The Wheel of Time: An Age, Period, and Cohort Analysis of Voter Behavior in a Post-Socialist Context

Due to availability of longitudinal survey data, most studies of generational differences in voter behavior have so far been focused on Western Europe. This paper fills this gap by placing generational voter research in a specific post-socialist context of the so-called ‘New Europe’. The data used in this paper stems from Croatian National Electoral Studies (CroNES) which include surveys conducted from 1995 to 2020. The paper tests intergenerational differences regarding two voter behavior outcomes – voting for the center-right Croatian Democratic Union (HDZ, EPP member, government party 1990-2000, 2003-2011, 2016-) and the center-left Social Democratic Party of Croatia (SDP, S&D member, government party 2000-2003, 2011-2015). To assess different conceptualizations of generations in the context of voter behavior, the paper builds three models – one based on birth cohorts, one based on sociological generations, and a third one, based on political generations. The third approach was specifically designed for this paper, and groups voters in generations according to their political socialization in the pre-socialist, early socialist, late socialist, and post-socialist periods. All three models use gender, education, religiosity, political interest, party identification, and left-right self-identification as control variables. The paper deals with the problem of disentangling age, period, and cohort effects by coding age in three large groups (18-29, 30-64, and 65-99) and treating period (survey year) as a continuous variable. We test binary logistic regression models which are supplemented by generalized additive models (GAMs) for a smoothed-out birth year (age) effect. In addition, differences between individual cohorts, sociological, and political generations were examined using Wald tests. Finally, the generational effects were further highlighted through the assessment of marginal effects. The results point to a large probability of voting for HDZ among the earliest cohorts, i.e., members of the Greatest Generation, socialized before socialist rule (before the 1945-1989 period), as opposed to very low probabilities among the newest generations, i.e., Millennials and Generation Z, socialized after democratic transition. This generational gap goes together with differences in education and religiosity levels, with the earliest cohorts more religious and less educated than newer generations. Wald tests demonstrated that differences across all three models (cohorts, sociological generations, and political generations) can be observed regarding voting for HDZ, whereas voting for Social Democrats could be mostly explained by sociological generations. Voters born in the 1940s can be expected to be least likely to vote for the Social Democrats, while earliest cohorts, those born at the beginning of the 20th century, have a higher likelihood of supporting the Christian Democrats.