Aaron Bernstein / Reuters

This story was updated on July 31, at 4:12pm.

On the same morning that the United States government reported the steepest economic collapse in U.S. history, President Donald Trump mused on Twitter about postponing the 2020 election. Trump is getting desperate, more desperate by the day. What might he do? What should Americans fear?

Earlier this summer, 67 former government officials and academic students of government gathered over four sessions of the nonpartisan Transition Integrity Project to analyze those questions. They included Michael Steele, a former chair of the Republican National Committee; John Podesta, the former White House chief of staff who chaired Hillary Clinton’s 2016 campaign; former Republican members of Congress; and a host of former elected officials, government staffers, consultants, and even journalists. I joined two of the sessions.

The sessions began with scenarios of what might happen on Election Day—a big Biden win, a narrow Biden win, a Trump win in the Electoral College coupled with a loss in the popular vote—and then played war games to ponder what might follow. The goal was not to make predictions, but rather to test scenarios and identify potential weak points in the system. The approach is common in the national-security world, but has not often before been applied to domestic politics.

The organizers of the event will in time produce a formal report on the results. But in light of the president’s ominous tweet yesterday, it’s worth summarizing some of what we found, while respecting the rules under which the event was held—which allowed for the disclosure of the substance of the exercise, but not what individual participants said.

The good news is that Trump cannot postpone the election or the next presidential inauguration; he has no means to do either of those things. Those dates are set by law or in the text of the Constitution.

Nor can Trump somehow cling to power after Inauguration Day once the electoral vote is certified against him. If the Electoral College certifies Joe Biden the winner when its votes are counted in Washington, D.C., on January 6, then at noon on January 20, Donald Trump ceases to be president. His signature loses all legal effect, the officer carrying the nuclear football walks away, the chairman of the Joint Chiefs of Staff does not take his call.

The bottom line: There do exist outer legal boundaries to the mischief that can be done by even the most corrupt president.

The bad news is that there is a lot of mischief that can be done within the legal boundaries by a determined president, especially with the compliance of the attorney general and enough political allies in the state capitals.

The worst news is that, faced with presidential lawlessness, few of the participants at the Transition Integrity Project found effective responses. The courts offered only slow, weak, and unreliable remedies. Street protests were difficult to mobilize and often proved counterproductive. Republican elected officials cowered even in the face of the most outrageous Trump acts. Democratic elected officials lacked the tools and clout to make much difference. Many of the games turned on who made the first bold move. Time after time, that first mover was Trump.

And even in the scenarios in which Biden’s team eventually won—that is, secured possession of the White House at noon on Inauguration Day, 2021—Team Trump by then had thoroughly poisoned the political system.

It diverted public resources to Trump personally.

It preemptively pardoned Trump associates and family members, and tried to pardon Trump himself from criminal charges including money laundering and tax evasion.

It intentionally tried to cause long-term economic damage so as to prevent early economic recovery—and boost Republican chances in the 2022 elections.

It destroyed, hid, or privatized public records.

It tried to sabotage the census to favor Republican redistricting after 2020.

It refused to cooperate with the incoming administration during the transition period, in ways that aggravated both the pandemic response and economic recovery.

And it sowed pervasive mistrust in the integrity of U.S. elections in ways that would polarize and embitter U.S. politics long after 2020.

Despite the president’s personal unpopularity as measured by polls, Trump’s side possessed—and used—important tactical advantages.

Those advantages start with the institutional powers of the presidency, notably the power to federalize the National Guard and take military control of state voting sites. They include also the asymmetry of the U.S. party system, and especially the fiercer team-mindedness of Trump loyalists and pro-Trump media.

The most persistent and powerful advantage, however, was the overconfidence of the legally minded Biden team that the Trump team would respect some norms and limits on its behavior. That expectation was again and again refuted by experience.

