle the union and observing that these might have been avoided by some model of federalism. Addressing Scotland as our case study, there is a lack of formally guaranteed representation at the centre for Scottish governmental institutions. In fact, centralizing dynamics operate within the settlement. And this has negative implications for two aspects of central-territorial relations in particular: inter-governmental relations and inter-parliamentary relations. The Scotland Act actually has little to say about what in federal systems is known as intra-state federalism, with a lack of detail on how institutions to co-ordinate policy for the UK as a whole would be set up, far less about how these should be designed or how they should operate. And the result is that there is not a formalized, and legally protected, set of mechanisms in place for occasions where serious competence disputes arise. Instead institutions operate largely at the behest of the centre and, therefore, depend upon the goodwill of the central government and parliament for their continuation. 18 We see this in the informal arrangements for inter-executive co-operation (Rawlings, 2000). But it is also notable in inter-parliamentary relations. The potential for tensions to be exacerbated are also evident at the level of interparliamentary relations between the Scottish Parliament and Westminster. One issue concerns a lack of clarity in the division of competences between legislatures, and the second is the lack of protection of the competences of the devolved legislature from the risk of central retrenchment; both of these seem to be particular to the UK's unitary model in contrast to a more highly developed federal system. The former has led to the infamous West Lothian question. 19 In the absence of an English Parliament, an argument persists that Scottish MPs exert too great an influence within the House of Commons,z° The contention is that since certain matters which are devolved to the competence of the Scottish Parliament are dealt with for England and, in certain cases, other parts of the UK, by Westminster, it is unfair that MPs who are returned to the House of Commons from Scotland can vote on matters which affect other parts of the UK but not Scotland. This has become particularly controversial in situations where the UK government has relied upon Scots MPs to pass legislation which will not have effect in Scotland. The consequence of this unhappy state of affairs, which is no more than a consequence of an inchoate process of decentralization, is that it has led to political difficulties that seem to have been entirely avoidable. In light of this, it can be seen to have the potential to stoke up resentments across the Union. Voters in England might justifiably feel aggrieved that their preferences are not being met; whereas attempts to resolve the issue, for example, by establishing a Grand

59

Committee within the House of Commons whereby only English MPs could vote on Bills concerned only with England,21 might suggest to Scots that the UK Parliament is now increasingly an English Parliament and that their interests might be better served by gaining more powers for the Scottish Parliament. The West Lothian issue displays how, in terms of its constitutional culture, the Westminster model is a lo~g way from recognition of the UK as a plurinational, multi-level state. The UI<: Parliament has clearly not developed the persona of anything akin to a federal legislature, able and willing to act as a forum for negotiating different territorial interests. One opportunity for change seemed to lie in reform of the House of Lords. This does not perform the role as a second, territorial chamber so familiar to federal. syste,ms. Changes made to the House of Lords since 1997 have not led to such a reOrIentation, and nor are there government plans to move reform in this direction. Concluding Remarks It may be that there is an inherent paradox in projects of constitutional decentralization

designed to accommodate sub-state nationalism in that they can serve to el~ha~ce the conditions for, rather than ward off, secession. However, such accommodation IS unavoidable if states are to remain true to the demands of multicultural and plurinational justice. Elsewhere the author has argued at length that this is a conclusion we can reasonably reach after some two decades of debate by political theorists on the compatibility of nationalism with liberal democracy (Tierney, 200~: chapters 2, 3 aI~d 9). In this time, liberal theorists in particular have made considerable advances II1 showing both the importance of sub-state national identities to many people and, therefore, that the achievement of individual goals for these people is partly contingent upon host states adapting their constitutions in order to offer these sub-state nations full recognition and the opportunity to nourish (Tully, 1995; Kymlicka, 2001; Moore, ?001; Requejo, 2005). Also, from a more pragmatic perspective, the denial of meanmgful self-government to sub-state nations seems to be ill-conceived since it may hastenyr.ocesses towards secession in less happy ways (as in the Fonner USSR and SOCialist Federal Republic of Yugoslavia) or lead states (e.g. Canada) into protracted constitutional debates that can prove to be a burden on state resources for decades. What the UK experience has illustrated is that theorists of federalism may need to re-think the assumption that federalism is the only vehicle by which telTitorial diversity might be constitutionally managed. At the same time, the lack of federal elements in the UK model, particularly in terms of representation at the centre, suggests that there are structural elements of a federal model which, if missing in a unitary or quasi-federal settlement, may in the long term prove to be dangerous. ~he solution need not be a fully formalized and constitutionally entrenched federalism. Such a model remains unpopular across the UK and, in any case, might remove the constitutional flexibility that has hitherto been so successful in responding to nationalist demands. Instead, a more structured system of inter-governmental and interparliamentary arrangements, that maintains flexibility whilst introducing some formal patterns of UK-wide co-operation and consultation, may go a long way to maintaining the idea of the British state as meaningful to, and popular among, the peoples of its sub-state nations.

60

S. Tierney

Federalism in a Unital)' State: a Paradox too Far?

61

Aclmowledgement

S (1994) Le Fedcralisme Fortement Asymetrique: Improbable et Indesirable, in L.F, Seidle (ed.), D ·Ion,. , . M I IPPR Seekillg a Nell' Calladiall Partllership: Asymmetrical alld COllji,deral OptiollS. ontrea : ,

The author is grateful to the editors and anonymous referees for helpful comments.

pp.133-152. .. . fR' I Edwards, O. D. (ed.). (1989), A Claim (!( Rightfor Scotlalld. Edinburgh: Polygon, (cIting A Clall11 0 Ig It

Notes IAlthough there IS dispute concerning Parliament's supremacy in the context of the European Communities Act 1972 ss 2 and 3, it is still generally held that devolution does not detract from the authority of Parliament as matter of legal principle. Alldersoll. Doherty alld Reid v. Scouish Millisters [20001 UKHRR 439. 2For exmnple, as set out by Wheare (1963: 10): "By the federal principle, I mean the method of dividing powers so that the general and regional governments are each, within a sphere, co-ordinated and independent", The enduring supremacy of the UK Parliament means that the devolved territories lack 'independence' in this federal sense. J According to 2006 estimates by the Office of National Statistics. "Government of Wales Act 2006. 552% voted for devolution on a 64% turnout, failing to meet the requirement in the Scotland Act! 978 that at least 40% of the entire electorate should vote in favour. "On the first question: I agree that there should be a Scottish Parliament, 74% voted Yes; and on the second: I agree that a Scottish Parliament should have tax-varying powers, 63.5% voted Yes. Referendums (Scotland and Wales) Act 1997. 7T he term 'extra-parliamentary' is, therefore, somewhat uncomfortahle. However, it was extraparliamentary in that it was unrecognized by the UK Parliament which, dominated by the Conservative Party, did not accept its legitimacy to engage in this process. HThe House of Commons has passed a resolution recognizing that Quehec forms a nation "within a united Canada". 'Resolution Respecting the Recognition that the Quebecois form a nation within a united Cunada', House of Commons,Journals, I st session, 39th Parliament, no.87, 27 November 2006. "Art 2 states: "The Constitution is based on the indissoluble unity of the Spanish Nation, the common and indivisible homeland of all Spaniards". Although other 'nationalities' are referred to they are not named, thus implicitly relegating them in terms of import~lllce behind Spain itself, while also running the reference to 'nationalities' together with a reference to 'regions'. I..http://www.parlament-cat.net/porteso j estatutj estatut_angles_1 00506. pdf liThe 'Sewel' convention named after Lord Sewel who, speaklllg for the government, suggested tillS process in parliamentary debate on the Scotland Bill. Hansard, H.L. Vol. 592, col. 791 (21 July 1998). 121n 2004 such an order was made under section 30(2) of the Scotland Act. Under section 63 further powers have also been extended to the Scottish Executive. iJ'Growing sense of Englishness explains why less than half of country feel British', Guardiall, 24 January 2007. I"For example, 'Exclude Scots MPs from English votes', The Telegraph, 30 October 2007. 15'Opposition parties hold talks on greater powers for parliament', The Scotsmall, 28 August 2007. I"Speech by Alex Salmond, SNP Party Conference, Aviemore, 30 October 2007. I7Hansard, vol. 312 Debates 12 May 1998 col. 257; Hansard, vol. 312 Debates 19 May 1998 col. 789. IH'Memos show how Blairand Brown "ignored" McConneli', SUllday Times (Scottish edn.), I 6 September 2007, at 7. I"The anomaly was raised in the late 1970s by Tam Dalycll, MP for West Lothian. 2°Despitc reduced representation by way of the Scottish Parliament (Constituencies) Act 2004. 21 'Tories will hand crucial powers to English MPs'. Observer, 28 October 2007.

References Bond, R. and McCrone, D. (2004), The Growth of English Regionalism? Institutions and Identity, Regiollal alld Federal Studies, Vol.l4, No.1, pp.I-25.

for Scotland, 1988, para 4.8). Erk, J. and Anderson, L. (2009), The Paradox of Federalism: Does Self-Rule Accomodate or Exacerbate Ethnic Divisions?, Regiollal alld Federal SllIdies, Vol. 19, No.2, pp. 191-202. ESRC Programme (2004), Interim Findings of the Devolution and Constitutional Change Programme, 'Devolution: What Difference Has it Made?'. . . , II R (2007) Out of Court: Why Have the Courts Played No Role in Resolving DevolutIon Disputes III H '" ~,. , . ~ the United Kingdom, Puhlills: the Journal of Federaltsm, Vo1.37, No.4, pp.578-5 . Henderson, A. (2007), Hierarcilies (!( Belollgillg: Natiollal IdellIity alld Politll:al Culture ill Scotlalld alld Quehec. Montreal: McGill/Queens UlIlversity Press. Keating, M. (2001 a), Plurillatiollal Democracv: Stateless Natiolls ill a Post-sovereiglllv Era. Oxford: Oxford University Press. . Keating, M. (2001 b), Natiolls Agaillstthe State-The Nell' Politics (!( Natiollalism ill Quehec, Cawloll/a alld

Scotlalld (2nd edn). UK: Palgrave. , . Kiely, R., McCrone, D. and Bechhofer, F. (2005), Whither Britishness'? English and Scottish people III Scotland, Natiolls alld Natiollalism, Vol.Il. No.1, pp.65-82. . Kymlicka, W. (1998), Filldillg Our Wa\': Rethillkillg Etl/l/oculllIral Relatiolls III Callada. Toronto: Oxford University Press. . . Kymlicka, W. (2()())), Minonty Nationalism and Multinational Federalism, III PolitiCS III the Vernacular: Natiollalism. Multiculturalism. alld Citizellship. Oxford: Oxford University Press, ~p.91-120. Leverhulme/ESRC (2007), Scottish Elecllon Study, conducted by the Scottish Centre for SOCial Research. MacCormick, N. (2000), Is there a Constitutional Path to Scottish Independence?, ParliamellIary Af(mrs, Vol.53, No.4, pp.72 1-736. McFadden, J. (1995), The Scottish Constitutional Convention, PuMic Law, Vo1.215. Moore, M. (2001), Ethics ,,(Natiollalism. Oxford: Oxford UniversityPress. Moreno, L. (1988), Scotland and Catalonia: The Path to Home Rule, III The Scottish Governmel1l Yearhook 1988. Edinburgh: Unit for the Study of Government in Scotland, pp.166-181. Moreno, L. (20()J), The FederalizatilJll (!( Spaill. London: Casso . Park, A. et al. (2005), Devolution, Identity and Public Opinion in Scotland. ESRC Programme. Avmlable at http://www.devolution.ac.ukjPark2.htm (accessed 26 February 2(09). Rawlings, R. (2000), Concordats of the Constitution, Law Quarterly ReView, Vol. I 16, pp.257-286. , F ('JOOI) Political Liheralism in Multinational States: the Legitimacy of Plural and Asymmetncal . RequeJo, , , . . " d 'C' b'l 'U'Federalism, in A. Gagnon and 1. Tully (eds), MultlllatuJIlal Democracll's. Camhn ge. ,1111 fI( ge III versity Press, pp.11 0-132. Requejo, F. (2005), Multillat/ollal Federalism alld Value Plu:-alism. L~lIldon: Routledge. Scotlalld's Claim, Scotlalld's Right (1995), Edinburgh: ScottIsh Conslitutlonal Convention. Scotland's Parliament (1997), Cm. 3658, p.v. . . Scottish Executive (2007), White Paper, CIWO.lilli( Scotlalld's FUllIre: A Natiollal COIII'ersatuJIl. Edillburgh: Scottish Executive. Tierney, S. (2004), COllstitutiollal Law alld Natiollal Pluralism. Oxford: Oxford University Press. S (2005) Reframing Sovereignty: Sub-state National Societies and Contemporary Challenges to T ·l e rney,.". 161 181 the Nation-state, IllIernatuJIlal alld Comparative Law Quarterly, Vol.54, No.1, pp. " - _. . Tully, J. (1995), Strallge Multiplicity: COllstillltiollalism ill all Age (d Diversit\'. Camhndge: Cambndge University Press. Wheare, K. C. (1963), Federal Gm'OW/lellt (4th edn). Oxford: Oxford UniverSIty Press.

The Partisan Logic of Decentralization in Europe

The Partisan Logic of Decentralization in Europe JASON SORENS Department of Political Science. University at Buffalo. USA

Introduction The paradox of territorial autonomy is that, while it should render full independence less appealing to minority nationalist voters, it also provides nationalist parties the institutional tools with which to hold a future referendum on independence and can strengthen voters' confidence that collective self-govemment can work (Kymlicka. 1998; Erk and Anderson, 20(9). It is not clear. a priori which of these effects should predominate: does decentralization reduce or increase the risk of secession? The fact that Britain, Spain, Italy and Belgium have all devolved substantial powers to regional governments over the past three decades. while countries such as Germany, Austria, Denmark, Norway, Sweden. Finland and the Netherlands have not, suggests at first glance that central governments have little doubt about the matter: decentralization appears intended to reduce the peripheral nationalist threat. The former four countries all have strong secessionist movements in the periphery. while the latter seven do not, if one excludes overseas dependencies. Apart from its overseas holdings, France has comparatively weak secessionist movements and has made little progress on decentralization, although the one region that has seen persistent nationalist connict, Corsica. received its elected regional assembly first and has

63

also received the most generous offers of autonomy. I Portugal granted autonomy to the Azores and Madeira shortly after democratization in the mid-1970s, when these island territories had seen nationalist agitation (now virtually non-existent). In short, the positive relationship between national secessionism and subsequent decentralization in Europe is extremely strong. Statistical studies have confirmed the relationship. but have also found that increases in autonomy do not result in immediate decreases in secessionist electoral support (Sorens. 2(04). Furthermore, the level of existing autonomy is not related at all to secessionist party strength (Sorens. 20(5), while legislative autonomy, defined more narrowly, is correlated positively with the growth of regional and regional-nationalist parties (Brancati, 2006; Sorens, 20(8). If devolution drastically reduced the prospects for independence, then we should expect independentist parties, such as the Scottish National Party (SNP), to reject it, but in fact the SNP consistently campaigned for devolution? If regional autonomy is intended to defuse secessionist politics, it does not seem to work. That fact in itself is interesting and worthy of explanation, but this paper addresses a further puzzle. If regional autonomy does not 'work' in the above sense, then why do central governments try it? One facile explanation might be that in democracies. governments simply give the people what they want. If voters want decentralization. then they will vote for federalist. autonomist or secessionist parties and, when they get into power, those parties will enact their favoured policy. Of course, regions with nationalist parties are almost always part of the periphery and represent a minority of the population of the country (Flanders in Belgium is an instructive exception).} Minority nationalist parties rarely get into power at the central level and, when they do, it is typically as a fairly weak junior partner in a coalition government (Rudolph and Thompson, 1985). In some cases, the m,~ority of the country can be quite hostile to decentralization proposals, particularly if they confer additional powers on minority regions. Democracies never give all the people what they want, and one might even start from the presumption that institutions of majority rule should tend to disfavour minority groups. Of course, the fact of the matter is quite different: democracies tend to be more generous toward ethnic and national minorities than are autocracies (GlIlT, 2000; Elbadawi and Sambanis, 2(02). The puzzle deepens. Rationalist political science offers a more sophisticated approach to the problem. If we assume that parties are motivated by the 'spoils of office'. then we should expect that govemments will support decentralization whenever it increases their prospects of re-election and oppose it when it does not. One could identify two approaches to the problem: one assumes that parties are motivated by the prospect of gaining ofllce at the central level (the 'central govemment office seeking' or 'CGOS' argument), and the other assumes that parties are motivated by the prospect of gaining office at the regional level (the 'regional government office seeking' or 'RGOS' argument). Under the CGOS argument, parties support decentralization in order to increase their vote share in countrywide elections and thus their prospects of forming or joining a government at the central level. Under the RGOS argument. by contrast. parties support decentralization because they expect that decentralization will increase the value of holding regional oftlce, and they expect to obtain more value from holding regional office than from holding central office. Since decentralization reduces the

64

The Partisall Logic of Decelltralizatioll ill Europe

.I. S()/'ells

value of holding oflice at the central level, but holding central ofllce is still more valuable than holding regional oflice under almost all feasible decentralization schemes, a party could expect to obtain more value from regional than central ofllce only if the party expects to join at least one regional government but does not expect to join or remain in govemment at the central level in the near future. The RGOS argument thus implies that opposition parties are more likely to support decentralization than parties in the central government and that this disparity between opposition and government preferences grows larger the less competitive are countrywide elections, and the more regionally concentrated is the opposition's support (Hypothesis I). If asymmetrical decentralization is feasible, then both government and opposition parties will tend to support autonomy for those regions where they are strong and oppose it for those regions where they are weak (Hypothesis 2). To my knowledge, this argument was first made by O'Neill (2003: 1075): "If [a] patty does not expect to compete strongly at the national level under a centralized system but expects to do well in subnational elections, then it faces powerful incentives to decentralize". O'Neill found support for the theory in five South American countries, none of which had strong secessionist movements during the period analysed, but tested it differently from this paper, using actual decentralization rather than party support for decentralization as the dependent variable and using data from presidential elections. None of the cases examined in this paper are presidential democracies, and all of them have significant secessionist movements. The CGOS argument implies that parties will support decentralization whenever doing so is a net vote-winner in countrywide elections. Since decentralization typically becomes an issue when a nationalist party promotes it, this paper limits its empirical purview to those European cases where a minority nationalist party exists. In these cases, so the argument goes, countrywide parties will adopt pro-autonomy programmes to the extent that doing so allows them to win votes from these nationalist parties. On the other hand, endorsing autonomy could lose the votes of anti-autonomy voters, who can be located within either the peripheral region where a nationalist party operates or the 'core' region of the country.4 To generate testable hypotheses from the CGOS argument, I rely on Gary Cox's (1990) theoretical analysis of centripetal and centrifugal incentives under different electoral systems and numbers of competitors. Two of Cox's results are particularly significant here: (I) the less proportional the system, the more centripetal is candidate competition; and (2) the fewer the candidates, the more centripetal is candidate competition. In the idealized case of countrywide two-party competition in a plurality single-member-district (SMD) system with a single dimension of voter ideology, both parties would seek the ideal point of the median voter in the median (most marginal) district. In such a case, both parties would support decentralization if and only if a majority of voters in a m,~ority of districts supported decentralization. Since peripheral regions with nationalist parties generally represent less than half of the voting population of the country, there would need to be some districts in the core that support decentralization by a majority, an unlikely scenario. Since voters in peripheral regions are generally more favourable to decentralization than voters in core regions, we should generally expect that only when very large majorities of voters in peripheral regions support decentralization will it even be possible that parties would find

65

supporting decentralization worthwhile. When these conditions (Ire satisfied, we should expect that both countrywide parties will endorse decentralization in an effort to win over the median voter. While this scenario is highly stylized, we should expect the following comparative statics to hold: the less proportional the electoral system, and the fewer the number of competing parties, the closer the countrywide parties will be on the issue of decentralization and, under such circumstances, they will generally not support decentralization until significant majorities in the periphery support it, and voters in the core are not significantly opposed (Hypothesis 3). Under more proportional and multiparty systems, we should expect countrywide parties to diverge in terms of their programmes for regional autonomy. Some parties may target anti-autonomy voters and some may target pro-autonomy voters, while others may downplay the issue altogether. The CGOS framework does not yield strong predictions about which parties will support and oppose autonomy, unlike the RGOS hypotheses. However, we might expect that in a strongly proportional and multiparty system, parties will concentrate on cultivating their own constituencies and adding to them on the margins. Therefore, parties with a constituency that already is favourably disposed to autonomy are more likely than parties dependent on anti-autonomy constituencies to try to broaden their appeal to voters who would otherwise be tempted to vote for nationalists. Historical sympathies and antipathies and ideological arguments about the effects of decentralization on society and economy are likely to make the difference in the policy strategy of central-government-office-seeking parties in these systems. Regarding asymmetrical decentralization, the CGOS theory has no strong expectations, but, in general, parties should support asymmetrical decentralization when it is a vote-winning strategy compared to symmetrical decentralization. This scenario seems possible only when some regions decide they do not want autonomy for themselves. Table I compares the different assumptions and expectations of the CGOS and RGOS theories of decentralization. Notably missing is the notion that decentralization reduces violence or support for secession. Despite the fact that central governments often offer such an explanation for decentralization when it happens, the fact that decentralization does not generally accomplish those things implies that partisan political calculations might be behind the process after all. Nevertheless, the empirical study in this paper will show that there are a few cases of decentralization that seemingly can be explained only by the more traditional hypothesis. The empirical method of this paper is to examine the major cases of recent secessionist challenges in European democracies in order to develop predictions from the hypotheses about whether (and which) non-nationalist parties will support decentralization and to compare the actual outcomes to the predictions. The cases are: Scotland and Wales in the UK, Catalonia and the Basque Country in Spain, Northern Italy and Sardinia in Italy, Flanders in Belgium, and Corsica in France.

United Kingdom In Britain, autonomy for Scotland and Wales initially became an issue in the nineteenth century after Ireland achieved home rule. With the First World War, the issue faded

66

Tile Partisan Logic (~f Decentralizatio/l in Europe

1. Sorells Table 1. CGOS and RGOS theories: assumptions and expeetations CGOS theory

RGOS theory Assumption

Decentralization's value Differences in party support

Support for asymmetry Trigger for support

Parties motivated by prospeet of regional ofilce Increases value of holding regional ofilce Parties expecting to be competitive in countrywide elections oppose. parties not expecting to be competitivc SUppOlt Decentralize only where party is strong Greater ehance of party sueeess in regional than countrywide elections

Majoritarian/two-party systems

Proportional/ multiparty systems

Parties motivated by prospeet of central office Captures support of median voter None

Captures support of 'nearby' voters Support where 'natural.' otherwise opposition

If some regions do not want autonomy Strong majority for decentralization in region. little opposition olltside

Unclear

away and was not revived on a wide scale until the Scottish Convention of 1947 -50. which obtained two million signatures to a covenant advocating a Scottish parliament with full legislative and IIscal powers (Webb. 1977). The Parliament of Wales Campaign from 1950-56 received 250.000 signatures (Osmond. 1997). Nevertheless. the two largest British parties. Conservatives and Labour. did not take notice until the Scottish and Welsh nationalist parties began to make electoral gains in the 1960s. 5 The SNP. formed in 1932, forthrightly advocated independence for Scotland. while Plaid Cymfll. the Welsh nationalist party, advocated 'self-government' for Wales with a seat in the UN but did not explicitly endorse independence until 2003. 6 By the late 1970s, both parties had adopted left-of-centre ideological stances, consistent with the median position in each region. The Liberal Party. perpetually in opposition, long favoured parliaments for Scotland and Wales. a commitment contained in every 7 post-war election manifesto. and there is some evidence that this position improved their electoral results in Wales especially (Davies. 1985). Before the 1990s. there was no notion of an assembly for England or its regions: devolution was to be asymmetrical. The Conservatives actually came to support Scottish devolution before Labour, endorsing the concept in opposition prior to the 1970 general election, which they won. In that election, the Conservatives finished just six and a half percentage points behind Labour in Scotland (38.0% versus 44.5%). In Wales. Labour maintained a huge lead over the Conservatives: 51.6% to 27.7%:' The Conservatives did not mention devolution for Wales. The Heath government published a White Paper on Scottish devolution but was distracted by economic decline. labour unrest and the Northern Ireland Troubles. and passed no legislation (Esman. 1994).

