For the last 10 years, we’ve been hearing that President Donald Trump will preside over the end of democracy in America. In liberal circles, that assertion is often accepted as fact. For many, the proof is in the evidence from other countries’ democratic declines.
A whole genre of American political writing is issuing this warning. Perhaps the best known entrant is How Democracies Die, by the Harvard political scientists Steven Levitsky and Daniel Ziblatt. The authors mention a few democracies that fought off authoritarianism but overwhelmingly recount just what the title says: how democracies die. There are many other examples of the genre. “I Watched It Happen in Hungary. Now It’s Happening Here,” announces a former U.S. ambassador. Having shaken hands with “democracy killers” from Thailand to Zambia, Brian Klaas warns in The Atlantic, “American democracy is dying.”
But the United States is different from many of the countries that feature prominently in the “death of democracy” literature. And for Americans concerned about what Trump will do in his second term, the ways other democracies have died isn’t the central concern. Those accounts are a bit like detailing how Covid can kill people but not assessing the chances, depending on age and risk factors, that the disease will kill you.
The real question is whether U.S. democracy will survive or not. The genre hardly asks that question, let alone answers it.
To be sure, Trump does all the same things as the authoritarians Levitsky and Ziblatt studied: He has refused to accept electoral defeats; called political opponents criminals and tried to jail them even while backing his own violent supporters; and lashed out at opponents and the media as “enemies of the people” — a chilling phrase that echoes Soviet dictator Joseph Stalin and the Nazi propagandist Joseph Goebbels.
But Trump’s authoritarianism also resembles that of dangerous populists who failed to kill democracy. Careful studies that never seem to get much press find that only about a fifth of dangerous populists actually kill democracy, including in different regions and across different time spans. If you’re serious about weighing the Trump threat, you should be asking what makes the difference between countries where democracy died and countries where it survives.
A significant difficulty in answering this question arises from the structure of the political science discipline itself. When I got my doctorate some 23 years ago, American politics occupied one subfield and comparative politics (the study of all other nations, often via comparison) occupied another — and never the twain should meet. Levitsky and Ziblatt are comparativists. This division has only begun to soften. It’s difficult to apply lessons from elsewhere to Trump’s America because the U.S. political system does differ starkly from others, but this disciplinary divide has made that problem even harder.
I’m also a comparativist, but where Levitsky and Ziblatt focused on Latin America and pre-World War II Europe (plenty of failed democracies in both), I have looked further and wider, from the old Mexican ruling party to the Greek financial crisis, from how economists think to the divisions splitting American liberals. I have found research outside of the literature on dying democracies that asks relevant and important questions. What can we learn, for instance, from the advanced capitalist democracies that survived the brutal 20th century — two world wars, the Great Depression in the 1930s, stagflation in the 1970s, Soviet and Chinese threats? Only two fell to internal authoritarian threats: Italy and Germany between the world wars.
This is a dire moment for U.S. democracy. To make it worse, the assassination of Charlie Kirk threatens to ratchet up not only the recent surge of political violence but to reanimate the poisonous tradition of political assassination that runs through Martin Luther King., Jr., President John F. Kennedy, the many Black leaders murdered during Jim Crow, and the Civil War. Deranged individuals perpetrated the recent attacks, but shared fear that U.S. politics is at an existential moment as the 2026 and 2028 elections approach surely portends more widespread and equally damaging threats of violence.
What if Americans didn’t see the dangers to our democracy as existential? If we had more faith that our democracy would survive Trump 2.0, it might take some venom out of this political atmosphere. It might provide some institutions more confidence to fight unconstitutional Trump demands — to be more like the law firm Perkins Coie, which fought back and not cave like another law firm, Paul, Weiss. It might make ABC and its affiliates more willing to stand up to threats from the administration to police what comedians like Jimmy Kimmel say on the air. The Democratic Party might pay more attention to what they can do for voters instead of harping on the Trump menace — a theme that obviously didn’t work for them in 2024. All Americans might better see beyond this fraught moment to focus more on solving the problems of a democracy that was already in grave need of repair before Trump made the situation far worse.
A careful comparison with countries that fought off autocratic attempts, as well as those that succumbed, suggests that American democracy might be more resilient than you think. At a minimum, it has crucial advantages over democracies that failed. Three main things stand out: None was nearly so rich. None was nearly so long-lived. And none had a legal establishment tracing its genealogy back to the Magna Carta in 1215.
