The coronavirus pandemic threatens the 2020 U.S. election. Many fear that the pandemic could deter many people from voting — or expose them to infection if they do. And so officials and observers around the country are calling for a nationwide expansion of voting by mail, giving Americans the opportunity to vote safely from their homes.

But while both Republican and Democratic local, state and national officials are among those calling for expanding voting by mail, other national party leaders argue that doing so would help their opponents. President Trump claimed that if the entire nation voted by mail, “You’d never have a Republican elected in this country again.” And some Democratic leaders worry that some Democratic-leaning voters might even be less likely to vote by mail.

In a new study, we evaluate these claims. We find that voting by mail does not give an advantage to either party.

How we did our research

We focused on California, Utah and Washington, three U.S. states with many or all counties voting entirely by mail. These states implemented their systems in a staggered fashion across counties, a few at a time. While all counties in Utah and Washington have now fully adopted voting by mail, California is still rolling out its program, with 14 out of 58 counties choosing to implement it so far. We collected election results and turnout numbers from each county in these states from 1996 through 2018. Our goal was to see whether switching to mail-in ballots changes who votes.

The states that have shifted to voting by mail differ systematically from those that have not; in particular, they’re more liberal. That means that if we compared elections in states that primarily vote by mail to states that don’t, we’d find big differences in voting patterns. But mail-in balloting didn’t cause states to be more liberal, although being more liberal might have led them to adopt widespread voting by mail.

To get around this, we collected data on how switching to voting by mail changed (or didn’t change) voting patterns for each county over time. We then compare how voting patterns change in counties that do move to voting by mail with counties within the same state that had not yet adopted it. This lets us be more confident that we are isolating the true effect of mail-in voting, rather than measuring preexisting differences between places that adopted it and those that did not.

Voting by mail does not give an advantage to either party

We find that voting by mail does not result in more votes for Democratic candidates; nor does it lead to having Democrats make up a larger fraction of the electorate, as measured by voters’ party registration.

Our best estimates show that when a county shifts to voting by mail, the change in the share of Democratic voters increases by about 0.1 percentage points. That’s a very small effect.

We also find that voting by mail does not increase either party’s share of the vote in elections. Our estimates make it quite clear that voting by mail would not lead to a major, permanent electoral shift toward either party.

Voters seem to appreciate the opportunity to vote from home

We also find that voters seem to appreciate the opportunity to cast a ballot by mail. In California, the 14 counties that have adopted voting by mail send each registered voter a ballot by default. But if voters want to, they can instead still vote in person in their county on Election Day. Mailing each voter a ballot increases the share of voters who vote from home by roughly 17 to 20 percentage points. While not surprising, this shows that large numbers of voters take advantage of the option of voting at home and mailing in their ballots.

What does this tell us about voting by mail during this pandemic?

Before interpreting our findings, a few caveats are in order. First, our evidence comes from counties that chose to launch voting-by-mail programs during normal times, and we only compared their results to those from a normally administered in-person election. Expanding voting by mail during a pandemic could be quite different; results could vary based on whether fear of in-person voting during the pandemic disproportionately deters Democrats or Republicans.

Second, we studied the effects of rolling out voting by mail in three states. Our results may not generalize to all states and to the nation at large. U.S. states and localities administer their elections very differently, making it hard to know for sure how the shift would affect all those elections in all those places. That said, however, California, Utah and Washington are three very different states with very different politics — and in all three states, shifting from in-person to by-mail voting had remarkably neutral partisan effects.

Finally, there may be other reasons to worry about shifting to voting by mail nationwide. Those might include the expenses associated with administering the program; potential disparate impacts on minority voters; and the technical implementation challenges recently detailed here at TMC by Barry Burden, Robert M. Stein and Charles Stewart III.

Implications for 2020

Despite these caveats, our study has a clear takeaway: Voting by mail does not give either party a fundamental advantage over the other. Based on our data at least, voting from home modestly increases participation roughly equally for both parties. That’s consistent with the conventional wisdom of election administration experts.

Daniel M. Thompson (@danmthomp) is a PhD candidate in political science at Stanford University.

Jennifer Wu (@imjenwu) is a predoctoral fellow at the Stanford Institute for Economic Policy Research.

Jesse Yoder (@jesselyoder) is a PhD candidate in political science at Stanford University.

Andrew B. Hall (@andrewbhall) is a professor of political science at Stanford University and a senior fellow at the Stanford Institute for Economic Policy Research.