67

The SNP made massive gains in the two 1974 general elections on the strength of the 'It's Scotland's Oil' campaign. Reacting to the threat, the new Labour government, which after several by-election losses depended for its survival on the co-operation of Liberal and Scottish and Welsh nationalist MPs. drew up plans for limited Welsh and Scottish devolution. The 1979 advisory referendums had a 'poison pill' attached: devolution would not be enacted unless 40% of the electorate voted yes, as well as a majority of those voting. The Welsh referendum failed anyway. but the Scottish result gave a majority to the bill but only 33% of the estimated electorate. When the House of Commons then failed to pass the devolution bill. the nationalist MPs voted to bring down the Labour government, triggering the general election that brought Margaret Thatcher's Conservatives to power. During the 1980s, widespread opposition to Conservative policies in Scotland and Wales strengthened SUppOlt for devolution and independence on the left (Levy, 1990). Conservative Party vote share shrank significantly in Scotland. By the 1997 election, Conservative support was down to 17.5% in Scotland and 19.6% in Wales. Under those circumstances. it is little wonder that successive Conservative governments refused to consider devolution. or that Labour activists in both countries IInally came to support devolution sincerely, if more overwhelmingly in Scotland than in Wales (Mitchell, 1998). In the 1997 referendums on devolution, all major parties except the Conservatives campaigned for 'yes' votes, but the evidence suggests that in Wales, unlike Scotland, rank-and-fIle Liberal Democrats voted overwhelmingly 'no', while Labour identifiers were split (Jones and Trystan, 1999). Since the referendums, the Conservative Patty has suggested adopting a parliament for England as a whole, while Labour tentatively explored assemblies ror the regions of England. A 2004 referendum on an elected assembly for England's North East, a Labour stronghold. went down to defeat by a margin of almost four-to-one. In summary, the RGOS-derived hypotheses explain patterns of party SUppOlt for and opposition to devolution much better than the CGOS-derived hypothesis in the UK. The RGOS theory explains why devolution plans were typically adopted by patties in opposition. or in one case. by a governing party that needed nationalist votes in parliament, and why, since 1974, Labour supported devolution for Scotland and Wales much more strongly than the Conservatives, while the Conservatives have broached the idea or an English parliament. Unlike the RGOS theory, the CGOS theory fails to explain why Conservatives so forcefully resisted devolution for Scotland, even though opposition clearly cost them votes in Scotland and won them few, if any, in England. Another explanation for Conservative opposition to Scottish devolution would be their historical unionist stance, going back to the Irish home rule question in the nineteenth century, but this argument does not explain Conservative support ror Scottish devolution in the early 1970s. The case of Wales also supports the RGOS theory more strongly than the CGOS theOlY. Labour did not adopt devolution ror Wales because of overwhelming local support ror the idea, or because or an imminent electoral threat from Plaid Cymru. Rather, Labour constructed a Welsh devolution campaign from scratch and, ultimately, the referendum narrowly passed (McCrone and Lewis, 1999). Since 1999, Labour has led every coalition government in Wales. The reason why Hypothesis 3 performs so poorly in the UK is, perhaps, that despite its plurality SMD electoral system, the

68

J. Sorens

UK has had unstable multiparty competition in a number of districts in Scotland and Wales. Under these conditions, the Conservatives may have been maximizing their vote share after all by adopting a centrifugal position on the autonomy issue. Since Britain's system is neither clearly two-party nor clearly multiparty, the CGOS theory does not yield strong predictions for this case. The electoral institutions adopted for the Scottish Parliament and the Welsh Assembly pose a puzzle for the RGOS theory. If Labour supported devolution out of hopes of gaining power in the devolved institutions, then why did they not endorse single-member districts (SMD) with plurality rule as the electoral system, a system likely to have given Labour an absolute majority of seats in both legislatures at the time? In fact, initial plans for devolution adopted SMD plurality rule, but vigorous opposition from the Liberal Democrats and even some Labour activists helped change the minds of Labour leadership (Keating and Jones, 1995; Osmond, 1995). Fortunately for Labour, proportional representation under the additional-member system has made it virtually impossible for secessionists to obtain a majority of seats in either assembly. As of this writing, the SNP is leading a minority government in the Scottish Parliament after winning a plurality of votes and seats in the 2007 Scottish election. Had SMD plurality rule been in place, it is quite possible that the SNP could have translated its slim plurality of votes into a majority of seats. In such a scenario, a referendum on independence would be more or less inevitable. The experience of Quebec would have been available to Tony Blair and the rest of the Labour leadership. The Parti Quebecois has never won a majority of votes in Quebec, but they were able to hold referendums due to their seat nu~orities manufactured by the electoral system. By adopting proportional representation (PR), Labour satislied the Liberal Democrats on their core issue and paved the way for centre-left, unionist, LabourLiberal Democrat coalitions that took power in both Scotland and Wales in 1999,

Spain In Spain, decentralization has been fairly symmetrical by contrast with Britain, although regional autonomy in Spain remains asymmetrical by contrast with traditional federal systems, such as the USA or Canada, and it proceeded quickly after democratization. Under Franco, Spain had been a highly centralized state. During the transition, some measure of regional autonomy was a foregone conclusion; it was a key demand of the anti-Franco opposition, a legacy of the political cleavages that divided Spain during the 1936-39 Civil War. The Spanish Constitution set forth two alternative paths for the creation of new 'autonomous communities'. The historic nationalities were originally allowed to 'opt in' to a wide range of competencies, while the second-track autonomies had a specific and unalterable list of competencies, with the residual power left to the central state (Fossas and Colome, 1993). Catalonia, Basque Country, Galicia and Andalusia obtained their autonomous institutions first and still possess greater powers in a few areas, but by 1983 all of European Spain had been broken up into autonomous communities. Besides the Basque Country and Navarre, Spanish autonomous communities still depend for their budgets on unrestricted grants from the centre and thus do not constitute a liscally federal system. Spain is often characterized as a "semi-federal" system (Conversi, 2(00).

The Partisan Logic

(~lDecentralizatioll

ill Europe

69

After the initial wave of regionalization, further controversy remained over whether the historic nationalities should retain more powers than the other autonomous communities, and over the extent of autonomy that all autonomous communities should enjoy. Lower-income regions, such as Andalusia, defended autonomy in regulatory policy and cultural affairs but generally favoured centralization of liscal powers, while higher-income regions, such as Catalonia and the Basque Country, demanded more fiscal autonomy from the centre (Keating, 1988: 212). The deepening of EU institutions has provoked calls from some of the autonomous communities for direct representation at the EU level. This provision was part of the Ibarretxe Plan for 'free association' between the Basque Country and Spain, rejected by the Spanish Cortes in 2005,'1 The conflict of interest between centralist and autonomist regions implies that support for autonomy even for strongly nationalist regions, such as the Basque Country and Catalonia, is not necessarily a net vote-winner for countrywide parties (partidos de 1I11lbito estatal). Due to its small district magnitUdes, Spain has one of the most highly consolidated party-list PR systems in Europe, with two dominant parties and a small countrywide third party (United Left). Thus, Hypothesis 3 predicts ambivalence or opposition from the two largest parties toward further decentralization. In the Basque Country and Catalonia, nationalist parties usually form the govemment at the regional level. Of the two largest non-nationalist parties, the Socialists (PSOE) and Popular Party (PP), the PSOE is generally more successful than the PP in both autonomous communities. Thus, Hypothesis 2 might suggest that the PSOE would be more sympathetic to further powers for Catalonia and the Basque Country than would the PP. However, there are not many weakly nationalist voters that the Socialists could win over. In both the Basque Country and Catalonia, the local branches of the PSOE receive a significant proportion of their SUppOlt from 'immigrants' born elsewhere in Spain, who are overwhelmingly anti-nationalist (Shafir, 1995: 65-68). This consideration should mitigate the Socialists' friendliness toward further powers for these two autonomous communities. In fact, neither the PSOE nor the PP has been very sympathetic to the full extent of Catalan and Basque demands. While in opposition, the Socialists endorsed LOAPA, a 1981 law intended to limit the powers both of the historic communities that had already obtained autonomy and of new autonomous communities yet to be created. LOAPA allowed the central government veto power over regional laws and required regional civil servants to be appointed by the central government's civil service. Over the objections of the now-Socialist government, the Constitutional Court voided much of the law in August 1983, while the provisions that remain have generally worked to raise the autonomy of the non-historic communities to the standard enjoyed by the historic communities (Agranofr. 1996; Conversi, 1997). Nevertheless, the PSOE has, as expected, generally treated nationalist concerns more favourably than the PP. When the president of the Basque Country, Juan Jose Ibarretxe, proposed a referendum on his plan for a new autonomy statute, the PP government of the time described the plan as 'treason' and passed a law authorizing imprisonment for any political leader who holds a referendum, which they argued to be illegal in Spain. 10 The PSOE government elected in 2004 ultimately rejected Ibarretxe's plan but scorned talk of arresting him if he called a referendum. This PSOE government did approve a more modest change in Catalonia's autonomy statute, which had been approved by referendum in

70

1. S(J/'ens

Catalonia. The most widely reported aspect of Catalonia's new statute was the fact that the preamble describes Catalonia as a 'nation' rather than a 'nationality'. terminology reserved for Spain itself in the Spanish Constitution. II The PP. by contrast. has taken the new Catalan statute to the Constitutional Court. seeking to have it overturned. Thus, the CGOS theory helps to explain the uneven and highly contested progress of decentralization in Spain following the initial creation of the autonomous communities. The RGOS theory is not very relevant for explaining the initial decentralization. which was a foregone conclusion during democratization. but it perhaps does explain why the Socialists have been a little more supportive of regional autonomy than the Popular Party. Furthermore. nationalists have led most coalitions in the Basque Country and Catalonia. making it unlikely that any countrywide palty could govem alone in either autonomous community. Of course, the Catalan Socialists did enter a coalition with the Catalan Republican Left (ERC) in 2003 and helped to develop the new autonomy statute for the community. while bearing in mind the concerns of the PSOE leadership (in fact. the modest nature of the new statute even caused ERC to leave the coalition and force an election in 2(06). Finally. the fact that the antiFranco opposition held to their plans for decentralization through the long decades of the dictatorship is certainly consistent with the RGOS theory. Italy

Apart from the regions with 'special autonomy' (Sicily, Sardinia. Aosta Valley. FriuliVenezia Giulia and Trentino-Alto Adige). Italian regions began the post-war period with only administrative, not legislative. authority and were unelected. despite constitutional provisions specifically authorizing elections and legislative power for the regions (Article 117). The special-status regions had legislative autonomy only over cultural affairs and subject to framework laws from the central government. Thus. Italy was essentially a unitary state. In 1971. the Italian government linally implemented elected regional government. but their powers remained strictly administrative. During those years, the Communists. who were perpetually in opposition. were the main supporters of decentralization (Gobetti. 1996). Initially. regions had some minor tax-raising powers. but the Christian Democrat (DC)-dominated central government abolished these powers in 1979. The issue of federalism for Italy quickly came to the fore with the Northern League's meteoric rise in 1990-96. With distrust of the central government at an all-time high due to the Tangentopo/i bribery scandals. the relatively high-income North in particular favoured greater autonomy over economic policy. and roughly a lifth of Northern voters have consistently favoured independence for the North (Gobetti. 1996; Woods and Beirich. 20(0). The Northern League joined the electoral alliance of the right with Forza Italia and the National Alliance in the 1994 elections. and the alliance was victorious. In government. Prime Minister Berlusconi. leader of Forza Italia. ignored the NOlthem League's demands for a federal reform of the state. League leader Umberto Bossi began a campaign of opposition within the government and ultimately brought the government down. supporting a short-lived government of technocrats until the 1996 election. in which the League ran on its own on a more radical platform of autonomy for the North (Tarchi. 1998). In that election the

The Partisan Logic of Decentralization in Europe

71

League scored its best result ever and began a short-lived campaign for the immediate secession of Northern Italy under the name of 'Padania' (Cento Bull and Gilbert, 2(00). As party in-fighting developed, the League turned away from secessionism to radical right-wing popUlism beginning in 1998, and its electoral results plummeted. In oflke. the centre-left government supported modest decentralization to the regions, including the granting of legislative powers in specific areas. A constitutional amendment enshrining these changes was passed by referendum in 200 I. Meanwhile, the Northern League rejoined the centre-right alliance. and Forza Italia this time made a serious commitment to a federal reform ofItaly. The alliance ('House of Freedoms') won the 2001 election. but the new constitutional proposals took years to work out, mostly due to the recalcitrance of the National Alliance, a party with its core strength in the relatively low-income. centralist South. By the time a referendum on the proposed amendment took place. in 2006, the alliance was out of power and widely unpopUlar, having lost a general election earlier that year. Unsurprisingly, the reforms were voted down by more than 60% of those voting. The amendment would have truly federalized Italy, granting the regions direct representation in a strong Senate and explicitly guaranteeing them exclusive jurisdiction in a variety of policy areas. The demand for decentralization has clearly been strongest in Northern Italy. especially Lombardy and Veneto, the strongholds of the Northern League. 12 These are wealthy and relatively conservative regions, having long voted for the DC before Tangentopoli. Since the League's hard right turn. they have gone solidly for Forza Italia. The evidence suggests that a majority of NOltherners would like to see greater autonomy, and thus Hypothesis 3 predicts that in Italy's two-alliance system the leaders of both alliances would try to compete for Northern votes by offering autonomy, provided that the issue did not lose them votes in the South. Hypothesis 2 would predict, by contrast. that Forza Italia would be more amenable to decentralization than the other countrywide parties. particularly the members of the 'Olive Tree' and 'Union' coalitions of the left. of which the Democrats of the Left constitute the largest member. (Hypothesis I is not very relevant here. since Italian elections have been highly competitive since 1994.) Since their stinging rebuke from the NOlthern League in the ill-fated 1994 government. Forza Halia has indeed come to SUppOltapparently sincerely-federalism for Italy. Meanwhile. the centre-left has vigorously campaigned against further decentralization. especially of taxation powers. The CGOS theory is consistent with the fact that all major countrywide parties came to support some decentralization in the I 990s, but only the RGOS theory can explain the differential support for further decentralization from Forza Italia and the centre-left. The RGOS theory also can help us understand why the DC governments that domimlted Italy from 1945 onwards delayed constitutionally mandated decentralization for so long. The DC feared that decentralization would give the Communists a regional base (Hypothesis I), allowing them to use patronage to build a bigger electoral machine or perhaps even to foment political instability. In the end, their fears were partly justified: the 'red centre' of Italy did indeed elect Communist-led governments. but as a general rule, Communist regional governments were actually better run than those in the rest of Italy (Putnam. 1993).

72

J. Sorens

The Partisan Logic of Decentralization in Europe

Belgium As in Italy, peripheral nationalism made a late entry on the Belgian political scene. In the 1950s and 1960s, Walloon Rally (RW) and People's Union (VU) focused on, respectively, Walloon and Flemish linguistic rights and regional economic interests and were not nationalist or secessionist in any sense. The Brussels-based Democratic Front of Francophones (FDF) focused on French speakers' linguistic interests in Brussels and its suburbs and was formed comparatively late, in 1964, as a reaction to the establishment of Dutch-only services in the Brussels suburbs. By the late 1970s, VU developed a radical wing pushing for extensive autonomy or even independence for Flanders, the relatively high-income region of Belgium. Thereafter, Walloons and Brussels Francophones split on the issue of decentralization, with relatively lowincome Wallonia favouring maintenance of the unitary state and French-speaking 'Bruxellois' generally favouring autonomy for Brussels, with full regional status equivalent to that contemplated for Wallonia and Flanders in the 1970 constitutional revision. That amendment began the process of decentralization in Belgium and represented a response by the major parties, now broken up into Francophone and Flemish wings, to the success of the explicitly linguistic parties. It established Francophone, Flemish and German non-territorial 'communities', which would have the right to legislate on linguistic issues, including education, for those communities. The 1978 Egmont Pact established limited territorial autonomy for the lirst time in the form of the regions of Flanders, Brussels and Wallonia. The status of Brussels was controversial, with Flemings demanding its inclusion within the Flemish region in accord with its historical sta~us. Ho~ever, the majority of Brussels residents are Francophone and they strongly re~ls~ed tillS demand: Moreover, there were disputes about which suburbs belonged wlthm the borders 01 the Brussels region. It was 1989 before the Brussels region was actually implemented. The Egmont Pact foresaw elected regional legislatures, but these were not initially implemented due to outstanding matters of dispute. Instead, parliamentary members from each region comprised the regional bodies. After 1978, further decentralization was more of a Flemish than a Walloon demand. In 199 L the lirst elections to the regional legislatures took place. In 1993, a n.ew constitutional amendment established Belgium as a federal state, allowing l:eglOns to decide their own institutional forms, giving regional law the same legal force as federal law, and requiring regional consent to future constitutional amendments. Regional governments are also allowed to make foreign treaties in areas under their specilied competencies. However, regions still lack robust fiscal autonomv and derive over 80% of their revenue from federal grants. There are no electorally signilicant 'countrywide' parties in Belgium, because all of the major parties have split along linguistic and regional lines and contest elections at both the regional and central levels. However, there are major non-nationalist parties (or parties that did not initially adopt explicitly nationalist commitments) that have had to respond to the nationalist challenge, particularly in Flanders. As a general rule, after the initial regionalization, Walloon and, to a lesser extent, Francophone parties have opposed further decentralization, particularly of taxation, while Flemish parties have generally supported further decentralization. Name changes in the main Flemish parties exemplify this dynamic. The 'Christian People's Party' (CVP) became J

73

'Christian Democratic and Flemish' (CD& V), while the 'Party for Freedom and Progress' (PVV) became 'Flemish Liberals and Democrats' (VLD). The mainstream Flemish pmties also consistently revised their manifestos toward greater emphasis on Flemish autonomy as support for the notion rose among Flemish voters. All the mainstream parties have had outright secessionist factions, and CD& V has officially changed its constitutional stance to favour a loose confederation among the Belgian regions. These trends are consistent with Hypothesis 3. Hypotheses I and 2 both predict that Flemish parties that expect to hold a stronger role in government at the federal level than at the regional level would be less enthusiastic about confederation or independence, either of which would render the federal level much less important. Since Flanders is, on balance, a conservative region, while Wallonia is left of centre, left-wing panies in Flanders have less hope of a leading role in Flemish regional government than of a leading role in the Belgian government. At the Belgian level, Flemish and Francophone parties of the same ideological family generally join government together due to the constitutional requirement of linguistic parity in the cabinet. Thus, before the 2007 election, VLD and the Francophone Reform Movement (MR) formed the Belgian government, together with both Flemish and Francophone Socialists. If Flanders were to separate from Belgium, the Socialist tendency in Flanders would be much weaker than it is currently in Belgium, and the Christian Democratic tendency in Wallonia would be much weaker than it is currently in Belgium. Thus, we should expect that parties of the right will be more sympathetic to confederation or independence in Flanders, while parties of the left will be more wary of these options. By the same token, we might suppose that the Democratic Humanist Centre (cdH) would have the most political anxiety among the Francophone parties about the possibility of Flemish secession. These predictions hold up rather well. CD&V has formed electoral cartels in 2004 and 2007 with New Flemish Alliance (N-V A), an explicitly secessionist, conservative pmty formed out of the collapse of VU. The left-wing breakaway from People's Union, Spirit, has, by contrast, moderated its nationalist platform and formed an electoral cartel with Socialist Party: Different (SP.a), arguably the least nationalist of the major Flemish parties. VLD have a moderate stance on the nationalist issue, which has led to the departure of their secessionist (and most conservative) factions, most notably the new party, Dedecker List. There is an almost perfect correlation between left-right ideology and radicalism on the national question among Flemish parties. VB is the most right-wing major Flemish party and is explicitly independentist. A radical anti-immigration party, VB took its current name after a judicial determination that the previous incarnation, 'Vlaams Blok,' could not receive public funding as a racist party (Erk, 2005). On the Francophone side, Hooghe (1991) noted that the cdH was lighting federalization for fear of domination by the Socialists in the regional government of Wallonia.

France France remains a unitary state, but small regionalist and secessionist movements have cropped up from time to time over much of the country's territory: Brittany, Alsace, Occitania (southern France), Savoy, Northern Catalonia (Roussillon), Northern

74

1. Sarens

Euskadi and Corsica. Of these, Corsica has had by far the most successful nationalist parties, generally garnering about one-fifth of the regional vote in recent regional elections. It has also suffered significant violence and banditry, including assassinations, kidnappings, aImed robberies and bombings, most of it coming from self-styled nationalist groups, most notably the Corsican National Liberation Front (FLNC), founded in 1976. Hypothesis 3 predicts that French parties should support autonomy for Corsica to the extent that the policy gains them more votes there than it loses them in France as a whole. It is doubtful that this condition is satisfied. Autonomy is a highly polarizing issue in Corsica, and there are not enough 'moderate nationalist' voters to entice the French parties to appeal to their interests, especially since France's majoritarian electoral system discourages targeting of minority ideological constituencies. While between a fifth and a qUaI1er of Corsicans vote for nationalists, there is also a significant antinationalist constituency, concentrated among French civil servants, pieds /lairs resettled from Algeria, and others tied financially to the central government, of whom there are many on the island. The most recent proposal to transfer some legislative competencies to the Corsican regional government was narrowly defeated in a referendum, 50.98% opposed to 49.02% in favour. The plans had been drawn up by Socialist Prime Minister Lionel Jospin and then promoted heavily by the centre-right government of Jean-Pierre Raffarin with the blessing of President Jacques Chirac. Thus, the evidence suggests that further devolution to Corsica has been a multipartisan affair, motivated not by electoral considerations but by an interest in stopping the violence. However, the initial establishment of the French regional assemblies does fit Hypothesis I. The Socialists had been in opposition since the beginning of the Fifth Republic (1958) but won control of both the Presidency and the National Assembly in 1981. In 1982 the Socialist-Communist coalition agreed to establish an elected assembly for Corsica and, in 1986, elected regional councils throughout France were established. Moreover, these regional councils used proportional representation for elections, which the Socialists also implemented briefly for the French National Assembly before it was again reversed by the centre-right coalition that won the 1986 election. PR helped reduce the centre-right's seat strength by allowing the National Front to win seats. Since the Front was and remains anathema as a coalition partner, PR made it more difficult for the centre-right to form governments at the regional level. Again, decentralization made sense for the centre-left as a way of reducing the risk associated with future countrywide elections. Defeat in National Assembly elections, which soon came, would not-and did not-preclude some political control and patronage in certain regions where the centre-left was strong. Corsica itself has a Gaullist minority, but is generally dominated by the slightly left-of-centre Left Radicals (PRG), who are a minor force in France as a whole but generally align themselves with the Socialists in the National Assembly. Because Corsican politics is dominated by a minor party, the Gaullists, UDF and Socialists have no reason to support additional powers for the island for patronage reasons. Discussion This section summarizes the evidence for the three hypotheses. Hypothesis 3, derived from CGOS assumptions, is broadly consistent with the evidence on the degree of

The Partisan Logic of Decentralization in Europe

75

central government parties' support for autonomy from all cases except Scotland and Wales. The prediction of complete convergence between the competing parties of government at the centre obviously does not hold up, even in majoritarian systems, with the possible exception of France, where right and left together supported more autonomy for Corsica. However, the failure of the referendum in Corsica suggests that the proposed new powers had little to do with electoral considerations. The divergence between the largest countrywide parties on autonomy is strongest in the UK, where the Conservatives steadfastly opposed devolution for Scotland and Wales, while Labour eventually came to support it, long after the Liberals had. RGOS-derived hypotheses I and 2, by contrast, predict differences in non-nationalist parties' support for decentralization, as well as the timing of changes in that supp0l1. Those predictions are broadly borne out in every single case examined in this paper. The only minor inconsistencies are the fact that the Conservatives proposed devolution for Scotland before Labour, despite being slightly behind Labour electorally in Scotland at the time, and that the centre-left government in Italy did implement modest decentralization reforms, rather than ignoring the concerns of the NOith completely. We can speculate that in these two cases the relevant parties hoped to make some inroads into moderate nationalist voters in peripheral regions, efforts that ultimately did not seem to payoff. By and large, the RGOS logic explains patterns and timing of decentralization in Europe beller than the CGOS logic. The rationale behind regional-government-office-seeking decentralization might seem cynical or at the very least insincere. These findings cast some doubt on the notion that competitive parties adapt their programmes to the 'will of the people', suggesting instead that, when they can, they will reshape institutions so as to ensure their long-term interests in gaining and wielding power. Since institutions set the rules of the political game for the foreseeable future, parties should light fiercely over how those rules are set (Tsebelis, 1990). Regional autonomy, as an institution, permanently reduces the value of governing at the central level and increases the value of holding regional government office. Thus, competing parties are more likely to see decentralization as a fundamental change in the rules of the game, rather than just another 'policy output' tailored to win votes. For an out-oF-sample robustness check of this paper's lindings, we might consider the case of Canada. However, because provincial paIties in Canada are organizationally independent of the federal parties, we should expect RGOS motivations to play no role in federal government-led decentralization. CGOS motivations may playa large role, however, because the most nationalistic province, Quebec, elects a large proportion of the seats in Canada's parliament and is split about evenly between secessionist and autonomist voters. The federal party that most convincingly mticulates an autonomist programme for Quebec generally comes in second in federal elections there (behind the secessionist Bloc Quebecois). The Progressive Conservatives won a number of seats in Quebec in the 1980s with their promise of constitutional negotiations. After those negotiations failed to produce a successful reform, the Progressive Conservatives were wiped out in Quebec and replaced with the Bloc Quebecois. During the lead up to the 1995 referendum, the Canadian Liberal government promised a "renewed federalism" for Quebec if the 'No' side won, and the Liberals remained the second largest party in Quebec until 2005, when the Conservatives surprised observers

76

1. Sorens

by defeating the Liberals in Quebec, having shed their earlier image of hostility to FrancopilOnie. In November 2006 the Conservative govemment passed a symbolic measure in Parliament recognizing Quebec as a "nation within Canada". The sheer size of Quebec makes it a plum prize in federal elections and perhaps explains why federal parties have not completely ignored Quebeckers' demands for further decentralization.