Rich democracies rarely die. As the political scientist Seymour Martin Lipset wrote in 1959, “The more well-to-do a nation, the greater the chances that it will sustain democracy.” In fact, he added, the idea goes back to Aristotle. Whether high per capita income can make a nation democratic is not so clear, and definitions of democracy, let alone calculations of income, can be squishy, but the general claim that wealth sustains democracy has repeatedly held up.
The prominent democracy theorist Adam Przeworski of New York University, born in Poland under the Nazis and raised under the Communists, famously observed in 2005 that while democracy had fallen again and again in developing nations, economies richer than Argentina when the military seized control in 1976 had survived a thousand years all told. None had failed, despite “wars, riots, scandals, economic and governmental crises, hell or high water.”
Turkey and Hungary have since broken the “Argentina 1976” threshold. Still, a 2020 study identified only three types of states where per capita income topped $36,000: democracies, petrostates and the wealthy trading hub of Singapore. Longevity matters, too. A 2005 statistical tour de force found that if a presidency survived more than 50 years and had a per capita income more than $23,000, it had a zero percent chance of failing. (Those studies used different base years for the income levels; I have adjusted them for inflation to approximate 2025 equivalents. By comparison, current U.S. GDP per capita is about $86,000.) Statistical studies do not guarantee there will never be outliers, as Przeworski learned. Still, such broad agreement over so many years is powerful.
Why rich, long-lived democracies are strong is actually puzzling, and despite a large literature on why nations democratize and (lately) why they fail, incredibly little has been written on it. Since the result holds for both parliamentary and presidential regimes, political scientists assume it arises less from the structure of political systems than it does from the underlying civil society.
Przeworski suggested that while lower-income societies can erupt in all-or-nothing struggle over scarce resources, wealthy societies develop an existential fear of losing the rule of law, dreading that fate as worse than any particular electoral loss. In fact, a commitment to the rule of law binds sectors of society in advanced capitalist democracies that might otherwise be in conflict. Educated workers are rewarded in those systems, so it makes sense that they want to protect it. But they are often seconded by those who hope they or their children can join this group. Recent immigrants find in developed democracies what their own nations may lack: a credible legal system, better schools, even drinkable water.
We’ve seen plenty of serious damage to democracy lately. But mostly such backsliding doesn’t kill democracy. Using the V-Dem database, which provides literally hundreds of indicators of governance on a scale from democratic to autocratic, researchers at the University of Texas find that in 30 instances of backsliding between 2000 and 2019 (including Trump 1.0), only eight led to democratic breakdown. That’s a failure rate, even once backsliding has occurred, of about 25 percent.
Civil society — big business, law firms, nonprofits, universities, media — fights backsliding. When Orban’s government creates businesses by showering them with contracts and keeps friendly media outlets afloat as the largest advertiser in the nation, they either submit or face ruin. A wealthy society provides such plentiful resources to independent institutions of civil society, and they span so far and wide across the political landscape, that they are effectively impossible to quash. Longevity surely contributes here, too, because the longer those institutions have to mature, the more established they become.
U.S. democracy has another strength: It’s presidential. The conventional wisdom used to suggest the opposite. Concerned more with the emergence and consolidation of democracy than its survival, theorists saw parliamentary systems as more secure. That’s because prime ministers often lead coalitions of multiple parties, giving diverse voices some say, and when things go wrong, they can be ousted in a vote of no confidence far more easily than a president can be impeached. Also, a few decades ago, presidencies looked unstable partly because the military routinely overthrew them in Latin America.
Recently, military coups are mostly out as ways to seize power. Instead, the current challenge to democracy has been autocrats who win election and try to grab authoritarian power from within to stay in office, and this is easier in parliamentary systems. Prime ministers by definition command a majority vote of parliament, so the opposition can’t even hope to check them by controlling an upper or lower legislative chamber like in the U.S. Congress. And when multiple parties split the vote, a small majority or even a plurality can sometimes win a supermajority in parliament. Notably, in 2010, with just 53 percent of the vote, Viktor Orban won a two-thirds majority in Hungary’s Assembly. It allowed him to gain domination over the Constitutional Court, rewrite the Constitution, and wily step by wily step, cinch autocratic control without even flaunting the letter of the law.
Presidential checks and balances also complicate the lives of aspiring authoritarians by making it harder to replace judges. Kurt Weyland of the University of Texas identifies 17 dangerous populists since 1980 in presidencies of at least middling strength, such as Argentina or Brazil. Only two succeeded in killing democracy, and those two had an important thing in common.
