The Election That Could Break America

Illustrations by Guillem Casasús / Renderings by Borja Alegre

There is a cohort of close observers of our presidential elections, scholars and lawyers and political strategists, who find themselves in the uneasy position of intelligence analysts in the months before 9/11. As November 3 approaches, their screens are blinking red, alight with warnings that the political system does not know how to absorb. They see the obvious signs that we all see, but they also know subtle things that most of us do not. Something dangerous has hove into view, and the nation is lurching into its path.

The danger is not merely that the 2020 election will bring discord. Those who fear something worse take turbulence and controversy for granted. The coronavirus pandemic, a reckless incumbent, a deluge of mail-in ballots, a vandalized Postal Service, a resurgent effort to suppress votes, and a trainload of lawsuits are bearing down on the nation’s creaky electoral machinery.

Something has to give, and many things will, when the time comes for casting, canvassing, and certifying the ballots. Anything is possible, including a landslide that leaves no doubt on Election Night. But even if one side takes a commanding early lead, tabulation and litigation of the “overtime count”—millions of mail-in and provisional ballots—could keep the outcome unsettled for days or weeks.

If we are lucky, this fraught and dysfunctional election cycle will reach a conventional stopping point in time to meet crucial deadlines in December and January. The contest will be decided with sufficient authority that the losing candidate will be forced to yield. Collectively we will have made our choice—a messy one, no doubt, but clear enough to arm the president-elect with a mandate to govern.

As a nation, we have never failed to clear that bar. But in this election year of plague and recession and catastrophized politics, the mechanisms of decision are at meaningful risk of breaking down. Close students of election law and procedure are warning that conditions are ripe for a constitutional crisis that would leave the nation without an authoritative result. We have no fail-safe against that calamity. Thus the blinking red lights.

“We could well see a protracted postelection struggle in the courts and the streets if the results are close,” says Richard L. Hasen, a professor at the UC Irvine School of Law. “The kind of election meltdown we could see would be much worse than 2000’s Bush v. Gore case.”

A lot of people, including Joe Biden, the Democratic Party nominee, have mis­conceived the nature of the threat. They frame it as a concern, unthinkable for presidents past, that Trump might refuse to vacate the Oval Office if he loses. They generally conclude, as Biden has, that in that event the proper authorities “will escort him from the White House with great dispatch.”

The worst case, however, is not that Trump rejects the election outcome. The worst case is that he uses his power to prevent a decisive outcome against him. If Trump sheds all restraint, and if his Republican allies play the parts he assigns them, he could obstruct the emergence of a legally unambiguous victory for Biden in the Electoral College and then in Congress. He could prevent the formation of consensus about whether there is any outcome at all. He could seize on that un­certainty to hold on to power.

Trump’s state and national legal teams are already laying the groundwork for postelection maneuvers that would circumvent the results of the vote count in battleground states. Ambiguities in the Constitution and logic bombs in the Electoral Count Act make it possible to extend the dispute all the way to Inauguration Day, which would bring the nation to a precipice. The Twentieth Amendment is crystal clear that the president’s term in office “shall end” at noon on January 20, but two men could show up to be sworn in. One of them would arrive with all the tools and power of the presidency already in hand.

“We are not prepared for this at all,” Julian Zelizer, a Prince­ton professor of history and public affairs, told me. “We talk about it, some worry about it, and we imagine what it would be. But few people have actual answers to what happens if the machinery of democracy is used to prevent a legitimate resolution to the election.”

Nineteen summers ago, when counterterrorism analysts warned of a coming attack by al‑Qaeda, they could only guess at a date. This year, if election analysts are right, we know when the trouble is likely to come. Call it the Interregnum: the interval from Election Day to the next president’s swearing-in. It is a temporal no-man’s-land between the presidency of Donald Trump and an uncertain successor—a second term for Trump or a first for Biden. The transfer of power we usually take for granted has several intermediate steps, and they are fragile.

The Interregnum comprises 79 days, carefully bounded by law. Among them are “the first Monday after the second Wednesday in December,” this year December 14, when the electors meet in all 50 states and the District of Columbia to cast their ballots for president; “the 3d day of January,” when the newly elected Congress is seated; and “the sixth day of January,” when the House and Senate meet jointly for a formal count of the electoral vote. In most modern elections these have been pro forma milestones, irrelevant to the outcome. This year, they may not be.

“Our Constitution does not secure the peaceful transition of power, but rather presupposes it,” the legal scholar Lawrence Douglas wrote in a recent book titled simply Will He Go? The Interregnum we are about to enter will be accompanied by what Douglas, who teaches at Amherst, calls a “perfect storm” of adverse conditions. We cannot turn away from that storm. On November 3 we sail toward its center mass. If we emerge without trauma, it will not be an unbreakable ship that has saved us.

Let us not hedge about one thing. Donald Trump may win or lose, but he will never concede. Not under any circumstance. Not during the Interregnum and not afterward. If compelled in the end to vacate his office, Trump will insist from exile, as long as he draws breath, that the contest was rigged.

Trump’s invincible commitment to this stance will be the most important fact about the coming Interregnum. It will deform the proceedings from beginning to end. We have not experienced anything like it before.

Maybe you hesitate. Is it a fact that if Trump loses, he will reject defeat, come what may? Do we know that? Technically, you feel obliged to point out, the proposition is framed in the future conditional, and prophecy is no man’s gift, and so forth. With all due respect, that is pettifoggery. We know this man. We cannot afford to pretend.

Trump’s behavior and declared intent leave no room to suppose that he will accept the public’s verdict if the vote is going against him. He lies prodigiously—to manipulate events, to secure advantage, to dodge accountability, and to ward off injury to his pride. An election produces the perfect distillate of all those motives.

