The events between Election Day 2020 and Inauguration Day 2021 posed a stress test for America’s republican institutions. Some behaved well; others revealed weaknesses that represent possible breakpoints in some future crisis. The lessons of this brush with constitutional extremity should inform proposals for election law overhaul. Reforms that shore up what were revealed as critical weaknesses deserve high priority; changes that would have made no difference may seem less urgent; and proposals that might destabilize the system or open up new risks of constitutional crisis are unlikely to be right for the moment.

Most constitutional actors resisted pressure to stray from their constitutionally prescribed duties during Donald Trump’s attempt to overturn his electoral defeat. They included the vice president, state election officials and legislators, and most members of Congress. No more than a smattering of lawmakers in any state legislature buy into the wild (and legally ineffectual) scheme of replacing already-chosen electors. Less inspiring of confidence was the performance of the House of Representatives, 139 of whose 435 members voted against certifying at least one state.

Policy responses to these tremors should aim, where possible with bipartisan buy-in, to strengthen the institutions that secure the peaceful electoral transfer of power and narrow the range of factual and legal questions that might in the future enable an election loser to throw into doubt the winner’s right to take office.

The most suitable area for congressional action is in clarifying and tightening up the confusing Electoral Count Act of 1887, which lays out rules for Congress’s handling of certified electoral votes following a presidential election.

The Constitution gives state legislatures broad authority over the method of selecting presidential electors. All have chosen popular election, a fact that is unlikely to change. Whatever the method, they must choose it beforehand by process of law: Election Day completes the act of elector selection, foreclosing further choices as to method.

Although the Constitution accords states the power to certify results, it does accord to Congress the much more limited role of ruling on irregularities in the submitted certifications themselves, such as errors in date, absent signatures, or claims of forgery. If multiple certifications have been submitted claiming to speak for a given state, it must also resolve which authentically does so.

Because the 1887 law does not exhaustively define proper grounds for objections, partisans in Congress have sometimes sought, wrongly, to use the occasion to relitigate the underlying election. Congress should also clarify the scope of language that permits state legislatures to devise methods for later selection if Election Day balloting has “failed,” an exception suited perhaps to a hurricane or similar disaster (ideally as defined by law in that state beforehand). Beyond that, revision of the Electoral Count Act should place a number of points beyond any possible cavil, such as that a vice president does not have discretion to reject duly certified slates.

Republican institutions rest on a less secure footing if political factions regularly portray their opponents’ election victories—or even their participation in politics—as illegitimate and reject the idea of “rotation in office,” which posits that it is normal for voters to periodically replace one set of incumbents with another.

Evidence in recent years has failed to substantiate sensational claims of either voter fraud and “rigging” on the one hand or widespread “voter suppression” on the other. Extensive audits and integrity controls indicate that the volume of fraudulently cast votes is unlikely to be high enough to affect many races. As for “voter suppression,” that epithet has been aimed at a variety of practices—many innocuous, widespread, or long-familiar—that have very little to do with preventing willing electors from casting ballots. For example, voter ID laws appear to have no detectable effects on turnout, minority or otherwise, and no detectable effects on fraud either.

In short, the alleged conflicts between ballot security and voter access are overstated. Conservatives should not act as if something is wrong with the goal of making voting more convenient and consumer-friendly; people like convenience, and not everyone has the same schedule, time demands, or car access. Liberals should concede that a practice like “ballot harvesting”—in which a single operative can be paid to collect hundreds of absentee ballots— does raise genuine concerns relating to voter privacy, risks of undue pressure, and security.

The fraud charges leveled by the 45th president following his loss are but one symptom of a rising unwillingness on both sides of the aisle to concede the legitimacy of election results. Following the 2016 election, for example, a significant share of Democratic voters polled credited an evidence-free theory that Russians had swayed the election by tampering directly with voting machine tallies, a theory recalling the enthusiasm some circles had shown in 2004 for the speculation that Diebold voting machines had delivered Ohio to George W. Bush.

Election administration at all times calls for methods that are secure against fraud and bad practice; an era of rising public distrust calls for methods that are also visibly so. Some moves in this direction have already gone forward with little controversy, as with anti-hacking safeguards and the principle of generating a paper and not simply electronic record for each ballot.

More is possible. Genuine professional-grade audits—as well as various recount methods and what are known as logic and accuracy tests to detect suspicious patterns—are already commonly used and can be adopted more widely. The same is true of transparency measures that—without allowing partisans or amateurs to interfere with counting or to violate security themselves—can serve to reassure doubters by making key stages of the count visible to party leaders and others. States should also mandate participating in what is known as the Electronic Registration Information Center, which allows them to compare their voter registration database with that of other member states to improve detection of moves, duplicates, and so forth.