Trump’s support may dismay his opponents, but it’s far lower than he needs to kill off democracy. In presidential systems of at least middling strength, populists who defeated democracy since 1980 had approval ratings of 80 percent or more — nearly twice Trump’s approval rating, which has been stuck in the 40s both in his first term and so far in his second.
The two presidents who in succeeded in killing Weyland’s democracies were Alberto Fujimori in Peru and Nayib Bukele in El Salvador. Both rose to power by solving major crises that helped them rack up public approval ratings of 80 to 90 percent. When a leader is that popular, centers that might check his power — the military, the courts, congress — may support him (they’re comprised of people too) or at least hesitate to coordinate in opposition.
In Fujimori’s case, one crisis was 10,000 percent inflation; the money that had bought a luxurious house five years earlier could purchase a tube of toothpaste when he took office in 1990. He drastically brought the rate down to much more normal levels. He also halted a civil war waged by Shining Path guerrillas, in which tens of thousands of Peruvians had died.
For his part, Bukele won El Salvador’s presidency in 2019 as a populist outsider who pledged to crack down on gangs that had overrun the county. Covid gave him that chance. He locked the nation down, which held the pandemic death rate to less than a fifth that of Mexico or the United States and also suppressed the homicide rate to a fifth of its 2015 peak. Securing a supermajority in Congress, he purged the judiciary, and that was pretty much that.
While Fujimori stars in the death-of-democracy literature as a cautionary example, President Jorge Serrano of Guatemala barely gets a mention. But Serrano’s experience is more common. Having accomplished little and seen his approval rating drop to 20 percent, in 1993 he tried copying Fujimori, closing Congress and the courts. Massive protests erupted, the Constitutional Court declared his coup illegal, he lost military support and he fled the country. Something similar happened last year in South Korea when an unpopular president’s attempt to impose martial law was thwarted dramatically by parliament and protesters.
The United States isn’t supplying the kind of crisis that could generate for Trump a Fujimori-style level of popular support. Advanced nations rarely do. And it’s doubtful the distractable Trump could solve a good crisis if he lucked into one. After all, he didn’t manage Covid particularly well in his first term, ending his presidency with an approval rating of just 34 percent. His attempts to manufacture crises in his second term, such as by sending troops into Los Angeles or Washington, so far seem more likely to hurt him.
Trump is also failing to expand his base of support. While competent autocrats built alliances with supporters, he keeps alienating them, from Elon Musk, still the richest man in the world, to Ileana Garcia, co-founder of Latinas for Trump. Pardoning Jan. 6 rioters who assaulted the police might fire up Trump’s base but is opposed by two thirds of Americans. Likewise, attacking officials his fertile paranoia imagines to be plotting against him; loosing new tariffs daily; passing a budget bill that will kick many of his supporters off Medicaid while cutting taxes for the wealthy — none of this is likely to help him build support. As job creation stagnates and inflation slowly rises, the very issues Trump campaigned on could weaken him. With around 8 percent more Americans disapproving than approving him, there’s little sign he is on track to reach the soaring level of popularity of successful populists-turned-autocrats.
Experience shows that, if all else fails, the judiciary is the last bulwark of democracy. So far, lower courts have repeatedly blocked Trump’s excesses. Still, the Supreme Court has the last say, and Trump is openly counting on its justices to endorse his expansions of presidential power. Granted, it is dominated by conservative Federalist Society justices who have long been committed to the idea of the “unitary executive.” That idea has led them to issue some awful emergency decisions: letting Trump replace agency heads with sycophants, fire government officials by the hundred thousand and block billions of dollars for programs Congress approved.
But it remains a Federalist Society Court, not a Trumpist Court. In his first term, Trump had the worst record at the Supreme Court in the modern era. Federalist Society judges threw out his ridiculous challenges to the 2020 election again and again. The acting U.S. attorney who recently resigned rather than obey orders to drop corruption charges against New York Mayor Eric Adams belongs to the Federalist Society. Federalist Society lawyers fired from the Justice Department are hanging out shingles to accept clients suing him.
There’s also less than meets the eye to some of the court’s pro-Trump decisions. Trump v. United States, which grants presidents immunity for any acts within their core responsibilities, would let a president order Navy Seals to assassinate a political rival, Justice Sonia Sotomayor charged in her dissent. But that decision does not immunize the Seals. Assassinating a Trump rival would still be murder one for anyone who took part. Or suppose Trump tried to assemble fake electors. As Justice Amy Coney Barrett pointed out in her concurring opinion, managing elections is not a core presidential responsibility — actually, not a responsibility at all — so Trump could well be prosecuted for that.