Pathology may exert the strongest influence on Trump’s choices during the Interregnum. Well-supported arguments, some of them in this magazine, have made the case that Trump fits the diagnostic criteria for psychopathy and narcissism. Either disorder, by its medical definition, would render him all but incapable of accepting defeat.

Conventional commentary has trouble facing this issue squarely. Journalists and opinion makers feel obliged to add disclaimers when asking “what if” Trump loses and refuses to concede. “The scenarios all seem far-fetched,” Politico wrote, quoting a source who compared them to science fiction. Former U.S. Attorney Barbara McQuade, writing in The Atlantic in February, could not bring herself to treat the risk as real: “That a president would defy the results of an election has long been unthinkable; it is now, if not an actual possibility, at the very least something Trump’s supporters joke about.”

But Trump’s supporters aren’t the only people who think extra­constitutional thoughts aloud. Trump has been asked directly, during both this campaign and the last, whether he will respect the election results. He left his options brazenly open. “What I’m saying is that I will tell you at the time. I’ll keep you in suspense. Okay?” he told moderator Chris Wallace in the third presidential debate of 2016. Wallace took another crack at him in an interview for Fox News this past July. “I have to see,” Trump said. “Look, you—I have to see. No, I’m not going to just say yes. I’m not going to say no.”

How will he decide when the time comes? Trump has answered that, actually. At a rally in Delaware, Ohio, in the closing days of the 2016 campaign, he began his performance with a signal of breaking news. “Ladies and gentlemen, I want to make a major announcement today. I would like to promise and pledge to all of my voters and supporters, and to all the people of the United States, that I will totally accept the results of this great and historic presidential election.” He paused, then made three sharp thrusts of his forefinger to punctuate the next words: “If … I … win!” Only then did he stretch his lips in a simulacrum of a smile.

The question is not strictly hypothetical. Trump’s respect for the ballot box has already been tested. In 2016, with the presidency in hand, having won the Electoral College, Trump baldly rejected the certified tallies that showed he had lost the popular vote by a margin of 2,868,692. He claimed, baselessly but not coincidentally, that at least 3 million undocumented immigrants had cast fraudulent votes for Hillary Clinton.

All of which is to say that there is no version of the Interregnum in which Trump congratulates Biden on his victory. He has told us so. “The only way they can take this election away from us is if this is a rigged election,” Trump said at the Republican National Convention on August 24. Unless he wins a bona fide victory in the Electoral College, Trump’s refusal to concede—his mere denial of defeat—will have cascading effects.

The ritual that marks an election’s end took its contemporary form in 1896. On the Thursday evening after polls closed that year, unwelcome news reached the Democratic presidential nominee, William Jennings Bryan. A dispatch from Senator James K. Jones, the chair of the Democratic National Committee, informed him that “sufficient was known to make my defeat certain,” Bryan recalled in a memoir.

He composed a telegram to his Republican opponent, William McKinley. “Senator Jones has just informed me that the returns indicate your election, and I hasten to extend my congratulations,” Bryan wrote. “We have submitted the issue to the American people and their will is law.”

After Bryan, concession became a civic duty, performed by telegram or telephone call and then by public speech. Al Smith brought the concession speech to radio in 1928, and it migrated to television soon afterward.

Like other rituals, concessions developed a liturgy. The defeated candidate comes out first. He thanks supporters, declares that their cause will live on, and acknowledges that the other side has prevailed. The victor begins his own remarks by honoring the surrender.

Concessions employ a form of words that linguists call performative speech. The words do not describe or announce an act; the words themselves are the act. “The concession speech, then, is not merely a report of an election result or an admission of defeat,” the political scientist Paul E. Corcoran has written. “It is a constitutive enactment of the new president’s authority.”

In actual war, not the political kind, concession is optional. The winning side may take by force what the losing side refuses to surrender. If the weaker party will not sue for peace, its ramparts may be breached, its headquarters razed, and its leaders taken captive or put to death. There are places in the world where political combat still ends that way, but not here. The loser’s concession is therefore hard to replace.

Consider the 2000 election, which may appear at first glance to demonstrate otherwise. Al Gore conceded to George W. Bush on Election Night, then withdrew his concession and fought a recount battle in Florida until the Supreme Court shut it down. It is commonly said that the Court’s 5–4 ruling decided the contest, but that’s not quite right.

The Court handed down its ruling in Bush v. Gore on December 12, six days before the Electoral College would convene and weeks before Congress would certify the results. Even with canvassing halted in Florida, Gore had the constitutional means to fight on, and some advisers urged him to do so. If he had brought the dispute to Congress, he would have held high ground as the Senate’s presiding officer.

Not until Gore addressed the nation on December 13, the day after the Court’s decision, did the contest truly end. Speaking as a man with unexpended ammunition, Gore laid down his arms. “I accept the finality of this outcome, which will be ratified next Monday in the Electoral College,” he said. “And tonight, for the sake of our unity as a people and the strength of our democracy, I offer my concession.”

We have no precedent or procedure to end this election if Biden seems to carry the Electoral College but Trump refuses to concede. We will have to invent one.

Trump is, by some measures, a weak authoritarian. He has the mouth but not the muscle to work his will with assurance. Trump denounced Special Counsel Robert Mueller but couldn’t fire him. He accused his foes of treason but couldn’t jail them. He has bent the bureaucracy and flouted the law but not broken free altogether of their restraints.

A proper despot would not risk the inconvenience of losing an election. He would fix his victory in advance, avoiding the need to overturn an incorrect outcome. Trump cannot do that.

But he’s not powerless to skew the proceedings—first on Election Day and then during the Interregnum. He could disrupt the vote count where it’s going badly, and if that does not work, try to bypass it altogether. On Election Day, Trump and his allies can begin by suppressing the Biden vote.

