Estimating the unintended participation penalty under top-two primaries with a discontinuity design

Participation in U.S. elections lags behind most of its developed, democratic peers. Reformers seeking to increase voter turnout often propose changes to the electoral system as means of addressing these shortcomings. One such reform, the top-two blanket primary, has been adopted in California and Washington in part to boost voter participation. Despite the promises of reformers, however, observers disagree as to its efficacy. In this paper, I estimate the participation penalty generated by top-two primaries using a regression discontinuity design (RDD). I estimate that general elections featuring two members of the same party – the arrangement reformers contend would increase turnout – actually decrease voter participation. I find that approximately 7% of voters “roll off” the ballot in the absence of party competition while overall turnout is unaffected. These results suggest that top-two primaries are likely to exacerbate rather than ameliorate trends in participation.