Registration and Turnout Effects of Automatic Voter Registration

• In Event: Making It Easier to Vote

Abstract

While much of the focus in the 2020 election was on mail voting as a way to address the pandemic, another reform saw a significant expansion: automatic voter registration. Automatic voter registration (AVR) laws take advantage of transactions at government agencies (such as the DMV) where applicant information can be captured and repurposed to register citizens to vote. Six states implemented AVR for the first time, on top of the 12 that had implemented already.

AVR raises fundamental questions about the participatory potential of America’s large unregistered population. Much scholarship has questioned whether modern registration reforms can truly expand the electorate, or if those who are not registered to vote are instead disengaged from politics in a way that is difficult to address. These theoretical questions have significant practical implications, as there is now serious conversation around legislation at the federal level to mandate AVR for all states. Understanding its impact on registration and turnout is essential.

In this paper we test the effect of AVR on registration and turnout with a difference-in-differences research design. In contrast to most of the existing literature, which works with state-level data, we dramatically increase statistical power with county turnout and registration aggregated to PUMAs to permit using the eligible population as a denominator. We find that AVR has a positive effect on both registration and turnout. Moreover, when we change the analysis to turnout as a share of the registered population, we find no evidence that registrants in AVR states are significantly less likely to vote. However, there are some signs that AVR was less effective in 2020, when the pandemic made it difficult to access the DMV.

Authors

• Eric McGhee, Public Policy Institute of California

• Mindy Romero, USC