In mid-March, as businesses shuttered, hospitals filled up and Americans hunkered down at home, election officials faced a difficult question: Was it even possible to hold safe and accessible elections in the age of COVID-19? More than four months later, we’re not that much closer to answering that question.
Since the World Health Organization declared an official pandemic on March 11, 37 states plus Washington, D.C., have held statewide primaries1 for president or state-level office. And while those that have gone poorly have tended to grab the headlines, there have been success stories too. Ultimately, it’s been hard to assess how well our democracy has adapted to the pandemic. So here’s a snapshot of all 38 statewide elections since the pandemic started and what macro trends we’ve been able to observe so far.
First, most states — and almost all those that actually made an effort to do so — were wildly successful at getting people to vote by mail (or at least vote before election day). In 24 out of 35 states for which we have this data, a majority of ballots were cast absentee.2 In addition, every state but one3 made more use of absentee ballots than it did in the equivalent election in 2016.4 Considering what a massive logistical undertaking it is to switch to a predominantly mail election, this is an impressive achievement by election officials. (Of course, as we’ll cover below, it didn’t always go off without a hitch.)
|Jurisdiction||Mail Ballot Access||Absentee||Change from 2016|
|Idaho||Mailed ballot applications||100||+86|
|North Dakota||Mailed ballot applications||100||+73|
|Ohio||Mailed instructions for requesting ballot||99||+85|
|Arizona||Anyone could vote absentee||89||+8|
|Nebraska||Mailed ballot applications||84||+63|
|Rhode Island||Mailed ballot applications||83||+79|
|Iowa||Mailed ballot applications||78||+59|
|Wisconsin||Anyone could vote absentee||75||+65|
|Kentucky*||Mailed instructions for requesting ballot||73||+70|
|Washington, D.C.||Mailed ballot applications||71||+64|
|New Mexico||Mailed ballot applications||63||+56|
|South Dakota||Mailed ballot applications||58||+44|
|Indiana*||Anyone could vote absentee||51||+35|
|Pennsylvania||Mailed instructions for requesting ballot||51||+48|
|West Virginia*||Mailed ballot applications||50||+49|
|Georgia||Mailed ballot applications||48||—|
|Florida||Anyone could vote absentee||46||+16|
|Delaware||Mailed ballot applications||45||+42|
|South Carolina||Anyone could vote absentee||22||+10|
|Virginia||Anyone could vote absentee||21||+20|
|Oklahoma||Anyone could vote absentee||14||+10|
|Illinois||Anyone could vote absentee||9||+6|
|Alabama*||Anyone could vote absentee||5||—|
Unsurprisingly, states that mailed every voter a ballot saw the highest share of their votes cast absentee, although it’s hard to definitively say that was the reason, as these states also offered few polling places — or, in Alaska, Hawaii, Kansas, Utah and Wyoming, no polling places at all.
However, other ways of encouraging mail voting had more mixed results. For instance, in the states that just mailed voters absentee-ballot applications or instructions for how to apply for an absentee ballot, there was anywhere from 45 percent absentee participation in Delaware to 84 percent in Nebraska. (But don’t read too much into the 100 percent absentee rates in Idaho and North Dakota — they, too, eliminated in-person voting.)
What we do know is the states that did not mail voters anything to nudge them toward voting absentee (such as Illinois, with 9 percent absentee participation, and Oklahoma, with 14 percent absentee participation) tended to have the lowest shares of absentee voters, and the smallest increases from 2016. The same was true of states like Louisiana and Texas, which still required voters to provide an excuse to vote absentee. A notable exception was Wisconsin, where 75 percent of votes were cast absentee despite nothing being mailed to them. Most likely, the intense news coverage predicting doom and gloom for Wisconsin’s primary caused Wisconsinites to heed the state government’s advice and request absentee ballots; both Joe Biden’s and Bernie Sanders’s campaigns also encouraged their supporters to vote by mail.
Of course, just because a state conducted its primary predominantly by mail does not mean it will be able to turn the same trick in November. For starters, as the next table makes clear, most of these elections were low-turnout affairs: Anywhere from 3 percent to 46 percent of the voting eligible population turned out to vote in them, while presidential general elections have VEP turnout rates around 60 percent. (And, needless to say, processing mail ballots for 60 percent of the population is a lot more challenging than processing mail ballots for 20 or 30 percent.) But the good news is that the coronavirus doesn’t seem to be putting a dent in turnout rates. Just as turnout in the presidential primaries before the pandemic hit was a bit above 2016 levels, 22 of the 34 elections for which we have the relevant data actually saw their VEP turnout race increase over the equivalent election in 2016.
|Date||Jurisdiction||2016 Turnout||2020 Turnout||diff|
|June 9||South Carolina||11||20||+8|
|June 2||New Mexico||23||28||+5|
|June 2||South Dakota||20||24||+4|
|June 9||North Dakota||24||28||+4|
|June 2||Washington, D.C.||20||22||+2|
|June 9||West Virginia*||34||32||-2|
|June 2||Rhode Island||24||16||-8|
Admittedly, turnout did plummet in many states this year — but for most of them, there is a compelling reason unrelated to the coronavirus. States like Delaware, Indiana, Louisiana, Ohio, Pennsylvania, Rhode Island and Wisconsin went from hosting contested presidential primaries on both sides (Democratic and Republican) in 2016 to being virtually irrelevant to the delegate math this year (both President Trump and Joe Biden sewed up their respective nominations by the end of March). And Arizona did not even hold a Republican presidential primary this year, which explains its huge drop-off in turnout.