There is no truth to be found in dancing around this point, either: Trump does not want Black people to vote. (He said as much in 2017—on Martin Luther King Day, no less—to a voting-­rights group co-founded by King, according to a recording leaked to Politico.) He does not want young people or poor people to vote. He believes, with reason, that he is less likely to win reelection if turnout is high at the polls. This is not a “both sides” phenomenon. In present-day politics, we have one party that consistently seeks advantage in depriving the other party’s adherents of the right to vote.

Just under a year ago, Justin Clark gave a closed-door talk in Wisconsin to a select audience of Republican lawyers. He thought he was speaking privately, but someone had brought a recording device. He had a lot to say about Election Day operations, or “EDO.”

At the time, Clark was a senior lieutenant with Trump’s re­election campaign; in July, he was promoted to deputy campaign manager. “Wisconsin’s the state that is going to tip this one way or the other … So it makes EDO really, really, really important,” he said. He put the mission bluntly: “Traditionally it’s always been Republicans suppressing votes … [Democrats’] voters are all in one part of the state, so let’s start playing offense a little bit. And that’s what you’re going to see in 2020. That’s what’s going to be markedly different. It’s going to be a much bigger program, a much more aggressive program, a much better-funded program, and we’re going to need all the help we can get.” (Clark later claimed that his remarks had been misconstrued, but his explanation made no sense in context.)

Of all the favorable signs for Trump’s Election Day operations, Clark explained, “first and foremost is the consent decree’s gone.” He was referring to a court order forbidding Republican operatives from using any of a long list of voter-purging and intimidation techniques. The expiration of that order was a “huge, huge, huge, huge deal,” Clark said.

His audience of lawyers knew what he meant. The 2020 presidential election will be the first in 40 years to take place without a federal judge requiring the Republican National Committee to seek approval in advance for any “ballot security” operations at the polls. In 2018, a federal judge allowed the consent decree to expire, ruling that the plaintiffs had no proof of recent violations by Republicans. The consent decree, by this logic, was not needed, because it worked.

The order had its origins in the New Jersey gubernatorial election of 1981. According to the district court’s opinion in Democratic National Committee v. Republican National Committee, the RNC allegedly tried to intimidate voters by hiring off-duty law-enforcement officers as members of a “National Ballot Security Task Force,” some of them armed and carrying two-way radios. According to the plaintiffs, they stopped and questioned voters in minority neighborhoods, blocked voters from entering the polls, forcibly restrained poll workers, challenged people’s eligibility to vote, warned of criminal charges for casting an illegal ballot, and generally did their best to frighten voters away from the polls. The power of these methods relied on well-founded fears among people of color about contact with police.

This year, with a judge no longer watching, the Republicans are recruiting 50,000 volunteers in 15 contested states to monitor polling places and challenge voters they deem suspicious-looking. Trump called in to Fox News on August 20 to tell Sean Hannity, “We’re going to have sheriffs and we’re going to have law enforcement and we’re going to have, hopefully, U.S. attorneys” to keep close watch on the polls. For the first time in decades, according to Clark, Republicans are free to combat voter fraud in “places that are run by Democrats.”

Voter fraud is a fictitious threat to the outcome of elections, a pretext that Republicans use to thwart or discard the ballots of likely opponents. An authoritative report by the Brennan Center for Justice, a nonpartisan think tank, calculated the rate of voter fraud in three elections at between 0.0003 percent and 0.0025 percent. Another investigation, from Justin Levitt at Loyola Law School, turned up 31 credible allegations of voter impersonation out of more than 1 billion votes cast in the United States from 2000 to 2014. Judges in voting-rights cases have made comparable findings of fact.

Nonetheless, Republicans and their allies have litigated scores of cases in the name of preventing fraud in this year’s election. State by state, they have sought—with some success—to purge voter rolls, tighten rules on provisional votes, uphold voter-­identification requirements, ban the use of ballot drop boxes, reduce eligibility to vote by mail, discard mail-in ballots with technical flaws, and outlaw the counting of ballots that are postmarked by Election Day but arrive afterward. The intent and effect is to throw away votes in large numbers.

These legal maneuvers are drawn from an old Republican playbook. What’s different during this cycle, aside from the ferocity of the efforts, is the focus on voting by mail. The president has mounted a relentless assault on postal balloting at the exact moment when the coronavirus pandemic is driving tens of millions of voters to embrace it.

This year’s presidential election will see voting by mail on a scale unlike any before—some states are anticipating a tenfold increase in postal balloting. A 50-state survey by The Washington Post found that 198 million eligible voters, or at least 84 percent, will have the option to vote by mail.

Trump has denounced mail-in voting often and urgently, airing fantastical nightmares. One day he tweeted, “mail-in voting will lead to massive fraud and abuse. it will also lead to the end of our great republican party. we can never let this tragedy befall our nation.” Another day he pointed to an imaginary—and easily debunked—scenario of forgery from abroad: “rigged 2020 election: millions of mail-in ballots will be printed by foreign countries, and others. it will be the scandal of our times!”

By late summer Trump was declaiming against mail-in voting an average of nearly four times a day—a pace he had reserved in the past for existential dangers such as impeachment and the Mueller investigation: “Very dangerous for our country.” “A catastrophe.” “The greatest rigged election in history.”

Summer also brought reports that the U.S. Postal Service, the government’s most popular agency, was besieged from within by Louis DeJoy, Trump’s new postmaster general and a major Republican donor. Service cuts, upper-management restructuring, and chaotic operational changes were producing long delays. At one sorting facility, the Los Angeles Times reported, “workers fell so far behind processing packages that by early August, gnats and rodents were swarming around containers of rotted fruit and meat, and baby chicks were dead inside their boxes.”

