Strict voter identification laws and turnout: Differential effects by election type and adoption timing

Abstract

Voter identification (ID) requirements remain a contested policy issue in American elections, with existing research producing mixed findings about their causal effects on voter turnout. Using state-level turnout data from 1984 to 2020 and synthetic difference-in-differences to control for time-varying confounders and increase statistical power, I examine the impact of strict voter ID laws across presidential and midterm elections and between early-adopting and late-adopting states. In aggregate, the laws have no significant effect on turnout. However, the disaggregated analysis reveals heterogeneity that may explain the conflicting findings in the literature. In presidential elections, late-adopting states—those passing laws after 2008—experience a 2.7 percentage point decrease in turnout, while early-adopting states—those passing laws before 2008—are not significantly affected. For midterm elections, there is no evidence of decreased turnout; counterintuitively, the main analysis suggests an increase of 2.9 percentage points, with larger effects among early-adopting states, though this finding is not robust across all specifications. Overall, these results indicate no compelling evidence that strict voter ID laws consistently suppress or boost voter turnout. Rather, the modest and opposing impacts across contexts appear to net out to minimal aggregate effects on participation, consistent with research showing that electoral reforms often produce small equilibrium effects.