But if the pandemic significantly suppressed turnout, it was probably in Illinois, where VEP turnout went from 40 percent in the 2016 primary to 25 percent in the 2020 primary. The pandemic seems like a good explanation for at least part of the drop-off: Illinois experienced long lines and other problems on election day (see below), and because the primary was held just days after the country locked down, voters didn’t really have time to request absentee ballots — nor did the state have time to mail them anything. (The lack of a competitive GOP presidential primary this year also explains a lot of the drop-off but not all.)
Similarly, there were non-pandemic-related reasons for the turnout increases in many states. For instance, Colorado opened its primaries up to unaffiliated voters; Oklahoma and Utah simply had more interesting races on the ballot in 2020 than in 2016. There were also states like Kentucky and Nebraska that didn’t hold one party’s presidential primary in 2016, and states like Alabama, Texas and Virginia, where there were no statewide races in 2016 (so many residents just didn’t have an election to vote in).
But there were several elections — in states like Idaho, Montana (on the Democratic side), Nevada, New Mexico, North Dakota and South Dakota — where turnout did increase even though those states held equally or less competitive primaries in 2020 than in 2016. You could arguably add Iowa to that list too; its 2020 down-ballot primaries were pretty competitive, but not enough to explain a 14-percentage-point turnout increase from 2016.
It’s plausible that these states’ embrace of mail voting helped expand the electorate; all of them either mailed ballots or sent absentee-ballot applications to voters. In addition, Kansas also saw turnout spike after the Kansas Democratic Party, spurred in part by the coronavirus, switched from a caucus in 2016 to an all-mail primary in 2020 (caucuses are notorious for their low turnout).
There’s an important caveat here, though: Just because the coronavirus did not lower turnout overall does not mean it didn’t disenfranchise individual voters. We know that at least some voters were unable to vote because of the pandemic. Indeed, the primaries conducted during the pandemic have been far from the smooth sailing suggested above. Instead, an unusual number of problems have occurred — some serious enough that they have made it harder for people to vote and even put them at a heightened risk for infection.
Two specific problems have come up again and again. At least 16 jurisdictions5 saw long lines at polling places, which had often been consolidated due to a shortage of poll workers. The unluckiest voters in Milwaukee waited two and a half hours to vote; in Washington, D.C., five; in Las Vegas, seven.
And some voters in at least 12 places6 reported never receiving the absentee ballots they had requested — sometimes weeks in advance. In some states not used to a heavy volume of mail voting, understaffed election offices were receiving ballot requests faster than they could respond to them. In six Ohio counties, about 4,500 people who requested an absentee ballot were not sent one because their request lacked essential information. In one Pennsylvania county, officials simply ran out of time to fulfill 400 ballot requests and sent out 6,000 ballots the day before the primary, when they had little chance of being received in time to be voted. In Maryland, 1 million ballots were delivered either late or not at all.
Other problems were less rampant — but no less concerning. In Illinois, multiple polling places lacked the proper voting equipment or cleaning supplies. In Georgia, new voting machines did not work correctly, either because they malfunctioned or poll workers were not adequately trained. In Florida, Illinois and New York, truant poll workers caused some polling places to open late. In Idaho, the state’s online absentee-ballot request portal got overloaded and crashed right before the deadline (which was extended as a result). In Pennsylvania and Washington, D.C., police wrongly told voters waiting in line to disperse because of a curfew even though voters were exempt from it. In South Carolina, polling-place consolidation and new, untrained poll workers apparently led to some voters being given the wrong ballots. In Texas, some poll workers walked off the job after other poll workers refused to wear face masks.
The coronavirus has put American democracy to the test — and by our reckoning, election officials have made big strides in a short period of time. But a lot of work must still be done.
Some states, like Idaho, Iowa, Nebraska, New Mexico, North Dakota and South Dakota, have risen to the challenge, greatly expanding their use of absentee voting and reporting few problems at the polls. These states also saw increases in turnout relative to the equivalent election in 2016. However, it may be hard to replicate their success nationally, as they all are mostly rural and sparsely populated, so their task may be easier than a state like New York or Michigan. Similarly, Arizona, Colorado, Montana, Oregon and Utah also scored highly because they already conduct their elections mostly via absentee ballot and thus were able to avoid major problems.
On the other hand, states like Illinois, Indiana, Pennsylvania and Wisconsin bungled the execution of their primaries. They did little to expand mail voting (although many voters cast absentee ballots anyway), and their turnout fell (which may or may not be related to the first two points). Georgia, Maryland and Washington, D.C., are also guilty of ineptitude that surely disenfranchised some voters, although their turnout rates didn’t reflect it — perhaps because their efforts to expand mail voting also brought new people into the electorate.
So where does this leave us headed into November? It’s hard to say. On the one hand, it is worth keeping in mind how states handled their primaries. But on the other hand, don’t assume a state that performed competently in the primary will do so in the general, or the inverse. They may, but the general election is also a very different beast — and, with that higher turnout, one that is much more difficult to tame. States may also learn from a bad experience during the primary and resolve to do things differently in the fall; they may have more or less funding available for November than they did for the primary, and they may tweak rules surrounding absentee-ballot or in-person voting access. Unfortunately, nobody really knows what the 2020 general election will look like — and how each state will fare.