Conclusion The results of this paper help to resolve a particular paradox of regional autonomy: that central govemments in advanced democracies have offered autonomy to peripheral regions with secessionist movements, despite the fact that autonomy does not usually dampen future secessionist electoral support. In general, parties in opposition support decentralization, particularly to regions where they are electorally strong, because they hope to gain power somewhere and enjoy the spoils of govemment at the regional level. In a sense, decentralization is an insurance policy for political parties facing winner-take-all electoral competition at the centre. The evidence from European parliamentary democracies confronting secessionist movements in the periphery confirms the basic theory for which evidence has been found among Latin American presidential democracies without secessionist movements (O'Neill, 20(3). If this logic accurately describes govemments' motivations, then monopolization of a region's politics by a secessionist or other regional party would reduce the prospects for decentralization. A party or coalition with the power to implement autonomy (i.e. controlling the central govemment) would have no regional-ortlce-seeking incentive to implement autonomy, because they would not enjoy a reasonable chance of gaining power at the regional level. For instance, we might expect that the Canadian govemment is less willing to offer new autonomy to Quebec than they would be if federal parties competed in Quebec provincial elections and did well. Similarly, Danish govemments lack strong reasons to concede further powers to the Faroe Islands and Greenland, and Finnish govemments have no electoral incentive to satiate Aalanders' demands for wider self-govemment. U The central-govemment-ortlce-seeking theory, which argues that decentralization follows popular demands as a way of winning votes at elections to the central parliament. is not so much wrong as it is usually inapplicable. To begin with, there are probably few cases in which decentralization is a net vote-winning strategy. since the votes won in the periphery can be cancelled out by votes lost in the core. Even if decentralization does net a few votes in countrywide elections. the gains may not outweigh the costs of a likely-permanent alienation of power from the centre to regional govemments. While there is some evidence of CGOS motivations in the literature. particularly in the UK and Italy. pattems of party support for decentralization seem to hew more closely to the predictions of the RGOS theory. If central govemments are interested in reducing the likelihood of secessionist referendums in their peripheral regions, then the appropriate strategy is not necessarily further decentralization. but redesign of the regional electoral system. so that nationalists are less likely to hold a majority. This strategy may not work in places such as the Basque Country and Catalonia, where a solid majority of regional voters support

Tile Partisan Logic of Decentralization in Europe

77

nationalist parties. but it does reduce the risk of secessionist parties' winning manufactured majorities in a winner-take-all system. Future research on the 'paradox of autonomy' might focus on the electoral system as a critical factor in nationalist politics at the regional level. The evidence presented in this paper helps us to understand why so many regional autonomy institutions seem badly designed from the point of view of the mitigation of secession risk. Institutions that help countrywide parties compete consistently for power at the regional level even while locked out of power at the central level may also allow secessionists the opportunity to gain power in certain regions. By the same token. if regional autonomy dramatically increased secession risk. then we would expect decentralization to be much rarer in democracies than it actually is. The fact that secession is extraordinarily dirtlcult in established democracies has created space for governments to use regional autonomy for partisan political purposes.

Acknowledgement The author would like to acknowledge helpful suggestions from Larry Anderson. Joshua Dyck, Jan Erk. Gregg B. Johnson. Harvey Palmer and Joost Van Spanje. All remaining errors are the author·s.

Notes IThe most recent proposal for legislative powers for the Corsican assembly was narrowly defeated in a referendum on the island. See BBC News Online. 'Corsicans reject autonomy offer', 6 July 2003, available at http://news.bbc.co.uk/2/hi/europe/3049986.stm (accessed 22 August 2007). 2Some SNP leaders did fear precisely that devolution would cause voters to reject independence (Keating and Bleiman. 1979; Levy, 1990) . .lIn this case. the constitutional requirement of linguistic parity within the cabll1et and the cort/oll .\'(llIilaire against Flemish Interest (VB) mean that all central governmcnt coalitions will include Francophone parties: the Flemish parties cannot simply "rule Belgium", "O'Neill (2003: 1076) argued that decentralization was nearly universally popular in her South American cases, perhaps because regional nationalism was not an issue. 50smond (1997) argued that Labour opposed home rule dUring the 1940s and 1950s because they thought that it would require a reduction 111 Welsh and Scottish parliamentary representation. 6BBC News Online, 'Plmd commits to independence: 21 September 2003, available at http://news. bbc.co.uk/2/hi/uk_news/wales/31244I R.stm (accessed 21 January 200S). 7 Historical Liberal. SOP and Liberal Democratic manifestos can be found at http://www. Iibdemmanifesto.com (accessed 21 January 200R). HFigures come from Bryn Morgan and Joseph Connelly, 'UK Election Statistics: 1945-2000: 29 March 2001, House of Commons Library. "BHC News Online, 'Basque independence plan rejected', 2 February 2005, available at http://news. bbc.co.uk/2/hi/europe/4228297.stm (accessed 23 August 20(7). iOBBC News Online, 'Spain to block Basque plan', 31 October 2003, available at http://news.bbc.co. uk/2/hi/europe/3232397.stm (accessed 23 August 2(07). II BBC News Online, 'Spain MPs back Catalonia autonomy', 30 March 2006, available at http://news. bbc.co.uk/2/hi/europe/4860244.stm (accessed 23 August 20(7). 12The Sardinian nationalist movement has been much weaker and countrywide parties of the left and right are roughly equally strong on the island, generating no strong predictIOns from Hypotheses I and 2. I.1Greenland, the Faroe Islands and Aaland all enjoy quite broad powers currently, but, in all three cases, thcre are strong political parties demanding further decentralization or even independence.

78

1. Sorens

References Agranoff. R. (1996), Federal Evolution in Spain. Illtematiollal Political Sciellce Rel·iew. Vo1.l7. No.4. pp.385-40 I. Brancati. D. (2006), Decentralization: Fueling the Fire or Dampening the Flames of Ethnic Conflict and Secessionism? Illfematiollal Orgallizatioll, Vo1.60. No.3. pp.65 1-685. Cento Bull. A. and Gilbert. M. (2001). The Lega Nord alld the Northem Questioll ill Italiall Politics. Basingstoke. UK: Pal grave. Conversi. D. (1997). The Basques. the Catalall.I·, alld Spaill: Altemative Routes to Natiollalist Mobilizatioll. Reno: University of Nevada Press. Conversi. D. (2000). Autonomous Communities and the Ethnic Settlement in Spain. in Y. Ghat (ed.), AWOIlomy alld Et/Illicity: Negotiatillg Competillg Claims ill Multi-Ethllic States. Cambridge: Cambridge University Press. pp.122-146. Cox. G. W. (1990). Centripetal and Centrifugal Incentives in Electoral Systems. Americall .loumal (!/' Political Sciellce. Vo1.34. No.4. pp.903-935. Davies. J. (1985). Plaid Cymru in Transition. in J. Osmond (ed.). The Natiollal Questioll Agaill: Welsh Politic'alldelltity ill the /9S0s. Llandysul: Gomer Press. pp.124-154. Elbadawi. I. and Sambanis. N. (2002). How Much War Will We See? Explaining the Prevalence of Civil War• .louma/ of COllflict Resolutioll. Vo1.46. No.3, pp.307-334. Erk. J. (2005). From Vlaams Blok to Vlaams Belang: The Belgian Far-Right Renames Itself. West Europeall Politics. Vo1.28. No.3. pp.493-502. Erk. J. and Anderson. L. (2009). The Paradox of Federalism: Does Self-Rule Accommodate or Exacerbate Ethnic Divisions? Regiollal alld Federal Studies. Vo1.l9. No.2. pp. 19 1-202. Esman. M. J. (1994). Ethllic Politics. Ithaca. NY.: Comell UllIversity Press. Fossas. E. and Colome. G. (1993). Political Parties alld IllstitutiOllS ill Catalollia. Barcelona: Institut de Ciencies Politiques i Socials. Gobetti. D. (1996). La Lega: Regularities and Innovation in Italian Politics. Politics alld Soclety. Vo1.24. No.1. pp.57-82. Gurr. T. R. (2000). Peoples Versus States: Millorities at Risk ill the New Celltury. Washington. DC: United States Institute of Peace Press. Hooghe. L. (1991). A Leap ill the Dark: Natiollalist COllflict alld Federal Rej(mll ill Belgium. Ithaca. NY.: Cornell University Press - Western Societies Program. Jones. R. W. and Trystan. D. (1999). The 1997 Welsh Referendum Vote. in B. Taylor and K. Thomson (cds). Scotlalld alld Wales: Natiolls Agaill? Cardiff: University of Wales Press. pp.65-93. Keating. M. (1988). State alld Regiollal Natiollalism: Territorial Politics alld the Europeall State. New York: Harvester-Wheatsheaf. Keating, M. and Bleiman. D. (1979). Labour alld Scollish NatlOllalism. London: Macmillan. Keating, M. and Jones. B. (1995). Nations. Regions. and Europe: The UK Experience. in B. Jones and M. Keating (eds). The Europeall Ullioll alld the Regiolls. Oxford: Clarendon Press, pp.89-114. Kymlicka. W. (1998). Is Federalism an Alternative to Secession? in P. Lchning (ed.). Theories (!/,SecesslOll. London: Routledge. pp.109-148. Levy. R. (1990). Scollish Natiollalism at the Crossroad.I·. Edinburgh: Scottish Academic Press. McCrone. D. and Lewis. B. (1999). The Scottish and Welsh Referendum Campaigns. in B. Taylor and K. Thomson (cds). Scotlalld alld Wales: Natiolls Agaill? Cardiff: University of Wales Press. pp.17-40. Mitchell. J. (1998). The Evolution of Devolution: Labour's I-lome Rule Strategy in Opposition. GOl'emmellt alld Oppositioll. Vo1.33. No.4. pp.479-496. O·Neili. K. (2003). Decentralization as an Electoral Strategy. Comparative Political Studies. Vo1.36. No.9. pp.1 068-1 091. Osmond. J. (1995). Welsh Europea/lS. Bridgend: Poetry Wales Press. Osmond. J. (1997). Creative COllflict: The Politics of Welsh Del'olutioll. Liandysul: Gomer Press. Putnam. R. D. (1993). Makillg Democracy Work: CiI'ie Traditiolls ill Modem Italy. Princeton. NJ.: Princeton University Press. Rudolph. J. R.• Jr. and Thompson. R. J. (1985). Ethnoterntorial Movements and the Policy Process. Compa/'{/til'e Politic.I·. Vo1.l7. No.3. pp.291-311.

The Partisan Logic of Decentralization in Europe

79

Shafir. G. (1995). Immigrallfs alld Natiollalists: Ethllie COllfliet alld Accommodatioll ill Catalollia. the Basque Coulllly, Latvia. alld Estollia. Albany. NY.: SUNY Press. Sorens. J. (2004). Globalization. Secessionism. and Autonomy. Electoral Studies. Vo1.23. No.4. pp.727-752. Sorens. J. (2005). The Cross-Sectional Determinants of Secessionism in Advanced Democracies. Compa/'{/til'e Political Studies. Vo1.38. No.3. pp.304-326. Sorens. J. (2008). Regionalists Against Secession: The Political Economy of Territory 111 Advanced Democracies. Natiollalism alld Ethllic Politics. Vo1.l4. No.3. pp.325-360. Tarchi. M. (1998). The Lega Nord. in L. De Winter and H. Tursan (cds). Regiollalist Parties ill Westem Europe. London: Routledge. pp.143-157. Tsebelis. G. (1990). Nested Games. Berkeley: University of California Press. Webh. K. (1977). The Growth (!/' Natiollalism ill Scotlalld. Glasgow: Molendinar Press. Woods. D. and Beirich. H. (2000). Globalization. Workers. and the Northern League. West Europeall Politics. Vo1.23. No.1. pp.130--143.

The Paradox

The Paradox of Ethnic Partition: Lessons from de facto Partition in Bosnia and Kosovo ERIN K. JENNE Intelllational Relations and European Studies DepaJ1mcnt. Central European University. Budapest. Hungary

Introduction Despite the near-universal condemnation of nationalism and ethnic cleansing by the international community. ethnic partition has gained resonance since the end of the Cold War as a 'method of last resort' for dealing with nationalist conflicts I in Kosovo. Sri Lanka. Palestine. Kashmir. East Timor, Aceh, Chechnya. Transdniestria. South Ossetia. Abkhazia. Nagorno-Karabakh, Iraq, Cyprus and Sudan. Ethnic partition draws on the logic of deterrence theory, which was a military strategy developed during the Cold War to respond to state threats. Very generally, such theories hold that the absence of supranational authority creates a condition of anarchy at the international level under which all states are forced to ensure their own security against the transgressions of others. Security dilemmas emerge when one state defensively arms to make itself more secure. leading other states to arm in response. The resulting arms spiral paradoxically renders all states less secure. When the prevailing military technology can be used offensively. every state has an incentive to strike first in its conflicts

(~lEth/lic

Partitioll

81

with other states, making militarized conllict more likely. To deter external attacks, each state therefore has an incentive to signal that it has a significant defellsive military capacity, in which case it is unlikely to strike first but can retaliate against a first strike with overwhelming force (Schelling, 1960; Jervis. 1978). The ethnic security dilemma extends this logic to the sub-state level. Here, state failure or collapse reproduces the condition of international anarchy on the domestic level. In the absence of state authority. individuals mobilize along the default category of ethnicity to ensure their own security. Conflict occurs in this scenario because the military technology used in guerilla warfare can be used for both offensive and defensive purposes, making it difllcult for groups to determine whether their opponents have aggressive aims. This uncertainty leads to spirals of conflict as each side operates on worst-case assumptions concerning the other's motives (Posen. 1993). Neither side will disarm unless there is an external security guarantee (Walter. 1997) or until one side achieves an outright military victory (Licklider, 1995). For these reasons. some have argued that separating the groups into distinct. defensible territories through border adjustments and supervised population transfers is the only viable solution to deeply entrenched ethnic conflicts. Kaufmann. the best-known advocate of ethnic partition. argues that ethnic civil wars call for just such extreme solutions. as they create internal security dilemmas so intense that they cannot be resolved without segregating 2 the warring groups into "defensible enclaves".' According to this logic. de jure partition is not necessary so long as the combatant groups are partitioned into territorial enclaves. allowing them to defend themselves against attacks by the other side. In this view, de facto ethnic partition should be sufficient for resolving the internal security dilemma. Scholars have criticized ethnic partition by arguing that it entails unacceptable costs: population transfers usually lead to significant loss of life, and territorial partition risks transforming internal conflicts into even deadlier inter-state wars. such as the nuclear stand-off between India and Pakistan following the partition of the subcontinent (Schaeffer, 1990). It can trigger new internal conflicts by producing new minorities that mobilize for independence. Territorial partition is also likely to encourage secessionist efforts elsewhere by similarly-situated minorities hoping for their own state. Finally, it is not even clear that partition delivers on one of its key selling points: preventing the re-emergence of conllict. In his analysis of 125 civil wars. Sambanis (2000) concluded that wars terminated by partition are no less likely to recur than wars ended through other means. 4 This article focuses on neither the humanitarian costs nor the morality of partition nor the effectiveness of this approach for preventing the re-emergence of violent conflict. These critiques have been made adequately elsewhere. 5 I instead examine whether de facto partition has helped to reduce nationalist conflict in post-war settings. This moves us beyond the more limited concern of whether partition can prevent the re-emergence of violence to the question of whether partition can be used to rebuild peaceful state-societies and stabilize the region following sectarian war. I conclude that ethnic partition. like ethnofederalism. actually pel]Jetuates nationalist conflict in post-war states as the leaders of territorial enclaves consolidate their sovereign power by marginalizing or persecuting ethnic minorities. At the same time. segregationist

82

E. K. Jenne

state institutions give national elites the means of aggrandizing political and economic control over their territories-power that may be projected outward to destabilize the region as a whole. I assess the validity of this argument by examining the effects of de facto partition on post-war Bosnia and Kosovo. These conflicts are examined in parallel because they are clear examples of de facto partition that occurred under international administrations that had full control over the protectorates. If partition can resolve nationalist conflict anywhere, it should do so in cases where international administrators have near complete discretionary power over postwar institutions. In this sense, the Bosnia and Kosovo conflicts serve as crucial tests for the predictions of partition theory. An additional reason for selecting these cases is that pm1ition advocates hold that Bosnia and Kosovo are actually cO/~firmat01y cases, arguing that the residual nationalist conflicts in the two protectorates are the product of inadequate ethnic partition. According to Downes, the integrationist elements of the peace settlements (minority returns and power-sharing) prolonged the conflicts by perpetuating ethnic insecurities and defensive nationalism on all sides. He recommended instituting a more complete partition by creating homogeneous nation-states in place of Bosnia and Kosovo (Downes, 2004: 265-266, 276). Contrary to this asscessment, the following analysis suggests that the ongoing conflicts in Bosnia and Kosovo are driven more by the institutional empowerment of nationalist elements than they are by intractable security dilemmas resulting from memories of war. If true, this indicates that partition is not the panacea for ethnic war that its advocates claim it to be, and that-at least in the Balkans-it has had the paradoxical effect ofpel7Jefllating conflict, albeit at a low level.

Why Partition Theory Gets it Wrong Many of the problems that plague ethnofederalism also adhere to ethnic partition. Ethnofederalism. particularly its consociational variant. is designed to persuade separatist minorities to remain in a common state. It does so by instituting minority vetoes, a system of regional autonomies and power-sharing arrangements, which are believed to render the minorities less apprehensive about their future in a majoritydominated state. Empowering rebellious minorities is thus seen as a means of ethnic de-mobilization. 6 Similarly. partition aims to de-mobilize warring groups by implementing structures that can provide them with a maximal means of defence should the other side violate the terms of the peace agreement. Like ethnofederalism. however. ethnic partition has the opposite effect of prolonging nationalist conflict. First. segregationist state institutions tend to undermine the state by weakening or even eviscerating the central government. Divided decision making degrades central governance by promoting legislative deadlock and reducing the government's ability to respond strategically to changing social and economic conditions (Rothchild and Roeder. 2(05). Zahar (2005) provides empirical support for the counterproductive impact of consociationalism on divided societies in unstable regions. She uses the case of Lebanon to demonstrate how power-sharing arrangements fail to secure peace when one or more groups has a powerful external ally. By conferring institutions of selfgovernment upon the combatant parties. consociational arrangements actually increased sectarian conflict in Lebanon by fragmenting the central government into ethnic factions

The Paradox of Ethnic Partition

83

that entered into alliances with outside patrons; the result was successive foreign interventions. This danger is amplified with defacto pm1ition, as the govemment has already been stripped of much of its sovereign power and is. therefore. even more vulnerable to foreign intrigues. Second. partition can promote connict by providing groups with resources for collective action. Indeed. the possession of autonomous institutions fuelled the secessionist movements in the Caucasus through a variety of social. economic and political mechanisms (Cornell. 20(2). Scholars of nationalism have shown that the possession of republican status under the Soviet Union inadvertently reinforced national identities and provided republican elites with the mobilizational resources to launch successful independence movements against Moscow (Roeder, 1991,2007; Brueilly. 1994; Slezkine. 1994; Suny. 1993; Bunce. 1999; Hale. 20(0). Segregationist institutions thus serve as potent resources for future independence movements once opportunity structures emerge (Bunce. 1999; Cetinyan. 2002; Jenne, 2004. 2(07). Third. ethnically segmented electoral systems promote extremism by inducing politicians to compete for the votes of a single ethnic group. Because of this, political contenders are unlikely to campaign on a platform of ethnic moderation. Instead, politicians will compete with one another over who is the strongest champion of the group's national interests. This dynamic leads to cycles of "ethnic outbidding" and electoral victories for ethnic extremists who refuse to co-operate across ethnic lines (Nordlinger. 1972; Shepsle and Rabushka. 1972; Horowitz. 1985; Harris and Reilly, 1998). It also creates a climate of ethnic fears and hostilities toward minorities. particularly during election cycles. Finally. ethnic partition places state resources in the hands of ex-militia leaders, rebel groups and warlords who stand to benefit from perpetuating nationalist conflict. It does so by conferring ethnic ownership over a given territory to militant leaders who maintain their grip over the region's economic and political institutions by cultivating ethnic tensions and fears (Rothchild and Roeder. 2005). Ceding post-war institutions to combatant groups. if only implicitly. elevates wm1ime militia elites to positions of political power in the post-war peliod. permitting them to prolit from trafficking, smuggling, money laundering and other forms of organized crime in frozen conflict economies (King, 200 I). To consolidate their power. nationalist elites are likely to encourage discrimination against. and outmigration of. ethnic minorities who implicitly challenge their claims to territory. We now turn to the record of partition in Bosnia and Kosovo to determine whether the pattern of ongoing conflict in each case matches the expectations of partition theory. If partition theory holds. then the sectarian conflicts in Kosovo and Bosnia should map onto ethnically intermixed areas as embattled minorities strike out pre-emptively in response to threatening local majorities. Post-war conflict should also centre around attempts by intemational administrators to reintegrate the groups through refugee return, army and police reintegration or efforts to strengthen the central government. Most importantly. violent conllict should occur primarily in ethnically-mixed rather than partitioned areas. If conflict is instead fuelled by institutionally empowered elites. then violence should be associated with struggles over ten-itory as well as efforts by local authorities to ethnically cleanse their territories and win greater autonomy from the state. If these

84

E. K. .lelllle

connicts fears of down in conflicts

are shown to be more a function of institutional empowerment than mutual victimization, then the security dilemma premise of partition theory falls both cases, suggesting that de facto partition has done little to resolve the in Bosnia and Kosovo and may have even perpetuated them.

Lessons from Ethnic Partition in the Balkans Bosnia As late as 1992, Bosnia's institutions were remarkably multi-ethnic. The republican armed forces were headed by a Muslim commander, a Serb deputy commander and a Croat chief of stafr. The Cyrillic and Latin alphabets enjoyed equal standing in the republic, and the three groups were represented proportionately in the army ranks. In June 1990, a large majority of Bosnians reportedly believed that nationalist parties should not be allowed to compete in the elections (Woodward, 2000: 160). Even as the initial Serbian offensive swept across the repUblic, "leJthnically mixed municipalities tended to remain passive in the forlorn hope of avoiding connict" (Hoare, 2004: 53). External factors played a key role in both the timing and the outcome of the Bosnian war. The declaration of independence by the Bosnian Assembly in October 1991 followed similar declarations by the Siovenian and Croatian leadership earlier that year. Indeed, military operations in Bosnia were initiated by connict spillover from neighbouring Croatia and were supervised by Yugoslav People's Army (JNA) forces. With Zagreb and Belgrade intervening on the side of their Bosnian counterparts, the connict soon morphed into a three-way struggle over Bosnian territory. The war ultimately produced 1.3 million refugees and I million internally displaced persons (IDPs), transforming an ethnically integrated republic into ethnically cleansed enclaves (Heimerl, 2005: 371). After three years of lighting, the international community brokered a deal among the combatant parties at an air-force base near Dayton, Ohio. The resulting Dayton Peace Agreement (DPA) established a post-war government with a mix of integrationist and segregationist elements. Although provisions for reintegration were built into the settlement, these were poorly implemented and the result was ethnic partition in all but name.

De Facto Pa,.tition unde,. Dayton Dayton created two powerful entity governments-the autonomous BosniakiCroat Federation (FBI H) for Bosniaks and Croats (under which most powers were devolved to ten cantons, eight of which had Muslim or Croat mL~orities) and the "radically autonomous" Republika Srpska (RS) for Serbs (Bose, 2005: 326). The entities were placed under a weak central government headed by a three-member presidency representing Bosniaks, Serbs and Croats.? The Bosnian Constitution gave each group an effective veto over any legislation deemed harmful to its "vital interests" (Article IV, 3.e). The enclaves enjoyed extensive self-government, including executive and legislative powers, canton and entity-based electoral systems and-most striking of all-ethnically segregated armies and law enforcement systems. The Bosnian Constitution also

The Paradox of Ethnic Pa,.tition

85

permitted the entities "special parallel relationships with neighboring states" (Article III, 2.a), allowing the Bosnian Croats and Serbs dual citizenship. To ensure the integrity of the state and prevent violations of Dayton, the Oflice of the High Representative (OHR) was created to oversee the protectorate, and a NATO force of 60,000 was stationed on the ground to implement the settlement. s This arrangement is a close (~lcsimile of the ethnically homogeneous, militarily defensible enclaves envisioned by partition theorists as a solution to the ethnic security dilemmas associated with internecine violence. Rather than laying the foundation for ethnic peace, however, de facto partition reproduced Bosnia's wartime divide in the post-war period, effectively imperiling the remaining ethnic minorities. The power-sharing institutions of Dayton were a dead letter from the beginning due to successive electoral victories of ultra-nationalists on all sides. The dominance of extremist politicians was due to Bosnia's segmented electoral system. Indeed, the ethnically homogeneous political systems have led to precisely the kind of ethnic outbidding anticipated by Horowitz, ensuring the success of ultra-nationalist paI1ies in 1996, 1997, 1998,2000,2002,2004, 2006 and 2008. At the same time, Dayton's system of ethno-territorial autonomies conferred considerable powers upon ex-warlords and militant groups who proceeded to establish top-down control over their national territories. This. in turn. promoted lawlessness, corruption and human rights abuses. Former warlords occupied high political positions, corruption nourished at every level of government, the police and law enforcement openly discriminated against members of ethnic minorities. indicted war criminals remained at large and foreign I1ghters (including Mujahadeen and Al Qaeda) enjoyed sanctuary in their respective national territories. According to one analyst, "Foreign governments continue to give direct support to ethnically controlled militaries. All this ... cements the wartime ethnic cleansers in power within monoethnic political frameworks" (Lyon, 2000: 110). Significantly. RS-the renegade Serbian republic during the war-had been incorporated whole-cloth into the new Bosnian government: War entrepreneurs adapted well to the post-conflict period and international intervention I ... andl were inadvertently assisted in this project by the priorities of peacebuilding. Not only did the international agencies underwrite the nationalists' grip on power (through ethnically strati lied elections in Bosnia), but the economic priorities of the post-connict 'protectorates' presented opportunities for the winners to reap further riches (Pugh, 2004: 55). In contrast. ethnic reintegration was nowhere in evidence. Lacking effective enforcement mechanisms. the international administrators failed to prevent ultranationalist politicians from paralyzing power-sharing institutions at the federal level. With the help of national vetoes. the perennial dominance of extremist leaders on all sides created "a de facto stalemate in the chambers of parliament and [prevented] any substantial decision making, not only on issues directly pertaining to the sensitivities of the national groups ... " (Bieber. 2006: 19). In view of the persistent gridlock in the Bosnian assembly. the High Representative was forced to govern through legislative lIat. Although the Oflice had a mandate of only one

86

E. K. Jenne

year. its powers were extended to three years and then expanded and extended indefinitely. This, in turn, undermined the development of Bosnia's democratic institutions. According to one observer. "[t]he powers and authority of the state have been subsumed by external actors, sucking out the life from the elected bodies which were initially to have taken over government responsibilities following a year's transitional period" (Bose, 2005: 326). Neither was there significant progress on minority returns until years later. Immediately after Dayton. the United Nations High Commissioner for Refugees (UNHCR) worked with the World Bank to facilitate refugee returns in Bosnia. However, the UNHCR had no enforcement powers and was only able to achieve returns that were neither disputed by new occupants nor opposed by local authorities. This was really only the case for returnees who belonged to the local ethnic majority; minority returns were very rare in the early years. UN officials noted that the authorities in the separate entities were "discouraging retUl11ees in the interests of creating ethnically pure blocs" (Gumbel. 1995). The international community had naively assumed that they could coax minority returns using "a mix of bribes. threats and other leverage available" (Heimerl. 2005: 379). As one Organization on Security and Cooperation in Europe (OSCE) official observed, the problem was that they "had no authority to coerce returns, [they] needed a legal system to implement these rights"Y A former OSCE official, Massimo Moralli, concurred: "In the first two years. there was no success in property restitution; only undisputed cases were resolved ... Until 1997, people didn't even have the right to reclaim [their property] because the municipalities had fixed the law". 10 Minority returns only got underway after the UN Security Council gave the High Representative the power to sanction and even replace officials who were blocking the process. Today, I million of the original 2.2 million refugees and IDPs have reportedly returned to their homes, including approximately 446,000 minority returnees. I I Strikingly, over 90% of refugee property claims have been adjudicated. However, this did not necessarily translate into ethnic reintegration on the ground (Heimerl, 2005: 384-387). According to one OSCE official involved in the returns, "property restitution happened too late; people were already settled in their new homes [and had new jobsl". 12 The IDPs had already rebuilt their lives elsewhere and were reluctant to pull up roots. Many were also concerned about their safety if they were to return. They, therefore. sold their property and remained where they were. UNHCR Protection Assistant Vesna Vignjevic pointed out that although the processing of property claims was nearly complete. "property restitution is just the first step in this process; sustainable return is the final goal". 13 Due to the unchallenged obstructionism of entity authorities in the early years, Bosnia missed a critical window of opportunity for ethnic reintegration after the war.14 With the militant leaders in control of homogeneous national territories, a nonconfrontational international administration and toothless central government. the prolonged national conllict in Bosnia can hardly be attributed to the fearful response of ordinary people who were forcibly reintegrated with their former enemies or to the preemptive strikes of beleaguered minorities who feared victimization. Nor can reintegration be blamed for the prolonged national divide: the federal power-sharing arrangements have been inoperative due to unwieldy national vetoes, and sustainable

The Paradox of Ethnic Partition

87

minority retull1 has been disappointing. Rather than ethnic fears, the conllict instead derives from Dayton's segregationist elements, which have empowered militants and warlords with an interest in continuing the conllicl. The system of ethno-territorial autonomies has given enclave authorities the tools to aggrandize power and wealth and the incentive to prevent minority returns and undermine the federal government. The minorities who have retul11ed to their homes have reacted to pervasive harassment not by mobilizing defensively. but by enduring the persecution or moving away. Had Dayton not institutionalized the results of wartime ethnic cleansing with segmented political and security structures, and had the international administrators pursued minority returns and sanctioned obstructionist authorities more aggressively in the early years. Bosnia might be closer to a peaceful multi-ethnic state today. There would also be no Croatian and Serbian statelets to tempt the territorial ambitions of Croatia or Serbia in the future. One analyst observed. "As a result of these incredibly careless moves today the international community is stymied and the ordinary citizens of Bosnia are paying the price" (Lippman, 2004: 38-39).

