Does Voting by Mail Increase Participation? Using Matching to Analyze a Natural Experiment

Abstract

This paper examines how the growth in vote-by-mail and changes in voting technologies led to changes in the residual vote rate in California from 1990 to 2010. We find that in California's presidential elections, counties that abandoned punch cards in favor of optical scanning enjoyed a significant improvement in the residual vote rate. However, these findings do not always translate to other races. For instance, find that the InkaVote system in Los Angeles has been a mixed success, performing very well in presidential and gubernatorial races, fairly well for ballot propositions, and poorly in Senate races. We also conduct the first analysis of the effects of the rise of vote-by-mail on residual votes. Regardless of the race, increased use of the mails to cast ballots is robustly associated with a rise in the residual vote rate. The effect is so strong that the rise of voting by mail in California has mostly wiped out all the reductions in residual votes that were due to improved voting technologies since the early 1990s. A decade ago, the nation became aware that voting machines are not simple ciphers through which voters cast their ballots. Palm Beach County, Florida provided the best illustration of how machine malfunction-exemplified by "hanging" and "pregnant" chad-and poor ballot design-exemplified by the "butterfly ballot"-could result in a vote being miscounted, if counted at all (e.g., Sinclair et al. 2000; Smith 2002; Wand et al. 2001). The Florida fiasco resulted in a strong public demand for improved voting technology, and led to a flurry of new research into the causes of "lost votes" due to voting technologies. 1 Within political science, this research has focused on explaining the residual vote rate in presidential elections as a function of the type of voting technology used by voters. (The residual vote rate is the percentage of ballots cast that either contain an over-or undervote for a particular race.) Making this research particularly important was the simultaneous emergence of public demand for improved voting technology to reduce the residual vote rate. This was made explicit by the Help America Vote Act which mandated retirement of older technologies and resulted in billions of state and federal dollars being spent to retire old voting machines.

Convenience Voting Can Affect Election Outcomes

“…The first approach follows Kousser and Mullin (2007) and uses the full set of precincts but preprocesses the data using Mahalanobis distance matching (Ho et al 2007). The second approach, which is based upon a regression discontinuity design (RDD), compares the outcomes we observe in precincts that are just above the 250-registrant threshold with outcomes we observe in precincts that are just below the 250-registrant threshold.…”

“…In this study, we leverage an election administration policy in California, previously identified by Kousser and Mullin (2007), which produces some variation in the use of VBM that approximates such an experiment. 5 Briefly, the California Election Code permits county elections officials to require all voters to use VBM if a precinct contains 250 or fewer registered voters (hereafter, ''VBM-only'' precincts).…”

Estimating Voter Registration Deadline Effects with Web Search Data

“…The effects of electoral rules have acquired new relevance in the wake of the Supreme Court decision in Shelby County v. Holder (2013), which invalidated a key section of the 1965 Voting Rights Act, allowing many more jurisdictions to amend electoral rules without federal oversight. As Kousser and Mullin (2007) observe, the United States is cross-cut by boundaries for elections at the local, state, and federal levels. Besides variation in registration deadlines, U.S. elections differ in myriad ways, such as the presence or absence of voter ID laws (Bentele and O'Brien 2013), the availability of mail ballots, sample ballots, and other information (Wolfinger, Highton, and Mullin 2005;Kousser and Mullin 2007), or the accessibility of polling places (Brady and McNulty 2011).…”

“…As Kousser and Mullin (2007) observe, the United States is cross-cut by boundaries for elections at the local, state, and federal levels. Besides variation in registration deadlines, U.S. elections differ in myriad ways, such as the presence or absence of voter ID laws (Bentele and O'Brien 2013), the availability of mail ballots, sample ballots, and other information (Wolfinger, Highton, and Mullin 2005;Kousser and Mullin 2007), or the accessibility of polling places (Brady and McNulty 2011). The onus is on the voter to find out about the details, and searching online is now the obvious way to do so.…”

“…The mean estimate in publications through the 1990s was that allowing EDR or keeping registration open through Election Day would produce a 6.4% point increase in turnout, but the mean estimate since 2000 is 3% points. Keele and Minozzi (2013) provide a lucid discussion of the difficulty of estimating the effects of voter registration laws (see also Kousser and Mullin 2007). Early studies used cross-sectional comparisons of state turnout, with the effects of registration laws estimated with a dummy variable for EDR states or a measure of the number of days when registration was closed (Rosenstone and Wolfinger 1978;Nagler 1991).…”