The partisan effects of low turnout: Analyzing vote abstention as a missing data problem

We propose a novel method for estimating the partisan effects of low voter turnout. Treating non-voters as missing data points on a variable measuring vote choice, we use a statistical model of multiple imputation to estimate the missing values on reported party choice at 28 elections in 25 countries in the Comparative Study of Electoral Systems dataset. Calculating the differences between observed and 'complete' (100%) turnout, we identify party and country-level conditions under which changes in turnout affect electoral outcomes. We also provide simulations of election outcomes for incremental changes in turnout. While our findings suggest few systematic gains from increased turnout for left-of-centre parties, we show that small parties and non-incumbents would benefit if everybody voted.