‘Why did they vote for those guys again?’ Challenges and contradictions in the promotion of political moderation in post-war Bosnia and Herzegovina

Abstract

Party assistance in Bosnia and Herzegovina has focused on encouraging multi-ethnic and moderate mono-ethnic parties by placing funding and political restrictions on ethnonationalist parties while providing technical assistance and public support to those parties with less nationalist agendas. However, multi-ethnic parties have met with very limited electoral success and moderate mono-ethnic parties have found electoral success primarily by radicalizing their political discourse to match that of nationalist parties. This study provides evidence for the effect of international intervention drawn from patterns of electoral support for multi-ethnic and moderate mono-ethnic parties in canton, entity and national elections from the 2006 general elections in Bosnia and Herzegovina. The analysis shows that even voters who are predisposed to vote for multi-ethnic parties are much less likely to choose non-nationalist parties in elections where parties representing other ethnicities also compete. As a result, there is a clear trade-off between promoting Bosnia's integration and promoting parties and candidates who refrain from using often incendiary nationalist rhetoric.

Keywords:

• Bosnia and Herzegovina
• party assistance
• ethnopolitics
• political parties
• nationalist parties

Acknowledgements

Earlier versions of this analysis were presented at a workshop on International Dimensions of Party (System Development) held at the University of Amsterdam, the Annual Meeting of the Midwest Political Science Association as well as the Working Group on the Political Economy of Democratic Sustainability at the Workshop on Political Theory and Policy Analysis, Indiana University. I would like to thank the participants in each for their valuable feedback. I am particularly thankful to Peter Burnell, André Gerrits and two anonymous reviewers for their contributions to improving both the content and delivery. The funding for field work supporting this research was provided by IREX-IARO.

Notes

For convenience, this account uses ‘Bosnia and Herzegovina’ and ‘Bosnia’ interchangeably. Where appropriate, ‘Bosnia and Herzegovina’ or ‘BiH’ is used to denote state-level institutions as opposed to those of Bosnia's entities or cantons.

The analysis is concerned primarily with the performance of multi-ethnic and non-nationalist parties. These differ from ethnic and nationalist parties in the composition of their leadership, the breadth of voters to which they appeal and their use of incendiary rhetoric that explicitly or implicitly threatens members of other nations. The task of categorizing parties is made more difficult by the fact that party behaviour changes over time and that there is a frequent disconnect between the official party platforms and statements by party leaders. Few parties in Bosnia would describe themselves as nationalist parties, but most engage in rhetoric that can clearly be categorized as such. The wartime, nationalist parties SDA, HDZ and SDS make up one natural group of parties, in that they were the chief protagonists of the war in Bosnia that accompanied the collapse of Yugoslavia. Each explicitly represents one ethnic group. Radovan Karadzic's SDS stands out from this group for its ultranationalist rhetoric before, during and after the war as well as its initiation of ethnic cleansing in Bosnia. SBiH, SNSD and HDZ-1990 are also ethnic parties, but they emerged after the war and initially presented a more moderate alternative to the wartime, nationalist parties. This paper defines the SDP, NSRzB, and LDS as non-nationalist, multi-ethnic parties because of their rejection of nationalist rhetoric and attempts to build constituencies from voters of more than one ethnic group. However, even this characterization of the SDP can be challenged based on the fact that the preponderance of its support comes from urban Bosniak voters. Despite this fact, the SDP is included as a multi-ethnic party because its leadership is multi-ethnic and it clearly seeks and receives votes beyond its core constituency of Bosniaks. Future research will be expanded to include newer parties with non-nationalist, multi-ethnic profiles (chiefly Naša Stranka) that have emerged since the last general election.

The classic piece on consociationalism is Lijphart, Democracy in Plural Societies. For discussions of Bosnia as a consociational system see Bose, Bosnia after Dayton; Belloni, State Building and International Intervention in Bosnia; Bieber, Post-War Bosnia.