In the name of efficiency, the Postal Service began de­commissioning 10 percent of its mail-sorting machines. Then came word that the service would no longer treat ballots as first-class mail unless some states nearly tripled the postage they paid, from 20 to 55 cents an envelope. DeJoy denied any intent to slow down voting by mail, and the Postal Service withdrew the plan under fire from critics.

If there were doubts about where Trump stood on these changes, he resolved them at an August 12 news conference. Democrats were negotiating for a $25 billion increase in postal funding and an additional $3.6 billion in election assistance to states. “They don’t have the money to do the universal mail-in voting. So therefore, they can’t do it, I guess,” Trump said. “It’s very simple. How are they going to do it if they don’t have the money to do it?”

What are we to make of all this?

In part, Trump’s hostility to voting by mail is a reflection of his belief that more voting is bad for him in general. Democrats, he said on Fox & Friends at the end of March, want “levels of voting that, if you ever agreed to it, you’d never have a Republican elected in this country again.”

Some Republicans see Trump’s vendetta as self-defeating. “It to me appears entirely irrational,” Jeff Timmer, a former executive director of the Michigan Republican Party, told me. “The Trump campaign and RNC and by fiat their state party organizations are engaging in suppressing their own voter turnout,” including Republican seniors who have voted by mail for years.

But Trump’s crusade against voting by mail is a strategically sound expression of his plan for the Interregnum. The president is not actually trying to prevent mail-in balloting altogether, which he has no means to do. He is discrediting the practice and starving it of resources, signaling his supporters to vote in person, and preparing the ground for post–Election Night plans to contest the results. It is the strategy of a man who expects to be outvoted and means to hobble the count.

Voting by mail does not favor either party “during normal times,” according to a team of researchers at Stanford, but that phrase does a lot of work. Their findings, which were published in June, did not take into account a president whose words alone could produce a partisan skew. Trump’s systematic predictions of fraud appear to have had a powerful effect on Republican voting intentions. In Georgia, for example, a Monmouth University poll in late July found that 60 percent of Democrats but only 28 percent of Republicans were likely to vote by mail. In the battleground states of Pennsylvania and North Carolina, hundreds of thousands more Democrats than Republicans have requested mail-in ballots.

Trump, in other words, has created a proxy to distinguish friend from foe. Republican lawyers around the country will find this useful when litigating the count. Playing by the numbers, they can treat ballots cast by mail as hostile, just as they do ballots cast in person by urban and college-town voters. Those are the ballots they will contest.

The battle space of the Interregnum, if trends hold true, will be shaped by a phenomenon known as the “blue shift.”

Edward Foley, an Ohio State professor of constitutional law and a specialist in election law, pioneered research on the blue shift. He found a previously un­remarked-upon pattern in the overtime count—the canvass after Election Night that tallies late-reporting precincts, un­processed absentee votes, and provisional ballots cast by voters whose eligibility needed to be confirmed. For most of American history, the overtime count produced no predictably partisan effect. In any given election year, some states shifted red in the canvass after Election Day and some shifted blue, but the shifts were seldom large enough to matter.

Two things began to change about 20 years ago. The overtime count got bigger, and it trended more and more blue. In an updated paper this year, Foley and his co-author, Charles Stewart III of MIT, said they could not fully explain why the shift favors Democrats. (Some factors: Urban returns take longer to count, and most provisional ballots are cast by young, low-income, or mobile voters, who lean blue.) During overtime in 2012, Barack Obama strengthened his winning margins in swing states like Florida (with a net increase of 27,281 votes), Michigan (60,695), Ohio (65,459), and Pennsylvania (26,146). Obama would have won the presidency anyway, but shifts of that magnitude could have changed the outcomes of many a closer contest. Hillary Clinton picked up tens of thousands of overtime votes in 2016, but not enough to save her.

The blue shift has yet to decide a presidential election, but it upended the Arizona Senate race in 2018. Republican Martha McSally seemed to have victory in her grasp with a lead of 15,403 votes the day after Election Day. Canvassing in the days that followed swept the Democrat, Kyrsten Sinema, into the Senate with “a gigantic overtime gain of 71,303 votes,” Foley wrote.

It was Florida, however, that seized Trump’s attention that year. On Election Night, Republicans were leading in tight contests for governor and U.S. senator. As the blue shift took effect, Ron DeSantis watched his lead shrink by 18,416 votes in the governor’s race. Rick Scott’s Senate margin fell by 20,231. By early morning on November 12, six days after Election Day, Trump had seen enough. “The Florida Election should be called in favor of Rick Scott and Ron DeSantis in that large numbers of new ballots showed up out of nowhere, and many ballots are missing or forged,” he tweeted, baselessly. “An honest vote count is no longer possible—ballots massively infected. Must go with Election Night!”

Trump was panicked enough by the blue shift in somebody else’s election to fabricate allegations of fraud. In this election, when his own name is on the ballot, the blue shift could be the largest ever observed. Mail-in votes require more time to count even in a normal year, and this year there will be tens of millions more of them than in any election before. Many states forbid the processing of early-arriving mail ballots before Election Day; some allow late-arriving ballots to be counted.

Trump’s instinct as a spectator in 2018—to stop the count—looks more like strategy this year. “There are results that come in Election Night,” a legal adviser to Trump’s national campaign, who would not agree to be quoted by name, told me. “There’s an expectation in the country that there will be winners and losers called. If the Election Night results get changed because of the ballots counted after Election Day, you have the basic ingredients for a shitstorm.”

There is no “if” about it, I said. The count is bound to change. “Yeah,” the adviser agreed, and canvassing will produce more votes for Biden than for Trump. Democrats will insist on dragging out the canvass for as long as it takes to count every vote. The resulting conflict, the adviser said, will be on their heads.