All of this, again, was just a tabletop exercise, specifically designed to test extreme scenarios—not a prediction of how things will play out. Perhaps everything will go smoothly. But as the president suggests postponing the election, it’s important to understand the hazards ahead, and the timelines and decision points that may prove crucial.

The voting period

The days of early voting, Election Day itself, and then the period of vote-counting that will follow offer fruitful possibilities for mischief.

In one of our scenarios, the attorney general sent federal marshals backed by the National Guard to seize vote-by-mail ballots, triggering a constitutional catastrophe that delayed the outcome of the count for weeks.

Local Republican officeholders have wide scope to burden voting by what they deem undesirable voters, especially ethnic minorities. The Civil Rights Division of the Justice Department has more or less entirely abandoned the field of voting rights. In the Trump era, the division has shifted its effort toward litigating in support of claims of religious discrimination.

In the exercises, when the vote went against Trump, his team tried to convince his supporters that they had been robbed—and that they were therefore entitled to take extreme, even violent, actions. In our exercises, however, the game-winning strategy was to goad the other side into violence. This was particularly true for Team Trump, whose supporters already fear violence from anarchists and antifa.

The meeting of electors in the states

Under current law, all disputes over vote-counting are supposed to be resolved by December 8, 2020. The electors are supposed to convene on December 14 in their state capitals, where they will sign their electoral ballots. The days from December 8 to December 14 offer Team Trump the last clear chance to alter the outcome.

In some of our scenarios, local Republican officeholders sowed enough confusion to justify sending two slates of electors to Washington to be adjudicated. That was a high-risk tactic that did not usually pay off, but could tempt some pro-Trump state governments.

The meeting of electors in Washington, D.C.

This normally ceremonial event is scheduled for January 6, 2021. It will be presided over by the incumbent vice president, Mike Pence. We tested what might happen in a close result—one in which the Republicans hold on to the Senate and Trump falls short of an Electoral College majority by just a single state’s vote—if Pence somehow tried to insist that the pro-Trump slate of electors was valid.

This did not usually work. Pence was a weak link in the Trump team, too concerned about his own future and his own reputation to go all-out in the way the core Trump team wanted.

Generally, once we got past the December 8 date, the Trump team’s options for keeping power dwindled to zero. What was left then was scorched-earth self-enrichment, self-protection, and spite.

The transition of power

The Obama administration took office amid a national crisis in January 2009, after what is generally regarded by experts as the smoothest and most successful transition in presidential history. The outgoing Bush team kept the Obama team closely informed about decision making after the financial crisis struck in October 2008—and the incoming Obama team scrupulously followed the “one president at a time” rule of crisis management.

Nothing like that can be expected this winter. Instead, we are likely to see a recurrence of 1932–33, when the defeated Herbert Hoover tried to sabotage the incoming Roosevelt administration in hopes of preparing his own comeback in 1936. Trump will soon be fantasizing about running again in 2024. If his health does not permit it, his children may envision a dynasty of their own. These are not realistic plans. The Trump brand will be toxic in U.S. politics after the catastrophes of 2020. But the Trump inner circle will not believe that—and its members may hope that if they can cause Biden to stumble out of the gate, they will benefit.

The Bush administration helped the Obama administration to be ready on day one. The Trump administration may not return that courtesy. In one of our scenarios, Trump moved permanently to Mar-a-Lago the day after the election and never returned to the White House again. The whole government had to operate around a lame-duck president who simply refused to do any work at all.

But we also discussed whether Trump’s need to satisfy his ego and his desire for money might not cause him to foment a transition-season crisis—especially one that would gain him some credit with Russia or the oil states. Postpresidential Trump will face extreme legal and business troubles, including the ruin of the hospitality industry. The flow of payments to his businesses from U.S. taxpayers, from Republican campaigns, from favor-seeking corporations, and from foreign governments will all cease.

What would Trump do to maximize his cash flow before it stops? As lurid as our imaginations were over the four days of disaster planning, on this question, at least, we probably underestimated the dangerous possibilities.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.