For a discussion of criticisms of power-sharing and consociational regimes, see Norris, Driving Democracy, 27–31. See also the lively debate on the behavioural effects of consociationalism in Northern Ireland, especially Tilley, Evans and Mitchell ‘Consociationalism and the Evolution of Political Cleavages in Northern Ireland, 1989–2004’; Mitchell, Evans and O'Leary, ‘Extremist Outbidding in Ethnic Party Systems is Not Inevitable: Tribune Parties in Northern Ireland’; Garry, ‘Consociationalism and its Critics’.

For example, the three members of the presidency of Bosnia and Herzegovina, each of whom explicitly represent and are elected by members of one of Bosnia's three ‘constituent peoples’. Specifically, Belloni, State Building and International Intervention in Bosnia, 74 argues that electoral engineering aimed at encouraging multi-ethnic parties failed because of Bosnia's consociational system.

Other analyses of support for non-nationalist and multi-ethnic parties in Bosnia found support for political attitudes as well as prewar ethnic distribution as explanatory factors favouring support for non-nationalist parties. Pickering, ‘Explaining Support for Non-nationalist Parties’, finds that, among Bosniaks, left-leaning survey respondents were much more likely to support non-nationalist parties than right-leaning respondents. Pugh and Cobble, ‘Non-Nationalist Voting’, find that support for non-nationalist parties depends on the pre-war demographic distribution, especially the presence of a Bosniak majority in multi-ethnic municipalities.

Bosnia's electoral system is complex and defies simple categorization. The elections used in the empirical section of this paper use proportional representation with open party lists. Their salient characteristic for the purpose of this analysis is whether their geographically-defined constituency contains one or more than one of Bosnia's three ‘constituent peoples’. The most detailed and current account of attempts at electoral engineering in Bosnia can be found in Belloni, State Building and International Intervention in Bosnia.

See, for example the New York Times editorial ‘Bosnia Unravelling’, 22 January 2009.

The empirical portion of this study includes the National Work for Betterment Party (Narodna Stranka Radom za Boljitak or NSRzB) as well as the Liberal Democratic Party (Liberalna Demokratska Stranka or LDS). NSRzB and LDS are much smaller parties and of more recent provenance but they share the SDP's resistance to the use of nationalist electoral rhetoric and policy of putting forward candidates from multiple ethnic groups. NSRzB especially enjoys support in Croat areas of Bosnia and Herzegovina. It is the combination of refraining from using nationalist electoral rhetoric and putting forward candidates from multiple ethnic groups that differentiates SDP, NSRzB and LDS from the other parties.

For an account of non-nationalist parties' attempts to contest the elections, see Pejanović, Through Bosnian Eyes. In pre-war elections, the most significant political divide was between nationalist parties from all ethnicities on the one side and the non-nationalist, socialist parties on the other. SDA, SDS and HDZ cooperated in attacking non-nationalist parties during the election campaign. This nationalist coalition collapsed once in office, after successfully cooperating to defeat the non-nationalist coalition.

The tenuous nature of this victory is apparent in the titles of the two International Crisis Group reports detailing the election results, Bosnia's November Elections: Dayton Stumbles, December 2000 and Bosnia's Alliance for (Smallish) Change, August 2002.

International Crisis Group, Bosnia's Alliance for (Smallish) Change.

Author interviews with Bosniak SDP members suggest the former interpretation: that Bosniak SDP supporters preferred to use their vote to elect an SDP of any ethnicity rather than choose between an SDA and SBiH candidate in the Bosniak election.

