To Ask or To Simulate? Estimating the Political Effects of Low Turnout at the 2009 European Parliament Elections

Abstract

The political effects of low turnout at elections are a recurring theme in politics and political science. Particular interest concerns the possibility that certain parties might gain and others lose from low voter turnout. Different methods have been proposed to estimate these effects. Some authors ask self-reported non-voters in surveys how they would have voted had they voted. Others predict these hypothetical vote choices statistically. We contribute to the discussion by simulating the vote choices of abstainers with the help of multiple imputation, using data from the 2009 European Election Study. The simulations show that numerous parties would have incurred considerable gains and losses compared to the actual results if the turnout at the 2009 EP election had reached the levels recorded at the preceding first-order election in each member state. We also compare our counterfactual election results with findings based on the reported hypothetical vote choices of non-voters in the 2009 European Election Study. Estimates of turnout effects based on multiple imputation are on average larger than those based on non-voters’ reported hypothetical vote choices. The differences depend neither on the size of the turnout effects nor on the number of parties.