“They are asking for it,” he said. “They’re trying to maximize their electoral turnout, and they think there are no downsides to that.” He added, “There will be a count on Election Night, that count will shift over time, and the results when the final count is given will be challenged as being inaccurate, fraudulent—pick your word.”

The worst case for an orderly count is also considered by some election modelers the likeliest: that Trump will jump ahead on Election Night, based on in-person returns, but his lead will slowly give way to a Biden victory as mail-in votes are tabulated. Josh Mendelsohn, the CEO of the Democratic data-modeling firm Hawkfish, calls this scenario “the red mirage.” The turbulence of that interval, fed by street protests, social media, and Trump’s desperate struggles to lock in his lead, can only be imagined. “Any scenario that you come up with will not be as weird as the reality of it,” the Trump legal adviser said.

Election lawyers speak of a “margin of litigation” in close races. The tighter the count in early reports, and the more votes remaining to count, the greater the incentive to fight in court. If there were such a thing as an Election Administrator’s Prayer, as some of them say only half in jest, it would go, “Lord, let there be a landslide.”

Could a landslide spare us conflict in the Interregnum? In theory, yes. But the odds are not promising.

It is hard to imagine a Trump lead so immense on Election Night that it places him out of Biden’s reach. Unless the swing states manage to count most of their mail-in ballots that night, which will be all but impossible for some of them, the expectation of a blue shift will keep Biden fighting on. A really big Biden lead on Election Night, on the other hand, could leave Trump without plausible hope of catching up. If this happens, we may see it first in Florida. But this scenario is awfully optimistic for Biden, considering the GOP advantage among in-person voters, and in any case Trump will not concede defeat. This early in the Interregnum, he will have practical options to keep the contest alive.

Both parties are bracing for a torrent of emergency motions in state and federal courts. They have already been skirmishing from courthouse to courthouse all year in more than 40 states, and Election Day will begin a culminating phase of legal combat.

Mail-in ballots will have plenty of flaws for the Trump lawyers to seize upon. Voting by mail is more complicated than voting in person, and technical errors are common­place at each step. If voters supply a new address, or if they write a different version of their name (for example, by shortening Benjamin to Ben), or if their signature has changed over the years, or if they print their name on the signature line, or if they fail to seal the ballot inside an inner security envelope, their votes may not count. With in-person voting, a poll worker in the precinct can resolve small errors like these, for instance by directing a voter to the correct signature line, but people voting by mail may have no opportunity to address them.

During the primaries this spring, Republican lawyers did dry runs for the November vote at county election offices around the country. An internal memo prepared by an attorney named J. Matthew Wolfe for the Pennsylvania Republican Party in June reported on one such exercise. Wolfe, along with another Republican lawyer and a member of the Trump campaign, watched closely but did not intervene as election commissioners in Philadelphia canvassed mail-in and provisional votes. Wolfe cataloged imperfections, taking note of objections that his party could have raised.

There were missing signatures and partial signatures and signatures placed in the wrong spot. There were names on the inner security envelopes, which are supposed to be unmarked, and ballots without security envelopes at all. Some envelopes arrived “without a postmark or with an illegible postmark,” Wolfe wrote. (Watch for postmarks to become the hanging chads of 2020.) Some voters wrote their birthdate where a signature date belonged, and others put down “an impossible date, like a date after the primary election.”

Some of the commissioners’ decisions “were clear violations of the direction in and language of the election code,” Wolfe wrote. He recommended that “someone connected with the party review each application and each mail ballot envelope” in November. That is exactly the plan.

Legal teams on both sides are planning for simultaneous litigation, on the scale of Florida during the 2000 election, in multiple battleground states. “My money would be on Texas, Georgia, and Florida” to be trouble spots, Myrna Pérez, the director of voting rights and elections at the Brennan Center, told me.

There are endless happenstances in any election for lawyers to exploit. In Montgomery County, Pennsylvania, not far from Wolfe’s Philadelphia experiment, the county Republican committee gathered surveillance-style photographs of purportedly suspicious goings-on at a ballot drop box during the primary. In one sequence, a county employee is described as placing “unsecured ballots” in the trunk of a car. In another, a security guard is said to be “disconnecting the generator which supplies power to the security cameras.” The photos could mean anything—­it’s impossible to tell, out of context—but they are exactly the kind of ersatz evidence that is sure to go viral in the early days of the Interregnum.

The electoral combat will not confine itself to the courtroom. Local election adjudicators can expect to be named and doxed and pilloried as agents of George Soros or antifa. Aggressive crowds of self-proclaimed ballot guardians will be spoiling to reenact the “Brooks Brothers riot” of the Bush v. Gore Florida recount, when demonstrators paid by the Bush campaign staged a violent protest that physically prevented canvassers from completing a recount in Miami-Dade County.

Things like this have already happened, albeit on a smaller scale than we can expect in November. With Trump we must also ask: What might a ruthless incumbent do that has never been tried before?

Suppose that caravans of Trump supporters, adorned in Second Amendment accessories, converge on big-city polling places on Election Day. They have come, they say, to investigate reports on social media of voter fraud. Counter­protesters arrive, fistfights break out, shots are fired, and voters flee or cannot reach the polls.

Then suppose the president declares an emergency. Federal personnel in battle dress, staged nearby in advance, move in to restore law and order and secure the balloting. Amid ongoing clashes, they stay to monitor the canvass. They close the streets that lead to the polls. They take custody of uncounted ballots in order to preserve evidence of fraud.

“The president can’t cancel the election, but what if he says, ‘We’re in an emergency, and we’re shutting down this area for a period of time because of the violence taking place’?” says Norm Ornstein of the American Enterprise Institute. If you are in Trump’s camp and heedless of boundaries, he said, “what I would expect is you’re not going to do one or two of these things—you’ll do as many as you can.”

There are variations of the nightmare. The venues of intervention could be post offices. The predicate could be a putative intelligence report on forged ballots sent from China.

