Abstract

Political behavior has been changing all over Western Europe and electoral volatility is one of the facets of politics in which this change is apparent. Theories on stabilization of political attitudes and behavior in lifetime and the slow rate at which change in the electoral arena is found to proceed, have led to the assumption of generational replacement as the mechanism driving change. The Netherlands, however, provide a remarkably different case of this trend in electoral volatility. The country has shifted from an example of how cleavages stabilize politics to one of the most electorally volatile countries in Europe. The Dutch surge in electoral volatility thus contrasts with expectations of a slow process driven by generational replacement. Starting from this apparent contradiction between the evolution of volatility in the Netherlands and theories on generational replacement, this article investigates time effects of electoral volatility. The study is based on an age, period and cohort analysis on the repeated cross-sectional data of the Dutch Parliamentary Election Studies, 1971–2010. Based on characteristics of such repeated cross-sectional data, individuals are cross-classified in birth cohorts and election years respectively, which overcomes the identification problem inherent in cohort analyses. Results of a Cross-Classified Random Effects Model (CCREM) indicate that, contrary to the hypothesis of new generations causing the increase in volatility, the Dutch change can be attributed primarily to period effects. As such, the analyses indicate that a general shift in the Dutch electorate has caused the growth in volatility and that supply-side factors should probably be analyzed when trying to explain electoral volatility.