Were there Integrationist Alternatives? Although the ongoing conllict in Bosnia is largely due to Dayton's segregationist institutions, a more integrated state has always been a possibility. In 1995, the Bosnian Serbs and Croats were fiercely opposed to joining a common Bosnian state, but understood that they had lillie choice in the matter as the USA and its NATO allies had already co-opted their erstwhile patrons-Belgrade and Zagreb. Thus, a settlement was reached that entailed a signilkant degree of post-war integration. However. these policies were not implemented in the early years because the international community preferred to cede power to the warlords and ultra-extremists rather than bear the costs of confrontation. Dayton also suffered from an "enforcement gap". "Whereas the provisions of the Accords were highly ambitious and designed to build a lasting peace, the means for implementing them were severely limited in both scope and authority" (Daalder. 2000: 174). There is evidence to suggest that the implementation of Bosnia's integrationist principles was not only possible, but likely to have yielded a peaceful multi-ethnic society over the long run. Once the OHR received authorization in 1997 to enforce minority returns, sack obstructionist officials and arrest indicted war criminals, there was an immediate and marked increase in the rate of minority returns. There has also been a perceptible movement toward recentralization of the state since Bosnia began accession negotiations with the EU. The entity governments have gradually ceded powers to the central government and. in 2007, the new High Representative. Miroslav Lajcak, pushed through constitutional reforms to reduce minority veto powers in the federal government, thus strengthening the centre at the expense of the entity authorities. Most remarkably, in mid-2005 the entity governments agreed to unite their separate armies and defence ministries to meet NATO's condition for joining the Partnership for Peace Agreement pursuant to full membership (lCG Report, 2007: 15-17). The fact that the entity leaders voluntarily gave up their means of self-defence suggests that reintegration of former combatant parties is viable following internecine warfare, in contrast to the expectations of partition theory.

88

E. K. Jenne

Kosovo

A former autonomous province in the Serbian Republic of Yugoslavia, by 1991 Kosovo had a population of 1.8 million and was 90% ethnically Albanian. Although Kosovar Albanians had endured repression and expatriations after World War II, the situation had largely normalized by the 1960s and, under the 1974 constitution, Kosovo acquired de facto republican status within the federation. In the 1980s, Kosovo re-emerged as a touchstone issue as Yugoslavia's economy weakened and the federation grew increasingly fragile; the Serbian public came to believe the once-marginal view that ethnic groups, such as the Albanians, had gained at the expense of the Serbian nation. In the late 1980s. Siobodan Milosevic (by then leader of the Serbian Communist Party) played on this hot button issue to enhance his political power, pushing through a raft of laws to change the demographic balance in the province in favour of Serbs. In 1989, Belgrade revoked Kosovo's provincial autonomy and dissolved its assembly. It continued to 'Serbianize' the province throughollt the 1990s, dismissing ethnic Albanians from positions in hospitals, state-run enterprises, the judiciary, the police, media outlets and schools. In response to these and other policies, a shadow Albanian government set up parallel social and economic institutions in Kosovo. Meanwhile, the militant Kosovo Liberation Army (KLA) instigated a violent insurgency against Serbian forces with the support of the Homeland Calling Fund and allies over the border in Albania. Following failed peace negotiations in early 1999, NATO commenced a 78-day bombing campaign to force the Serbian government out of the province. As part of the peace agreement, the UN Security Council Resolution 1244 mandated that Yugoslav forces withdraw from Kosovo and that a protectorate be established under international auspices. NATO put a force on the ground to monitor implementation of the agreement, and the United Nations Mission in Kosovo (UNMIK) was formed to oversee the reconstruction of the country's infrastructure, economy and public institutions.

De Facto Partitioll under UNM1K

UN Resolution 1244 mandated an ethnically integrated Kosovo. The resolution committed the international community to "ensure conditions for a peaceful and normal life for all inhabitants of Kosovo" as well as the "safe and free return of all refugees and displaced persons" (Annex 2, Articles 4 and 5). It further stipulated that Kosovo remain an integral part of the federation; UNMIK was tasked with forming transitional political institutions, "taking full account of [... J the principles of sovereignty and integrity of the Federal Republic of Yugoslavia" (Annex 2, Article 9). These institutions were to be implemented by an international peacekeeping force-a condition vital for gaining the acquiescence not only of the Serbs, but also the Russians in the UN Security Council. Despite these directives, the settlement was implemented in such a way that de facto partition was the inevitable result. Although the agreement explicitly called for an international peacekeeping force, the Kosovo Forces (KFOR) was established under NATO command. This was a far cry from the UN-led peacekeeping force

The Paradox (~f Ethnic Partitioll

89

envisaged by Belgrade when it signed the agreement. As champion and wartime ally of the Kosovo Albanians, NATO was perceived as anything but a disinterested arbiter. Consequently, UNMIK and KFOR troops "were received as liberators by the Kosovo Albanians and as an occupying force by the Serbs" (Yannis, 200 I: 43). Fearing for their safety now that the ethnic tables had turned, up to 200,000 Kosovar Serbs !led the province as the war came to a close. In contrast to Bosnia where ethnic cleansing occurred during the war itself, most Serbs !led or were driven out ({fter the war had ended-in full view of tens of thousands of KFOR troops. The remaining 100,000 Serbs quickly became ghettoized into enclaves (Beaumont, 1999). A little less than one-half now reside in small enclaves south of the Ibar River surrounded by Albanian territory; the remainder lives north of the river where they comprise the dominant ethnic group in several municipalities close to the Serbian border. The province was thus partitioned on KFOR/UNMIK watch, contrary to the stipulations of the peace agreement. The international administrators did not attempt to prevent Kosovo's de facto partition, nor did they try to block Belgrade from reasserting its in!luence over Serbian regions in northern Kosovo. Soon after the war had ended, the Serb National Council-headed by the Serb separatist, Milan Ivanovic-was established in north Mitrovica with its own military defence. The aim of the renegade government "was to consolidate the ethnic divide along the Ibar River in Mitrovica as a way to safeguard the survival of the Serbs in the northern part of Kosovo, and to prepare for possible [de jure1 partition" (Yannis, 200 I: 41). Since that time, Kosovo's de facto partition has grown ever more entrenched. According to a recent ICG report, "Kosovo Serbs, with the support of Belgrade, have where possible quit, disabled or seized key institutions-municipal authorities, police, courts, customs and railway-and strengthened parallel institutions" (ICG Report, 2008: 3). With the backing of their kin state, the Serbs of north Kosovo have dug in their heels and refused to co-operate with Kosovo authorities. Besides permitting the establishment of a Belgrade-backed Serbian statelet in northern Kosovo, UNMIK further reWed the ethno-ten'itorial divide by giving control over Kosovo's political institutions to Kosovar Albanian elites, most of whom had ties to the former KLA. The elevation of ex-KLA militants to positions of leadership led to the infiltration of the government by organized crime, which had economic interests in perpetuating ethnic hostilities. Even after the war had ended, the KLA military police (the black-shirted Policia Ushtarake) continued its operations with impunity, extorting 'contributions' from businesses and individuals, expropriating property and burning houses belonging to minorities (King and Mason, 2006: 56-57). COlTlrption remained entrenched at the highest levels of the government. As the director of a local NGO put it, "The political elite as a whole is closely connected [to] organized crime. If they get benefits from a situation of conflict, why should they address the issues that separate the groupS?",15 An International Crisis Group (lCG) analyst concurTed, "There are political and economic interests in keeping the status quo [of partition] in the north-it is in the interest of businesses to keep things 'as is"'. 16 This is why ethnic partition after war tends to have conflict-promoting effects-because partitioned territory is usually conferred upon warlords and guerilla leaders who stand to gain from perpetuating the con!lict that brought them to power in the first place.

90

E. K. Jenne

Kosovo's new leadership held onto much of its KLA weaponry and internal command structures. They formed the so-called Kosovo Protection Corp (KPC) as a civilian defence organization that recruited heavily from the former KLA. The commanders of the six 'regional task forces' were the very same KLA commanders of these zones during the war (King and Mason, 2006: 58). One analyst explained UNMIK's unstinting support for the militia leaders: "UNMIK officials chose to coopt the armed Albanian elite because they had no stomach for confrontation" (Pugh, 2004: 58). Although the Resolution mandated that the KLA be disarmed, serious decommissioning did not take place. KLA armaments were instead trafficked by former insurgents, many ending up in the hands of the Liberation Army of Presevo, Medvedja, and Bujanovac (UCPMB) in Serbia and the National Liberation Anny (NLA) in Macedonia, as ex-KLA leaders took advantage of their political dominance in Kosovo to launch violent insurgencies over the border into Serbia and in Macedonia. Ethnic Albanians in southern Serbia and western Macedonia responded to these crossborder campaigns by mobilizing for annexation. Not surprisingly, the Serbian minority has been the target of continual harassment and discrimination in post-war Kosovo. Although UNMIK maintains that inter-ethnic crimes have levelled off in recent years, polls conducted by UNDP/Riinvest show that half of the Serbian population considers ethnically motivated attacks to be their great17 est threat. Serb villagers in south-eastern Kosovo have reported systematic harassment by Albanians trying to force them to sell their homes to ethnic Albanians. Tellingly, ethnic Serbs continue to limit their movement outside their enclaves. The pogrom launched against Serbs in March 2004 and the upsurge in shootings and bombings since August 2005 convinced many Serbs that the Albanians planned to expel them once Kosovo attained independence (ICG Report, 2006: 9). This pattern of violence suggests that, rather than being driven by mutual ethnic fears and contlict spirals, Albanian militant leaders and extremists have responded opportunistically to their post-war leverage by driving out the Serbs to establish an even stronger claim over Kosovo territory. Meanwhile, the provisions for ethnic reintegration are essentially a dead letter. Although Serb refugees have the right to return to their homes and reclaim their property under Resolution 1244, there are no effective means of enforcement. According to Human Rights Watch, fewer than 18,000 of the 250,000 Serbian, Roma and other minority refugees have returned to their pre-war homes (Human Rights News, 2(08), and the international administration has made no concerted effort to increase the rate of return. Although minority return in Bosnia failed to deliver pre-war levels of integration, the success it did achieve was due to a well-elaborated and enforced regime of property restitution-something that is largely absent in Kosovo. According to the head of a local NGO, "UNMIK [has] not focused on property restitution at all, in contrast to Bosnia. Here, there is a much softer, more hands-off approach ... there are no sanctions for non-cooperation by municipal authorities". 18 The integrationist elements of Resolution 1244--refugee return and property restitution-remain largely unimplemented. On this evidence, the ongoing nationalist contlict in Kosovo cannot be attributed to mutual insecurities and fears resulting from memories of fratricidal violence during the 1999 war. Rather than spirals of connict driven by mutual fears of

The Paradox of Ethnic Partition

91

victimization by the other group, the Albanian militants-the presumptive winners of the post-war peace-have perpetrated nearly all the post-war harassment and violence against small pockets of ethnic Serbs, who hardly pose a threat to the Albanian majority. Nor can the continuing nationalist contlict be attributed to the integrationist elements of the peace arrangements. Since minority refugees have not returned to their homes and the remaining Serbs (who are largely confined to their enclaves) do not participate in Kosovo' s governing structures, there is little in the way of reintegration in Kosovo. Although prolonged uncertainty over Kosovo's status undoubtedly generated anxieties on both sides, the group responsible for the greatest number of ethnically motivated attacks-the UNMIKlKFOR-backed ethnic Albanians-have the least to fear from the other group, as the weakened and ghettoized Serbs have tled the province rather than take up arms. The continuing hostilities in Kosovo are a direct consequence of de facto partition, which has given the Serb and Albanian elites the incentive to pursue policies of ethnic exclusion in their respective territories. Ex-KLA leaders established control over the institutions of the new state, while Serb separatists in the north used their alliance with Belgrade to consolidate control over northern Kosovo. The Ibar River is the central battleground for these clashing agendas. The Kosovo contlict is thus far more a story of consolidation of power by ethnic militants over their respective territories than it is of mutual fears and insecurities derived from memories of war. Were there Integrationist Alternatives?

Partition advocates believe that the continuing conflict in Kosovo is largely due to incolllplete ethnic partition of the post-war state. In their view, the Kosovo war created mutual fears so strong that co-existence in a unitary state was untenable in the post-war period. According to partition theory 'ethnic islands' in enemy territory generate irresolvable ethnic security dilemmas, creating incentives to strike first in order to avoid being vanquished. If true, then most incidents of violence in post-war Kosovo should be concentrated in and around the Serbian enclaves in southern Kosovo. In contrast, violence should be minimal in the north where the Ibar river divides the largely self-governing Serbs from the Albanians in the south. Contrary to this prediction, ethnic clashes in post-war Kosovo have occuned precisely where ethnic partition is the 1110st complete-along the Ibar River in Mitrovica. Indeed, Mitrovica has been the site of significant communal violence, including the ethnic riots that injured 300 and left eight dead in 2004 as well as ethnic clashes following Kosovo's declaration of independence in 2008. That ethnic violence in Kosovo is concentrated along the fault line between the Albanian and Serbian territories suggests that the hostilities are driven more by post-war competition over tenitory than by the ethnic security dilemma, which predicts greater violence in areas of ethnic internlingling in the Serbian enclaves in the south. 19 Also telling is the fact that the Serbs in north Mitrovica put forward more extreme demands in the status negotiations than the Serbs in the south, who proved more amenable to compromise over Kosovo independence. Since Kosovo's declm'ation of independence, the Serbs in the north "have used force or its threat to expel

92

E. K. Jenne

or reshape Kosovo institutions ... [whereasJ the Serbs in the non-contiguous enclaves south of the river have been more cautious" (lCG RepOlt, 2008: 25). Given the fact that the Serbs in the south are in a far more perilous position vis-ti-vis their Albanian neighbours than Serbs in the north, the greater intransigence of Serbs in the north can only be due to their comparatively greater bargaining leverage against the Kosovar Albanian leadership?O Although the Serbian enclaves have attempted to maintain their solidarity with their more intransigent co-ethnics in the north, their geography restricts their cooperation with Belgrade, leading them to adopt a more conciliatory approach toward the Albanian-dominated state. As a result, integrationist policies have enjoyed greater (albeit still minimal) success in the south than in the north, suggesting that reintegration is possible if separatist minorities are offered strong economic incentives, and perceive no viable alternatives, to joining the state.

Conclusion The notion that partition (de facto or de jure) is the only viable solution to protracted sectarian conflict is a seductive claim that is not, however, borne out by empirics. Partition theory is premised on the notion that ethnic civil wars generate intractable ethnic security dilemmas, according to which warring groups cannot be reintegrated because neither party can trust the other not to take advantage of it once it has disarmed. Post-war Bosnia and Kosovo are routinely cited as examples of ethnic security dilemmas that necessitate partition as a 'last resort' solution. However, this article shows that the prolongation of these conflicts is more a function of institutional empowerment of ethnic extremists that occurred as a result of de facto partition than it is of grass roots fears and insecurities due to ethnic intermingling, memories of war or fears of ethnic reintegration. De facto partition has ensured the electoral success of nationalist parties and policies; impeded property restitution and refugee return; permitted rent-seeking and corruption by nationalist elites; and segregated security and police forces along ethnic lines-creating a climate of extreme insecurity for ethnic minorities residing in the 'wrong' territory. It was not that integrationism was impossible in Bosnia and Kosovo; the peace arrangements included strong mandates for reintegration in both protectorates, and OHR and UNMIK had full control over implementation. Post-war Bosnia and Kosovo were instead partitioned by default due to the passivity of the international administrators in the face of fierce resistance by local ethnic elites. The evidence in these cases suggests that there were integrationist alternatives to de facto pmtition in both protectorates. In Bosnia, the international administrators might have gradually dispensed with the entities and cantons by incrementally transferring powers from the autonomies to the central government or by creating multiple overlapping jurisdictions. Refugee return and property restitution could have been implemented more effectively in both cases through third party coercion. reversing much of the wartime ethnic cleansing. In Bosnia, the power-sharing government might have avoided gridlock through incremental introduction of majoritarian decision making or limitations on the use of national vetoes.

The Paradox (~f Ethnic Partition

93

The case of Prijedor in Bosnia illustrates how effective third-party intervention can yield meaningful ethnic reintegration in post-conflict settings. Prior to the war. Prijedor was a mix of Bosnian Muslims, Serbs and Croatians, with Muslims and Serbs as the two largest groups. Under Serb control, Prijedor became the site of detention camps and mass slaughter during the war; the Muslim population was expelled from the municipality. Following years of stalled return-despite guarantees of return and property restitution under Dayton-Muslims began to return en masse in the late 1990s after NATO removed (and in one case killed) indicted war criminals and forcibly dismissed numerous police officers who had been involved in the wartime atrocities. This. according to one analyst, gave Muslims a "sufficient sense of security" to attempt to return to their homes in Prijedor. Today, about half of the pre-war Muslim population has returned; Bosniaks now participate in every area of public life, including the municipal government (Belloni, 2005: 440-441). Indeed, the infeasibility of reintegration after sectarian violent may be a myth, the credibility of which is bolstered by the fact that integrationist solutions are so rarely tried. Where they have been applied. policies of reintegration appear to work so long as they are paired with a credible guarantee of protection from ethnicallymotivated attacks. which may be a critical precondition for success. Indeed, ethnic reintegration achieved reasonable success in post-World War I Czechoslovakia, where the previously irredentist Sudeten German leadership ultimately wound up in coalition governments with the main Czechoslovak parties from the mid-1920s to the mid-1930s. The brief ethnic Albanian insurgency in Macedonia in 200 I was resolved under the Ohrid Framework Agreement, which de-commissioned the Albanian rebels, gave the Albanian language official status and provided for minority returns as well as in ethnic integration in law enforcement and the civil service. Perhaps the most striking example in recent history, Rwanda has managed to achieve a remarkable degree of ethnic reintegration in the wake of the 1994 genocide. At the end of the war, a National Reconciliation Commission was established to rebuild a common Rwandan identity. special courts were setup to address ethnic grievances associated with atrocities and successive governments have included both Hutus and Tutsis. If these societies were able to reintegrate so soon after having lost their friends and l~lInilies in internecine conflict or where one side feared significant ethnic retribution by the other, there is every reason to believe that such institutions might be implemented to good effect in post-conflict settings elsewhere. The use of reintegration to rebuild post-conflict societies, as well as the conditions that favour the success of such strategies, stand as promising and vitally important avenues for flllther research in the field of peace studies.

Acknowledgement The author gratefully acknowledges the generous support of the Carnegie Corporation, which provided the funding to conduct research for this article. Thanks also go to LaiTY Anderson, Jan Erk, one anonymous reviewer as well as Charles Philpott, Rhodri Williams, Florian Bieber, Matteo Fumagalli and Anil Duman, who offered useful feedback on earlier drafts.

94

E. K. Jenne

Notes INationalist conflict is defined here as struggles between two or more ethnic communities over control of state territory. institutions or other resources-such conflicts may be violent or non-violent. Nationalist conflict is most easily identified in its violent forms. but varies in intensity from more moderate conflicts Involving discrimination against or persecution of ethnic minorities; to collective rebellion or secessionism; to the most extreme forms such as genocide or full-blown civil war. Ethnic or sectarian war IS thus a subset of the larger category of nationalist conflict. 2Segregationist solutions consist of dividing the warring groups into distinct territories and giving them separate state status or extensive autonomy at the substate level. Population transfers may be used in either case to homogenize the ethnoterritorialunits. In contrast. imegrationist solutions do not create or recognize ethnoterritorial autonomy at any level of government. 'For arguments in favour of partition as a solution to ethnic war. see Kaufmann (1996. 1998.20(7). See also Johnson (2008). Chapman and Roeder (2007). Downes (2004. 20(6), Mearsheimer and Van Evera (1995) and Mearsheimer (1993). 4Laitin (2004) corroborated Sambanis' findings using the Minorities at Risk (MAR) dataset. More recently, Chapman and Roeder (2007) re-ran Sambanis' statistical analysis using his own data after dropping all non-ethnic conflicts from the dataset. Although germane to the debate over ethnic partition, these scholars deal more narrowly with whether partition is the best means of preveming the re-emergence of violent COl!flict in the wake of civil war. In contrast. the present article assesses the value of partition for resolving nationalist conflict-both violent and non-violent-and rebuilding state societies in two cases where it has been most assiduously applied. 5For these and other critiques of partition theory, see Horowitz (1985), Kumar (1997a. 1997b), Schaeffer (1990), Etzionl (1992-93), Gottlieb (1994), Carment and Rowlands (2004) and Fearon (2004). 6For more on the consociational approach to conflict resolution, see Lijphart (1990). See also Horowitz (1985) and McGarry and O'Leary (1993). 7According to the 1991 census. the Bosnian Republic was 17% Croat. 31% Serb and 43% Muslim. HThe US-led Implementation force (IFOR) becmne SFOR (Stabilization Force) in 1996, which was finally replaced by a European force (EUFOR) of 7000 in 2004. 9 Author interview, Halisa Skopljak, Legal Assistant, I-Iuman Rights Dept., OSCE Mission to Bosnia and Hercegovina, San~jevo, Bosnia, 29 May 2006. JO Author interview, Massimo Moratti, Sarajevo, Bosnia, 31 May 2006. IIUNHCR Representation in Bosnia and Hercegovina, Statistics Package, 30 June 2008, available at http://www.unhcr.ba/updatejune08/SP_06_2008.pdf (accessed 7 November 200S). 12lnterview, Halisa Skopljak, 29 May 2006. J3lnterview, Vesna Vignjevic, UNHCR Protection Assistant, Sarajevo, Bosnia, 31 May 2006. 14Entity and local authorities even succeeded in restricting the freedom of movement for ethnic minorities to move about the protectorate. 15Author interview, director of local NGO, Pristina, Kosovo, 18 September 2006. 16 Author interview, ICG analyst, Pristina, Kosovo, 21 September 2006. 17http://www.kosovo.undp.org/publications/publications.asp, as cited in ICG Report (2006). IH Author interview, Kristin Griffith, MercyCorps Mission Director, Pristina. Kosovo, 19 September 2006. Experts on ethnic reintegration stress that property restitution is critical to sustainable minority return. 191 am grateful to Florian Bieber for this observation. 2°Author interview. EU diplomat, Brussels, Belgium. 6 December 2007.

