Generational replacement and voter turnout in Japan

A large body of research shows that generational replacement has contributed to long-term declines in voter turnout by combining cohort-specific participation patterns with changes in population composition. This research note uses a different approach to understand how high turnout would be in a given election if entering cohorts voted as actively as the cohorts they replaced, conditional on age and electoral context. Using age-by-election official turnout data from Japan, we propose an approach that estimates cohort-specific turnout functions varying across both the life course and elections using Generalized Additive Models. These estimates are combined with observed demographic change to construct age-matched, election-conditioned counterfactual turnout rates for individual elections. The results show that generational turnout gaps in Japan are large but not constant. They vary systematically by age and election year. Accounting for this variation yields more precise assessments of the electoral impact of generational replacement and indicates that cohort differences explain a substantial share of Japan’s recent turnout decline. The findings underscore the value of modelling generational effects dynamically when evaluating turnout change in aging democracies.