Dr. Elliott Fullmer, Associate Professor of Political Science, answers questions about the research behind his most recent book, “Tuesday’s Gone: America’s Early Voting Revolution.”

Your research on the subject of early voting—and the trend toward early voting—actually began many years ago. Tell us more.

The first early voting laws emerged in the late 1970s and 1980s, though they did not immediately attract many citizens. After the 1992 elections, we started to see gradual increases; by 2004, about 20% of the country’s voters were casting ballots before Election Day. Much of this happened with little fanfare, as early voting received little national attention until the 2008 election, when Barack Obama’s presidential campaign so effectively used it to mobilize supporters.

Due to its growing importance, I became interested in early voting in 2009, and it became my dissertation topic at Georgetown University. I am glad it did, as it has only become more of a factor in U.S. elections since.

What is the goal of early voting?

Most early voting laws were adopted with bipartisan state legislative majorities, and the reasons were typically twofold. Reformers hoped that early voting could improve turnout. Secondly, there was pressure on states to ease the election administration process for county officials who oversee elections. By spreading out voters over a longer period of time, administrators are better able to control lines and reduce wait times. In addition, administrators can identify potential problems during the early voting period before a flood of voters arrives on Election Day.

Has it worked?

Yes, but it’s complicated! Much of the initial literature on early voting suggested that it was actually having little effect on turnout. One of the principal findings of my research is that early voting does increase voter participation when counties offer a sufficient number of locations. It turns out that even within states that offer early voting, offerings at the county-level can vary quite a bit. Controlling for a number of other known turnout predictors, I consistently find that an additional early voting site per 1,000 voting-age residents increases county turnout by at least 2-3 percentage points. These findings are consistent across the 2012, 2014, 2016, and 2018 U.S. elections. The size of the effect is not earth-shattering, but it is important to note that three percentage points equates to over 4 million additional voters in a presidential election. 

Furthermore, I find that turnout effects are even greater—sometimes as much as eight points—in urban counties.
On this point, it is worth acknowledging that there are many reasons why people vote or do not vote. Early voting, like any electoral reform, should not be expected to bring every non-voter to the polls. But when casting a ballot is burdensome for an interested voter, it can make the difference between participating and not participating.

Regarding election administration, county officials—and I’ve spoken to over 100 of them—generally affirm that early voting does help identify problems and ease the burdens of Election Day. But they will also acknowledge that it adds new stressors. For example, counties with small staffs can find it difficult to manage multiple voting locations for several consecutive weeks. All administrators seemingly agree, however, that voters like early voting and become accustomed to it once it is offered. 

But you write that there are unintended consequences. What are they?

Anytime you change an election law, there is the potential for unintended consequences. The most notable such consequence that I find is that early voting sites are inequitably distributed across the U.S. on the basis of race. Simply put, counties with larger Black populations have fewer sites per capita. This is concerning for two reasons. First, it means that early voting is widening disparities in voting access between racial groups. Second, it suggests that early voting is not yet reaching its true potential. Given that we know early voting sites bring higher turnout, more sites in heavily Black counties (where too few currently exist) should bring additional increases in participation.

Another unintended consequence that I examine involves down-ballot roll-off, which occurs when voters cast a vote for high-profile offices (e.g. president or governor), but leave lesser known offices (e.g. county commissioner) blank. I find that early voters are a bit more likely to engage in roll-off, likely because they are casting ballots before down-ballot candidates (in many cases) even begin campaigning. The effect is much greater in non-partisan elections where there is no party cue to guide uncertain voters.

In your opinion, could broader early voting be good for democracy? How would it need to be implemented, or what would need to change to make it so?

Absolutely. Early voting has already increased turnout and made voting easier for millions of regular voters. It has also meaningfully improved election administration across the U.S. Although there are unintended consequences, I believe they can be addressed. For example, the federal government can and should take a proactive role to ensure that early voting sites are equitably distributed across the country. Debates between candidates should probably begin earlier (for general elections) to accommodate the fact that more citizens are casting early ballots. I acknowledge early voting’s unintended consequences not because I believe they suggest that programs are futile. Rather, I do so in order to identify issues that demand attention to ensure that early voting can achieve its true potential.