References Beaumont, P. (1999), 'Killings Blamed on KLA', The Observer, 25 July. Belloni, R. (2005), Peacebuilding at the Local Level: Refugee Return to Prijednr, Imemational Peacekeeping (Special Issue), Vo1.l2, No.3, pp.434-447. Bieber, F. (2006), After Dayton, Dayton? The Evolution of an Unpopular Peace, Etlll1opolitics, Vo!.5, No.1, pp.15-31.

The Paradox of Ethnic Partition

95

Bose, S. (2005), The Bosnian State a Decade after Dayton, Illternational Peacekeeping. Vo1.l2, No.3, pp.322-335. Breuilly, J. (1994), Natiollalism alld the State. Chicago: University of Chicago Press. Bunce, V, (1999), Subversive Illstitlltiolls: The Desigll and the Destruction (!( Socialism and the State. Cambridge: Cambridge University Press. Carment, D. and Rowlands, D. (2004), Vengeance and Intervention: Can Third Parties Bring Peace Without Separation?, Security Studies, Vo!.13, No.4, pp.366-393. Cetinyan, R. (2002), Ethnic Bargaining in the Shadow of Third-Party Interventlon.lllfel7lational Organization, Vol.56, No.3, pp.645-677. Chapman, T. and Roeder, P. G. (2007), Partition as a Solution to Wars of Nationalism?, American Political Science Review, Vol.IOI, No.4, pp.677-69I. Cornell, S, (2002), Autonomy as a Source of Conllict: Caucasian Conllicts in Theoretical Perspective, World Politics, Vo!.54, No.2. pp.245-276. Daalder, I. (2000). Getting to Dawon: The Making ,,(America's Bosnia Policy. Washington DC: Brookings Institution Press. Downes, A. B. (2004), The Problem with Negotiated Settlements to Ethnic Civil Wars, Security Studies, Vo1.l3, No.4, pp.230-279. Downes, A. B. (2006), More Borders, Less Conflict: Partition as a Solution to Ethlllc Civil War, SAIS Review, Vo!.26, No.1, pp.49-61. Etzioni, A. (1992-93), The Evils of Self-Determination, Foreign Policy, Vo!.H9 (Winter), pp.21-35. Fearon, J. D. (2004), Separatist Wars, Partition, and World Order. Security Swdies, Vo1.l3, No.4, pp.394-415. Gottlieb, G. (1994), Nations without States, Foreign Affairs, Vo!.73, No.3, pp.100-112. Gumbel, A. (1995), 'Mutual Mistrust Mires Refugees in the Misery of Homelessness'. The Indepelldellf, 6 August. Hale. H. E. (2000), The Parade of Sovereignties: Testing Theories of Secession in the Soviet Setting, British Joumal (d'Political Science, Vo1.30, No.1, pp.31-56. I-Iarris. P. and Reilly, B. (cds) (199S), Democracy alld Deep-Rooted Conflict: Options for Negotiators. Stockholm: International Institute for Democracy and Electoral Assistance. Heimerl, D. (2005), The Return of Refugees and Internally Displaced Persons: From Coercion to Sustainability? Intematiollal Peacekeepillg, Vo1.l2, No.3, pp.377-391. Hoare. M. A. (2004), How Bosllia Armed. London: Saqi Books. 1-loroWilZ, D. (1985), Ethnic Groups ill COI!flict. Berkeley: University of California Press. Human Rights News (200S), Kosovo: Build New State on Rule of Law, 18 February. Available at http:// www.hrw.org/en/news/2008/02/ 14/kosovo-build-new-state-rule-law (accessed 9 November). International Crisis Group Report (2006), Kosovo: The Challenge of Transition, Europe Report, Vo1.170, 17 February. International Crisis Group Report (2007), Ensuring Bosnia's Future: A New International Engagement Strategy, Europe Report, Vo1.80, 15 February. International Crisis Group Report (200S), Kosovo's Fragile Transition, Europe Report, Vol.l96, 25 September. Jenne, E. K. (2004), A Bargaining Theory of Minority Demands: Explaining the Dog That Did Not Bite in 1990s Yugoslavia, llIfematiollal Studies Quarterly, Vo1.48, No.4, pp.729-754. Jenne, E. K. (2007), Ethnic Bargailling: The Paradox (d'Minority Empowermellf. Ithaca: Cornell University Press. Jervis, R. (l97S), Cooperation under the Security Dilemma, World Politics, Vo!.30, No.2, pp.167-214. Johnson, C. (200S), Partitioning to Peace: Sovereignty, Demography and Ethnic Civil Wars, Ime/'llational Security, Vo1.32, No.4, pp. I 40-170. Kaufmann, C. (1996), Possible and Impossible Solutions to Ethnic Civil Wars, International Security, Vo1.20, No.4, pp.136-175. Kaufmann, C. (1998), When All Else Fails, Iml'mational Security, Vo1.23, No.2. pp.120- 156. Kaufmann, C. (2007), An Assessment of the Partition of Cyprus, International Swdi".\' Per.lpectives, VoI.S, No.2. pp.206-223. King. C. (2001), The Benefits of Ethnic War: Understanding Eurasia's Unrecognized States, World Politics, Vo!.53. No.4. pp.524-552.

96

E. K. Jenne

King. I. and Mason. Wo (2006). Peace at allY Price: HoII' the World Failed Kosovo. London: Hurst & Company. Kumar. R. (1997a). DiI·ide alld Fall? Bosllia ill the Allllais (!(Partitioll. London: Verso. Kumar, R. (1997b). The Troubled History of Partition. Foreigll Affairs. Vo1.76. No.1, pp.22-34. Laitin. D. D. (2004). Ethnic Unmixing and Civil War, Secllrity SllIdies. Vo1.13. No.4. pp.350-365. Licklider. R. (1995). The Consequences of Negotiated Settlements in Civil Wars, 1945-1993. Americall Political Sciellce Review. Vo1.89. No.3. pp.68 1-690. Lijpilart. A. (1990), The Power-Sharing Approach. in J. Montville (ed.). COI!tfict alld Peacemakillg ill Mllitiethllic Societies. Lexington, MA: LexlIlgton Books, pp.491-509. Lippman. p, (2004), Big Changes in Bosnia, or Just Another Drama? Washillgtoll Report OIl Middle East Affairs. Vo1.23, No.7, pp.38-39. Lyon. J. M. B. (2000). Will Bosnia Survive Dayton? Cllrrellt HistO/:v. Vo1.99, No.635, pp.1 W-I17. McGarry, J. and O'Leary, B. (eds) (1993), The Politics of Ethnic Conflict Regulation: Case Studies of Protracted Ethnic Connicts, London: Routledge. Mearsheimer, J. J. (1993). 'Shrink Bosnia to Save it', Nell' York Times. 31 March. Mearsheimer, J. J. and Van Evera, S. (1995). When Peace Means War, Nell' Repllblic. 18 December. pp.16-21. 18 December. Nordlinger. E. (1972), COIltfict Reglliarioll ill Divided Societies. Cambridge: Cambridge University Press. Posen. B. R. (1993). The Security Dilemma and Ethnic Conflict, in M. Brown (ed.). Ethllic ClIIltfict alld llllerlIatiollal Secllrity. Princeton: Princeton University Press, pp.I03-124. Pugh. M. (2004). Rubbing Salt into War Wounds: Shadow Economics and Peacebuilding in Bosma and Kosovo, Problems (if Post-COl1l11ll11lislll, Vo1.51, No.3. pp.53-60. Roeder, p, G. (1991), Soviet federalism and Ethnic Mobilization, World Politics. Vo1.43. No.2, pp.196-232. Roeder. Po G. (2007). Where NllIioll-States Come From: IllStitlltiollal Challge ill the Age (!( Natiollalism. Princeton: Princeton University Press. Rothchild, D. and Roeder. P. G. (2005), Power Sharing as an Impediment to Peace and Democracy. in P.G. Roeder and D. Rothchild (eds), SlIstaillable Peace: POlI'er alld Democracv qtier Civil Wars. Ithaca: Cornell University Press, pp.29-50. Sambanis, N. (2000), Partition as a Solution to Ethnic War: An Empirical Critique of the Theoretical Literature, World Politics. Vo1.52, No.4, pp.437-483. Schaeffer. R. (1990). W(lIpaths: The Politics (!( Partitioll. New York: Hill and Wang. Schelling. T. (1960), The Strategy (!( ClIIltfict. Cambridge: Harvard University Press. Shepsle. K. and Rabushka. A. (l972). Politics ill Plllral Societies: A Theon' (if Democratic III stability. Columbus: Charles Merrill. Slezkine, Yo (1994), The USSR as a Communal Apartment. or How a Socialist State Promoted Ethnic Particularism, Sllll·ic Review. Vo1.53. No.2, I1p.414-453. Suny, R. G. (1993), The Revellge (if the Past: Natiollalism. Revollltioll. alld the Collapse of the S(JI'iet Ullioll. Stanford: Stanford University Press. Walter. B. F (1997). The Critical Barrier to Civil War Settiement.llltematiollal Orgalli:,atioll. Vo1.51. No.3, pp.335-364. Woodward. S. (2000), Costly Disinterest: Missed Opportunities for Preventive Diplomacy in Croatia and Bosnia and Herzegovina, 1985-1991, in B.W Jentleson (cd.), OpporllIllities Missed. Opporlllllities Sel:,ed: Prevelllive Diplomacy ill the Post-Cold War World. Lanham: Rowman & Littlefield. pp.133-172. Yannis. A. (20()l), Kosovo under International Administration. SlII'I'ival. Vo1.43, No.2. pp.31-48. Zahar, M.-J. (2005). Power Sharing in Lebanon: foreign Protectors, Domestic Peace, and Democratic failure. in P.G. Roeder and D. Rothchild (cds). SlIstaillable Peace: POII'er alld Democracv qfter Civil Wars. Ithaca: Cornell University Press, I1p.219-240.

State, Society and Separatism in Punjab KRISTIN M. BAKKE Department of Political SCience. Leiden University, Leiden, The Netherlands

Introduction Recent years' policy discussions about conflicts in countries as different as Afghanistan, Iraq, and Sri Lanka show that a hotly debated institutional 'fix' to intrastate conflicts is federalism or decentralization. In academic circles, these institutions have received mixed reviews. While some consider federalism a compromise solution, others see the same institutions as a step towards further conflict and even state disintegration. I This study draws attention to the interaction between institutional design and what the editors of this Special Issue call "uncodified factors", I move away from the debate about the pros and cons of federalism, maintaining that an understanding of federal institutions' divergent effect on intrastate conflicts calls for a consideration of how these institutions are not isolated from the societies they govern. I argue that the effects of policy and fiscal decentralization on separatism are conditioned by any given region's ethnic composition and wealth. I consider separatism to include movements "seeking a separate region within an existing state, as well as those seeking a separate and independent state" (Horowitz, 1981: 169). After situating my argument in the literature, I examine separatism in Punjab in northern India. A Sikh-majority region with a significant Hindu minority. Punjab has twice found itself in a struggle with the central government since India's

98

State. Society and SeparatislII in Punjab

K. M. Bakke

independence in 1947. In the I 960s. the Sikh community engaged in the Punjabi Suba movement. requesting a separate Punjabi-speaking homeland. In the late 1970s the Sikhs began demanding greater autonomy. culminating in a nine-year-Iong violent confrontation with Delhi (1984-93). Case studies are useful for identifying how one independent variable's impact on the dependent variable depends on other variables (e.g. George and Bennett. 2(04). Exploring such interactions is at the heart of this study. As federalism is typically seen as helping the ethnically diverse Indian state hold together. it is instructive to understand the dynamics of one of the country's more troublesome centre-region relations.

99

affecting people's grievances and assessments about the value of staying put in the state. s If the institutions enable separatist leaders to blame the central government for people's everyday problems. the institutions may. perhaps unintentionally. help nurture anti-centre collective action (cf. Goodwin. 200 I). The more intense the grievances, the fewer the perceived benefits of staying put in the federation, and the more clearly the centre is to blame, the more likely people are to pursue separatist goals. Understanding the degree to which the central government is going to be considered unfair or unbeneficial requires that we know something about the society that these institutions govern. I now turn to such state-society interactions as they relate to policy and fiscal decentralization.

Federalism and Separatism The conflict literature remains divided about federalism's "peace-preserving" potential. 2 While some argue that such institutions can accommodate the aspirations of both national and regional actors (e.g. Gurr. 2000; Hechter. 2000; Stepan. 200 I; Bermeo. 2002; Saideman et a/.. 20(2). others see the same institutions as a step towards conflict or state disintegration (e.g. Roeder. 1991; Bunce. 1999; Eaton. 20(6). While important, the discussion about the pros and cons of federalism does not capture the very divergent record of existing federal states in avoiding conflict. Not only is there variation among federal states-consider the peaceful Canadian federation versus the more conflict-ridden Russian federation-but also within them. Russia. for instance. has peaceful relations with most of its 83 regions but does poorly in maintaining peace with Chechnya and a handful of others. Thus, this paper is not asking whether federal states are a better fit than unitary states to contain conflicts. Rather. the question is what are the conditions under which federalism can help contain separatist conllicts (Amoretti and Bermeo. 2004; Hale. 2004; Sambanis and Milanovic, 2004; Bakke and Wibbels. 2006; Brancati. 2(06)? While some scholars have focused on the role of federal institutions irrespective of their cultural context (e.g. Filippov et al .. 2004: 161), others have argued that there is no single federal formula independent of contextual factors (e.g. Erk. 2(07). Acknowledging that there is no cookie-cutter federal solution to conflicts in divided states. I suggest that the degree to which federal institutions can contribute to preserving peace depends on how these institutions respond to characteristics of the societies they govern. The conllict literature posits that ethnicity and wealth are key factors 3 likely to influence any given region or group's separatist aspirations. Hence. an important step in assessing federalism's peace-preserving capacity is to examine the ways in which the workings of specific federal institutions may be affected by these societal traits (Horowitz. 1985: 613-628). The story that explains why regional actors pursue greater autonomy or independence is a dynamic story about the interactioll of societal factors and the institutions that govern the relationship between the centre and the regions of the state. Building on the federalism literature. the institutions I focus on are those that govern policy autonomy (the degree to which regional governments make policy decisions) and fiscal autonomy (the degree to which regional governments fund their own public goods provision).4 These institutions can. depending on how they respond to ethnicity and wealth. shape the perceived legitimacy of the state by

Policy AutonolllY

In the vein of the classic fiscal federalism literature (Tiebout, 1956; Oates. 1972), scholars have argued that federal institutions can help contain ethnic conflicts if they decentralize policy-making capacity over issues central to ethnic minority groups' recognition, such as language, education and culture (e.g. Gurr. 2000; Hechter. 20(0). Local provision of public goods that are valued by only segments of the population is seen as "superior because it increases the likelihood that the right mix of goods will be produced-that mix which is most consistent with the distinctive values of the national group" (Hechter. 2000: 143). Underpinning these arguments is the notion that policy autonomy can contain popular grievances based on ethnic discrimination or enhance the beneHts of remaining part of the state. making it difficult for regional elites to 'play the ethnic card'. If, however, the struggle is in the name of a region and not only an ethnic group, the ethnic composition of that region is likely to condition the kind of policy autonomy that can help preserve peace. Assuming that a separatist challenger is a stronger challenger when the separatists are in control of a regional government, to the degree that ethnicity is insufficient to mobilize a critical mass of supporters. elites are unlikely to mobilize people based on ethnicity alone. As such, 'ethnic' policy autonomy might be insufficient or unimportant for containing separatist tendencies. Whether ethnicity is sufficient to establish a critical mass in the region depends hU'gely on the ethnic group's share of the regional population (cf. Posner, 2(04). The smaller the share of a region's population is a single ethnic minority group (and the more ethnically heterogeneous the region is), the less likely separatist elites in the region are to mobilize people around ethnicity alone. In turn. the less likely 'ethnic' policy autonomy matters for preventing or containing separatism. 6 If. however, a large share of the regional population belongs to the same ethnic minority group, or if mobilization happens in the name of the ethnic group alone, 'ethnic' policy autonomy may mitigate separatist opposition, while the lack thereof does the opposite. Moreover, while it is important to consider whether ethnicity is used to mobilize people, it is equally important to consider how ethnicity is used to mobilize people. Arguments about the subversive nature of federalism along ethnic lines focus on how these institutions politicize an ethnic identity dimension. such as language. Such institutionalization does not, however. explain whether ethnicity encourages moderate separatist demands, such as greater autonomy. or more radical demands,

100

State. Society alld Separatism ill PUlljab

K. M. Bakke

such as independence. If the state is or has been considered a threat to the existence of the ethnic group, the demands raised are likely to take on a radical form. If there is a history of state repression or violence, the ethnic group may not be as concemed about institutions that allow policy autonomy over 'typical' ethnic concerns, such as culture, education, language and religion. Rather, the concern is about checking the centre and ensuring the group's physical safety (cf. Lake and Rothchild, 1996: Weingast, 1998). Thus, it is reasonable to expect that if a history of a threatening centre is key to mobilize an ethnic group, the group may want control over issues that help protect its physical security, including defence, even independence. In such a situation, policy autonomy with respect to typical ethnic concerns within the federation may be an insufficient peace-preserving means. Fiscal Autonomy While rederalism can help contain intrastate cont1icts ir the regions are granted policy autonomy over issues central to a minority group's recognition or sarety, such autonomy means little in the absence or money to spend on decentralized priorities. Regional govemments can either raise revenues on their own to cover most of the expenses for these tasks (through taxes, rees, loans), or they can rely on transfers from the central government (loans, grants). The peace-preserving potential of these options are likely to be conditional on any given region's wealth. On the one hand, if a region is relatively resource-poor, reliance on its own source revenues will probably impair its ability to implement policies rather than empower it, fuelling both grievances related to the lack or public goods provision and a sense that the region is not receiving its fair share from the central government. Thus, in relatively poor regions, it is likely that central transfers are the more peace-preserving option. On the other hand, relatively resource-rich regions can arford to rund public goods provision from their own revenues. The population in such regions is unlikely to want to Ilnance the state's poorer regions and is, therefore, likely to prefer Ilscal autonomy, which also enables policy autonomy.? In general, to the degree that intergovernmental Ilscal relations hurt a region's economy, the public's perceived benellts of staying put in the state are likely to decrease.

State, Society and Separatism in Punjab Scholars have argued that the 'Punjab crisis' was a result of 'relentless' centralization under Indira Gandhi (Brass, 1991: 210), the disintegration of the Congress party (Kohli, 1991), ethnic identity 'hardened' by interactions with the state (Singh, 2(00), or economic conditions (0' Souza, 1985; Singh, 1985). The argument in this article acknowledges that both institutional and societal factors played a role. Its contribution is to theorize and trace the ways in which institutional efrects were conditioned by ethnicity and wealth. The Punjabi Suba agitation of the I 960s was a quest for a separate Punjabi-speaking state in India. x While the rest of India's major ethno-Iinguistic groups were granted their 'own' states within the federation with the States Reorganization Act in 1956, Delhi gave in to the Punjabi Sikhs' demands only in 1966, when Punjab was

101

divided into present-day Haryana and Punjab. The reorganization carved out a majority-Punjabi speaking state in which Sikhs dominated, but it also left several questions unresolved. First, it excluded the city that had served as the capital of greater Punjab, Chandigarh, transforming it into a federally governed territory that would serve as the joint capital of Punjab and Haryana. Secondly, the reorganization stipulated that the centre regulate how much of the water rrom Punjab's rivers were to go to Punjab and its neighbours. In addition, several Punjabi-speaking areas were left out or the new state. These issues, particularly the water question, contributed to the cont1ict that emerged in the 1970s. In 1973, the major Sikh party, the Akali Dal, issued the Anandpur Sahib Resolution (ASR), requesting that India become a 'real' rederation where the states be granted autonomy over all government functions except defence, foreign relations, currency, railways and communications. While the demands based on the ASR could have been met within the Indian federation, Punjab ended up in violent conrrontation with the centre. In the early I 980s, separatist militant groups emerged and began to use Sikh temples as sanctuaries, which resulted in Operation Bluestar-also known as the Golden Temple Massacre. On 6 June 1984, the Indian army stormed the Golden Temple at Amritsar, which is Sikhism's holiest temple, killing hundreds. The attack was not immediately followed by a rise in separatist violence in Punjab, but from 1986, violence escalated in a spiral or insurgency and counter-insurgency campaigns. Since 1992-93, life in Punjab has returned to 'normal', occasionally interrupted by deadly clashes. Policy AlIlollomy The question that guides this section asks how ethnicity has int1uenced policy autonomy as a peace-preserving means in Punjab. Sikh identity is based on religion, memories of past struggles and language (Brass, 1974). Sikhism is a religion that emerged in reaction to Hinduism more than 500 years ago. Until the turn of the nineteenth century, the boundary between the two religions was t1uid, but gradually Sikhs came to emphasize their identity as separate from Hindus. Central to this identity are memories of struggles and past kingdoms tied to the territory of Punjab. Added to these pillars of religion and historical memories is language. Between the 1920s and the I 960s, language and religious identity became congruent in Punjab, as Hindus increasingly declared Hindi as their mother tongue, and Sikhs, Punjabi. Sikh politicians were concemed about their minority status in Punjab even prior to India's independence, so when Delhi in 1953 set up the States Reorganization Commission to consider organizing the country along linguistic lines, the Akalis pushed for a Punjabi-speaking state-a Punjabi Suba-within India, where the Sikhs would constitute a larger share of Punjab's popUlation. At the time, the Sikhs constituted about 35% of Punjab's population, while the majority was Hindus. The Akali Dal's proposal was based on language, although it highlighted that a Punjabi-speaking state would satisfy the Sikh community (Shanna, 1969: 192). The Akalis claimed that their quest was based on language and not religion, but opponents to the Punjabi Suba argued that the demand was a poor disguise for a demand in favour of a majority-Sikh homeland (see Sarhadi, 1970: 198-242). The history of the demand

102

K. M. Bakke

points to its religious basis, mainly the Sikhs' concern about being dominated by another religious group (Brass, 1974: 320; Grewal, 1990: 171-174). The Akalis' demand for a Punjabi-speaking state was not out of the ordinary at the time. Ethnolinguistic groups elsewhere in India mobilized in pursuit of greater autonomy, but the reorganization of the states in 1956 took the steam out of most of these movements.') In contrast, the States Reorganization Commission, which had the partition of India along religious lines fresh in mind and saw the Akalis' call for a Punjabispeaking state as religiously based, rejected the request. The Akalis saw the commission's rejection as an example of the centre's discrimination of the people of Punjab, particularly the Sikhs, emphasizing that of all the 14 ofllcial languages in the constitution, Punjabi was the only one left without a state of its own. As linguistic identity was tied to religious identity, the rejection came to be seen as discrimination of the Sikhs as a religious group as well. The Akali leader Sant Fateh Singh stated that, "If non-Sikhs had owned Punjabi as mother tongue, then the rulers of India would have seen no objection in establishing a Punjabi state" (quoted in Brass, 1974: 325-326). As other attempts at meeting the request for a Punjabi-speaking state failed in the 1950s, by 1960, the Akalis launched a mass agitation in the name of a Punjabi SlIba, focusing almost exclusively on language concerns and avoiding references to religious grievances (Brass, 1974: 321-327; Purl. 1983: 53; Grewal, 1998: 77 - 78). In response to this campaign, the central government passed the Punjab Reorganization Act in 1966, dividing the state into Punjabi-speaking (and Sikh-majority) Punjab and Hindi-speaking (and Hindu-m~ority) Haryana. The reorganization largely met concerns about language and almost put an end to ethno-linguistic conflict with the centre (see Sarhadi, 1970: 462). For the purposes of understanding the conditions under which ethnic policy autonomy can help preserve peace, the Punjabi Suba agitation shows that when the struggle was waged in the name of the Punjabi Sikhs as a group-aimed at carving out a new state for that group-it was about explicitly ethnic reasons, related to the lack of policy autonomy in areas we typically think about as central to an ethnic group's recognition. Indeed, a key concern was education. Worried that younger generations were abandoning Sikhism, some Sikhs thought their traditions could more easily be kept alive in a Sikh-dominated state that allowed the teaching of Sikh history and traditions in public school (Singh, 1992: 39-40). While not formally giving in to a religious demand, the linguistically-based reorganization in 1966 also granted the Sikhs a Sikh-majority state. In contrast to the Punjabi Suba agitation, in the I 970s-80s, the struggle was in the name o I' Punjab as a state. At the time, Punjab was a majority Sikh state (about 60%), but close to half of the popUlation was Hindus. Many of the Sikhs' ethnic demands had been met with the 1966 reorganization, although there were remaining issues, including Punjabi-speaking areas left out of the new state. Moreover, the years following the granting of the Punjabi Suba witnessed a centralizing tendency across India. In 1978, several policy areas were moved from the state list to the concurrent list, including education, giving the centre the power to make decisions regarding textbooks and curricula. Among the Akalis, this move was seen as yet another strategy of Delhi to limit the expression of Sikh culture, history and religion (Govel11ment of Punjab, 1998:

