Mail-In Voting in the U.S.
By. Dr. Clara Suong, Lab Deputy Director for Computational Social Science Initiatives
Recently, mail-in voting has been under the media spotlight. Due to the COVID-19 pan- demic, demands for mail-in voting are expected to be significantly higher for this election. Political elites on both sides of the aisle have been busily reacting to this trend. President Trump has railed against mail-in voting, arguing that voting by mail will lead to fraud and admitting that he intends to disrupt mail-in voting (from which he later backtracked). Commentators on the left as well as the U.S. Post Office are fretting or stating that the mail-in voting may not be available as widely as possible and that many mail-in ballots and votes will not arrive in time to be used or counted, due to possibly slow delivery by the U.S. Postal Service. This may be exacerbated by the expected volume of last-minute requests and state laws incongruent with the Postal Service’s expectations as well as cost-cutting measures at the Postal Service,some of which are to be implemented before and others delayed to be implemented after the election. In fact, few states are ready for the increase in demand for mail-in voting, according to the Brookings Institute.
The unsaid but obvious assumptions here are: that Republicans prefer to vote in person whereas Democrats prefer to vote by mail; and that availability and timely delivery of mail- in ballots are crucial for the GOP’s failure and the Democratic Party’s success because they would affect voter turnout and, consequently, the electoral outcome. In other words, there are two parts to the underlying logic: first, citizens’ partisanship affects their preferred mode of voting; second, different modes of voting affects overall turnout, enough to result in different electoral outcomes. What are political scientists’ takes on these statements? How are these issues discussed in the academic literature inpolitical science?
Voter Turnout and Voting Costs
First, let’s examine why voter turnout is important. It is not surprising that the key issue is the voter turnout (or whether the votes get counted), rather than the vote choices, in these debates about mail-in voting, given the lowvoter turnout and the cost of voting. In fact, voting itself may not be necessarily “profitable” for a self-interested, hyper-rational homo economicus voter in elections with non-zero voting costs. Voting costs time (which can be used to generate income), that is, time to register, deliberate, go to the polls, and mark the ballot (Downs, 1957, 265). As Anthony Downs illustrated in his seminal book An Economic Theory of Democracy, “when there are costs to voting...rational abstention becomes possible even for citizens who want a particular party to win...[and]...even low voting costs may cause many partisan citizens to abstain” (265).
Many political scientists have lamented the low and unequal turnouts in the U.S. The U.S. has one of the lowest turnout rates, with only about 56% of the U.S. voting-age population voting in the 2016 presidential election, compared to Belgium (87.2%), Sweden (82.6%) and Denmark (80.3%). (note: scholars have even proposed voting to be made mandatory (Hill, 2006).) This implies that electoral outcomes often hinge on which party can mobilizemore voters more effectively than their counterpart, i.e., voter turnout by party.
Efforts to Lower the Costs of Voting and Election Administration
In this context, mail-in ballots can boost the voter turnout in principle by lowering the costs of voting, saving the voters time spent going to the polls and waiting in line to mark the ballot (hence the income lost) on Election Day, which is a weekday in the U.S. and not a holiday like many other countries.
In fact, mail-in ballots are just one of the many options intended to cut voting costs and boost voter turnout. Both developed and developing countries have adopted innovative alternatives in election administration. Some have adopted, or experimented with, mail-in voting for all and made voting on the internet (i-voting, in which voters can cast ballots remotely) an option for voters. While i-voting is limited in the U.S., some countries have embraced or are experimenting with it. Estonia uses it in all national and local elections. Canada, Australia, France, Brazil, and Switzerland have run i-voting trials in local or regional elections (Germann and Serdu¨lt, 2017, 1).
In the U.S., governments have adopted e-voting (electronic voting), digitalizing some election administration processes in their efforts to cut costs and make them more efficient. For example, states are required to maintain computerized voter registration lists in accordance with the Help America Vote Act (HAVA) of 2002. The law also pushed states to adopt direct-recording electronic (DRE) systems (also known as e-voting machines) (Choi and Kim,2012, 433). Similarly, India, Brazil, the Philippines and Venezuela have extensively implemented e-voting (Alvarez et al., 2013, 117).