This paper addresses only the most significant Bosniak, Croat and Serb parties in Bosnia, as well as the three parties that are most clearly multi-ethnic parties. The SDA (Stranka demokratske akcije) is the most established Bosniak (Bosnian Muslim) party and was led by Alija Izetbegović during and after the war. The next largest Bosniak party is SBiH (Stranka za Bosnu i Hercegovinu) which is led by Haris Silajdžić, currently the most prominent Bosniak politician. HDZ-BiH (Hrvatska demokratska zajednica) is the dominant Croat party in Bosnia, but a more moderate splinter party, HDZ 1990, has emerged to challenge it in recent election cycles. The dominant Serb party after the end of the war was the SDS (Srpska demokratska stranka), which continues to be powerful in many municipal governments in Republika Srpska; however, SDS has largely been replaced by SNSD (Savez nezavisnih socialdemokrata), led by Milorad Dodik at higher levels of the state. The most significant multiethnic party is the SDP-BiH (Socijaldemokratska partija Bosne I Herzegovina). Two smaller multi-ethnic parties NSRzB (Narodna stranka Radom – Za boljitak) and LDS (Liberalna demokratska stranka) are also included in the analysis, primarily to better extend the analysis to Croat-dominated areas where SDP does not perform well.

These ‘Bonn Powers’ of the OHR understandably are controversial. Belloni, State Building and International Intervention in Bosnia, 20–5 gives a balanced overview. For a critique see Knaus and Martin, ‘Travails of the European Raj’.

Office of the High Representative, ‘Decision removing Ante Jelavic from his position as the Croat member of the BiH Presidency’, 7 March 2001, http://www.ohr.int/decisions/removalssdec/default.asp?content_id=328 (accessed August 15, 2010).

Office of the High Representative, ‘Decision removing Ivan Damjanovic from his position as Mayor of Glamoc’, 7 September 2001, http://www.ohr.int/decisions/removalssdec/default.asp?content_id=304 (accessed August 15, 2010).

Office of the High Representative, ‘Decision to remove Mr. Dragomir Vasic from his position as member of the Republika Srpska National Assembly and as councilor in the Zvornik Municipal Assembly’, 7 March 2003, http://www.ohr.int/decisions/removalssdec/default.asp?content_id=30263 (accessed August 15, 2010).

Office of the High Representative, ‘Decision removing Mr. Bosko Lemez from his position in the Management Board of Elektroprivreda Republika Srpska’, 26 February 2003, http://www.ohr.int/decisions/removalssdec/default.asp?content_id=29338 (accessed August 15, 2010).

For a more detailed analysis of international efforts and the limits of their effects see Manning and Antic, ‘The Limits of Electoral Engineering’. Information on specific actions of the Office of the High Representative was obtained from its official website, http://www.ohr.int. Data on the distribution of OHR decisions were compiled by the author.

This situation did not hold for the Croatian presidential election in 2009 in which an independent candidate, Milan Bandić received more diaspora votes than the HDZ candidate. Republic of Croatia Election Commission, http://www.izbori.hr/2009Predsjednik/indexr.html (accessed August 15, 2010).

See International Crisis Group, Bosnia's November Elections: Dayton Stumbles, 18–19.

See Todorova, Imagining the Balkans.

For an even-handed discussion of the debate over Bosnian mentality as it relates to interethnic cooperation, see Bose, Bosnia after Dayton.

For the seminal discussion of consociational democracy see Lijphart, Democracy in Plural Societies. For a discussion of the application of consociationalism to the Bosnian case see Mujkic, ‘Bosna i Herzegovina i izazovi konsocijacije’.

This homogeneity is in many cases the direct result of successful ethnic cleansing. Serbs were a minority in the following municipalities (percentage of Serbs in 1981 is in parentheses): Bosanski Brod (33% of Serbs), Bosanski Šamac (41%), Bratunac (40%), Čajniće (48%), Derventa (40%), (Doboj 39%), Foča (42%), Kotor Varoš (41%), Modriča (38%), Prijedor (42%), Rogatica (37%), Srebrenica (28%), Višegrad (33%), Vlasenica (44%), Zvornik (41%). Source: Bertić, Veliki atlas Jugoslavije, 228–30.

Washington Agreement, 1994, Section II, http://www.usip.org/files/file/resources/collections/peace_agreements/washagree_03011994.pdf (accessed August 15, 2008).

