The United States is in the grip of an era of violent populism. Threats and acts of political violence have been on the rise for roughly a decade, affecting a wide variety of victims, including Republican Representative Steve Scalise, Democratic Governor Gretchen Whitmer, then Speaker of the House Nancy Pelosi, and U.S. President Donald Trump. In September 2024, I argued in Foreign Affairs that Americans must be prepared for an even more “extraordinary period of unrest” involving “serious political assassination attempts, political riots, and other instances of collective, group, and individual violence.” Sadly, this prediction has been borne out in 2025. An arsonist attempted to burn down Pennsylvania Governor Joshua Shapiro’s home (while he and his family were inside), an assassin killed Minnesota House Representative Melissa Hortman—and in September, a shooter murdered the commentator and activist Charlie Kirk in the most significant assassination in the United States since the 1960s.
Kirk’s death, in particular, has prompted bitter arguments among partisans about which political “side”—the left or the right—is to blame for the turn toward political violence. The truth is that neither is most responsible. Because it is notoriously difficult to assemble a comprehensive list of incidents of political violence and then accurately categorize them by their ideological motivation, the Chicago Project on Security and Threats (CPOST), a University of Chicago research center I run, studied threats to members of Congress prosecuted by the Department of Justice. By focusing on a discrete, well-defined group of potential targets, this study largely avoids the subjectivity that muddies much research on political violence. We determined that, since 2017, the total number of threats to lawmakers has risen markedly, and Democratic and Republican members have been equally targeted.
This finding supports other research that shows that political violence in the United States now stems from both the left and the right, a rare and unusually dangerous phenomenon. Left to its own momentum, political violence is likely to escalate further, with major consequences for American liberal democracy: it drives fear in communities and among leaders who perceive themselves to be under threat and, in turn, a willingness to accept constraints on civil liberties or wield government power to suppress the danger. That only increases the likelihood that the legitimacy of future elections will be questioned. But the broad nature of the threat also suggests that if political leaders join forces to condemn political violence, they could push back the tide.
HE SAID, SHE SAID
Violent populism—a phase of politics characterized by high levels of political violence and broad support for it—now represents a greater risk to American democracy than any competition with another country or any menace by a foreign terrorist group. The United States’ democratic foundations have, of course, been threatened by political violence in the past. During the 1920s, for instance, the Ku Klux Klan and nativists carried out terror campaigns against Black people, Catholics, and immigrants. In the 1960s and 1970s, urban riots and political assassinations were a more regular feature of American life.
But unlike other waves of violent populism over the past century, the new surge is defined by historically high levels of political violence motivated by both left- and right-wing ideology. In the 1960s, analysts broadly agreed that left-wing instigators were responsible for the preponderance of American political violence—for example, the Weather Underground’s “Days of Rage” protests in 1968. Likewise, there is a scholarly consensus that between the early 1970s until roughly 2015, people motivated by right-wing ideology carried out most acts of political violence in the United States, peaking with the 1995 Oklahoma City bombing that killed 168.
Since Kirk’s assassination, U.S. leaders and commentators have argued over which political faction is more responsible for the rise in political violence. Trump and others in his administration have insistently claimed that the “radical left” is now disproportionately to blame. Prominent writers and think tanks have asserted that the right is more at fault. On September 11, for instance, the Cato Institute released a study claiming that between January 1, 1975, and September 10, 2025, (and excluding the 9/11 attack, whose lethality was an outlier), terrorists motivated by right-wing ideologies have murdered more Americans than those motivated by left-wing views. Two weeks later, the Center for Strategic and International Studies released a study claiming that “in recent years, the United States has seen an increase in the number of left-wing terrorism attacks and plots.”
Death tolls cannot be a proxy for the threat posed by violent populists.
What aligns these two apparently opposed points of view is the conviction that one side must be most culpable—and that accurately identifying this perpetrator faction is key to reversing the rise of violence. But the reality is that the pattern of U.S. political violence has fundamentally shifted. Politicians on both the left and the right are now subject to an extraordinary degree of threat. Indeed, a close look at the data in each of the dueling studies reveals a pattern of rising attacks carried out by both right- and left-wing perpetrators starting about ten years ago.
