Evaluating Election Performance in the Era of the Big Lie

With the release of the Election Administration and Voting Survey by the U.S. Election Assistance Commission last week, updating the Elections Performance Index for the 2020 election is now in full swing.

A decade ago, the EPI was created to shift the debate about election policy in a fact-based direction. As designed by the Pew Charitable Trusts and now continued by the MIT Election Data and Science Lab, the EPI is grounded in the notion that elections in the U.S. should be convenient and secure. Objectively measured and presented, an election system’s vital signs provide a glimpse into how these twin goals are being met.

By now, with six federal elections under our belt, updating the EPI for a new election should be routine. This time, it’s not.

The Big Lie — the claim that the 2020 election was stolen from Donald Trump — looms over the effort.

How can an activity such as the EPI provide data-based commentary on the past election, knowing that some will take any adverse inferences as further evidence that the 2020 election is illegitimate?

As tough as it is to encourage evidence-informed debate about election policy in this environment, it gets worse. As we’ve heard over the past two weeks at the summer meetings of NASED and NASS, election officials themselves continue to be subject to physical threats for simply doing their jobs.

How can a project such as the EPI, which inevitably will cast a bad light on some and good light on others, proceed knowing that any criticism of election policy that follows will seem like gratuitous piling-on?

As Team MEDSL begins its sprint to get the 2020 version of the EPI out the door, I offer nine thoughts about the subject. As I do, keep in mind that the real purpose of the EPI is to encourage a broad empirical approach to election policy. Thus, some of my remarks will pertain to the EPI itself, but most of these remarks will instead be relevant to the larger goal: to assess the state of all election policies against metrics.

1. Objective evidence points to the 2020 election as being successfully administered.

I start with the obvious point that Nate Persily and I made in the Journal of Democracy article “The Miracle and Tragedy of the 2020 U.S. Election”: “Amid a pandemic posing unprecedented challenges, local and state administrators pulled off a safe, secure, and professional election.”

As we pull together the EPI’s indicators, we will likely see overall gains, even with the challenges of the election. We already know that turnout and voter registration reached record levels, provisional ballots were down, residual vote rates were down, and states continued extending their online presence and auditing programs. Every EPI indicator won’t see an across-the-board improvement; the rejected absentee ballot rate will be especially tricky. Still, when the EPI is finally published, I suspect that many more indicators will see improvements than not.

2. Evidence-based improvement of elections was hijacked on January 6, 2021.

The Capitol Riot did many things, one of which was to limit our ability to hold evidence-based debates about election administration. The riot took an issue that was already polarized around the enduring alternative views of public policy that Democrats and Republicans brought to the table and poisoned the well.

Luckily, many of the items in the EPI are not easily twisted into a left-right dimension. That’s why they were chosen. Are reducing provisional ballots, improving the efficiency of mail ballot operations, encouraging turnout among people with disabilities, and providing information to voters about elections online partisan issues? I don’t think so.

However, other issues that once enjoyed broad support now attract partisan framings. Among these are turnout, registration, and absentee ballot administration. So are other indicators that might be added to the EPI one day, such as the gap in turnout between whites and non-whites. Perhaps the best we can hope in this environment is to shine lights on issues favored by both parties and accept that everyone will criticize us for something.

3. A fact-based assessment of 2020 will help us decide what to keep from the pandemic adjustments.

Lots of changes were made to how Americans voted in 2020 to accommodate the challenges of the pandemic. Which should become permanent? One approach to this question is simply to ask if the accommodation had an “access” or “security” flavor, consult your partisan leanings, and not ask much more than that.

Another approach is to measure and weigh all the considerations and then decide based on a balanced review of all the evidence. I know this is an old-fashioned way of thinking. However, I’m still convinced it’s an approach that can help improve the experience for voters and honor the different values people have about what matters most in setting election policy.

A good example of how this could help to guide policy is the question of whether states should continue to allow absentee ballots to come in after Election Day, provided the voters mailed them before Election Day, of course. The reflexive partisan positions on this question are obvious. A metrics-based approach opens up a broader conversation once one explores such measurables as the number of ballots that arrived after Election Day in 2020 (not many) and how much allowing ballots to arrive after Election Day slowed down the count (a lot). In the end, partisanship might prevail, but an empirical exploration of the issues at least unearths what’s practically at stake.

4. The world moves on, and 2020 offers invaluable opportunities to learn lessons.

The 2020 election saw election officials experiment in a wide variety of ways. Obviously, they expanded opportunities to vote by mail and did so in different ways. But, they also had to innovate in less obvious ways, such as reconfiguring polling places and rethinking poll worker training. The squeeze put on election departments highlighted funding inadequacies in much of the country. The list could go on.

The U.S. Election Assistance Commission has published a series of fascinating video interviews with election officials about lessons learned from the 2020 election. An important thing that series does is show that it is possible to assess the pluses and minuses of the 2020 election without spiraling into partisan recriminations. John Fortier, Senior Fellow at the American Enterprise Institute, and I are finishing a report for the EAC on lessons learned, as well. We take on various topics, from registration to funding, to assess how the system performed using neutral criteria. As we move away from the 2020 election and its immediate aftermath, there are signs that not every critical look at the election process has to devolve to partisan sniping.

