A supporter of Jair Bolsonaro wears a mask of Donald Trump as he celebrates after Bolsonaro won Brazil's presidential election on Oct 28, 2018.Miguel Schincariol / AFP / Getty

When Jair Bolsonaro won Brazil’s presidential election in October to the consternation of the country’s traditional political elite, commentators were sharply divided about the implications. Some warned that Bolsonaro, a far-right populist who has openly expressed admiration for the brutal military dictatorship that ruled Brazil from 1964 to 1985, presented a clear and present threat to democracy. Others argued that Brazil’s strong institutions, including its aggressive press and fiercely independent judiciary, would rein in his authoritarian tendencies.

The fight over Bolsonaro echoes the academic debate over so-called populist figures around the world. Some scholars have warned that populists tend to be phenomenally corrupt, perpetuate their hold on power by delegitimizing the opposition, and inflict lasting damage on their countries’ democratic institutions. Others, including the historian Niall Ferguson, have suggested that populist governments are usually so incompetent that they prove short-lived. Yet others, including the political theorist Chantal Mouffe, have emphasized the positive potential of populism, and insinuated that critics of these movements are simply defenders of the failed status quo.

Right now, the four most populous democracies in the world are ruled by populists: Narendra Modi in India, Donald Trump in the United States, Joko Widodo in Indonesia, and Bolsonaro in Brazil. That makes it rather important to know which scholars are correct: Either democracy is in the midst of an unprecedented global retreat, or we’re witnessing a salutary course correction in which citizens are finally holding global elites to account for their failures. (Or, if Ferguson is right, nothing much will change.)

The most obvious way to settle this urgent matter is to look at the impact that populist governments have actually had on democracies in the past. To that end, we constructed a comprehensive database of populist governments. Doing so was an inherently fraught exercise: If you ask three scholars about the nature of populism, you are liable to get five different answers. Besides, populism is not like a light switch that is either on or off; some leaders exhibit certain (but not all) classic characteristics of populism.

Here’s how we formed our list: We selected 66 leading peer-reviewed journals in political science, sociology, and regional studies; identified all articles published in these journals on the subject of populism, as well as political leaders linked with populism; then vetted each potential case study, consulting with country and regional experts. Populist governments, in our working definition, are united by two fundamental claims: (1) Elites and “outsiders” work against the interests of the “true people,” and (2) since populists are the voice of the “true people,” nothing should stand in their way.

Ultimately, we identified 46 populist leaders or political parties that have been in power across 33 democratic countries between 1990 and today, giving us the ability to settle the theoretical debate about the tension between populism and democracy in a rigorous, empirical way, on a global scale, for the first time. The results were alarming: Populists are highly skilled at staying in power and pose an acute danger to democratic institutions.

On average, ordinary democratic governments remain in office for a brief span of time: three years. Six years from their first election, four in five non-populist governments have already been booted from power. Populist governments, by contrast, manage to sustain their hold on power for a significantly longer stretch; on average, they hold on for about six and a half years, or more than twice as long as their non-populist rivals.

Populists aren’t just more likely to win reelection once or twice; they are also much more likely to remain in power for well over a decade. Six years after they are first elected, populist leaders are twice as likely as non-populist leaders to still be in power; twelve years after they are first elected, they are more than five times as likely.

Arguably, these findings are not, in themselves, all that concerning: The longer survival rate for populists may simply reflect their efficiency or popularity. But among populist leaders who entered office between 1990 and 2015, only a small minority left office as a result of the normal democratic process.

In fact, only 17 percent of populists stepped down after they lost free and fair elections. Another 17 percent vacated high office after they reached their term limits. But 23 percent left office under more dramatic circumstances—they were impeached or forced to resign. Another 30 percent of all populist leaders in our database remain in power to this day. This is partially a function of the recent rise of populism: Thirty-six percent of those populist rulers who still remain in power were elected over the past five years. But even more of them have been in office long enough to raise serious concerns: About half have led their country for at least nine years.

The most important issue, however, is neither how long populists stay in office nor even how they ultimately leave, but what they do with their power—and, in particular, whether their tenure causes what political scientists call “democratic backsliding,” a significant deterioration in the extent to which the citizens enjoy basic rights.

Here, too, our findings were sobering, to say the least: In many countries, populists rewrote the rules of the game to permanently tilt the electoral playing field in their favor. Indeed, an astounding 50 percent of populists either rewrote or amended their country’s constitution when they gained power, frequently with the aim of eliminating presidential term limits and reducing checks and balances on executive power.