This is speculation, of course. But none of these scenarios is far removed from things the president has already done or threatened to do. Trump dispatched the National Guard to Washington, D.C., and sent Department of Homeland Security forces to Portland, Oregon, and Seattle during summertime protests for racial justice, on the slender pretext of protecting federal buildings. He said he might invoke the Insurrection Act of 1807 and “deploy the United States military” to “Democrat-run cities” in order to protect “life and property.” The federal government has little basis to intercede during elections, which are largely governed by state law and administered by about 10,500 local jurisdictions, but no one familiar with Attorney General Bill Barr’s view of presidential power should doubt that he can find authority for Trump.

With every day that passes after November 3, the president and his allies can hammer home the message that the legitimate tabulation is over and the Democrats are refusing to honor the results. Trump has been flogging this horse already for months. In July he tweeted, “Must know Election results on the night of the Election, not days, months, or even years later!”

Does it matter what Trump says? It is tempting to liken a vote count to the score at a sporting event. The losing coach can belly­ache all he likes, but when the umpire makes the call, the game is over. An important thing to know about the Interregnum is that there is no umpire—no singular authority who can decide the contest and lay it to rest. There is a series of lesser officiants, each confined in jurisdiction and tangled in opaque rules.

Trump’s strategy for this phase of the Interregnum will be a play for time as much as a concerted attempt to squelch the count and disqualify Biden votes. The courts may eventually weigh in. But by then, the forum of decision may already have moved elsewhere.

The Interregnum allots 35 days for the count and its attendant lawsuits to be resolved. On the 36th day, December 8, an important deadline arrives.

At this stage, the actual tabulation of the vote becomes less salient to the outcome. That sounds as though it can’t be right, but it is: The combatants, especially Trump, will now shift their attention to the appointment of presidential electors.

December 8 is known as the “safe harbor” deadline for appointing the 538 men and women who make up the Electoral College. The electors do not meet until six days later, December 14, but each state must appoint them by the safe-harbor date to guarantee that Congress will accept their credentials. The controlling statute says that if “any controversy or contest” remains after that, then Congress will decide which electors, if any, may cast the state’s ballots for president.

We are accustomed to choosing electors by popular vote, but nothing in the Constitution says it has to be that way. Article II provides that each state shall appoint electors “in such Manner as the Legislature thereof may direct.” Since the late 19th century, every state has ceded the decision to its voters. Even so, the Supreme Court affirmed in Bush v. Gore that a state “can take back the power to appoint electors.” How and when a state might do so has not been tested for well over a century.

Trump may test this. According to sources in the Republican Party at the state and national levels, the Trump campaign is discussing contingency plans to bypass election results and appoint loyal electors in battleground states where Republicans hold the legislative majority. With a justification based on claims of rampant fraud, Trump would ask state legislators to set aside the popular vote and exercise their power to choose a slate of electors directly. The longer Trump succeeds in keeping the vote count in doubt, the more pressure legislators will feel to act before the safe-harbor deadline expires.

To a modern democratic sensibility, discarding the popular vote for partisan gain looks uncomfortably like a coup, whatever license may be found for it in law. Would Republicans find that position disturbing enough to resist? Would they cede the election before resorting to such a ploy? Trump’s base would exact a high price for that betrayal, and by this point party officials would be invested in a narrative of fraud.

The Trump-campaign legal adviser I spoke with told me the push to appoint electors would be framed in terms of protecting the people’s will. Once committed to the position that the overtime count has been rigged, the adviser said, state lawmakers will want to judge for themselves what the voters intended.

“The state legislatures will say, ‘All right, we’ve been given this constitutional power. We don’t think the results of our own state are accurate, so here’s our slate of electors that we think properly reflect the results of our state,’ ” the adviser said. Democrats, he added, have exposed themselves to this stratagem by creating the conditions for a lengthy overtime.

“If you have this notion,” the adviser said, “that ballots can come in for I don’t know how many days—in some states a week, 10 days—then that onslaught of ballots just gets pushed back and pushed back and pushed back. So pick your poison. Is it worse to have electors named by legislators or to have votes received by Election Day?”

When The Atlantic asked the Trump campaign about plans to circumvent the vote and appoint loyal electors, and about other strategies discussed in the article, the deputy national press secretary did not directly address the questions. “It’s outrageous that President Trump and his team are being villainized for upholding the rule of law and transparently fighting for a free and fair election,” Thea McDonald said in an email. “The mainstream media are giving the Democrats a free pass for their attempts to completely uproot the system and throw our election into chaos.” Trump is fighting for a trustworthy election, she wrote, “and any argument otherwise is a conspiracy theory intended to muddy the waters.”

In Pennsylvania, three Republican leaders told me they had already discussed the direct appointment of electors among themselves, and one said he had discussed it with Trump’s national campaign.

“I’ve mentioned it to them, and I hope they’re thinking about it too,” Lawrence Tabas, the Pennsylvania Republican Party’s chairman, told me. “I just don’t think this is the right time for me to be discussing those strategies and approaches, but [direct appointment of electors] is one of the options. It is one of the available legal options set forth in the Constitution.” He added that everyone’s preference is to get a swift and accurate count. “If the process, though, is flawed, and has significant flaws, our public may lose faith and confidence” in the election’s integrity.

Jake Corman, the state’s Senate majority leader, preferred to change the subject, emphasizing that he hoped a clean vote count would produce a final tally on Election Night. “The longer it goes on, the more opinions and the more theories and the more conspiracies [are] created,” he told me. If controversy persists as the safe-harbor date nears, he allowed, the legislature will have no choice but to appoint electors. “We don’t want to go down that road, but we understand where the law takes us, and we’ll follow the law.”

