Estimating the Electoral Effects of Easing Ballot Access

• In Event: Making It Easier to Vote

Abstract

During the 2020 election, many states and localities eased ballot access through expanding early voting hours, facilitating voting by mail, and so forth, in response to the coronavirus. While these changes may have directly impacted turnout, it is plausible that these changes also impacted the aggregate vote shares. Hansford and Gomez (2010) introduced an instrumental variables method for estimating how changes in turnout impact changes in vote shares. They confirmed that Democrats tend to fare better in higher turnout elections. We extend their method and apply it to presidential election voting in 2020. We also contrast findings based on this method with two alternative approaches drawn from the study of compositional data. The models we selected have been applied to multiparty elections by Katz and King (1999), Tomz et al (2002) and Mebane and Sekhon (2004), and were applied to U.S. congressional elections by Basinger, Cann, and Ensley (2012, 2017) who treat non-votes as a third "party" alongside Democrats and Republicans, in order to examine how campaign spending impacts turnout and vote shares simultaneously. While the empirical analysis looks backward at the 2020 election, its' findings allow us to look forward to the potential impact of future restrictions on voting rights that Republican legislators in many states have already proposed since Trump's defeat.

Authors

• Scott J Basinger, University of Houston

• Michael J. Ensley, Kent State University