One particularly important objective for states in current circumstances is to facilitate and, where appropriate, mandate local reporting of complete or near-complete counts on Election Night. A good example is Florida with its early and efficient reporting of results on Election Night 2020. When states do the opposite—in particular, when they refuse to provide for some advance processing of mailed ballots—results will tend to arrive in two widely spaced waves: first, same-day in-person votes, and second, votes by other methods. When the partisan composition of the waves differs, as in 2020, the resulting pattern may be for one side to open up a lead and then be overtaken by the other, leading to claims that someone or other must have engaged in overnight “vote dumps.” There is a genuine national as well as within-state interest in ensuring that counts do not stretch into multiple days.

Timely and gracious acknowledgments of election losses are signs of a healthy democratic culture. Although candidates cannot be forced to concede, states often do structure financial incentives in ways that influence behavior, as with rules providing that when the reported margin of victory exceeds a stated margin, candidates that demand recounts must put up the associated cost themselves.

The Framers largely left the responsibility for elections to the states and localities. States are given the lead in regulating elections for the House and Senate, although the election clause empowers Congress to regulate the manner of election by law. (See “Redistricting.”) The Electoral College device is structured to hold to a minimum the capital’s influence on the selection of the president.

Amendments to the Constitution together with implementing legislation have added crucial national-level constraints of equal protection and noninfringement of the right to vote on such bases as that of race and sex. Still, election administration in America remains highly decentralized, relying on armies of local officials, community volunteers, and election judges. States are free to set their own course among a wide choice of methods. Even within a single state, communities may differ on which voting machines to use or how to handle voter correction (“curing”) of incomplete ballot submissions.

There are important reasons to be cautious about allowing Washington to displace or regiment this decentralized responsibility.

One is the practical value of decentralization in resilience. Innovations adopted piecemeal can be sorted out by local trial and error with less risk of mass failure. The novel voting machine design doesn’t cause havoc everywhere at once; communities considering a voting method like ranked-choice voting can learn from New York City’s difficult tryout.

But those are secondary benefits compared with the big one: no Washington official or agency can start bossing around or removing local election officials generally and on short notice. By not entrusting the running of elections to a single central agency, we avoid what economist Steven Landsburg calls “centralizing the power to decide who will yield power.”

After the 2020 election, we can see that it is especially vital to curb the risk of a succession crisis: a situation in which more than one candidate with broad support is tempted to claim legitimate control of the government, often because of an election outcome that is indeterminate in law or practice. Yet some widely lauded reform ideas would head us in the opposite direction.

Consider, for example, the scheme known as the National Popular Vote Interstate Compact (NPVIC). The NPVIC’s premise is to have states each pass identical legislation agreeing to award their electoral votes collectively to whichever candidate wins the national popular vote. As of this writing, the NPVIC has been adopted by 15 states and the District of Columbia with 195 electoral votes, more than halfway toward the 270 votes that by the terms of the compact would bring it into legal force.

But the drafters of the compact did not see fit to include workable definitions of how and when a national popular vote would be computed, nor any dispute resolution mechanism in case of disagreement or resistance. A national vote implies a national recount should results prove close, yet no law requires states to conduct a recount. The NPVIC simply takes it for granted that all states report their popular votes in a tidy and readily comparable fashion.

As Cato’s Andy Craig has demonstrated, a variety of plausible fact patterns could generate dangerous indeterminacy about results. For example, states are currently free under the Constitution to adopt, and have adopted at times in the past, voting procedures that baffle the hope of obtaining a uniform count per candidate. Alabama in 1960 used a system in which the names of neither John F. Kennedy nor Richard Nixon appeared on the ballot. Voters instead were free to pick and choose among electors, many of whom were unpledged. Because of these anomalies, respectable sources at the time differed as to whether it was Kennedy or Nixon who had edged the other in popular votes—not that it mattered, since it was clear Kennedy had won the Electoral College.

That’s not even counting the chances for deliberate sabotage by uncooperative states. Various bills in NPVIC-resistant state legislatures gesture in that direction; one that passed the North Dakota Senate proposes to withhold release of that state’s popular vote until after the Electoral College has voted.

The chief short‐​term goal in election reform should be to learn from and implement the lessons of late 2020 and early 2021. That suggests measures to shore up the legal and the factual certainty and transparency of election outcomes.

Election administration is an imperfect art at best with plenty of genuine tradeoffs. We should refrain from treating everyday disagreements as attempts to “rig” results or deprive others of the franchise.

Craig, Andy. “The Fatally Flawed National Popular Vote Plan.” Cato at Liberty (blog), November 17, 2021.

———. “How to Pick a President: A Guide to Electoral Count Act Reform,” Cato Institute Policy Analysis no. 931, June 28, 2022.

———. “What Changes Should Be Made to the Electoral Count Act?” Cato at Liberty (blog), January 12, 2022.

Muller, Derek T. “Invisible Federalism and the Electoral College.” Arizona State Law Journal 44 (2012): 1237.

Olson, Walter. “The Dos and Don’ts of Defending Democracy.” Cato Policy Report, November/​December 2021.

Schweers, Jeffrey. “Don’t Call It ‘Flori-Duh’—Sunshine State Breaks 20-Year Hex on Election Problems.” Tallahassee Democrat, November 5, 2020.