Electoral Turnout in Central and Eastern Europe: The Legacies of Communist Socialisation

Abstract

This article assesses the impact of individual-level experiences of political socialisation under communist rule on the propensity to vote in elections after 1989 in eight Central and Eastern European countries. Using CSES and CEU data, it sheds light on the heterogeneity of the effects of communist socialisation on electoral turnout, showing how these effects are jointly dependent on the coerciveness of the previous regime and individual religious observance. While exposure to authoritarian communist socialisation reduces the likelihood of electoral participation independently of religiosity, socialisation in totalitarian times only has a negative effect on the voting propensity of frequent churchgoers.

Notes

1 Lack of trust in political parties is often used as an example. Nonetheless, Pop-Eleches and Tucker (2011, p. 392) discovered that the simplest explanation for the trust gap in CEE—that living under communism made individuals less likely to trust political parties—was not supported by empirical evidence. These authors found that those who had lived longer under communism reported significantly higher party trust. Moreover, Tworzecki (2008) points out that if the communist legacy was behind the lack of trust in political organisations in Poland, it would be difficult to explain spontaneous massive and institutionalised civic actions such as the Solidarity movement in 1980–1981.

2 On political culture under communist regimes, see Brown and Gray (1979), Brown (1984); on political culture in postcommunist regimes see, Pollack et al. (2003).

3 For a more detailed list see, Ekiert and Ziblatt (2013, p. 106, fn. 50), Simpser et al. (2018).

4 Kitschelt et al. (1999, p. 12) expose the falseness of the ‘dichotomy of explanations for democratic process features between “legacies” of the old regimes and “institutions” of the new democracies’. Instead, they propose a sequential model in which institutions are initially endogenous to the political process but in time become exogenous.

5 This framework is based on the idea that ‘an additional year of exposure should be correlated with additional support for the pro-regime attitude’ (Pop-Eleches & Tucker 2017, p. 6).

6 For an account of the interesting case of the Polish referendum of 1987, in which Solidarity again campaigned for a boycott see, McManus-Czubińska et al. (2004).

7 This study will not focus on religious denomination as the majority of churchgoers included in the analysis are Catholics, mostly Roman but also some Eastern Orthodox. The German Democratic Republic is excluded from the analysis. In some studies of political behaviour in postcommunist countries, East Germany is included. Due to the specificity of its democratic transition, by incorporation of the GDR into the Federal Republic of Germany, East Germany is not optimal for broader comparative inquiries. This approach in the selection of countries is used in many studies, see, for example, Cześnik (2007, pp. 66–7) and Neundorf (2010b).

8 The case of Poland is particularly noteworthy, given the strength of the Polish Catholic Church. Even under communism, the Polish Church did not hesitate to use its moral authority to influence policy (Grzymała-Busse 2015), offering an alternative source of authority within society, dwarfing the mobilisation efforts of the communist regime during the totalitarian period and laying a moral foundation for opposition efforts during authoritarian rule (Casanova 1993). This was not the case for the hardline regimes of Romania and Czechoslovakia, where the Orthodox and the Catholic Church respectively did not provide a moral alternative to the communist regime (Judson 1999).

9 Still, it is necessary to acknowledge that the state-enforced secularisation during communism was at least partially successful and according to some accounts a slight religious revival can be observed in some postcommunist countries (Müller & Neundorf 2012).

10 The differences between reported and actual turnout tend to be considerable, but they are comparatively lower in post-electoral surveys carried out shortly after the elections (such as CSES), as there is evidence that over-reporting rates increase as time passes between the election day and the date the respondent is interviewed (Belli et al. 1999). Furthermore, it should be noted that the Slovenian elections of 1996 were omitted from the analysis due to under-reporting of over 15% since all other elections were characterised by over-reporting.

11 Elections included are Bulgaria (2001), the Czech Republic (1992, 1996, 2002, 2006, 2010, 2013), Hungary (1998, 2002), Latvia (2011, 2014), Poland (1991, 1993, 1997, 2001, 2005, 2007, 2011), Romania (1996, 2004, 2012), Slovakia (1992, 1994, 2010, 2016) and Slovenia (2004, 2008, 2011).

12 With regard to the latter, it has been put forward that the direction of the causal relationship is inverse—that voting boosts satisfaction with democracy rather than the other way around (Kostelka & Blais 2018). Nonetheless, for the purpose of this study—assessing the effects of communist socialisation—this is not a caveat, given that, independently of the direction of causality, including a measure of satisfaction with democracy controls for a relationship inherent to the democratic period, posterior to the exposure to communist socialisation.

13 Ideology is not included as control given that data for many respondents are missing, although it could be argued that, especially in the model testing for H4, it should be accounted for. I reran the models including a measure of left/right self-placement, and the results (available upon request) were robust.

14 See, for example, the special issue of Electoral Studies, 33 (2014) dedicated to this subject, especially its editorial, for a thorough description of the APC approach.

15 Given that CSES data are unbalanced, modelling APC cross-classified models would be feasible only for the cases of Poland and the Czech Republic, limiting the scope of the study.

16 As was already mentioned, due to data unavailability, I rely on self-reported church attendance at the time of the survey as a proxy for prior church attendance under communism.

17 The only changes are that the negative coefficient in model 1 in Tables A3 and A4 becomes significant; the negative effect of the interaction term of totalitarian socialisation and religious services attendance is significant at 0.1 (before 0.05), and the interaction term of authoritarian socialisation and religious services attendance becomes significant at 0.1.

18 The results controlling for ideology and both measures are available upon request.

19 The latter effect remains significant at the 0.1 level.

Additional information

Notes on contributors

Piotr Zagórski

Piotr Zagórski, Autonomous University of Madrid, Department of Political Science and International Relations, Faculty of Law, c/ Marie Curie, 1, Ciudad Universitaria de Cantoblanco, 28049 Madrid, Spain. Email: [email protected]