Socialized with “old cleavages” or “new dimensions”: An Age-Period-Cohort Analysis on electoral support for mainstream-left and mainstream-right parties in Western Europe

Across Western Europe, both parties on the center-left and parties on the center-right have suffered considerable electoral losses in past decades. Research has repeatedly demonstrated that these parties are less appealing to younger voters. However, it is still unclear how persistent these age patterns are and what causes them. Are they representing an age (i.e., life-cycle), a cohort (i.e., generational), or a time-specific (i.e., period) difference? In this paper, I thus look at the electoral decline of Western European mainstream parties over the past three decades by adding a cohort perspective to the downtrend of those parties. I analyze six Western European countries’ cohort differences and aim at detecting both country differences, as well as a general pattern. Having said this, I expect that the electoral crisis of mainstream parties is due to a continuous and persistent decline of appeal among birth cohorts that have grown up in more recent decades. These cohorts experienced their formative years in an era in which issues from the socio-cultural dimension of politics were much more the subject of political conflicts. The socialization experience within this political Zeitgeist hints that more recent cohorts have developed a different hierarchy of dimensional salience with issue bundles from the socio-cultural dimension ranking higher on that ladder. Western European mainstream parties, however, are less connoted to this dimension of politics and should be persistently less popular among newer cohorts, especially among those socialized in the 1980s, the 2000s, and 2010s. I thus expect that age differences in Western European electorates are representing cohort differences and not lifecycle or period differences. Furthermore, I argue that these differences are rooted in persistently diverging perceptions of salience that cohorts have developed. To disentangle the three different dimensions of age, I use data stemming from the European Values Study (1981-2018), as well as individual-level data from national election studies. I apply APC-logistic regressions relying on a generational scheme that fits the Western European context. I also test how the perceived salience of socio-cultural issue-bundles accounts for cohorts’ diverging party preferences. Finally, to test for the robustness of the generational scheme and to detect empirically existing boundaries, I use generalized additive models (GAMs).