Does Voting History Matter? Analysing Persistence in Turnout

Abstract

Individuals who vote in one election are more likely to vote in the next. Yet modelling the causal relationship between past and current voting decisions is intrinsically difficult, as this positive association can exist due to habit formation or unobserved heterogeneity. This article overcomes this problem using longitudinal data from the British National Child Development Study (NCDS) to examine voter turnout across three elections. It distinguishes between unobserved heterogeneity caused by fixed individual characteristics and the initial conditions problem, which occurs when voting behavior in a previous, but unobserved, period influences current voting behavior. It finds that, controlling for fixed effects, unobserved heterogeneity has little impact on the estimated degree of habit in voter turnout; however, failing to control for initial conditions reduces the estimate by a half. The results imply that voting in one election increases the probability of voting in a subsequent election by 13%.