And the administration has lost some key substantive cases in this term, as well. On March 15, Trump violated a district court order by sending 137 supposed members of a Venezuelan gang under the pretext of the Alien Enemies Act, along with 101 others under normal deportation procedures, to one of Bukele’s notorious jails. But when the administration tried to repeat that trick, the Supreme Court ruled nine to zero — without a single dissent — that the ancient right of habeas corpus requires that all of them be allowed to challenge in court both the constitutionality of the charges and their status as a gang member.
The bottom line is that, unlike Republicans in Congress who fear being primaried by the Trump political machine, federal judges hold life tenure and care how they’re seen by history. Even regarding that Trump v. United States decision, Justice Neil Gorsuch insisted, “We’re writing a rule for the ages.” And for all the court’s seeming willingness to strengthen Trump’s hand over the executive branch, the president’s chances of completely replacing all the justices, as Bukele did, or rewriting the Constitution, as Fujimori did, would seem to be close to zero.
Finally, violating a Supreme Court ruling would generate massive opposition. A Times/Siena College poll found that only 6 percent of Americans would support such a thing. As the examples of Guatemala and South Korea show, aspiring autocrats who grab for power as their popularity wanes are the ones who tend to land in jail or exile.
To be sure, U.S. democracy has some vulnerabilities most other advanced democracies do not. Inequality is a big one, contributing in no small measure to political polarization and democratic erosion.
High school graduates’ real wages rose a grand 27 cents over the past half century, and workers with less than four years of college did only slightly better. The United States is at the bad end of what political economists call the Great Gatsby Curve, where worse income inequality goes hand-in-hand with worse social mobility. Advanced firms abandoned flyover country for metropolitan areas, robbing public schools of funds, while deaths of despair — from drugs, alcohol and suicide — surged. The Biden administration’s immigration fiasco badly exacerbated existing racial tensions.
However, inequality need not lead to a democratic collapse. Latin American nations that foiled authoritarian bids — such as by Fernando Collor and Jair Bolsonaro in Brazil or by a small handful of Peruvian presidents since 2000 — struggle with equal or worse income inequalities, fierce racial and social divides, ideological antagonism from neoliberalism to reckless Marxism (sometimes alternately espoused by the same famous zealot), and weaponized social media. It’s hard to see why the United States should do worse than these nations.
Along with American democracy’s wealth, longevity and legal tradition, the people are its bulwark. Authoritarians don’t have to be political scientists to know that if they lose public support, they can fall — sometimes ousted at a sudden tipping point, as in Eastern Europe in 1989. To try to manufacture an appearance of strength, Trump exaggerates his support, broadcasts favorable lies, attacks critical media and tries to cow opponents.
This might help explain why Trump is unnerved by public protests. Drawing on a comprehensive database of civil disobedience, Erica Chenoweth of the Harvard Kennedy School has estimated that when 3.5 percent of the population has risen in nonviolent protest against autocrats, almost 90 percent of them have fallen. It’s a rule of thumb that mostly worked in the past, Chenoweth cautions, not a guarantee that a 3.5 percent protest rate will work now. Most of the protests in the database occurred before widespread social media, so the more sustained efforts needed to organize them might have been crucial.
Still, Chenoweth underlines citizens’ vital role in democracy. In fact, her analysis is that even 1 percent protests took down almost half of the autocrats they were directed against. Some 5 million Americans — 1.5 percent — joined the No Kings protest against Trump in July, according to organizers, which would make it the third largest protest in U.S. history.
Massive protests matter in part because they signal wider agreement. The drivers who passed No Kings protests across the nation honking in support weren’t among the 5 million, but they count too, as do those feeling discontent but taking notice as they were cleaning their kitchens or mowing their lawns. Living individual lives, citizens may not be aware of others’ thinking. Protest makes known resistance that would otherwise remain merely private and implicit.
Large, peaceful protests also signal to powerful elites in the military, business, courts, legislature, even the ruling coalition, that the current state of affairs is going badly. Elites are citizens, too, and often divided, even in military governments. Protests give some permission to powerful elites and allies to at least remain neutral, if not abandon support for a regime. Other coalition members, strategically calculating odds based on what they see, may keep an eye out for signs they should desert so as not to be caught on a sinking ship.
Despite his bluster, Trump is probably worried. He should be. American democracy is wounded. But, however awful its deterioration feels, it’s a long way from dead.
