War as an Enduring Determinant of Party Choice in Postwar Southeast Europe

Abstract

In postwar elections, voter choices are often shaped by the memory of past violence. Taking Bosnia & Hercegovina and Croatia as case studies, this study examines war as an enduring determinant of party choice among the age cohorts who lived through the wars of the 1990s, and the cohorts who were born after. Based on a representative survey of over 5,000 citizens, the results show that in Bosnia & Hercegovina, war-related issues and social divisions continue to inform party preferences in the postwar generation as much as they did in the generations that came before. In Croatia, by contrast, war-related issues are showing signs of diminishing political relevance.

Disclosure statement:

No potential conflict of interest was reported by the author(s).

Notes

1 See also Fabbe et al. (2019).

2 There is also the life cycle effect, which results in a change in attitudes and in the relevance of social divisions as people age. However, it is not believed to apply to the topic at hand and therefore is not discussed here.

3 Target users were first divided into different predefined population weights (based on age, gender, education and location), drawn from the official national statistics for each country. Users were selected and recruited using Facebook’s random algorithm—they logged into the online survey app via their Facebook account to verify a unique ID. Each user was selected at random in order to fill in the prespecified quota.

4 Specifically, 238 strata were identified in Bosnia & Hercegovina, and 294 in Croatia.

5 There are serious concerns that these respondents speeded through the survey without carefully reading and answering the questions, rendering their answers meaningless and uninformative.

6 This sample was achieved after excluding 1,118 respondents for not answering all necessary questions, giving false answers or answering too quickly.

7 See also, The Bosnian Book of the Dead (Sarajevo, Research and Documentation Center, 2007).

8 Figure A1 in the Appendix visualised this process (also see Van der Eijk et al. 2006).

9 After excluding respondents with missing values for the independent variables, the number of unique respondent*party combinations remaining was 25,866 in Bosnia & Hercegovina and 29,900 in Croatia.

10 The statements capturing nationalism among voters are based on the works of Blank and Schmidt (2003) and Davidov (2011), who argue that nationalism is the idealisation of the nation and belief that one’s own nation is superior to others. This is sometimes also referred to as ‘blind patriotism’ (Adorno et al. 2019).

11 ‘Population by Ethnicity, by Towns/Municipalities’, Croatian Central Bureau of Statistics, 2011 Census, available at: https://www.dzs.hr/Eng/censuses/census2011/results/htm/E01_01_04/e01_01_04_RH.html, accessed 2 December 2020; ‘Census of Population, Households and Dwellings in Bosnia and Herzegovina’, Agency for Statistics of Bosnia and Herzegovina, 2013, available at: https://www.dzs.hr/Eng/censuses/census2011/results/htm/E01_01_04/e01_01_04_RH.html, accessed 2 December 2020.

12 To further clarify the $\hat{Y}$ calculation process, the variable ‘war experiences’ contains the predicted PTV scores derived from an OLS regression model in which the actual PTV scores of a specific party are explained by the five independent variables explained in the main text: being a war veteran, having been in physical danger during the war, having been forced to emigrate because of the war, having close family or friends killed during the war, suffering from a war-induced physical disability, and the degree to which one is suffering from war-related trauma.

13 See also Džankić (2015).

Additional information

Funding

The research for this article was conducted at the Université du Luxembourg and was supported by the European Research Council under grant number 714589.

Notes on contributors

Christophe Lesschaeve

Christophe Lesschaeve, Faculty of Social Sciences, Department of Political Science, University of Antwerp, Sint Jacobstraat 2–4, 2000 Antwerpen, Belgium. Email: [email protected]