State. Society and SeparatislI/ in Pllnjab

103

875-876; Dhillon, 2004: 98-100). However, these ethnic concerns, while part of the Akalis' struggle, were not at the forefront. The struggle emerging in the 1970s was based primarily on-and justified based on-economic reasons, some of them a consequence of the 1966 reorganization. The Anandpur Sahib Resolution (ASR) of 1973 called for the Indian constitution to be "recast on real federal principles" (quoted in Grover, 1999: 317). The resolution's political goal is the "pre-eminence of the Khalsa", where Khalsa refers to the 'pure' Sikh community, which led some to see the ASR as a call for a separate Sikh state. According to Puri (1983), however, the 1973 resolution in its full format was never on the agenda. Indeed, the policy proposals that were developed based on the ASR in 1978 put socio-economic concel11S at the core. For example, they called for better procurement prices and more subsidies for the state's farmers. Granted, these issues were of greater concel11 to the state's Sikh population than the Hindu popUlation, as the Sikhs dominated the agricultural sector. However, the Akalis, knowing that the state in whose name they were fighting was only barely a Sikh-majority state-and that they would need a coalition partner if they were to rule the state-did not fight a struggle restricted to the Sikhs. While the Akalis' vote base is the Sikh popUlation, its main competitor, the Congress party, also draws a substantial share of the Sikh voters, forcing the Akalis to turn to the Hindu-dominated Jana Sangh or Janata Party as coalition partners (Brass, 1991: 176-177). The Akalis' election manifestos from 1967 played down religious concerns and focused on secular, economic and political programmes (Puri, 1999: 448-450). Religious concel11S were not entirely off the agenda (Samiuddin, 1985: 688-691; Brass, 1991: 193-195). The 1978 Ludhiana resolutions, for instance, called for proper representation of Sikh minorities in other states. Yet, appeasing the Akalis' demands was not about policy autonomy over areas that we typically think about as central to minority group's recognition. In fact, the only concessions that the centre offered in negotiations during 1982-83 were related specifically to the Sikh community, such as amending the constitution to clarify that Sikhs are not Hindus (Brass, 1991: 203), but such concessions turned out to be insufficient. In contrast, the accord negotiated between the Akali Dal and Delhi in 1985 included provisions of the transfer of Chandigarh and the Punjabi-speaking areas in Haryana, as well as a promise that the river water issue would be presented to a tribunal. The accord referred the ASR's call for 'real' federalism to a commission on centre-state relations. To the major faction of the Akali Dal, the accord was promising. Moreover, it was well-received in Punjab, which has led observers to the conclusion that had it been implemented, it would have been a key step towards stemming the conllict (Kohli, 1991: 368-369).10 The point here is that this accord was not only about ethnic concerns; it was also about centre-state relations more generally, Unlike the Akalis, the militant groups that emerged in the I 980s waged a battle primarily in the name of the Sikhs and not Punjab as a whole. In the early 1980s, the militant leader Sant Jar'nail Singh Bhindranwale voiced concerns about the fate of Sikh religious traditions, while some of the militant groups that emerged later in the decade called for an independent Sikh state, Khalistan, and implementation of a religious code of conduct (Tully and Jacob, 1985; Judge, 20(5). While the militants were also concerned with economic hardship (Telford, 1992), their struggle was

104

K. M. Bakke

explicitly about the Sikhs. Because Sikhism is based on equality, the plight of poor Sikh farmers came to be linked to religious autonomy and independence (Pettigrew, 1995). Indeed, both the Punjabi SII!Ja movement and the militant movement in the 1980s suggest that where the struggle was waged in the name of the ethnic group, 'ethnic' policy autonomy was a key concern (but, as I turn to next, the militants sought a more radical form of autonomy than the SuiJa activists-an independent Sikh state). In contrast, where, as in the case of the Akalis' campaign, the struggle was waged in the name of Punjab as a state, which was only barely a Sikh-majority state, the struggle was not about policy autonomy that focused exclusively on the Sikhs. The separatist struggle in Punjab also shows how the ties around which ethnic mobilization occurs influence separatist demands and, as such, policy aL\lonomy as peace-preserving. While both the Punjabi SlIiJa agitation in the I 960s and the militants' struggle in the 1980s focused on 'ethnic' policy autonomy, a key difference was that the militants in the 1980s also called for an independent state. In the SII/Ja agitation, mobilization happened along the dimensions of religion and language. The arguments were about protecting the status of the Punjabi-speakers and Sikhs within the federation. The militant groups that emerged in the mid- to late 1980s, however, sought to mobilize people based on a threatening centre, which enabled them to mobilize people in support of an independent Sikh state-a more difficult demand to negotiate than greater policy autonomy within the federation. While Punjab's population had not widely supported the idea of Khalistan prior to Delhi's Operation B1uestar in 1984, the operation boosted the very forces it was supposed to suppress. According to the leader of a militant group: Prior to June 1984 we used to talk about ... the right to self-determination within India. After 1984 we needed our own independent home. The government that could kill hundreds, send thousands to prison, rape our women and generally humiliate our people, there could never be a compromise with them! (quoted in Pettigrew, 1995: 149). Operation Bluestar set off a series of violent events, among them Prime Minister Indira Gandhi's assassination by Sikh members of her own bodyguard, which spurred anti-Sikh riots in Delhi, killing thousands. Towards the end of 1984, new militant groups began to emerge, some of which called for independence. Notably, the militant groups did not officially state that they wanted to create an independent Sikh state, Khalistan, until 1986, after the attack on the Golden Temple." As the militancy in Punjab was rising, the Punjab police under the command of K. P. S. Gill and the central security forces escalated the counter-insurgency campaign, which had a radicalizing effect (Pettigrew, 1995; Kumar et al., 2(03). Indeed, after the 1984 attack on the Golden Temple the centre could more credibly be 'blamed' for threatening the physical safety of the Sikhs, which boosted radical demands. Observers of Punjab politics have noted that the only way that the militants could have operated was if the popUlation provided support (others, however, have pointed out that such popular support may rather have been the result of a population fearing for their lives if they did not co-operate). In this environment, accommodation through 'ethnic' policy autonomy within the federation was less relevant than in the

State, Society alld Separatism ill Plllljab

105

Punjabi Suba agitation. Indeed, key to the centre's justification for using force rather than accommodation in the first place, in 1984, was what it considered "the maturing of a secessionist and anti-national movement" (quoted in Grover, 1999: 323). If the militant movement had not been secessionist prior to Operation Bluestar, it became so afterwards, making accommodation more difficult. Fiscal Allfollomy

TUl11ing to the intergovernmental fiscal system, I suggest that while poor regions are likely to welcome fiscal transfers, rich regions are likely to want to hold on to their riches and, as such, prefer flscal autonomy. Punjab has since the I 960s been one of India's wealthiest states. In terms of revenue expenditures covered by own-source revenues, it has also been one of India's most fiscally autonomous states (Bagchi, 2(03). Given my proposition, this is a peace-preserving combination. So why did separatist demands, many economically motivated, resurface in the 1970s? While the Akali Dal, as I turn to next, criticized the centre for jeopardizing the region's wealth, this allegation did not create univocal support for the party. In the state's flrst three elections after the 1966 reorganization, it gained less than 30% of the popular vote. The party's main vote base has always been the well-off landowning Sikhs, while lower-caste Sikhs have voted for the Congress pmty. The struggle fought by the Akalis has primarily reflected the concerns of the landowning Sikhs. Although Punjab in the 1970s was more flscally autonomous than any other Indian state, in the sense that it covered a large share of its own expenditures, the Akalis pushed for greater autonomy over decision making over taxation and spending. In 1978, Gurcharan Singh Tohra, an Akali leader, called for more taxation power to the states. '2 The idea was that if the state had more taxation powers, it could better manage its economy (see also Government of Punjab, 1998: 910-918, 952-955). While such demands can be considered a case of federalism encouraging separatism, a closer look at the demands and the centre's role in Punjab's economy suggests that the Akalis' concerns reflected aspects of centre-state relations that the popUlation of a rich region might find unfavourable. In the 1970s, the Akali Dal accused the centre of using transfers to control the states. In India, transfers are funnelled to the states through three channels, with varying strings attached. Statutory transfers based on the Finance Commission's recommendations are formula-based and non-discretionary. In contrast, plan transfers from the Planning Commission are based on the discretion of the centre, although a formula was introduced in 1969. Similarly, assistance to the states for implementation of centrally sponsored programmes comes in the form of discretionary transfers through the Planning Commission. Typically, these programmes concern policy areas on the state list, and implementation is required. The Akalis reacted against the ways in which the centre could use transfers to impede on Punjab's hard-won autonomy. The Akali leader Tohra argued that the Planning Commission "has reduced the states to a beggar-status by doling out grants-in-aid at the discretion of the centre". 13 The composition of transfers to the states in the 1970s and 1980s suggests that these complaints were not unfounded. In the late 1970s, a significant portion of central resources going to Punjab were discretionary grants and loans,

106

K. M. Bakke

which are the kinds of transfers with most strings attached (e.g. Rao and Singh, 2001).14 While regional politicians in other states, too, reacted against such strings, to the Akalis, it looked like no coincidence that the centre was tightening its grip on the state just when they had gained a state of their own. The Punjab case also demonstrates that concerns about wealth go beyond the workings of the IIscal system, encompassing other features of centre-region relations that affect a state's economy. In Punjab, central control of river waters and agricultural prices, as well as the lack of central investments in industries, precipitated the Akalis' demands that Punjab was not receiving its fair share of revenues from the centre. To understand why these issues mattered to Punjab's popUlation, one must consider the state's economy. With the partition of India in 1947, the most fertile areas of greater Punjab went to Pakistan, and Indian Punjab became a food-deficit state. However, Punjab rapidly turned into India's most prosperous state and, by the I960s, it topped the list when it came to per capita income and placed towards the bottom in terms of poverty. This growth would not have taken place without a state government dedicated to agricultural development, but it was also 1~lcilitated by the central government's willingness to provide grants to development of infrastructure. In the IIrst two flve-year plans (1951-61), the highest spending priority in Punjab was power and irrigation, helped by large grants from the centre (Bhalla, 1995). These initiatives helped pave the way for the Green Revolution of the I960s, which was a period of economic growth driven by agricultural developments that transformed Punjab into India's bread-basket. The point here, as it relates to the functioning of the federation, is that in the immediate post-independence period, Punjab was in poor shape, but the central government helped transform it into one of India's success stories. Once Punjab became prosperous, segments of the population came to see the centre as detrimental to the state's financial well-being. Central control of river waters is a key reason. The 1966 reorganization stipulated that both water and hydro-power from three of Punjab's live rivers should be controlled by a board of representatives from each of the states affected by these rivers but under central government supervision. Water, in'igation and hydro-electric power greatly affect the states' agriculture and industries. These matters are the jurisdiction of the states, although the constitution includes a provision for central regulation of inter-state rivers if it is in the 'public interest'. For the centre to reserve control over Punjab's rivers was seen as a violation of the new state's autonomy. According to the Sikh historian G. S. Dhillon (2004: 60-61), the sharing of river waters limited canal irrigation and forced farmers to opt for the more expensive tube-wells. About three-quarters of Punjab's river waters are allocated to other states (Kumar, 2(05) and, while it is dimcult to say whether this water rightfully belongs to Punjab, the Akalis have perceived it that way. According to one Akali politician in 1982, the sharing of river water meant that Punjab's wealth was "gifted away to others at the cost of our economy" (quoted in Jafar, 1988: 29). This sentiment was not just political rhetoric. Survey data from villages in 1988 indicate that among people's key economic concerns, particularly among small and medium farmers, was canal irrigation. 15 The river water sharing created a sense that Punjab was deprived of its wealth and treated unfairly. In Dhillon's (2004: 63)

State, Society and SeparatislI/ in Punjab

107

opinion, "there has been a calculated plan to denude Punjab and its people of its natural wealth and thereby seriously to jeopardize the economic, industrial and agricultural destiny of the State". Similarly, because agriculture was so important to Punjab's economy, it was highly dependent on government subsidies, low input costs and high output prices on farm produce, which were largely determined by the centre. There was a sense among the farmers in Punjab that the government's procurement prices were set too low (Grewal, 1990: 212-214), particularly when the Green Revolution began to level in the late 1970s. While Punjabi fanners, on average, were not poor, their wages declined in the early I 970s. Moreover, prices for rice and wheat gradually declined from 1960 to the end of the 1970s (McGuirk and Mundlak, 1991: 40). Many of the ASR demands spelled out these concerns of the state's farming population. For example, the third resolution from 1978 calls upon the Indian government to "bring about parity between the prices of the agricultural produce and that of the industrial raw materials so that the discrimination against such states that lack these materials may be removed". Again, 16 survey data reveal that these were sentiments that occupied Punjabi villagers. Punjab's Movement Against State Repression has sought to demonstrate that the large number of rural suicides in Punjab since the mid-1980s is a direct consequence of central policies harmful to the state's fanners, encouraging calls for greater autonomy over policies that affect the state's economy. Likewise, while Punjab's agricultural sector was growing until the early 1980s, the industrial sector lagged behind, fuelling the notion that Punjab was losing out. While Punjab in the mid-1960s topped the list of India's states when it came to per capita income and social development, it was bypassed by several states in terms of industrial development (Grewal, 1975: 64). Punjab had the capital to develop industries, but the central government was unwilling to provide industrial licences to the state (Gujral, 1985: 48; Bhalla, 1995: 103). Moreover, data from 1976 and 1981 show that of the centre's investments in the states, Punjab received less than 2% (Maini, 2004: 231). Both the Akalis' 1978 resolutions and their agitation campaign in 1982 called for the centre to invest in Punjab's industries. While the Akali supporters, primarily the landowning Sikhs, saw the lack of central investments in Punjab's industries as a justiflcation for greater autonomy, the same trend also boosted recruitment into the militant movement. Due to the small size of Punjab's industrial sector, it did not serve as an alternative source of income for the losers of the Green Revolution, the middle or small peasants. Nor did it provide employment chances for the increasingly educated Sikh youth from the rurallandowning class. These groups both formed the recruitment base of the militant movement that emerged in the I 980s (Telford, 1992; Pettigrew, 1995; but see also Puri et aI., 1999). Not only were they disillusioned with the central government, but also the Akalis, for doing too little, creating tensions within Punjab. Let me return to the proposition about the peace-preserving potential of flscal autonomy in rich versus poor regions. Consistent with expectations, the Sikh struggle in the 1950s-60s was not based on reasons related to wealth and flscal relations or about seceding from the Indian union, as Punjab, in a dimcult position after the partition, beneflted from central grants and loans to rehabilitation and development of the state's infrastructure. Similarly, nor in the 1970s-80s was the Akalis' struggle

108

K. M. Bakke

about seceding from India. Living in a rich region with significant fiscal autonomy, Punjab's population, while dissatislled with central intrusions, had reasons to stay put in the rederation. However, the case suggests that concerns about wealth go beyond Ilscal arrangements. The Akalis' concern with river waters was essentially about a rich region in which a relatively rich group, the landowning Sikhs, considered the centre's actions to diminish their wealth. These are dynamics consistent with my expectations about what a rich region might want. As the Green Revolution began to level in the late 1970s, the large landowners worried about their wealth. Indeed, the 1978 resolutions reveal a concern not only with greater autonomy but with getting more from the centre, including industrial investments. Indeed, the case shows that it is diflicult to talk about a state's wealth as ir it were unitary. In Punjab, the losers of the Green Revolution saw their grievances as resting with the centre in terms of, ror example, insuflicient subsidies for agricultural inputs, which is consistent with how I would expect a poor region to act. The less well-to-do Sikhs also saw their grievances as resting with the Akalis' inability to represent their interests, creating conllicts within the state that complicated its interactions with the centre (Bakke, 2(08). The effects of such internal divisions within a region merit further investigation.

Conflict De-escalation The common story about the 'Punjab crisis' is that it came to an end thanks to a war-weary population, an increasingly unpopular militant movement, and a successful counter-insurgency campaign. According to some observers in Punjab, the conllict's underlying causes remain unresolved, which is why we still see a rew pro-Khalistan parties. While there is much truth to this view, it is not the case that nothing changed. Since the I 990s, both Punjab's society and rederal institutions have changed in ways that have reduced the potential ror centre-region conllict. The ri ver water question has been on the agenda in Punjab since 1966 and, all along, the arguments of the Punjabi governments have been the same as when the dispute first emerged. More generally, Punjab's farmers are racing economic difllculties, contributing to high rarmers' suicide rates. Many of Punjab's rarmers are heavily in debt, which they attribute primarily to low (government-set) prices on agricultural produce (wheat and rice) and high costs on agricultural inputs (Jodhka, 2006; Singh et ai., 2(08). While these are concerns similar to those voiced in the 1970s, it is unclear that the population considers the centre responsible. Survey data rrom the 2002 state elections show that among the population aware or the state's increasing rate of rarmers' suicides, the bulk of the blame was put on the state government alone, or on both the central and state governments. I? Punjab is still among India's better-orr states, although the rural popUlation is becoming impoverished (Centre ror Research in Rural and Industrial Development, 2(02). In the 1990s, liberalization caused Punjab's (and other north Indian states') growth rates to slow down, while the trend in India as a whole was the opposite (Singh and Singh, 2(02). Whereas Punjab's f~trIners have been concerned about their economic lot since the 1970s, by the 1990s Punjab was racing an agricultural crisis (e.g. Chand, 1999). Thus, to a cel1ain extent, Punjab now needs the centre.

State, Society and Separatism in PllI(jab

109

The trend in intergovernmental Ilscal relations is that the Punjab state government is runding a smaller share or current expenditures rrom own-revenue receipts (Bagchi, 2(03). And, across India, the volume and number of centrally sponsored schemes remain high and bypass the state budgets (Garg, 2(06). Recall that the Akalis considered transrers, particularly transrers with strings attached, a major problem in the 1970s. Today, the Akali Dal is no longer emphasizing an anti-centre stand and has moved towards rhetoric of "cooperative rederalism" (Kumar, 20(5). In terms of ethnicity and 'ethnic' policy autonomy, Punjab remains a Sikh-majority state with a large Hindu minority, and the Akalis depend on coalition partners to rule the state. Indeed, in 1997 and 2007, the Akalis won the state elections on a ticket joint with the Hindu-dominated Bharatiya Janata Party and, in both elections, the Akali Dal sought to portray itselr as a Punjabi rather than Sikh party (e.g. Kumar, 2(07). While the central government has continued to make inroads into education, which is a statelevel responsibility, this issue has not been actively pursued by the Akalis. As for policy autonomy more generally, in 1988, the Sarkari a Commission on centre-state relations made several recommendations meant to solve centre-state conllicts and, by 2005, most were at some stage of implementation (Mathew, 2(06). While this process has been slow, it moves towards the kind of 'real' federalism that the Akalis called ror in the I 970s. There are still voices in Punjab that call for the creation of Khalistan, seeking to mobilize the population around the idea of a threatening centre. They use the Delhi riots of 1984 and human rights abuses in the counter-insurgency campaign in the late 1980s as justi Ilcation. While the mistreatment or Sikhs in these events is well documented (Kumar et al., 2(03), such fear-appeals no longer appear to aid mobilization because the militant groups in the 1980s alienated the very population in whose name they were lighting. In addition, the central government has begun to offer compensation to the victims or the 1984 Delhi riots. These developments have contributed to taking the steam out of rear-based appeals and calls for Khalistan, which makes accommodation within the federation a possibility. Thus, while the centre's counter-insurgency campaign and the war-weariness or Punjab's population certainly contributed to the demise or the 'Punjab crisis', there have been gradual changes in the state's society as well as the state's relationship with the centre that have helped reduce the potential ror centre-region conllict.

Conclusion Federalism as a means ror managing intrastate conflicts is a hotly debated policy alternative. Similarly, in the conllict literature, federalism has received mixed reviews. This paper seeks to overcome the pro and con nature or these debates by rocusing on the conditions under which rederalism can help preserve peace. I maintain that an understanding or rederal institutions' peace-preserving potential calls for close attention to the ways in which institutions interact with traits of the societies they govern. In particular, the efrects of policy decentralization depend on any given region's ethnic make-up, and the efrects of fiscal decentralization depend on any given region's wealth. The case study or separatism in Punjab also suggests, however, that intrastate struggles involve a number or actors and motives. Indeed, part of the story

110

K. M. Bakke

that explains why Punjab ended up in a violent conflict with the centre is about divisions within the region, which complicated centre-region bargaining. Thus, not only are centre-region relations important. but intemal struggles in the region itself are important as well. This is a research agenda wOlth further exploration (see Cunningham, 2008), but it, too. points to the importance of focusing on how states and societies interact. For policy makers, the key lesson of this research is the impOltance of in-depth knowledge about the societies in which hopes for peace are pinned on federalism. According to then-Chairman of the US Senate Foreign Relations Committee, Joe Biden, "The last best chance for a stable Iraq is federalism-giving the warring factions breathing room in regions with control over the fabric of their daily lives". t8 The "breathing room" needed, though, may vary from region to region. depending on their ethnic and economic characteristics.

Acknowledgement The author would like to thank the politicians, public officials, journalists and scholars in Amritsar, Chandigarh and Delhi who talked to her about federalism and separatism in Punjab. Harish Puri and Pramod Kumar were of particular help. In addition, thanks go to Joel Migdal, Erik Wibbels, Steve Hanson, Mike Ward and Paul Brass for helpful comments on earlier versions or parts of the argument presented in this article. The editors of this issue, Jan Erk and Lawrence Anderson, and two anonymous reviewers, have also provided useful feedback. The research for this article was supported by grants from the National Science Foundation (SES-0517153), the ChI'. Michelsen Institute (Norway), and the University of Washington's Graduate School.

State. Society alld SeparatislI/ ill P1Injab

I II

7Rich regions may also want central transfers (Trelsman, 2(01), but because such regions are likely to be financing poor provinces through transfers, It is reasonable to expect these regions to favour decentralized taxation as a means to fund region-level tasks. Hln India, the subunits below the federal level are called states. 9The 1956 reorganization meant that most states came to have a dominant regional language, typically hased on the spoken language of the state's nl<~jority ethnic group. which became the language in public administration and primary schools. As primary education was a state-level responsihility, linguistically hased states meant that India's major ethnic groups gained increased autonomy in that policy area. In general. the reorganization granted to the major ethnic groups more autonomy across all policy areas on the constitution's State List, which is the list of the states' responsihilites in the federation, as they now had their 'own' states. I"The accord was. however, never fully implemented (Kohli, 1991: 370-376). II Bhindranwale, who was killed in Operation Bluestar, never stated that he wanted an independent Sikh state, but in April 1986. a gathering of militant organizations officially adopted a Declaration of Khalistan. 12See Tohra's speech 'Fedeml Polity-The Question of Autonomy: Its Meaning, Necessity and Framework' (Singh. 1994: 146-157). I.1 lbid. See also Dhami (1975: 30). I"While. 111 the late I 970s. discretionary grants and loans on average made up around 65% of grants and loans going to the Indian states. the share in Punjab was 78% in 1977.90% In 1978 and 83% in 1979 (based on data from the Reserve Bank of India. provided by Jonathan Rodden). "Survey of 234 households, carried out by Pramod Kumar, Institute for Development and Communication. Chandigarh. 1('1bid. 17 Akali voters typically blamed the centre. while Congress voters blamed the state government (Kumar and Kumar. 2(02). IHQuoted in 'Senators: A Federal System is Last Best Chance for Iraq', States Neil's Service, 7 June 2007.

References Notes IArguments about federalism's effect on intrastate conflict may be applied equally well to states with decentralized governance structures but which are not considered federal. On definitions. see Treisman (2007: 21-27). ~he term "peace-preserving federalism" was coined by Benneo (2002). 'While the mechanisms vary (e.g. Gagnon. 1994/95: Lake and Rothchild. 1996: Brown, 2000: Kaufman. 2001: Petersen, 2(02), ethnic identity is typically hypothesized to help solve the collective action problem. especially when ethnic groups are territorially concentrated (Toft. 2(03). Others posit that it is not identities but access to resources and wealth that cause conflict. either through grievances or concerns on the part of the poorer (e.g. Muller and Seligson, 1987: Gurr, 2(00) or richer (e.g. Bolton and Roland. 1997: Alesina et al.. 2000: Sambanis and Milanovic. 2(04) groups and regions in the state. "Much research has emphasized the role of political parties in federations as well (e.g. Riker. 1964: Stepan, 2001: Filippov et al.. 2004; Brancati, 2(06). Elsewhere, I focus on the role of political (party) ties in the case of Punjab (Bakke, 2(08). 5Federalism can also strengthen subnational identities and resources for mobilization. which are important arguments among those opposed to such institutional arrangements. Such arguments assume that there are grievances and rationales for mobilization. and the focus in this study is how federalism affects those motives. 6 1n regions without democratic competition, the importance of forming a critical mass might be less important. but even 111 such regions. the larger the share of the population supporting the separatist cause, the stronger the challenge to the state. Thus. even in non-democratic regions. separatist elites may mobilize people around non-ethnic issues if the ethnic group only makes up a small share of the population.

Alesina. A .• Spolaore, E. and Wacziarg. R. (2000). Economic Integration and Political Disintegration. The Americall Ecollomic Review. Vo1.90. No.5. pp. 1276-1 297. Amoretti. U. M. and Benneo. N. (eds) (2004). Federalism alld Territorial Cleal'((ges. Baltimore: Johns Hopkins University Press. Bagchi. A. (2003). Rethinking Federalism: Changing Power Relations between the Center and the States, Pubiius. Vo1.33. No.4. pp.21-43. Bakke. K. M. (2008). The Turn to Violence in Chechnya and Punjab: Separatist Struggles in Decentralized States. Paper presented at the Annual Meeting of the American Political Science Association. Boston. Bakke. K. M. and Wihbels. E. (2006). Diversity. Disparity. and Civil Conflict 111 Federal States. World Politics, Vo1.59. No.1. Pl'. I-50. Bermeo. N. (2002), The Import of Institutions. joul'llal of Democracy. Vo1.13. No.2. pp.96-11 O. Bhalla. G. S. (1995). Agricultural Growth and Industrial Development in Punjab. in J.W. Mellor (ed.). Agriculture 0/1 the Road to Illdustrializatioll. Baltimore: Johns Hopkins UllIverslty Press. pp.67-112. Bolton. P. and Roland. G. (1997). The Breakup of Nations: A Political Economy Analysis. Quarterly Joul'llal (!f Ecollomics, V01.112. No.4, pp.1 057-1090. Brancati. D. (2006). Decentralization: Fueling the Fire or Dampening the Flames of Ethlllc Conflict or Secessionism. Illfel'llatiollal Orgallbaioll, Vol.60, No.3. pp.65 1-685. Brass, P. (1974).l..allguage, Religioll alld Politics ill North Illdia. New York: Camhridge University Press. Brass, P. (1991). Etllllicity alld Natiollalism: Them), alld COlllparisoll. New Delhi: Sage. Brown. R. (2000). Social Identity Theory: Past Achievements. Current Problems and Future Challenges. Europeall joumal (If Social Psvc/lOlogy. Vo1.30. No.6, pp.745-778. Bunce, V. (1999). Sui>versive Illstitutiolls: The Desigll alld the Destructioll (If Socialislll alld the State. New York: Camhridge University Press.