The Cato study does show that, since 1975, attackers that the researchers categorized as right-wing have killed more people than attackers described as left-wing. But death tolls cannot be a proxy for the threat posed by violent populists. The number of people killed in a politically violent incident is often a function of circumstances. In the attack on Democratic lawmakers in Minnesota, for example, three intended victims survived. When Ronald Troyke stalked the town of Arvada, Colorado, armed with an AR-15 in 2021—intent on killing law enforcement officers—he almost certainly intended to murder more than one, but a bystander shot him dead. And because lethal attacks are a small sample, it would take only one more mass shooting to dramatically alter what the study claims to reveal.
Even more important, although the Cato study shows that although right-wing attacks were more prevalent than left-wing ones between 1975 and 2015, after 2016, they become nearly equal. The CSIS study’s data reveals that acts of political violence committed by attackers it categorizes as left- and right-wing have both risen in the past decade. Establishing any kind of reliable count of incidents of political violence is a huge challenge. There is no official FBI definition for “political violence”; creating one would require congressional legislation, because only the U.S. Congress has the power to define what constitutes a federal crime. And it can be very hard to capture all incidents with certainty and accurately judge perpetrators’ motivations, leaving analysts of violent incidents open to accusations of bias. In its study of political violence, for instance, Cato categorizes the attacker who killed a student at Antioch High School in Nashville, Tennessee, in January 2025 as right-wing and the murderer of two Israeli embassy staffers in May 2025 as left-wing, while the CSIS study omits the first and describes the second attacker’s motivation as “ethnonationalist.” These ambiguities show the inherent difficulties of assembling a comprehensive data set and classifying political motives when the depth of evidence varies case by case and when incidents may be poorly covered by the media.
DOUBLE THREAT
Fortunately, there is a better way to systematically capture a picture of political violence in the United States: by considering threats against members of Congress that the Department of Justice deemed to be serious enough to prosecute. Studying this data has multiple advantages. It creates an objective standard for what counts as a serious threat, one determined by a relevant government institution and reliably identified with media reports and public court records. And it establishes a clearer set of incidents by focusing on readily identifiable political targets. Although using the partisan identity of targets as a proxy for the partisan motivations of perpetrators is not a perfect approach, it is a fairly good one (especially when considering whether an overall pattern changes over time), since cases motivated by personal issues are known to be rare. Annual numbers of threats can also be compared across years. Most important, the risk that the researcher will embed partisan bias in categorizing threats by political ideology is eliminated.
CPOST has comprehensively assessed threats made against members of Congress between 2001 to 2024. During this 25-year period, the Department of Justice prosecuted 377 threats to U.S. legislators, counting as a single threat cases in which a perpetrator threatened the same legislator more than once or multiple legislators in the same court case so as not to inflate the numbers. The threats included perpetrators repeatedly calling a senator’s office to threaten an assassination, sending menacing powder to a legislator’s office, carrying a weapon to a legislator’s office or home—or, of course, physically harming the legislator. It is possible that different administrations’ Justice Departments did not prosecute threats against the other party’s legislators as they did threats against their own, but the fact that our study period covered different parties’ tenures in the executive branch balances this risk.
There was a clear turning point in the nature and magnitude of the hazard. Every year starting in 2017, prosecuted threats increased more than fivefold from the previous year. Between 2001 and 2016, Democrats appeared to be generally more at risk (except during the first Bush administration, when all legislators were subject to a relatively low degree of threat). Since 2016, however, the threats to Republican and Democratic members of Congress have been roughly equal. And like the Cato and CSIS data, starting in that year, CPOST’s study shows a marked rise in political violence on both the right and the left.
By all crucial measures, the pattern is the same: political violence has been rising over the past decade, and it is high on both the right and the left. Continuing to emphasize relatively small differences in the balance only contributes to a dangerous blame game that may well make matters worse.