See Bielasiak, ‘The Institutionalization of Electoral and Party Systems in Post-Communist States’.

Grzymala-Busse, ‘Encouraging Effective Democratic Competition’.

Davidson, ‘Ambassador Douglas Davidson's Remarks at Prof. Dr. Mirko Pejanović's Book Presentation’.

Manning, ‘Elections and Political Change in Post-War Bosnia and Herzegovina’.

USAID and NDI restrict their party assistance, consisting primarily of technical support for campaigning, to multiethnic parties. Center for Democracy and Governance, USAID Political Party Development Assistance. See note 20 for attempts to hinder nationalist parties.

While Bosnia is in the earliest stages of the EU accession process, the Office of the High Representative is also the EU Special Representative in Bosnia and Herzegovina. Following the heavy use of the High Representative's power to remove officials and enact legislation through 2003, there was a shift toward conditionality as a means of exercising influence over Bosnian politicians. For example, police reform and constitutional reform have been put forward as pre-requisites for steps toward EU accession.

Grzymala-Busse and Innes, ‘Great Expectations: The EU and Domestic Political Competition in East Central Europe’.

In fact, most studies question whether such a divided yet stable society is possible at all, notably Rabushka and Shepsle, Politics in Plural Societies.

Horowitz, Ethnic Groups in Conflict.

Ibid., 318.

Ibid.

Ibid., 347.

This pressure took the form of party assistance to the SDP, sanctions on the nationalist parties, and media campaigns targeted at nationalist incumbents (see International Crisis Group Bosnia's November Elections: Dayton Stumbles, 2–3). Pressure continued after the election, as the American and British ambassadors to Bosnia and Herzegovina engaged in ‘energetic lobbying and arm-twisting’ in order to create a coalition that excluded the war-time nationalist parties. International Crisis Group, Bosnia's Alliance for (Smallish) Change.

For results see http://www.izbori.ba, the website of the electoral commission for Bosnia and Herzegovina.

See Lijphart, Democracy in Plural Societies.

In Lijphart's later work, consociational systems are grouped under a broader set of consensus-based systems, Lijphart, Patterns of Democracy.

See Mujkic, ‘Bosna i Herzegovina i izazovi konsocijacije’.

Norris, Driving Democracy: Do Power-Sharing Institutions Work?

Garry, ‘Consociationalism and its Critics: Evidence from the Historic Northern Ireland Assembly Election 2007’.

Mitchell, Evans and O'Leary, ‘Extremist Outbidding in Ethnic Party Systems is Not Inevitable: Tribune Parties in Northern Ireland’.

Horowitz, Ethnic Groups in Conflict.

All precinct-level election data was obtained by the author from the Electoral Commission of Bosnia and Herzegovina (http://www.izbori.ba).

Since neither individual-level electoral return data nor suitable individual-level survey data are available, the data analysis here makes use of aggregate, precinct-level returns, which are subject to ecological inference (EI) problems. Ecological inference problems emerge when attempting to draw individual-level conclusions based on aggregate data. In this case, it is necessary to estimate the proportion of voters who split their tickets while not directly observing individual votes. Rather, it is only possible to observe the aggregate proportion of non-nationalist voters in each election. This study makes use of EZI (Benoit and King, EZI: An Easy Program for Ecological Inference) to generate estimates of the proportion of voters who split their votes between non-nationalist parties in cantonal elections and nationalist parties in FBiH Assembly or BiH Assembly elections. The results used in this analysis follow the recommendations in Adolph et al., ‘A Consensus on Second Stage Analyses in Ecological Inference Models’, in that the covariates for the second stage analysis have been included in the extended EI model, and observations are weighted according to the standard errors of the EI point predictions (whereby less certain predictions are devalued in the analysis relative to more certain predictions).

The combined totals for SDP, National Party – Work for Betterment, and the Liberal Democratic Party.