Republicans control both legislative chambers in the six most closely contested battleground states. Of those, Arizona and Florida have Republican governors, too. In Michigan, North Carolina, Pennsylvania, and Wisconsin, the governors are Democrats.

Foley, the Ohio State election scholar, has mapped the ripple effects if Republican legislators were to appoint Trump electors in defiance of the vote in states like Pennsylvania and Michigan. The Democratic governors would respond by certifying the official count, a routine exercise of their authority, and they would argue that legislators could not lawfully choose different electors after the vote had taken place. Their “certificates of ascertainment,” dispatched to the National Archives, would say that their states had appointed electors committed to Biden. Each competing set of electors would have the imprimatur of one branch of state government.

In Arizona, Secretary of State Katie Hobbs, who oversees elections, is a Democrat. She could assert her own power to certify the voting results and forward a slate of Biden electors. Even in Florida, which has unified Republican rule, electors pledged to Biden could meet and certify their own votes in hope of triggering a “controversy or contest” that would leave their state’s outcome to Congress. Much the same thing almost happened during the Florida recount battle of 2000. Republican Governor Jeb Bush certified electors for his brother, George W. Bush, on November 26 of that year, while litigation of the recount was still under way. Gore’s chief lawyer, Ronald Klain, responded by booking a room in the old Florida capitol building for Democratic electors to cast rival ballots for Gore. Only Gore’s concession, five days before the Electoral College vote, mooted that plan.

In any of these scenarios, the Electoral College would convene on December 14 without a consensus on who had legitimate claims to cast the deciding votes.

Rival slates of electors could hold mirror-image meetings in Harris­burg, Lansing, Tallahassee, or Phoenix, casting the same electoral votes on opposite sides. Each slate would transmit its ballots, as the Constitution provides, “to the seat of the government of the United States, directed to the President of the Senate.” The next move would belong to Vice President Mike Pence.

This would be a genuine constitutional crisis, the first but not the last of the Interregnum. “Then we get thrown into a world where anything could happen,” Norm Ornstein says.

Two men are claiming the presidency. The next occasion to settle the matter is more than three weeks away.

January 6 comes just after the new Congress is sworn in. Control of the Senate will be crucial to the presidency now.

Pence, as president of the Senate, would hold in his hands two conflicting electoral certificates from each of several swing states. The Twelfth Amendment says only this about what happens next: “The President of the Senate shall, in the presence of the Senate and the House of Representatives, open all the certificates and the votes shall then be counted.”

Note the passive voice. Who does the counting? Which certificates are counted?

The Trump team would take the position that the constitutional language leaves those questions to the vice president. This means that Pence has the unilateral power to announce his own reelection, and a second term for Trump. Democrats and legal scholars would denounce the self-dealing and point out that Congress filled the gaps in the Twelfth Amendment with the Electoral Count Act, which provides instructions for how to resolve this kind of dispute. The trouble with the instructions is that they are widely considered, in Foley’s words, to be “convoluted and impenetrable,” “confusing and ugly,” and “one of the strangest pieces of statutory language ever enacted by Congress.”

If the Interregnum is a contest in search of an umpire, it now has 535 of them, and a rule book that no one is sure how to read. The presiding officer is one of the players on the field.

Foley has produced a 25,000-word study in the Loyola University Chicago Law Journal that maps out the paths the ensuing fight could take if only one state’s electoral votes are in play.

If Democrats win back the Senate and hold the House, then all roads laid out in the Electoral Count Act lead eventually to a Biden presidency. The reverse applies if Republicans hold the Senate and unexpectedly win back the House. But if Congress remains split, there are conditions in which no decisive outcome is possible—no result that has clear force of law. Each party could cite a plausible reading of the rules in which its candidate has won. There is no tie-breaking vote.

How can it be that Congress slips into unbreakable deadlock? The law is a labyrinth in these parts, too intricate to map in a magazine article, but I can sketch one path.

Suppose Pennsylvania alone sends rival slates of electors, and their 20 votes will decide the presidency.

One reading of the Electoral Count Act says that Congress must recognize the electors certified by the governor, who is a Democrat, unless the House and Senate agree otherwise. The House will not agree otherwise, and so Biden wins Pennsylvania and the White House. But Pence pounds his gavel and rules against this reading of the law, instead favoring another, which holds that Congress must discard both contested slates of electors. The garbled statute can plausibly be read either way.

With Pennsylvania’s electors disqualified, 518 electoral votes remain. If Biden holds a narrow lead among them, he again claims the presidency, because he has “the greatest number of votes,” as the Twelfth Amendment prescribes. But Republicans point out that the same amendment requires “a majority of the whole number of electors.” The whole number of electors, Pence rules, is 538, and Biden is short of the required 270.

On this argument, no one has attained the presidency, and the decision is thrown to the House, with one vote per state. If the current partisan balance holds, 26 out of 50 votes will be for Trump.

Before Pence can move on from Pennsylvania to Rhode Island, which is next on the alphabetical list as Congress counts the vote, House Speaker Nancy Pelosi expels all senators from the floor of her chamber. Now Pence is prevented from completing the count “in the presence of” the House, as the Constitution requires. Pelosi announces plans to stall indefinitely. If the count is still incomplete on Inauguration Day, the speaker herself will become acting president.

Pelosi prepares to be sworn in on January 20 unless Pence reverses his ruling and accepts that Biden won. Pence does not budge. He reconvenes the Senate in another venue, with House Republicans squeezing in, and purports to complete the count, making Trump the president-elect. Three people now have supportable claims to the Oval Office.

There are other paths in the labyrinth. Many lead to dead ends.

This is the next constitutional crisis, graver than the one three weeks before, because the law and the Constitution provide for no other authority to consult. The Supreme Court may yet intervene, but it may also shy away from another traumatizing encounter with a fundamentally political question.

