How Not to Defeat Authoritarianism

In 2024, democracy was “on the ballot.” That, at any rate, was the popular slogan. To which there was a popular response: Democrats must moderate! “Moderation” was the key to winning elections, it was said, and winning elections was the only way to turn back the existential threat to democracy.

If you watched on Election Night expecting Kamala Harris’s appeals to moderates would pay off, only to see them fall short, you weren’t alone. Political consultants bet heavily on the same calculation. Harris distanced herself from progressive positions (including some she herself had previously supported), emphasized kitchen-table issues, and courted Liz Cheney. She lost anyway.

While democracy faces an existential threat, pundits and strategists remain stuck in the past.

But the idea that moderation is the path to victory—and thus to preserving democracy—remains a matter of firm conviction among many political commentators and consultants. The New York Times editorial board recently declared that “moving to the center is the way to win,” dismissing research that finds otherwise as lost in “statistical complexities.” As prominent pundit Matthew Yglesias wrote in early 2025, “What the Democrats need … is not just more moderate candidates. They need a more moderate ideology.” Or, in political scientist Ruy Teixeira’s words from 2022, “Moderation = Democratic votes.” On this view, the primary reason that Democrats are losing is that voters perceive the national party as too extreme: embracing policy positions or activist groups on issues that too many people see as too far from the mainstream, from policing and climate change to immigration and transgender rights. And Harris’s efforts to repair those perceptions were too little, too late.

There are many dissenting voices. Zohran Mamdani, for one, warned against bowing “at the altar of caution” in his victory speech in November. But the Times, Yglesias, and Teixeira are hardly alone in preaching at this altar. And as proponents of moderation will be quick to say, Mamdani won in New York City, not the swing districts in California’s Central Valley and Pennsylvania’s suburbs that a national party needs to win real governing power, from the presidency to the Senate. Moderation, in short, remains the conventional political wisdom.

We think this view is wrong. In an era of intense partisanship, nationalized elections, and low trust in elites and institutions, the electoral benefits of moderation are at best small and inconsistent, at worst counterproductive—in any case, not a reliable path to electoral victory or to meeting the existential threats to democracy. Moderation is more conventional than wise, and the most prominent cases for it badly misunderstand how politics works.

Why do so many think moderation is the right way to win? The most familiar and intuitive argument is that it’s simply what voters say they want. This view is strongly associated with popularism, a way of thinking about politics that has flourished in Democratic circles since 2020.

Popularists argue that candidates who want to win elections need to focus relentlessly on what polls well and abandon what doesn’t. The claim is not that progressive policies are bad on the merits, though popularists do sometimes criticize them on this basis. The claim is that too many voters are turned off by progressive ideas and rhetoric, dooming Democrats’ chances to win office and pursue any agenda to the left of the GOP. Popularism’s strategic imperative is thus framed as a hardheaded corrective to progressive idealism.

Ezra Klein has summarized the view this way: “Democrats should do a lot of polling to figure out which of their views are popular and which are not popular, and then they should talk about the popular stuff and shut up about the unpopular stuff.” The job of a campaign, in other words, is to discover what voters already want and package it for sale. To win elections, Democrats need to treat voters as consumers with fixed preferences, not as citizens whose behavior is malleable and whose views are responsive to media, leadership, and persuasion.

This approach to politics is not new. The technical term is political pandering, and it is as old as democracy itself. Thucydides derided the Athenian demagogue Cleon as a leader guided not by wisdom but by whatever drew the assembly’s loudest applause. Political theorists have long treated Cleon as a warning: when leaders stop leading and instead mirror the crowd, democratic institutions decay. Popularism cloaks this same impulse in empiricism. The roar of the crowd has been replaced by polling data, but the emptiness remains. Worse, Cleon’s pandering at least stirred the crowd. The moderation that popularists recommend dampens enthusiasm, demobilizes or even alienates the base, and cedes the affective terrain to opponents who are happy to rile people up.

To be fair, popularism does correctly diagnose a bind that Democrats face. They are geographically concentrated in ways that create structural disadvantages in the Senate and Electoral College; in order to win a governing majority, they need to do well among relatively conservative districts and can’t afford to alienate most voters outside deep-blue urban areas. But popularism’s prescription is fatally limited. It confuses listening with leadership and measurement with meaning, assuming that elections are won by calibrating positions to match local preferences.