To participate in politics in a meaningful way, a country must have freedom of the press, so that citizens can make informed choices; protect civil liberties, so that citizens are free to voice their preferences and organize around their interests; and maintain political rights, so that most adults have the right to participate in free and fair elections. On all of these counts, populist governments fall short. Controlling for the many ways in which countries that elect populists may be different from countries that do not—including per capita income, recent economic performance, a country’s history with democratic institutions, and civil conflict—we found that populist rule is associated with a 7 percent decline in freedom of the press, an 8 percent decline in civil liberties, and a 13 percent decline in political rights.

Overall, 23 percent of populist governments initiate democratic backsliding, defined as at least a one-point drop in a country’s democracy score as defined by the Polity IV project. By comparison, only 6 percent of non-populist governments are responsible for this kind of deterioration. In all, a populist government is four times more likely than a non-populist one to damage democratic institutions. (And it is likely that we’re under-counting actual cases of democratic erosion because of status-quo bias in organizations that measure the robustness of democracies. Despite ample evidence of the erosion of rule of law and media freedoms in Hungary and Poland, for example, Polity IV had not yet registered democratic backsliding in these countries as of 2017.)

But are all populists equally dangerous? According to thinkers like Mouffe, scholars need to draw a sharp distinction between left-wing and right-wing populists. While right-wing populists victimize unpopular minorities and weaponize public anger for illicit goals, left-wing populists are supposedly far more likely to correct elite failures on behalf of the poor and downtrodden. The best response to right-wing populists, according to this camp, is not a preference for parties and candidates that respect long-standing democratic rules and norms—but rather the election of left-wing populists.

The data do not bear out this argument. Since 1990, 13 right-wing populist governments have been elected; of these, five brought about significant democratic backsliding. Over the same time period, 15 left-wing populist governments were elected; of these, the same number, five, brought about significant democratic backsliding. This suggests that left-wing populists are not likely to be a cure for right-wing populism; they are, on the contrary, likely to accelerate the speed with which democracy burns out.

In any case, traditional ideological measures may not do a particularly good job of capturing the nature of these movements. Also since 1990, 17 populist governments have come to power that cannot be easily classified as either left- or right-wing. Once again, five of these governments initiated democratic backsliding, suggesting that ideological hue is less important a predictor of the damage a government is likely to inflict on democratic institutions than the extent to which it is populist.

Populists often get elected on a promise to root out corruption. In Brazil, Bolsonaro soared in popularity by riding public anger against the “Carwash” scandal, a giant scheme of kickbacks from construction contracts that implicated much of the country’s political class, including the ex-president Luiz Inácio da Silva. In Italy, the populist Northern League has long railed against corrupt politicians in “thieving Rome.” In the United States, President Trump famously vowed to “drain the swamp.”

But far from draining the swamp, most populists have, as the economist Barry Eichengreen put it, simply replaced the mainstream’s alligators with even more deadly ones of their own. In fact, we found that 40 percent of populist heads of government are ultimately indicted for corruption. Since many populists amass sufficient power to hamper independent investigations into their conduct, it is likely that this figure actually underestimates the full extent of their malfeasance.

This suspicion is corroborated by a second piece of information: Our data show that populist governments have led their countries to drop by an average of five places on Transparency International’s Corruption Perceptions Index. Some cases are far more extreme than that: Venezuela, for example, dropped by an astounding 83 places under the leadership of Hugo Chávez.

Since populists often thrive on anger about all-too-real shortcomings—elites who really are too remote, political systems that really are shockingly corrupt—it is tempting to hope that they can help rejuvenate imperfect democracies around the world. Alas, the best evidence available suggests that, so far at least, they have done the opposite. On average, populist governments have deepened corruption, eroded individual rights, and inflicted serious damage on democratic institutions.

But it is also crucial to note what our results do not show. First, as advertisements for financial products so often put it, past performance is not necessarily indicative of future results. It is possible that changing circumstances, like the ideological evolution of populist movements or the growing influence of social media, make it either more or less likely that populist governments will undermine democratic institutions in the future.

Second, it is as yet unclear how easily the experience of past populist governments, which have mostly been concentrated in middle-income countries with some recent experience of authoritarian rule, will translate to rich countries with long democratic traditions. Thanks to the strength of its civil society and the widespread commitment to constitutional order, the United States, for example, may prove better able to withstand a populist president.

Finally, averages say little about individual cases. Citizens of countries that are governed by authoritarian populists should certainly be concerned that similar governments have eroded checks and balances in a large number of cases. But that is a reason to fight rather than a reason to grow fatalistic.

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