DEATH SPIRAL
Kirk’s assassination did not simply constitute another data point in a years-long trend. It reflected a recent, even sharper acceleration in political violence. And it set off its own cascade of aggravating events: a crackdown on free speech, a probable copycat attack on an Immigration and Customs Enforcement office in Dallas, and in late September, Trump’s order to U.S. generals to “handle . . . the enemy within” and treat American cities as “training grounds.” “They spit, you hit,” he commanded. Trump’s opponents, meanwhile, have amped up their rhetoric. “You’ve got to fight fire with fire,” 2024 Democratic presidential candidate Kamala Harris said in a late September speech.
Intentionally or not, U.S. political elites on both sides of the partisan spectrum have encouraged the mobilization of the country into two separate, fighting camps. This violent polarization is visible on the streets: this year, Tesla dealerships have been the target of nearly 100 politically motivated attacks and ICE officers are facing assaults; Trump has responded with increasingly aggressive threats to treat predominantly Democratic cities like “war zones.” And tens of millions of Americans who have not committed political violence now say they support it.
Kirk’s assassination reflected a recent acceleration in political violence.
For the past four years, every quarter, CPOST has surveyed Americans to gauge their support for political violence. In our most recent poll, conducted between September 25 and September 28, over a quarter of self-identified Democrats agreed that “the use of force is justified to remove Donald Trump from the presidency,” and over a quarter of Republicans agreed that the president “is justified in using the U.S. military to stop protests against the Trump agenda.” This is triple the proportion of respondents who agreed with similar questions we posed in September 2024.
Research by scholars such as the Massachusetts Institute of Technology’s Roger Petersen, the late Princeton economist Alan Krueger, and the University of Madrid’s Ignacio Sanchez-Cuenca has clearly shown that an increase in popular support for political violence often precedes real assassinations, bombings, and bloodletting. That is why it is crucial to understand the current breadth of support for political violence in the United States—as well as the fact that violent acts are perpetrated by people motivated by both right- and left-wing ideologies. Spirals of violence can take on their own momentum, generating reciprocal cycles of emulation and revenge.
In the absence of a major effort to forestall such a spiral in the United States, political violence’s momentum will not halt. And in light of Trump’s orders to the U.S. military, it is useful to remember that British troops entered Northern Ireland in August 1969 with the intent to de-escalate local violence. Instead, their presence led to the rise of the Provisional Irish Republican Army, which split from the official IRA over a dispute about violent tactics. The Provisional IRA’s explicit goal was to expel the British troops by force, and its terrorist campaign lasted for decades.
VIRTUOUS CYCLE
Our CPOST September survey did reveal a reason for optimism. It revealed that a large majority of Americans still abhor political violence—and that equal numbers of Democrats and Republicans agree that threats of violence against politicians constitute a serious problem. Furthermore, the study found that over 80 percent of Democrats and Republicans agreed that leaders in both parties “should make a joint statement condemning any political violence in America.” This contingent includes some of the respondents who supported political violence, suggesting they could turn against it if they had confidence that partisans from the other side would, as well.
It is crucial that U.S. leaders speak to this majority. Between now and the U.S. midterm elections in November 2026, there is virtually no chance that any grand bargain will truly close the United States’ cavernous partisan divide. A summit against political violence attended by top leaders such as Trump, California Governor Gavin Newsom, and Illinois Governor J. B. Pritzker—in which they all stressed that political violence is illegal, immoral, and un-American—would be very powerful. But even a more modest joint condemnation of political violence would be meaningful.
Scholarly research consistently shows that if the public is exposed to rhetoric from their leaders that threatens violence or characterizes their political opponents using dehumanizing metaphors, its support for political violence rises. There is good reason to think that calming statements can encourage the opposite trend. Since Kirk’s assassination, various Democratic and Republican governors have separately condemned political violence. Assembling a group of leaders to do so jointly at the same publicized event would send the strong signal that U.S. leaders can live with each other—and so should all Americans.