Sixty-four days have passed since the election. Stalemate reigns. Two weeks remain until Inauguration Day.

Foley, who foresaw this impasse, knows of no solution. He cannot tell you how we avoid it under current law, or how it ends. It is not so much, at this point, a question of law. It is a question of power. Trump has possession of the White House. How far will he push boundaries to keep it, and who will push back? It is the same question the president has posed since the day he took office.

I hoped to gain some insight from a series of exercises conducted this summer by a group of former elected officials, academics, political strategists, and lawyers. In four days of simulations, the Transition Integrity Project modeled the election and its aftermath in an effort to find pivot points where things could fall apart.

They found plenty. Some of the scenarios included dueling slates of electors of the kind I have described. In one version it was the Democratic governor of Michigan who first resorted to appointing electors, after Trump ordered the National Guard to halt the vote count and a Trump-friendly guardsman destroyed mail-in ballots. John Podesta, Hillary Clinton’s campaign chair in 2016, led a Biden team in another scenario that was prepared to follow Trump to the edge of civil war, encouraging three blue states to threaten secession. Norm-breaking begat norm-breaking. (Clinton herself, in an August interview for Showtime’s The Circus, caught the same spirit. “Joe Biden should not concede under any circumstances,” she said.)

A great deal has been written about the proceedings, including a firsthand account from my colleague David Frum. But the coverage had a puzzling gap. None of the stories fully explained how the contest ended. I wanted to know who took the oath of office.

I called Rosa Brooks, a Georgetown professor who co-founded the project. Unnervingly, she had no answers for me. She did not know how the story turned out. In half of the simulations, the participants did not make it as far as Inauguration Day.

“We got to points in the scenarios where there was a constitutional impasse, no clear means of resolution in sight, street-level violence,” she said. “I think in one of them we had Trump invoking the Insurrection Act and we had troops in the streets … Five hours had gone by and we sort of said, ‘Okay, we’re done.’ ” She added: “Once things were clearly off the rails, there was no particular benefit to seeing exactly how far off they would go.”

“Our goal in doing this was to try to identify intervention moments, to identify moments where we could then look back and say, ‘What would have changed this? What would have kept it from getting this bad?’ ” Brooks said. The project didn’t make much progress there. No lessons were learned about how to restrain a lawless president once a conflict was under way, no alternative moves devised to stave off disaster. “I suppose you could say we were in terra incognita: no one could predict what would happen anymore,” Brooks told me in a follow-up email.

The political system may no longer be strong enough to preserve its integrity. It’s a mistake to take for granted that election boards and state legislatures and Congress are capable of drawing lines that ensure a legitimate vote and an orderly transfer of power. We may have to find a way to draw those lines ourselves.

There are reforms to consider some other day, when an election is not upon us. Small ones, like clearing up the murky parts of the Electoral Count Act. Big ones, like doing away with the Electoral College. Obvious ones, like appropriating money to help cash-starved election authorities upgrade their operations in order to speed up and secure the count on Election Day.

Right now, the best we can do is an ad hoc defense of democracy. Begin by rejecting the temptation to think that this election will carry on as elections usually do. Something far out of the norm is likely to happen. Probably more than one thing. Expecting other­wise will dull our reflexes. It will lull us into spurious hope that Trump is tractable to forces that constrain normal incumbents.

If you are a voter, think about voting in person after all. More than half a million postal votes were rejected in this year’s primaries, even without Trump trying to suppress them. If you are at relatively low risk for COVID-19, volunteer to work at the polls. If you know people who are open to reason, spread word that it is normal for the results to keep changing after Election Night. If you manage news coverage, anticipate extra­constitutional measures, and position reporters and crews to respond to them. If you are an election administrator, plan for contingencies you never had to imagine before. If you are a mayor, consider how to deploy your police to ward off interlopers with bad intent. If you are a law-enforcement officer, protect the freedom to vote. If you are a legislator, choose not to participate in chicanery. If you are a judge on the bench in a battleground state, refresh your acquaintance with election case law. If you have a place in the military chain of command, remember your duty to turn aside unlawful orders. If you are a civil servant, know that your country needs you more than ever to do the right thing when you’re asked to do otherwise.

Take agency. An election cannot be stolen unless the American people, at some level, acquiesce. One thing Brooks has been thinking about since her exercise came to an end is the power of peaceful protest on a grand scale. “We had players on both sides attempting to mobilize their supporters to turn out in large numbers, and we didn’t really have a good mechanism for deciding, did that make a difference? What kind of difference did that make?” she said. “It left some with some big questions about what if you had Orange Revolution–style mass protest sustained over weeks. What effects would that have?”

Only once, in 1877, has the Interregnum brought the country to the brink of true collapse. We will find no model in that episode for us now.

Four states sent rival slates of electors to Congress in the 1876 presidential race between Democrat Samuel Tilden and Republican Rutherford B. Hayes. When a special tribunal blessed the electors for Hayes, Democrats began parliamentary maneuvers to obstruct the electoral count in Congress. Their plan was to run out the clock all the way to Inauguration Day, when the Republican incumbent, Ulysses S. Grant, would have to step down.

Not until two days before Grant’s term expired did Tilden give in. His concession was based on a repugnant deal for the withdrawal of federal troops from the South, where they were protecting the rights of emancipated Black people. But that was not Tilden’s only inducement.

The threat of military force was in the air. Grant let it be known that he was prepared to declare martial law in New York, where rumor had it that Tilden planned to be sworn in, and to back the inauguration of Hayes with uniformed troops.

That is an unsettling precedent for 2021. If our political institutions fail to produce a legitimate president, and if Trump maintains the stalemate into the new year, the chaos candidate and the commander in chief will be one and the same.

This article appears in the November 2020 print edition with the headline “The Election That Could Break America.”