112

K. M. Bakke

Centre for Research in Rural and Industrial Development (2002), Punjah Del'eioplllelll Report. Government of India: Planning Commission. Availahle at http://planningcommission.nic.in/plans/stateplan/sdr_ punjah/sp_sdrpunh.htm (accessed 25 Fehruary 20(8). Chand, R. (1999). Emerging Crisis in Punjab Agriculture. Econolllic alld Polilical WeekiI'. 27 March. Cunningham. K. G. (2008). Strategic Accommodation. Iowa State University: Department of Political Science. unpublished manuscript. Dhami. M. S. (1975). Millorill' Leaders IlIIage '!( Ihe Illdiall Polilical Sl'slelll: All ExploralOlT Srudy ,,( Ihe Auirudes ,!(Akali Leaders. New Delhi: Sterling. Dhillon. G. S. (2004). Illdia COlllllliiS Suicide (3rd edn). Chandigarh: Singh and Singh. D·Souza. V. S. (1985). Economy. Caste. Religion. and Population Distribution: An Analysis of Communal Tension in Punjab. in A. Samiuddin (cd.). The PIII(jah Crisis: Challellge and Respollse. Delhi: Mittal. pp.52-74. Eaton. K. (2006). The Downside of Decentralization: Armed Clientelism in Colombia. SeClirilV Siudies. Vo1.l5. No.4. pp.I-30. Erk. J. (2007). Explainillg Federalislll: Srale. Socielv alld COllgruellce ill Auslria. BelgiulII. Callada. GerlllallY alld SWllzer/alld. London: Routledge. Filippov. M .. Ordeshook. P. C. and Shvetsova. O. (2004). Desigllillg Federalislll: A Theon' ,,( Se(t~ SlIsraillah/e Federal Illsfiluriolls. New York: Cambridge University Press. Gagnon. V. P. (1994/95). Ethnic Nationalism and International Conflict: The Case of Serbia. IlIIe/'l/ariOllll1 Secllrity. Vo1.l9. No.3. pp.130-166. Garg. S. C. (2006). Transformation of Central Grants to States: Growing Conditionality and Bypassing State Budgets. Ecollolllic alld Polilical Weeklv. 2 December. George. A. L. and Bennett. A. (2004). Case SllIdies alld TheO/y Deve/oplllelll ill IIIl' Social Sciellces. Cambridge. MA: MIT Press. Goodwin. J. (2001). No Olher Way Oul: Slares alld Rel'ollliiollarv Movelllellis. 1945-1992. New York: Cambridge University Press. Government of Punjab (1998) [19881. Memorandum. Government of India. COllllllissioll 011 Celllre-SIaie Relariolls: Reporl. Parr If. New Delhi: Government of India Press. pp.863-985. Grewal. B. S. (1975). Cellire-Srale Fillallcial Relaliolls ill Illdia. Patiala: Punjab University Press. Grewal. J. S. (1990). The Sikhs ,,( Ihe Pllnjab. New York: Cambridge University Press. Grewal. J. S. (1998). Sikh Identity. the Akalis and Khalistan. in J. S. Grewal and I. Banga (cds). III Plllljab ill Prmperily alld Violellce 1947-1997. New Delhi: K.K. Publishers. pp.65-1 03. Grover. V. (ed.). (1999). The Srorv ,!(Pulljab Yeslerday alld Todav. Vol.3 (2nd edn). New Delhi: Deep and Deep. Gujral. l. K. (1985). The Economic Dimension. in A. Singh (cd.). Pllnjab ill Illdiall Politics: Issues alld Trellds. Delhi: Ajanta. pp.42-53. Gurr. T. R. (2000). Peoples verslls Srales. Washington. DC: United States Institute of Peace Press. Hale. H. (2004). Divided We Stand: InstitutIOnal Sources of Ethnofederal State Survival and Collapse. World Polilics. Vo1.56. No.2. pp.165-193. Hechter. M. (2000). COlllaillillg Natiollalislll. New York: Oxford Universlly Press. Horowitz. D. (1981). Patterns of Ethnic Separatism. COlllparalive SllIdies ill Sociely alld flisrory. Vo1.23. No.2. pp.165-195. Horowitz. D. (1985). Elllllic Groups III COlljlici. Berkeley: University of California Press. Jal[lr. G. (1988), The Sikh Volcallo. New Delhi: Atlantic Publishers & Distrihutors. Jodhka. S. S. (2006). Beyond 'Crises': Rethinking Contemporary Punjab Agriculture. Ecollomic alld Political Weekly. 22 April. Judge. P. S. (2005). Religioll. Idelllily alld Nariolll/(}od: The Sikh Miliralll MOl'emelll. New Delhi: Rawa!. Kaufman. S. J. (200 I). Model'll Halreds: The Symbolic Polilics ,!IElllllic War. Ithaca. NY: Cornell University Press. Kohli. A. (1991). Democracy alld Discolllelll: IlIdia's Growillg Crisis (!( Govel'llabilily. Cambridge: Cambridge University Press. Kumar. A. (2005). Electoral Politics in Punjab: 1966-2004 . ./oll/'llal '!I' PIIIIJab SlIIdies. Vo1.12. No.1. pp.III-128. Kumar. A. (2007). Punjab Elections: Exploring the Verdict. Ecollomic alld Political Weekly. 2 June.

State, Society alld Separatism ill PIlI{jab

113

Kumar. A. and Kumar. S. (2002). Assembly ElectIOns: Decline of Identity Politics. Ecollomic alld Political Weekly. 13 April. Kumar. R. N .• Singh. A .. Agrwaal. A. and Kaur. J. (2003). Redllced ro Ashes: The IlIsllrgencv alld HlIllIllll Rigills ill Punjab. Kathmandu: South Asia Forum for Human Rights. Lake. D. A. and Rothchild. D. (1996). Containing Fear: The Origins and Management of Ethnic Conllict. Illtematiollal Security. Vol.21. No.2. pp.41-76. Maini. G. (2004). The Gialll Awakells: PIIIIJab Illdustry alld Groll'Ih. New Deihl: India Research Press. Mathew. G. (2006). Repuhlic of India. in A. Majeed. R.L. Walls and D.M. Brown (cds). Dislribulioll '!( POII'ers alld Re.lpollsibilities ill Federal COlllliries. Montreal: McGill-Queen's University Press. pp.155-180. McGuirk. A. and Mundlak. y, (1991 l. Illcelllives alld COllsiraillls ill Ihe Transj(lI'Ilullio/l '!( Plllljab AgriC/t/Illre. Washington. DC: International Food Policy Research Institute. Research Report No.87. Muller. E. N. and Seligson. M. A. (1987). Inequality and Insurgency. Amencan Polilical Sciellce Rel'iel!'. Vo1.81. No.2. pp.425--452. Oates. W. (1972). Fiscal Federalislll. New York: I-Iarcour Brace JovanovIch. Petersen. R. D. (2002). Underswlldillg Elllllic Violellce: Fear. Harred, and Reselllllll'lII ill Tl!'ellllelh-Celllurl' ' Easlerll Europe. New York: Cambridge University Press. PeHigrew. J. (1995). The Sikhs (d'Ihe Plllljab: Unheard VOIces ,!(SIate alld Gllerilla Violellce. London: Zed Books. Posner. D. N. (2004). The Political Salience of Cultural Difference: Why Chewas and Tumbukas Are Allies in Zambia and Adversaries in MalaWI. Alllericall Political Sciellce Rel'ieIJ'. Vo1.98. No.4. pp.529-545. Puri. H. (1983). Anandpur Sahib Resolution: What Do the Akalis Really Want'? IlIIerdi.l·clplille. Vo1.15. No.2. pp.47-63. Puri. H. (1999). Akali Politics: Emerging Compulsions. in V. Grover (ed.). The Storv '!I' Plllljab Yesterday alld Todal'. Vol.2 (2nd edn). New Delhi: Deep and Deep. Puri. 1-1 •• Judge. P. S. and Sekhon. J. S. (1999). Terrorislll ill Plllljab: Ulldersllllulillg Grassroots Reality. New Delhi: I-Iar Anand. Rao. M. G. and Singh. N. (200 I). Tile Political ECOIIOIllY '!!' Celller-SWIe Fiscal Trallsfers ill Illdia. Stanford University: Center for Research on Economic Development and Policy Reform. Working paper No. I 07. Riker. W. (1964). Federalislll: Ongm. Operalioll. Sigllificallce. Boston: Lillle. Brown and Company. Roeder. P. (1991). Soviet Federalism and Ethnic Mohilization. World Polilfcs. Vo1.43. No.2. pp. I 96-232. Saideman. S .• Lanoue. D .. Campenni. M. and Stanton. S. (2002). Democratization. PolitIcal Institullons. and Ethnic Conflict: A Pooled Time-Series Analysis. 1985-1998. COlllparative Polilical SIt/dies. Vo1.35. No.l.pp.I03-129. Samballls. N. and Milanovic. B. (2004). Explaining the Demand lilr Sovereignty. Yale University: Department of Political Science. unpublished manuscript. San1Juddin. A. (1985). The PUlljab Crisis: Chal/ellge alld Respollse. Deihl: MiHal. Sarhadi. A. S. (1970). Plllljabi Suba: The Slory (!(tlw SIrliggle. Delhi: U.c. Kapur & Sons. Sharma. P. K. (1969). Polilical Aspecis '!!'Swtes Reorgallbllioll ill Illdia. New Deihl: Mohuni. Singh. G. (1985). The Socio-Economic Bases of the Punjab Crisis. III A. Samiuddin (ed.i. The Plllljab Crisis: Challellge alld Respollse. Delhi: MiHal. pp.75-89. Singh. G. (1994). Politics ,!(Sikh HOlllelalld. Delhi: Ajanta. Singh. G. (2000), Elllllic GJ/(flici ill Illdia: A Case Siudy ,d·Pulljab. London: MacMillan Press. Singh. K. (1992). My Bleeding PUI(jab. New Delhi: UBS Publishers' Distributors. Singh. L. and Singh. S. (2002), DeceleratIOn of Econon1Jc Growth in Punjab: Evidence. Explanation. and a Way-Out. Ecollolllic alld Polilical Weeklv. 9 February. Singh, S .. Kaur. M. and Kingra. H. S. (2008). Indebtedness among Farmers in Punjab. Ecollolllic alld Polilical Weekly. 28 June. Stepan. A. (2001). Toward a New Comparative Politics of Federalism. (Multi)Nationalism. and Democrucy: Beyond Rikerian Federalism. in Arguillg Comparative Politics. New York: University Press. pp.315-362. Telford. H. (1992). The Political Economy of PunJab: Creating Space for Sikh Separatism. Asiall Surl'el'. Vo1.32. No.ll. pp.969-988.

114

K. M. Bakke

Tiebout. C. (1956). A Pure Theory of Local Expenditures. jOllrnal of Political Ecollomy. Vo1.64. No.5. pp.416-424. Toft. M. D. (20()3). The Geography of Ethllic Violellce: [dl'llIity. [llIl'1"l'sts. alld the [Ildivisihilitv of Territory. Princeton: Princeton University Press. Treisman. D. (2001). After the Deluge: Regiollal Crisis alld Political COllsolidatioll ill RlIssia. Ann Arbor: Michigan University Press. Trcisman. D. (2007). The Architeclllre of Governmellt: Rethillkillg Political Decelltralizatioll. New York: Cambridge University Press. Tully. M. and Jacob. S. (1985). Amritsar: Mrs. Galldhi's Last Battle. London: J. Cape. Weingast. B. R. (1998). Constructing Trust: The Political and Economic Roots of Ethnic and Regional Conllict. in K. Sol tan. E. M. Uslaner and V. I'!auller (cds). [lIstillltiollS alld Social Order. Ann Arbor: The University of Michigan Press. pp.163-200.

The Paradox of Federalism: Some Practical Reflections DA VID CAMERON Department of Political Science. University of Toronto. Ontario. Canada

Introduction This collection of essays on the paradox of federalism tackles one of the central conundrums of federal theory. and one of the issues of greatest practical concel11 to those considering the federal option. Does federalism foster secession or inhibit it? Many of today's conflict-ridden political communities, considering the introduction of federal structures as a possible resolution of their problems. would dearly love to know the answer to the question. The opening paper by Jan Erk and Lawrence Anderson sets out the issue very clearly and reviews the scholarly literature on the subject. Erk and Anderson explore the paradox in the light of three dimensions: the will and capacity of the subunits (and presumably of the central govel11ment as well); institutional design (for example: the nature and number of subunits; the distribution of powers; representation in central institutions; the electoral and party systems; the judiciary); and what they call uncodified factors. the extra-constitutional socio-economic factors that provide the context within which the formal structures and processes of govel11ment exist. What I propose to do here is offer some reflections about federalism and secession that have arisen out of practical experience working on federal issues in existing, emerging

I 16

D. Cameron

and potential federations. For purposes of this discussion, I will concentrate primarily on three countries: Sri Lanka, Iraq and Canada. As I see it-at least in the short term-the institutional structure of federalism is a distinctly secondary matter when it comes to understanding why a country is or is not facing secession. It is not that it doesn't matter, but that it matters far less than a number of other forces and factors, which mostly fall into the Erk and Anderson uncodified category. This is not to deny that federalism gives subnational units political resources that can be deployed in the service of a secessionist movement. But it is to say that the active utilization of these institutional resources in the service of secession is unlikely in the absence of other more significant factors, such as a territorially concentrated minority's protracted experience of discrimination and exclusion. That is the short term. In the longer term, I am not so sure. It is clear that, with the passage of time, the formative power of institutions is considerable, and the political identities that are often at the root of secessionist movements can be as much the product of institutional structures and processes as the reverse. Just as nation states may become more distinctive and differentiated as the years pass, so may subnational units, especially if they are the home of ethnocultural minority groups (Young, 1995: 182-183). Differentiation does not always or necessarily happen, however, and even if it does, this does not automatically lead to increased tension and the risk of fracture. Indeed, as Quebec's experience shows, the reverse process can sometimes lead to increased tension. Historically, French Canada stood apart from the rest of North America, with a lifestyle and culture that were palpably different; the rise of the secessionist movement in Quebec in the 1960s was the product in part of a process of modernization, in which, according to several social indicators (urbanization, secularization, participation in the market economy), Quebecers became more like Canadians elsewhere in the country. Nevertheless, one can get a glimpse of the constitutive power of federal structures by looking at what sometimes happens when new subnational units are created in federations. In Canada, Saskatchewan and Alberta, for example, were carved out of the North West Territories in 1905. That they are the willed creation of an actor at a defined historical moment can be seen in their shape-rectangular, with borders as straight as a surveyor's line, except for the south-western frontier of Alberta which is defined by the Rocky Mountains, which the Province shares with British Columbia. The north-south boundary between Alberta and Saskatchewan is simply a meridian of longitude. Quite rapidly after their establishment as new provinces in the Canadian federation, Saskatchewan and Alberta began to develop their own distinctive character. Settlement patterns and the emerging political economy of the two communities led to a growing differentiation of one from the other. What had begun as little more than arbitrary lines drawn on a map by federal omcials quickly became the borders of distinctive communities. Politically, both were popUlist, but Saskatchewan developed a left-wing political culture, while Alberta generated a right-of-centre politics. I am not saying that the creation of these provincial institutions produced a secessionist movement, although at times during the post-war period it would have been only a slight stretch to speak of Albertan nationalism, as the Province's political leaders self-consciously went about constructing a provincial state and community defined by its oil and gas

The Paradox

(~f Federalism:

Some Practical Reflections

117

wealth, its integration into the mid-West North American oil and resource culture, and its commitment to an ideology of private enterprise and individualistic self-reliance-a far cry from the communitarian political culture of Saskatchewan next door, which proved to be the birthplace of Canadian socialism (and public healthcare in Canada). The observation of experiences such as these cannot but make one respectful of the creative power of institutions, operating over the long haul. Nevertheless, I would contend that the role of political structures such as these is primarily, although not exclusively, instrumental. They provide signilicant resources to politicians and communities within larger political systems to do what it is they wish to do. But only to a limited extent do the structures themselves lead provincial or subnational actors to wish to do something they would not otherwise want to do. Primarily, they assist in the realization of needs and aspirations that are shaped by forces that transcend the simple realities of federalism. While it is true, for example, that the systematic oppression of a region by a federal government dominated by another community can foster secessionist sentiments, it is equally true that the same result can occur for the same reasons in a unitary state; it is the fact of oppression, more than the availability of subnational political authority that makes the difference. Sri Lanka's experience illuminates this fact. The secessionist violence of the Liberation Tigers of Tamil Eelam (LTTE) broke out within a unitary state, and their chateau fort has always been the north, where the indigenous Tamil population has traditionally been concentrated. Ironically, until recently, the long-running civil war produced a kind of unconstitutional caricature of federalism in Sri Lanka, in which the writ of the national government stopped in many parts of the north, and the LTIE independently operated schools, hospitals, the police force and social services. (The territory which the LTTE controls is cUITently being signilicantly reduced as the military force of the Sri Lankan govemment pushes the Tigers back into smaller enclaves.)

Federalism Then and Now In classical liberal theory, I individuals escape from the state of nature and enter civil society by agreeing to a social contract that establishes legitimate political power from which all will benefit and by which all will be bound. Similarly, in classical federal theory, autonomous political communities freely join together to form a new, complex polity from which all will benefit and by which all will be bound. The constitution is federalism's social contract. Contemporary federal experience, however, is rather different. Instead of the image of free peoples coming together to build something better, the picture today is often of warring communities, locked in a political relationship from which they cannot escape. Federalism, in such melancholy situations as these, often presents itself as each community'S reluctant second choice-a system designed to make an unsatisfactory situation habitable. The challenges confronting the founders of contemporary federations of this sort are, therefore, rather different from what classical theory assumes. Instead of showing the federating communities the mutual benefits that justify coming together in a new federal union, federal lawgivers are often faced with the bleaker task of taking something apart, of replacing an existing political union, which has ceased to

I 18

D. Call/eroll

be just or viable, with a more complex political aSSOCiatIon constructed on the foundation of pluralism. Most of the papers in this collection are exploring this type of federal reality, in which the introduction of decentralized government is meant to respond to and contain the centrifugal forces the state is experiencing. In the contemporary world, the federal moment seems as often as not to arrive toward the end of acute civil conflict, when a grudging realization emerges among the combatants that the old regime cannot stand, but that the utter collapse of the state is not tolerable either. Federalism, then, often appeals to countries and to an international order struggling with ethno-cultural conflict, separatist movements and terrorism. It may be that the cradle of federalism in the twenty-first century will be found as much in countries such as Sudan, the Congo, Sri Lanka, Nepal and Iraq as in, let us say, the United Kingdom or Italy. What seems undeniable is that for the last decade or so, we have been living through an intense period of innovation and experimentation in federal and other forms of governance (Griffiths, 2005: xv, 3; Hueglin and Fenna, 2006: ch. I; Watts, 1999: 4-6). At first blush, it seems equally undeniable that in many, if not most, of the cases confronting the contemporary world of federal innovation the relevance of earlier experience associated with the constitution of federations from previously autonomous units seems to be of relatively limited value. Historically, the central issue confronting federal countries, such as the USA, Switzerland and Canada, at their founding seems to have been the costs and benefits of a new form of political association, rather than the political calculus of secession. Federalism by composition does not seem to raise the question of secession in as acute a way, as does federal creation by devolution. Yet, closer scrutiny of these 'classical' cases leads us to refine this observation. The US' federal union was adopted in 1789 on the ashes of the failed confederation which had been in existence since 1781. It took a brutal civil war in the I 860s to answer the question whether American states had a right to secede from the union. The Swiss federation was created in 1848 after the Sonderbund Civil War of the previous year. Through a series of referendums in the I 970s, the new canton of Jura was created out of part of the canton of Berne; the presence of the Swiss army was required at one point to keep the peace, so intense was the conflict over what might be termed a form of "internal secession". Two of the six British North American colonies that participated in the talks leading to the establishment of the Canadian federation in 1867 declined to join at the beginning, not being prepared to trust their community's welfare to this new. untried political organization. And Nova Scotia, one of the four founding members of the new federation, attempted unsuccessfully to secede the year after the federal bargain was struck. The contemporary Quebec sovereignist movement is, of course. well known internationally. The state of West Australia, too. tried unsuccessfully to secede from the Australian federation 32 years after it was formed in 190 I. So it would be a mistake to conclude that the challenges of integration and the question of secession are matters unknown to the classical, successful federations that blazed the trail for others. Each of them had to wrestle with the suspicion and fear that attends a new and untried political organization. All have known periods of significant tension and uncertainty; all, so far, have survived. The central question posed for this issue of Regional and Federal Studies implies the prior existence of a federation or decentralized political system from which

The Paradox (?l Federalisl1l: SOll/e Practical Reflections

119

secession is a possibility. In seeking to understand whether federal and decentralized systems foster or inhibit secession, the natural approach is to examine the conditions or circumstances that appear to encourage or discourage fracture. The papers in this issue display the complexity of the question and the variety of real life answers that may be offered in response. whether it be a question of Bosnia and Kosovo, Punjab. the UK, Spain, Italy or Belgium. In these concluding reflections, though, it may be of value to employ a somewhat different approach in tackling this question, namely. to reflect on the circumstances or conditions that pen11it a federal form of government to come into being in the first place. Certainly, this is an acute issue for many of the 'federation-seeking' countries emerging out of circumstances of violent conflict. If we can understand the factors that support the emergence of a federal regime, we may be in a better position to appreciate the forces that may lead to its collapse. If it is possible to identify the conditions that appear to be necessary for a federation to establish itself in the flrst place, one can then speculate about what happens when some or all of these conditions cease to exist in functioning federations. Secession might be understood, from this point of view, as the withdrawal of the critical conditions needed for the continued existence of the federal system. What can one say. then, about the prerequisite conditions for the construction of a federal regime?

Prerequisites There has been some significant discussion of this issue. although it would be wrong to conclude that a consensus or a satisfactory conclusion has been reached on the subject. I will refer to the writings of K. C. Wheare (1963) and Ronald L. Watts (2004) in this discussion. One of the frustrating-but, from our point of view, potentially significantcharacteristics of much of the consideration of federal prerequisites is the fact that it is very difficult to disentangle those features that are peculiarly related to the introduction and maintenance of a federal regime from those that underlie decent, constitutional, democratic systems of any kind. whether federal, unitary, consociational, parliamentary or congressional. The difficulty is this: many of the things that might be identified as prerequisites of a federal system also look pretty much like prerequisites of a properly functioning liberal democratic state of whatever kind. Let us take K. C. Wheare's list in Chapter 3 of Federal Government. "Some Prerequisites of Federal Government". His analysis is based on the assumption that federalism emerges out of the common desire of previously more or less autonomous units to associate; it does not really deal with devolutionary federal processes. where the considerations look quite different. Wheare's starting out position-that federalism requires a desire to live under a single government, combined with a desire to live under regional governments and a capacity to work under this dual system-seems sound for any circumstances, but, as he acknowledges, it does not get you very far. It is little more than the attitudinal detlnition of the federal principle itself. The factors expressing a desire for shared rule, according to Wheare (military insecurity/common defence, desire for independence from foreign powers. hope of economic advantage, some form of anterior political association, geographical neighbourhood,

120

D. Cameron

similarity of political institutions, political leadership) seem most easily related to integrative federalizing processes. They may have something to say about the globalizing and multi governance dynamics at the international level, but less about many of the most acute 'domestic' or 'within-state' federal challenges facing communities in the twenty-first century. In a presentation to a colloquium in Switzerland on the relevance of federalism as a tool of conflict management in the Near East, Ronald Watts (2004: 8) presented a list of the significant political processes and practices on which the effectiveness of different federations has depended: I. 2. 3. 4. 5. 6.

disposition to democratic procedures; non-centralization as a principle; checks and balances to limit the concentration of political power; open political bargaining for making collective decisions; genuine group power-sharing within central institutions, often consociational; respect for constitutionalism and the rule of law.

