The New Jim Crow: An Analysis of the Effects of State Voter Identification Requirements on Voter Turnout

Abstract

Voter identification requirements can be categorized using two classifications. The two classifications used to categorize the restrictiveness of voter identification requirements are not mutually exclusive and may overlap. First, voter identification requirements can be separated by whether state law requires a registered voter to present either photo identification or non-photo identification to a poll worker to vote on Election Day. As of September 2021, 17 states currently require voters to provide photo identification, while 19 require voters to provide non- photo identification to vote on Election Day. I have conducted an evaluation of how, if at all, changes in voter identification laws in selected states affected voter turnout. For my evaluation, I have conducted a quasi-experimental analysis similar to that found in the 2014 Report to Congressional Requesters (GAO) on issues related to state voter identification laws. However, my analysis goes further than that of the 2014 Report to Congressional Requesters (GAO). My analysis compares the changes in voter turnout between two states that implemented new voter identification requirements with changes in voter turnout in four states that did not implement new voter identification requirements between the 2012 and the 2016 general elections. I chose to analyze the effects of voter identification requirements on voter turnout between the 2012 and the 2016 general elections as there is a clear knowledge gap within the public policy field. Prior to this paper, zero public policy-oriented academics have attempted to analyze the effects of voter identification requirements on voter turnout between the 2012 and the 2016 general elections. In addition to closing the knowledge gap, my analysis also attempts to control for factors outside of changes in voter identification requirements. Based on the quasi-experimental analysis, voter turnout decreased in Mississippi but not Virginia from the 2012 and 2016 elections to a greater extent than in the comparison states- Connecticut, New Jersey, New York, and Pennsylvania. The analysis suggests that the decreases in voter turnout in Mississippi compared to the comparison states were attributable to changes in the states’ voter identification requirements. Voter turnout decreased in both New Jersey and New York, remained constant in Connecticut, and increased in Pennsylvania between 2012 and 2016. Among the treatment states, voter turnout in Mississippi declined by a larger amount than in the four comparison states, whereas voter turnout in Virginia remained relatively unchanged. Compared to the changes in turnout in all the comparison states combined, I estimate that turnout for eligible voters declined by an additional 6.4 percent and increased by 0.8 percent in Mississippi and Virginia, respectively.