5. The 2020 election was likely more successful in some places than others.

Much commentary about the 2020 election within the elections community reminds me of the children of Lake Wobegon, who are all above average. But, a quick look at all the data we now have about the election makes it clear that some succeeded better than others. In 2016, Georgia’s absentee ballot rejection rate was 6.4 percent, similar to New York’s 5.9 percent. In 2020, Georgia cut that rate to 0.4 percent, a 95 percent reduction. New York didn’t even halve its rate, which went down to 3.6 percent. What does Georgia know that New York doesn’t?

The election was most successful in states where the two sides of the election administration equation, policy and implementation, worked as hoped — policymakers took the appropriate adaptive changes, and administrators took advantage of the clearing they had been given. But, we also know, or should know, that at least one side of this equation was unbalanced in some places. Some legislatures didn’t update antiquated absentee ballot laws; some local election officials didn’t recalculate their space needs to reflect new realities; etc. If we are actually to learn lessons from 2020 about how to keep elections going during emergencies, then at some point, we’ll have to make fine distinctions about what worked best.

6. The 2020 election may have been less convenient and secure for some people.

Just as some places may have fared better in the 2020 election, it’s possible that some categories of people fared better, too. People with disabilities and those with low levels of education may have had greater problems adapting to the new realities of voting. Looking beyond the confines of the 17 EPI indicators, if COVID-19 hit minority communities especially hard, it is likely that it also hit the ability to adapt to voting in those communities especially hard, through recruiting poll workers, finding safe and convenient polling places, and communicating with voters. A metrics-based approach to assessing the 2020 election is critical for understanding how well the response was dispersed to voters with different abilities and resources.

7. Shortcomings in the 2020 elections mainly were a problem of policy and not performance.

State legislatures, and occasionally Congress, make election laws and then adjourn. This leaves election administrators, who have year-round jobs, to implement them and stick around. In the decade I’ve been associated with the EPI, I’ve learned that performance on the Index is driven primarily by the policies that state legislatures choose rather than the implementation strategies administrators choose. (There are exceptions, of course.) However, administrators get blamed because they’re the ones left to do the dirty work and then defend their actions, including when their hands were tied.

When the EPI reports that a state has done poorly on some measure or many of them, it will be important to remember that the cause is likely the policymakers, not the administrators. Low turnout? High levels of rejected absentee ballots? No post-election audits? Don’t blame a county election official; blame the legislature and governor.

8. Election administration is not the only way to assess the performance of American elections.

The EPI is about the nitty-gritty of elections. Informative websites. Accessible polls. Automated registration procedures. Auditing results. Checking those signatures. It is a voter-centric approach to assessing elections that looks mainly at the experience of those who showed up to vote.

This isn’t the only way to assess the performance of American elections. The Electoral Integrity Project, started by Pippa Norris and now managed by Holly Ann Garnett and Toby James, provides another approach. It’s broader-based and asks more about the cultural support for electoral institutions. Rather than assess elections via objective indicators, the EIP is based on surveys of experts. It goes beyond administration into topics such as gerrymandering, coverage of elections by the media, campaign finance, and barriers faced by minority ad women candidates. This is an approach that assesses American elections from an international perspective.

The value of an approach like the EIP’s is that it asks about the health of democracy in general. That gets beyond even legislative policymaking into issues of culture, economics, and constitutional legacies. If you think changing policy is hard, think about changing culture or the economic system. In any event, the EPI project has a defined lane, which is election administration. Parallel lanes have things to teach us, as well.

9. The judgment of policy change in 2021 will be rendered in 2022 and 2024.

To kick off this year’s Election Science, Research, and Administration Conference, I moderated a panel of election administrators and political scientists who talked about the 2020 election, administrator-academic collaborations, and how to move forward in improving elections. One of the quotes from an administrator in that panel hit me between the eyes:     

The 2022 election will have all the pressure but none of the help.

This quote can be interpreted in many ways. One interpretation is that election administrators will be on their own for 2022 and 2024 as the pandemic emergency slips behind us. Part of being “on their own” will be running these elections under the same policies and resources that govern for regular times. If we believe the evidence that the 2020 election was administered at a high level of proficiency, we must capture this success with objective measures. With that, we will be able to judge whether policymakers were in fact able to learn anything of lasting value from 2020.

Although I have a reputation as an optimist when it comes to the quality of election administration and American democracy, I also believe that the health of our democracy stands at a crossroads. I often remind my students of the quote from E.E. Schattschneider’s 1960 classic, The Semi-Sovereign People: A Realist’s View of Democracy in America, that “the struggle for democracy is ongoing,” so perhaps it’s always at a crossroads. Still, whether or not we believe that democracy is in peril these days, the promise of self-government cannot be fulfilled without good election laws implemented expertly.

As MEDSL rolls out the EPI over the next several months, we will do our best to be fair about best practices and where improvement can be had. We will also use the expanded capabilities of our EPI website to give voice to different views about election administration, including those who administer the laws. In this way, we hope to do our part in pulling the discussion about election administration in the U.S. back to the fact-based space where it belongs.