Clearly. the first and the sixth on this list sustain good government in general, whether federal or not. The practices identified in points 2-5 above are more obviously, although by no means exclusively, related to effective federal operations. If, for example, one takes the fifth-power-sharing within central institutionsclearly. that can be a vital component in the effective operation of a culturally diverse unitary state and, indeed, an incapacity to share power in this way can be a spur to introduce federal elements into a pluralist political community under stress. Unalloyed majoritarian democracy in a unitary state can create injustices, driving a political community towards secession movements or pressures for federal reform, which is why, perhaps, Watts' statement concerning the first process above is qualified ("a strong disposition to democratic procedures since they presume the voluntary consent of the different groups ofcitizells in the constituent ul/its"). A popular appreciation of the limits as well as the power of democracy, including a respect for minorities, seems essential in a federal regime, but it seems equally essential in a diverse community endowed with a unitary state, if it is to remain just and stable. Later in the same paper (15-16), Watts speaks directly of five necessary preconditions for federal or confederal solutions to operate effectively in moderating multicultural conllict: I. 2. 3. 4.

the will to federate; underlying shared values and objectives; trust; development of a political culture emphasizing sharing and co-operation and fostering respect for constitutional norms and structures and the rule of law; 5. supportive economic conditions, including complementary economies, common economic interests and the moderation of economic disparities. Again, what is striking about this list is the fact that all but the first (which is, in a sense, a formal requirement) apply as much to the workability of non-federal solutions

The Paradox of Federalism: Some Practical Reflections

121

as federal or con federal solutions. If preconditions 2-5 are sufficiently present in a society. it is perhaps a pardonable exaggeration to say that almost any political system can be made to work; if they are absent, then no constitutional or political regime is likely to work. or work very well. If this line of reasoning is broadly correct. then the issue for those countries emerging from serious conllict is not so much "What are the preconditions necessary for a federal solution?" but "How does one cope with the absence of some or all of the preconditions necessary for a constitutional. broadly consensual, democratic solution of any kind?". That is an awesomely difficult question, and one to which I would be the first to admit I do not have the answer, but it does not seem to me to be a specifically federal challenge. Sri Lanka and Iraq are two countries that have been actively discussing federal models in recent vears. The Sri Lankan efforts during the 2002-05 peace negotiations under the auspice~ of the Cease Fire Agreement were shelved with little accomplished, and civil war has broken out again. Iraq proceeded further faster. at least in constitutional development, negotiating a federal constitution in the summer of 2005 that is now being fitfully implemented. How do they line up with the Ilve preconditions outlined above? It seems to me that both are notably dellcient in meeting preconditions 2-5. Iraq, at the moment, dramatically fails test 5 (supportive economic conditions); Sri Lanka, less so. But, on the question of shared values, trust and a supportive political culture. both are in very bad shape, and this reality sets the terms that must be met in fashioning any workable political settlement, not just a possible federal alTangement. If the terms thought necessary for the introduction of a federal political system are absent, why has federalism been so central to the political debate in these two cases? A theoretical alternative would be to tackle directly the issues of trust. shared values and political culture, and leave the choice of constitutional arrangement for another day. I assume that the reason why factors such as these are not being focused on is because there is a recognition that trust cannot be established. shared values cannot be created, and political cultures cannot be changed directly and in the immediate term. These developments occur in the course of doing something else. as indirect by-products of other human activity. What other kind of human activity seems a plausible candidate here? In most cases, it is thought to be the peace process itself. It is the discussions, the negotiations. the trade-offs and compromises-it is this testing crucible of finding accommodations under severe pressure that, if successful, turns enemies into associates in a common enterprise, transforms enmity and suspicion into trust, and begins to reveal the common interests and purposes that lie beneath entrenched conflict. South Africa, buttressed by extraordinary leadership, seems to be one of the few instances in which a generally successful transition has been made. This offers an account of why we do not typically attempt to tackle these principled deficiencies head on, but it does not explain or justify the selection of federalism as the preferred institutional arrangement. Why federalism? As often as not, and certainly in the cases of Sri Lanka and Iraq, it is because there appears to be no other better alternative. It is not a first choice; it is everybody's second choice. The Sinhalese majority and the Government of Sri Lanka would prefer a unitary state; so would many of the Sunnis of Iraq. The Sri Lankan Tamils, at least as represented by the LTTE, and the Kurds of northern Iraq, would rather have an independent state. Neither of the parties to these

122

The Paradox (~f FederalislII: SOllie Practical Reflections

D. Cameron

conflicts can get what they want, if negotiation rather than force is to settle the issue. Federalism is what the parties fall back on. So the relevant question at this point is not, properly speaking: "Why federalism?" but rather "Is there any other alternative that is better?". If the answer to that question is no, one is left with the task of trying to work something out, despite the forbidding obstacles to its realization. The factors that seem to lead to the selection of federalism as the best-or the least bad-arrangement in the circumstances, include the following. • The parties to the conflict, or the participants in the process. begin from preferred positions that are mutually incompatible-secession and independence. on the one hand: and a centralized. unitary state, on the other. • Neither side, through force or diplomacy. is capable of imposing its will on the other. • International actors are flatly opposed to secession, but recognize the actual or potential injustice and discrimination embedded in the maintenance of the existing arrangements. • The pressures to make peace, or to achieve a negotiated settlement, are greater than the incentives for going-or going back-to war. These are hardly the preconditions for the establishment of a successful federation. but they are the brute pressures that lead the pmties in conflict to this negotiating ground. It is clear that, in these circumstances. the prospect of a successful outcome is modest. but it seems to me that it would be equally modest if one had some other political and constitutional end in view. The difficulty lies not so much with the absence of the necessary preconditions for federalism, but with the absence of the necessary preconditions for successful peacemaking and constitution building of any kind. Indeed. it is my view that consociational fOims and practices would have sufficed to sustain a unitary Sri Lankan state if they had been conscientiously introduced four decades ago: it is. as much as anything, because of the utter failure to do so that the country finds itself faced with a starker, simpler choice today: to negotiate an end to the violence and civil conflict. or to carryon like two scorpions in a bottle. If a negotiated termination of the civil conflict is to be achieved, it will in the present unhappy circumstances necessarily entail some form of federalism. This analysis cannot leave one optimistic about the prospects for federalism in Sri Lanka, but neither can it leave one optimistic about the prospects for peace and reconstruction in Sri Lanka more generally. Yet, it would seem perverse to set aside all discussion of federalism in that country because it will be so very difficult to achieve. If there was a plausible. preferable alternative. so be it: but if there is none. then faule de lIIieux the effort to construct some form of federal system in the country seems a worthwhile enterprise. A parallel, although not identical, stream of analysis seems to me to apply to Iraq as well. It possesses a federal constitution, although it is not by any means a fully federal country yet. If Iraq ceases to be a going concern at some point. it would seem obtuse to visit the blame on its federal structure. Surely. its failure to survive as a single political community. federal or not. will be the result of its sad history and toxic political culture, more than the product of its particular constitutional regime. As Ron Watts (1999: 110-11) said in a chapter on the pathology of federations, "it is not so much

123

because they are federations that countries have been difficult to govern but that it is because they were difficult to govern in the first place that they adopted federation as a form of government".

Canada and Quebec For many years. Canada has had one of the most powerful secessionist movements in the modern world. and some brief reflections on that experience may be of some use in considering the theme of this issue. The creation of Quebec as a subnational unit in 1867 permitted that predominantly francophone society to develop in substantial degree according to its own genius. first as a conservative. Catholic realm and, more recently. as a modern, urban. secular, French-speaking society. The Canadian federation has so far managed to accommodate these transformations, including the rise in the I 960s and I 970s of a very powerful secession movement. The sovereignty movement in Quebec displays a couple of distinctive features relevant to our story. First, it has been non-violent. Since the October Crisis of 1970. the Quebec sovereignists have been relentlessly democratic and peaceful in the pursuit of their ambitions. Secondly, it has been. paradoxically. an independence movement which has not really been in pursuit of full-scale independence. With the exception of a few hard-line separatists. the movement has consistently sought sovereignty plus continued, close association with the rest of Canada-sovereignty with a difference. As a Quebecois wag put it: "What do Quebecers want? They want a free and independent Quebec within a strong and united Canada". Why is this so? Why has the movement been peaceful and why has it aimed as much at an altered relationship with the rest of Canada as at outright independence? My suspicion is that these two distinctive features are related to the country's standing as a constitutional. democratic-and deeply federal-state. It has never been plausible to argue that the Quebecois are an oppressed people. And when advocates of the independence of Quebec emerged in the 1960s. there was never any inclination to brand them as traitors. Secession was seen to be within the ambit of free and legitimate expression and, implicitly. I think, other Canadians reacted to the challenge within the frame of reference of what might be called the double-consent theory. namely, that it was not just Canadians as individuals, but the two main linguistic communities of Canada that need to consent to the state for it to retain its authority. Not only that: political power in Canada. since the rise of the sovereignist movement in the 1960s has been flowing to the provincial governments, in pmt because of the demands of Quebec nationalists for greater autonomy. Both the demand for self-rule and the demand for shared rule-in this case. equitable francophone participation in the affairs of the Canadian government-have been addressed. at least in significant degree. And. just as Quebec has been transforming its society, so too has the rest of Canada been changing-especially in the cities-moving from a traditional. British view of itself. to a multi-racial, highly pluralistic society. powerfully shaped by waves of post-war immigration. If the country is free. democratic. prosperous and-however untidily and imperfectly-flexible in accommodating the evolving requirements of the national minority. how seriously will its members feel the need to leave? These features of Canadian society have sustained the country during its

124

D. Cameroll

worst moments, which have been when questions of national identity have been allowed to rise to the symbolic and constitutional level. The last four decades of national-unity debate have engendered another characteristic of Canadian society, and it too has shaped the country's approach to secession. 2 The Quebec sovereignty movement's sustained and powerful challenge to the very existence of the country has fostered an uneasy realization among many Canadians that their hold on Canada is a matter of cOl/tingent possessiol/. It has not been possible during these years to assume that the country in which Canadians live is effectively etemal, that it will always be there for its citizens. Canadians have had to recognize that it may not. With the assistance of the Supreme Court of Canada, Canadians have leamed that even the possible dismantling of one of the most successful countries in the world could be justified by the values they hold dear, and could be executed by democratic means. In its decision in the Quebec Secessioll Reference, the Supreme Court of Canada identified four fundamental, unstated principles that support Canada's constitutional order: federalism; democracy; constitutionalism and the rule of law; and respect for minority rights. It discovered a constitutional duty to negotiate with Quebec in the event that a clear majority of the citizens of Quebec in response to a clear question favoured the secession of the province. In doing so, it established in constitutional law the contingent nature of Canadian political affiliation. Not for Canada the ringing declarations of France, une republique indivisible, or Italy, La Repubblica. ul/a e il/divisibile, or Brazil, for/l/ada pela IIniiio indissohivel dos Estados e Municipios e do Distrito Federal. or Australia, one indissoluble Federal Commonwealth under the Crown of the United Kingdom of Great Britain and Ireland: Canada, for its part, will exist as long as Canadians will it. As Emest Renan said, a nation is a plebiscite de tous les jours. There are worse foundations on which to construct a political community. Canadians are leaming that the values for which the country stands can and should obtain even during its possible breakup. This lesson deprives the country of national glory, but it speaks to the practical functions the state is expected to perform-in Canada, to provide the blessings of peace, order and good government. It reminds Canadians that the state itself is a human artefact, reared up to serve the interests and needs of the people for whom it is responsible, not an entity endowed with intrinsic moral or spiritual value. What Canada asks of its citizens is adherence to the constitution and the values underlying the constitutional order, not doctrinal acceptance of a universal patriotic creed. It is within the framework of these values that the accommodation of diversity-even, possibly, the accommodation of secession-is to be found. Canadian experience encourages the 'desacralization' of the public realm. What Canadian today believes that it is the sacred duty of citizens to preserve the state and the national community and that, to seek the opposite, is treason? Canadians cannot afford that view. How, with that view, would the country do business with the substantial minority of the population who hold democratically to the conviction that the Canadian experiment has more or less failed and should be wound up? How would they treat with what is often a sovereignist Government of Quebec, or with a sovereignist political party in the federal Parliament? Here, it seems, is a case in which the accommodation of a national minority appears to have blunted the force of secession. In fact, the apparent willingness to

The Paradox of Federalism: Some Practical Reflections

125

accommodate the actual secession of Quebec, should that be necessary, appears to have been part of the context in which the power of the secessionist movement in Canada has, at least for the time being, waned. The situation is reminiscent of the paradoxical comment that was sometimes made by Quebec nationalists: "Recognize our right to national self-determination so that we won't use it". The daily plebiscite has so far been favourable to Canada. Conclusion This leads us back to where we began. Does federalism foster or inhibit secession? As with so many good questions in social science, the answer seems to be yes. It offers institutional resources to actors to help them achieve what they want to accomplish, but what they want to accomplish is only to a relatively limited extent shaped by the institutions of federalism. Federalism is unlikely to hold an unjust regime together; but. equally, it is unlikely to pull a just regime apart. Justice-rather than a glorified governing instrument-is surely closer to the heart of the matter than federalism. To the extent that, in a liberal democratic regime, the crude proxy of justice is consent, federalism introduces the communitarian dimension of justice often missing in unitary states. If, for example, there are national communities within the boundaries of a political regime, if there are territorially concentrated linguistic or ethnic groups. or regional religious communities, then there is a need to justify the political order, whether federal or not. on communitarian principles as well as individualistic ones. It seems reasonable to conclude that the likelihood of there being a secession movement in a given polity turns more on how people are treated than on whether or not they are federally governed. Notes IThese three paragraphs are drawn from Cameron (2007a). "The next two paragraphs are drawn from Cameron (2007b: 82-83).

References Cameron, D. (2007a), Making Federalism Work, in M. E. Bouillon, D. Malone and B. Rowswell (eds), Iraq: Prevelltillg {/ New Gelleratioll of COlljiict. Boulder and London: Lynne Rienner, pp.153-167. Cameron, D. (2007h), An Evolutionary Story. in J. G. Stein, D. R. Cameron, J. Ihhitson, W. Kymlicka, J. Meisel, I-I. Siddiqui and M. Valpy (eds), Ulleasy Partners: MlIlticllltllralislIl alld Rights ill Callada. Waterloo: Wilfrid Laurier University Press, pp.71-94. Griffiths, A. L. (cd.). (2005). Halldhook of Federal COllllfries, 2005. Montreal: McGill-Queen's University Press. I-Iueglin, T. O. and Fenna, A. (2006), Comparative Federalism: A Systematic IlIqlllrv. Peterborough: Broadview Press. Watts. R. L. (1999), Comparillg Federal Systems (2nd edn). Montreal: McGill-Queen's University Press. Watts, R. L. (2004), Federal Co-existence in the Near East: General Introduction. Paper presented at the Jean Nordmann Foundation Colloquium on "Federalism: A Tool for Conflict Management in Multicultural Societies with Regard to the Conflicts in the Near East", Fribourg. Wheare, K. C. (1963), Federal Govemmellt (4th edn). Oxford: Oxford University Press. Young, R. A. (1995). The Secessio/l of Quebec alld the Future (!f'Callada, Montreal: McGill-Queen's Press.

Index Croatia 31, 87 cultural heterogeneity 36, 40, 41 Czechoslovakia 14, 16,23

Index

Page numbers in Italics represent tables. accommodation 3 agriculture: Punjab 107, 108 Alberta 116 Anandpur Sahib Resolution (ASR) 101 anarchy 81 Andalusia 69 Anglo-Irish Treaty (1921) 33 Anglo-Scottish union 55 asymmetrical decentralisation 64, 65 asymmetrical federalism 26 Australia 118 Austria 62 Austro-Hungarian Empire 14 Austro-Hungary 23 autonomous homelands 14 autonomy 4, 16, 53, 63; arrangements 14, 15, 16, 17, 21, 23, 24, 27; paradox 77; regional 63, 75, 76; weaknesses 17,18 Azerbaijan 16, 23 Basque country 33, 69, 70, 76 Belgium 14, 16,23,24, 62; decentralisation 72-3 Bermeo, N. 3 Biden, J. 13, 110 Bieber, F. 85 Blair, T. 52 Boden Paper 14 Bose, S. 86 Bosnia 10, 14, 23, 27, 82, 84-7; constitution 84; post-war 92,93; refugees 86 British Empire 33 Britishness 54 Brussels 72

Buchanan, A. 5 Burma 14, 19,23 Canada 7, 11, 16, 23, 45, 75, 116, 118; containment strategy 56; socialism 117 cantonization 3, 27 Catalonia 69, 70, 76 central government office seeking (CGOS) 63, 75, 76; framework 64, 65,66,67, 70; Italy 71 centralisation 41 centre-periphery bargaining 21, 27 Chechnya 98 China 14,23 civil war 14,16,68,81,92,117,122 classical federal theory 117 Cold War 80 collective representation 2 common-state governments 22, 23, 24,26 common-state policy 22 comparative federalism 9 concessions 32, 33, 34, 35, 36, 37, 42; enforceability 36, 37, 38, 40 conflict management 2, 4, 5, 120 Congo 118 consociationalism 82 constitution 38, 117 constitutional accommodation 45 constitutional law 9 containment strategy 56 contingent possession 124 cooperative federalism 109 corruption: Kosovo 89 Corsica 74, 75 Cox, G. 64

Dayton Peace Agreement (DPA) 84 de facto partition 81, 83, 84, 92; Bosnia 82, 84-7; Kosovo 82, 88-91 de facto secession 16, 23, 26 de jure partition 81 de jure secession 16, 26 decentralisation 2, 5, 10,37,46,62,64, 65, 76; Europe 63 decolonisation 14 defensible enclaves 81 Delhi riots (1984) 109 democracy 14,46, 63, 120 Denmark 23, 62 deterrence theory 80 devolution 2, 4, 19,24,47,51,63, 118; referendum 67; Scotland 10, 45, 46, 49,50-2,57,58,67; UK 45,47,48, 49, 53, 54, 65-8; Wales 48, 67 devolved power: retaining model 53 Dikshit, R. 23.24 dilemma of recognition 2 distinct society 52 diversity 3, 124 dual federation 14 dual identity 54 dual nationality identities 20 Duffy-Toft, M. 37 economic disparity 8 economic reform 21 electoral competition 35 electoral systems 7,64,65; ethnically segmented 83; majoritarian 7; proportional 7; Quebec 56; Scotland 56 England 47, 52; nationalism 55 environmental protection 21 Ethiopia 14, 19,23 ethnic civil wars 81, 92 ethnic cleansing 80. 87, 89, 92 ethnic conflict 3, 4, 5, 9,10,11.31,81; regulation 3

127

ethnic demobilisation 82 ethnic identities 25 ethnic islands 91 ethnic minorities 81 ethnic mobilisation 4, 104 ethnic outbidding 83 ethnic partition 2, 80, 81, 83, 89 ethnic reintegration 93; Czechoslovakia 93; Rwanda 93 ethnic security dilemmas 92 ethnic wars 4 ethnicity 98, 99 ethno-cultural conflict 118 ethno-linguistic conflict 1,4 ethno-linguistic diversity 2 ethno-territorial partition 10 ethnocultural minority groups 116 ethnocultural pluralism 16 ethnofederalism 4. 9. 14. 19,21,24,27, 81; weaknesses 17, 18 ethnofederations: failure 23; surviving 23 European Union (EU) 55, 69 federal institutional design 7, 8 federal institutions 98 federal state 45 federalisation 3 federalism 2. 5. 11, 30, 49, 52, 59, 97, 117; Canada 45; conflict management 120; Germany 46; India 98; Italy 70; Nigeria 25; palliative potential 4; paradox 5; peace-preserving capacity 98, 109; prerequisites 119-23; USA 46 Finland 23. 62 fiscal autonomy 98, 100; peace-preserving potential 107; Punjab 105-8 tiscal decentralisation 99 Flanders 72 France 23, 62; decentralisation 73-4 Gandhi, I. 104 Georgia 16, 23 Germany 46, 62

128

Inde.x:

Ghai, Y. 4 Golden Temple Massacre 101, 104, 105 Gorenburg, D. 4 group identity 4 guerilla warfare 81 Gurr, T.R. 4 Hale, H. 5 Haryana 102, 103 Hechter, M. 5, 99 Hinduism 101 Hoare, M.A. 84 homeland elections 21 homeland governments 19,20,21, 22, 24, 26, 27; absence 20; voting rights 25 Horowitz, D. 3, 4, 97 Human Rights Watch 90 Ibarretxe Plan 33, 69 identity formation 39 imperial collapse 23 inclusion 2 India 10, 14, 16, 19,23,25,81. 97-110; partition 106 Indonesia 14, 19, 23 industry: Punjab 107 institutional accommodation 30, 31,

39,40,41 institutional design 9, 38, 115 integration 4, 92 international law 2, 5 international peacekeeping force 88 intra-state federalism 58 Iraq 6, 11, 13,27, 116, 118, 121, 122 Ireland 33 irrigation: Punjab 106, 108 Italy 23, 62; decentralisation 70-1 judiciary: role 8 Kashmir 10 Khalistan 104, 109 kin groups 8 Kosovo 9, 10, 82, 88-92, 92; corruption 89; post-war 90,91,92,93; war 91

Index Kymlicka, W. 16 Lapidoth, R. 16 Lebanon 82 liberal democracy 42, 59 Lippman, P. 87 Lombardy 71 Lyon, 1.M.B. 85 McGarry, 1.: and O'Leary, B. 3, 15 Malaysia 14, 15, 16,23 minority returns 86, 87, 90 minority rights 18, 25, 26, 124 mixed identities 47 Moldova 23 money laundering 83 Moreno national identity 55 Moreno test 54 mu Iticu Itural conflict 120 nation-state projects 22, 25, 27 nation-states 21 national diversity 49 national identity 27, 46, 47, 48, 54; Scotland 50; Soviet Union 83; UK 47,54 national identity conflict 19, 21 nationalism 4, 9, 41, 59, 80, 83; English 55 nationhood 46, 52; Scotland 51 NATO 88, 89, 93 negotiated autonomy 4 Nepal 118 Netherlands 62 Nigeria 3, 14, 15, 16, 19,23,27; federal ism 25 Nordlinger, E. 15 normative entitlement 46 Northern Ireland 47, 48, 49, 53 Norway 62 nullification crisis 32 Operation B1uestar 101, 104, 105 organised crime 83 Ottoman Empire 14,23

Pakistan 14, 15, 16, 19, 23, 81 Papua New Guinea 16, 23 partition 81, 92; Bosnia 83; India 106; Kosovo 83 Philippines 23 Plaid Cymru 66, 67 pluralism 45, 48, 118 plurinational constitution 47 pi uri national federalism 47 plurinational states 46, 49, 59; UK 47,

48,49,59 policy autonomy 98, 99-100; ethnic 99, 101, 104, 109; Punjab 101-5 policy decentralisation 99 political equilibrium 30, 31, 40, 41 political independence 35 political justice 11 political mobilisation 31, 32 political parties 7 political power 41 Portugal 63 power: distribution 19,32,41 power sharing 5 precedent-setting 37 preventive diplomacy 4 property restitution 86, 92 Pugh, M. 85, 90 Punjab 10, 97-110; agriculture 107, 108; industry 107; irrigation 106, 108; separatism 109, 110; violence 101 Punjab crisis 100, 108, 109 Punjab Reorganisation Act (1966) 102 Punjabi Suba 101, 102, 104, 105 Quebec 11, 33, 35, 52, 68, 116, 123, 125; sovereignty movement 123, 124 recentralisation 19, 23, 25, 27 recognition 2 refugees 86; Serbia 90 regional autonomy 63, 75, 76 regional government office seeking (RGOS) 63, 65, 75, 76; framework 64,66,67,68, 70; Italy 71 Renan, E. 124

129

Roeder, P. 4 Russia 23, 98 Russian Empire 14,23 Russian Federation 19 Rwanda 93 Saskatchewan 116 Scotland 10, 45-53, 55, 66, 75; autonomy 57 Scotland Act (1998) 45, 51, 53, 57, 58 Scottish devolution 49, 50 Scottish independence: referendum 57 Scottish National Party (SNP) 33, 56, 63,

66,67,68 secession 2, 5, 23, 34, 41, 54, 59, 115, 118; capacity 6, 7; political dynamics 30; prevention 36; Quebec 57; risk 62, 77; Scotland 53; timing 32; will 6, 7 secession proneness 9 secession proofness 9 secessionism 5, 8, 25, 31, 40; Europe 63 secessionist conflict 1 secessionist mobilisation 32 Second World War 14, 19 segregationist institutions 83 self-rule 2, 3, 4, 5, 6, 9, 123 semi-federal system 68 separatism 97, 105, 118; Punjab 109, 110 Serbia 87, 89, 90; refugees 90 Serbia-Montenegro 23 shared rule 2, 4,119,123 Sikhism 101, 102, 104 smuggling 83 Snyder, 1. 4 social cleavage structure 8 socialism: Canada 117 soft partition 13 South Africa 23, 121 sovereignty 5, 17,21, 24, 41, 51, 88, 123 Soviet federalism 4 Soviet Union 14,23,25,83; union-republic governments 23 Spain 7, 14,23,24,47,52,62; autonomous communities 68; civil war 68; decentralisation 68-70

130

Index

Sri Lanka 6, 11, 116, 117, 118, 121, 122; civil war 117 stable equilibrium 38,39 state disintegration 97, 98 state identity 47 state security 80 state-society interactions 10 sub-state autonomy 48 sub-state nation 47, 52, 59; autonomy 55 sub-state nationalism 59 Sudan 23, 118 Sweden 62 switchman role 20 Switzerland 14,46, 118 symbolic asymmetry 52, 53 Tajikistan 23 Tanzania 14, 19, 23 territorial autonomy 62 territorial decentralisation 16, 32, 34, 36 territorial diversity 45, 46, 47; Byzantine model 47 terrorism 118 trafficking 83 Treaty: Anglo-Irish (1921) 33 Uganda 16,23 Ukraine 23

uncodified factors 8, 97, 115 unilateral secession 57 Union of Soviet Socialist Republics (USSR) 20, 23 unitarism 3 unitary state 46, 70, 117 United Kingdom (UK) 6, 9, 10, 23, 46, 52, 59, 65-8; constitution 48, 49 United Nations High Commissioner for Refugees (UNHCR) 86 United Nations Mission in Kosovo (UNMIK) 88, 89, 90 United States of America (USA) 46, 118 Uzbekistan 23 Veneto 71 violence 122 violent conflict 83 violent separatist movements 3 voting rights 25 Wales 47,49,52,53,66, 75 Wallonia 72 Watts, R.L. 15, 119, 120, 122 wealth 98, 108 Wheare, K.c. 119 World Bank 86 Yugoslavia 4, 14, 15,23; civil war 5