Voting Modes’ Effect on Turnout in Practice
While these are intended to make voting easier for all citizens, political scientists have found that these options have only minuscule or null effects on the overall voter turnout. For ex- ample, Karp and Banducci (2000) found in their analysis of Oregon’s 1996 special election by vote only by mail to fill a U.S. Senate that mail-in voting did increase voter turnout overall. However, they noted its marginal effect, arguing that “[v]oting only by mail is likely to increase turnout among those who are already predisposed to vote, such as those with higher socioeconomic status” (Karp and Banducci, 2000, 223). In the U.K., researchers found that more availability of postal voting had only “a broadly neutral effect on turnout” noting that the increase in the numbers of postal voters in the 2005 British General Election did not result in a significant increase of voter turnout overall (Rallings, Thrasher and Borisyuk, 2010, 236).2 Moreover, scholars found no turnout effect of internet voting in the trials in Geneva and Zurich during the Swiss referendum votes in 2001-2014 and 2001-2011, respectively (Germann and Serdu¨lt, 2017).
Moreover, more mail-in voters had only a very weak effect on the electoral outcome; there is little evidence that the Labour Party, which was the most popular party among mail-in voters, “directly benefited from the increase itself rather than a change in the balance between the numbers of Labour supporters in the postal and in-person categories (Rallings, Thrasher and Borisyuk, 2010, 236-7).
The small and unequal effect of voting modes on voter turnout can be attributed not only to citizens’ pre-existing inclination to vote and socio-economic backgrounds but also to their comfort with the technology and trust (or lack thereof) of the system’s ability to prevent fraud and miscounts. Some studies have found that electronic voting can reduce voter confidence; one study in 2013 e-voting trials in Salta, Argentina found that voters were concerned about its (in)ability to maintain ballot secrecy (Alvarez et al., 2013). Other studies have found electronic voting to increase voter confidence; a study that examined attitudes towards e-voting toward a 2007 trial in Colombia found that participants of the study found “e-voting not only easier than the current voting system, but also substantiallymore reliable” (Alvarez et al., 2009, 75).
Among U.S. voters, scholars have found in 2011 that concerns about election fraud due to e-voting are not as prominent as previously thought (Beaulieu, 2016). However, they found that distrust of e-voting is contingent on (potential) voters’ age; older individuals are more concerned about e-voting as a source of election fraud (Beaulieu, 2016, 18).
Morever, previous studies found no (or a very weak) relationship between partisanship or political ideology and trust in alternative voting modes. In other words, it is difficult to say that Republican voters are inherently against e-voting or mail-in ballots and Democratic voters are for e-voting or mail-in ballots per se. Instead, scholars have argued that “voters do not see electronic voting as presenting clear opportunities for manipulation for either party” (Beaulieu, 2014) and that concerns about election fraud due to e-voting is not as prominent as previously thought (Beaulieu, 2016).
The Pandemic and the Future of Alternative Voting Modes
However, it is likely that the pandemic has reconfigured the relationship between partisanship and voting modes for the upcoming election. Mail-in voting have become a political issue, politicized by the current political climate. Lockhart et al. (2020) predict that Republicans’ preference for in-person voting and Democrats’ preference for mail-in voting, coupled with local health conditions and electoral rules, may result in different turnouts by party, potentially affecting the election outcome. They speculate that “for Republicans in areas with high infection rates, concerns about COVID-19 exposure combined with reluctance to cast a mail ballot may lead to lower participation” and “[f]or Democrats or independents who would now prefer to vote by mail, the laws in some states that limit vote-by-mail opportunities could decrease their turnout.” (Lockhart et al., 2020, 2).
It also remains to be seen whether the pandemic will accelerate states’ adoption of i-voting. It is possible that the pandemic may increase the need, hence the demand, for voting remotely. However, concerns about election security and the possibility of i-voting helping opponents may dissuade countries from adopting it (Germann and Serdu¨lt,2017, 1).