American presidents customarily leave office when voters reject them. Czars, emperors, and would-be prime ministers for life do whatever they can to hold on to power. To extend his rule until 2036, Russian President Vladimir Putin recently held a referendum to amend his country’s constitution. While some Russians publicly opposed the proposal, few had any doubt about the outcome. Two years ago, Chinese President Xi Jinping succeeded in eliminating term limits that had been established after Mao Zedong’s Cultural Revolution—a period of ideological madness that killed tens of millions of people, including family members of the country’s ruling class. In his grand presidential address of 2017, Xi stated specific objectives for his nation to achieve by 2025, 2035, and 2049, the centennial of the People’s Republic of China—suggesting that he intended to lead China until at least 2035 (when he would be 82).

Even in democracies, circumstances sometimes allow deeply compromised leaders to remain in office. Benjamin Netanyahu has already become the longest-serving prime minister in Israel’s history, having survived repeated scandals. He is currently facing an indictment for bribery and fraud brought by Israel’s attorney general. In the past year, even though he could not win a majority of the Israeli Parliament in three successive elections, Netanyahu successfully maneuvered to keep his position on each occasion. Now in the second month of a three-year “deal” with Defense Minister Benny Gantz that allows him to remain prime minister for 18 months, Netanyahu has made a commitment to swap roles for the subsequent 18 months. But many expect him to renege on those terms, or try to renegotiate them, before Gantz takes over.

In short, three leaders whom Donald Trump has praised have extended their tenure in office. On this canvas, what could Trump do? Undoubtedly, a whiff of paranoia is evident in many claims of potential skulduggery now swirling about. And yet the fear that Trump may not leave office, no matter what happens in November, has become mainstream. Joe Biden, the presumptive Democratic nominee, has called the prospect that “this president is going to try to steal this election” his “single greatest concern.

Putin, Xi, Netanyahu, and Trump all differ from one another in many ways, of course. But each has reasons to avoid relinquishing his hold on power. Putin, Xi, and Netanyahu genuinely have grand ambitions. Each wants to expand his country’s formal borders—Russia’s into Ukraine, Israel’s into the West Bank, China’s to reintegrate Hong Kong and Taiwan. Trump’s signature banner promises to “Make America great again,” and the thought that this could include territorial expansion—specifically the purchase of Greenland—has also occurred to him.

The most generous explanation for why these men might believe themselves indispensable is this: Who else can any of them trust to achieve their ambitions? The more cynical questions to ask are: If any of these four leaders leaves office, who could ensure his freedom from arrest and humiliation? Who would guarantee the wealth and well-being of his family and friends?

Historically, heads of state clung to power until they died or were overthrown. Of the 23 czars who ruled Russia from 1547 to 1917, how many voluntarily handed over power to a successor? Zero. Fifteen died of natural causes, six were overthrown, and two were assassinated. Over roughly the same four centuries, China had 16 emperors—all but one of whose reigns ended involuntarily: by death or forced abdication. Call it the czar’s dilemma: However daunting the challenges of holding power, the dangers of losing it are even greater.

America’s Founding Fathers recognized this dilemma. They created a republic, not a monarchy, with an elected chief executive. Acutely conscious of the abuses of “mad King George,” they designed a Constitution of what the presidential scholar Richard Neustadt called “separated institutions sharing power.” By dividing power among the president, Congress, and the courts—and by giving them separate sources of legitimacy—they created a complex system in which each checks and balances the others. While that has invited a continuous struggle for power that has made governing messy and often ugly, their purpose, as Justice Louis Brandeis famously explained, was “not to promote efficiency but to preclude the exercise of arbitrary authority.”

Buttressing constitutional limits on presidential power was the wise precedent that George Washington established when he stepped down after two terms. No one broke that tradition until Franklin D. Roosevelt, who was elected four times during the Great Depression and World War II. As one of his colleagues quipped, the only way he would ever leave the White House was in a coffin. And that’s the way he exited in 1945. Within six years of his death, Congress had passed and three-quarters of the states had ratified the Twenty-Second Amendment, which now limits presidents to two terms.

But the amendment does not close off every avenue by which a president could hang on. Lawrence Douglas, the author of Will He Go? Trump and the Looming Election Meltdown in 2020, argues that the current American electoral system has a “Chernobyl-like defect”: Nothing in the Constitution or federal laws guarantees a peaceful transfer of power from a sitting president to his successor.

Why would an elected head of state transfer the immense powers of the presidency to an opponent whom he believes could be a threat to his nation, his vision, and even himself? Just because some people claim that the opponent received more votes in an election in which both parties are pointing to irregularities and abuses?

Central to the democratic transfer of power is a norm of deference to process. All elections have many inconsistencies and aberrations. But successful democracies require a strong presumption of respect for the legitimacy of the electoral process—indeed, a willingness to accept outcomes that are messy, controversial, and perhaps even perverse. Yet as poisonous partisanship has spread into every aspect of government, claims of voter suppression, fraudulent mail-in ballots, and abuses in disqualifying voters and counting votes are eroding that presumption. And unfortunately, the labyrinthine process between a citizen’s vote and the outcome of a presidential election offers many opportunities for the suspicious or paranoid to claim that the election was rigged or the results a sham.

While citizens across the nation will vote for president on November 3, the winner is determined not by the popular vote but by the Electoral College. The process has four steps that, under normal conditions, produce a clear result. First, prior to the election, Electoral College votes are allocated among states according to the decennial U.S. census (which presumes that the census is completed properly). Second, citizens cast their vote for the candidate of their choosing (and the corresponding slate of electors), and each state counts votes according to its own procedures. Third, the electors for the winning candidate in each state meet, vote for the candidate to whom they are pledged, and send that report to Congress. Finally, on January 6, a joint session of Congress certifies states’ electoral votes and declares a winner.

To win, a candidate must have 270 electoral votes. But if no candidate reaches that threshold, because some states’ electoral votes are contested or disqualified, the Constitution says that the House of Representatives selects the president and the Senate the vice president. In making these choices, each state casts a single vote. That Democrats have the majority in the House today does not matter. What matters is that in 26 of the 50 states, Republicans hold a majority of the House delegation. If Republicans retain this advantage after the November election, they will elect Trump. Alternatively, if Democrats succeed in wresting a seat or two from Republicans in a few closely divided states, they would then have a majority and elect Biden. If Democrats flip only one state, neither party would reach the minimum 26 votes. The Constitution offers no guidance on what happens next.

This November 3 will be unlike any previous presidential election. Some states are requiring in-person voting, while others are allowing mail-in balloting at an unprecedented level. Many are experimenting with new ways to protect the vote from foreign interference. Amid the many possible uncertainties, the outcome of the November vote could be as confused as that of the Iowa caucus in February. Could we see failures in voting systems like the fiasco in the Georgia primary election last month, or the long wait for results like in the more recent congressional primaries in New York and Kentucky? Under a cloud of fraud and abuse allegations, could Americans see both candidates claiming victory?

For some readers, the thought of a presidential election without a clear victor may sound fanciful. Yet the United States has endured a number of seriously contested presidential elections, each of which highlights dangers that could arise again this year.

In the 1824 contest between John Quincy Adams and Andrew Jackson, Jackson clearly won the popular vote and had the largest number of votes in the Electoral College—though he fell 32 short of a majority. In what Jackson rightly denounced as a “corrupt bargain,” Adams made a deal with Henry Clay, the House speaker who had himself been a candidate for president but finished fourth in the Electoral College. For throwing his support to Adams, Clay became Adams’s secretary of state.* Although he accepted the loss, Jackson spent the next four years waging a political war against Adams, undercutting his authority in Washington and across America on the grounds that his presidency was illegitimate. And when the two faced off again in 1828, Jackson won by a large margin, ushering in the Jacksonian revolution.

A decade after the Civil War, the presidential election of 1876 was even more bitterly contested. The Democratic candidate, Samuel J. Tilden, beat the Republican, Rutherford B. Hayes, in the popular vote, but 20 decisive electoral votes remained in dispute. When neither side could agree on an outcome, the country wobbled on the brink of chaos. Outgoing President Ulysses S. Grant developed contingency plans for martial law, lest the country return to civil war. The election was not resolved until months later, when the parties struck a secret deal. According to the Compromise of 1877, Hayes became president—but in return he agreed to remove federal troops from the South, effectively ending Reconstruction.

In 2000, the presidential contest between George W. Bush and incumbent Vice President Al Gore came down to one swing state: Florida. Bush had a significant lead there as Election Night came to a close. But by morning the following day, television networks had to retract their announcements of the victor. Without a win in Florida, neither candidate had the required 270 electoral votes. As Florida began recounting its ballots, disputes arose about whether ballots with “hanging chads” should be disqualified and, eventually, whether and how the recount should continue. The matter finally went to the Supreme Court, which in a 5–4 decision sided with the Republicans—in effect, awarding victory to Bush. In his concession speech, Gore said: “Let there be no doubt, while I strongly disagree with the Court’s decision, I accept it.”

Historically, the difference between a would-be U.S. president and a czar has been that the former accepts the possibility of defeat at the ballot box. After a heated campaign one year before the nation dissolved into civil war, Stephen Douglas conceded to Abraham Lincoln with the words “Partisan feeling must yield to patriotism.” Yet that clear principle may wither in the face of a close election—especially if the current president is insistent on holding on to power. If our society remains discombobulated amid a continuing pandemic, if voter-suppression measures keep some Americans from having their say, if mail-in-ballot glitches and foreign interference affect vote counts, and if exit polls suggest that the election is close, nightmare scenarios become more probable. Under these conditions, America’s best hope for escaping the czar’s dilemma is for one candidate to win decisively.