Political scientists Christopher Achen and Larry Bartels systematically rebutted this assumption—and the more general folk theory that people mainly vote on the basis of freestanding policy preferences—in their 2016 book Democracy for Realists. Moreover, as Daniel Hopkins shows in The Increasingly United States (2018), contemporary U.S. politics are more nationalized and polarized than ever—trends that render preference-matching efforts especially ineffective. Tip O’Neill, Speaker of the House from 1977 to 1987, urged that “all politics is local.” That’s not true anymore. Elections have become national referendums, less impacted by candidates’ stated policy positions. Even the most finely tuned messaging is being drowned out by partisan tides. In short, it is much harder for candidates to outrun their party than it once was.

More fundamentally, the popularist fixation on polls undermines the movement energy required to confront authoritarian threats. Popularists see voters as fixed points—or at least extremely rigid ones—instead of as agents whose preferences can be mobilized and transformed. But in moments of democratic crisis, politics is a search for the right movement, not the right maneuver. Playing defense forfeits the moral clarity and collective purpose that have sustained successful anti-authoritarian movements worldwide.

So much for the wisdom of chasing polls. But can a case for moderation be made independently of the argument for pandering? Here, too, the popularist refrain contains a grain of truth. For decades, the electoral benefits of moderation were real and well-documented. Political scientists developed rigorous tools to measure candidate ideology and found that centrist candidates often outperformed their more ideological counterparts. Claims that moderation helped win votes had empirical support.

But that was a different era. As elections nationalized and polarization intensified, the old rules stopped applying. A growing body of research now demonstrates that the moderation advantage has largely disappeared. In the Trump era, candidate ideology has little consistent effect on vote share. The scholarly consensus has shifted. The popularist case for moderation has not kept up. Instead, it relies on analyses that ignore this research.

The median voter theorem made sense in an era of stable coalitions, low polarization, and localized elections. That era is over.

Perhaps the most prominent metric comes from Split Ticket, an organization of political data analysts that claims moderates dramatically outperform progressives. Their metric, called Wins Above Replacement (WAR), is borrowed from baseball, estimating how much a player contributes to a team’s success compared to a replacement-level alternative. In that context, WAR can provide some insight because each player generates data from over a hundred games per season. A congressional candidate, by contrast, faces voters once every two years. As political scientist and statistician Andrew Gelman has noted, this means political WAR estimates are far more uncertain and should not be taken too seriously.

Moreover, Split Ticket’s analysis is proprietary; the group does not disclose how it calculates estimates, so the assumptions built into its model cannot be independently evaluated. When we tested the group’s metric against a transparent baseline—how well a candidate performed relative to their party’s presidential nominee—we found that less than a third of Split Ticket’s WAR score reflected actual candidate overperformance; the rest came from undisclosed adjustments. We tried to replicate Split Ticket’s results, testing dozens of reasonable modeling approaches and different measures of ideology. None produced effects anywhere near as large. Other analyses, including data journalist G. Elliott Morris’s transparent and replicable WAR model, reached similar conclusions. Morris’s verdict echoes our own: moderation is “overrated” and “not a silver bullet.”

The Times editorial board recently made a case similar to Split Ticket’s on the basis of PAC endorsements, comparing candidates backed by centrist PACs against everyone else—including unfunded candidates in hopeless races. This comparison just doesn’t make sense. If you apply the same method to candidates backed by progressive PACs, you find that they also appear to outperform everyone else until you adjust for fundraising and incumbency, in which case the apparent advantage again vanishes. The Times dismissed academic research as too complex and disconnected from voter perceptions, yet when we measure ideology based on how voters perceive candidates, the moderation effect is the smallest of any measure tested.

At a fundamental level, both Split Ticket’s analysis and the Times’s confuse correlation with causation. Joe Manchin is the poster child for moderation’s supposed benefits, but his success reflects decades of personal brand-building in a state that realigned around him. His moderation was credible to West Virginia voters in a way a replacement Democrat’s would not be. Indeed, when Manchin retired and Democrats nominated Glenn Elliott, also a moderate, Elliott lost by 40 points. The lesson isn’t that moderation wins; it’s that being Manchin wins.

To illuminate the analytical challenges underlying these debates, consider a simple, transparent metric: how much better or worse a congressional candidate performs compared to their party’s presidential nominee in the same district. Being an incumbent generally provides a substantial electoral boost of 2 to 3 percentage points above the presidential nominee’s performance; in contrast, shifting dramatically to the ideological center yields at most negligible benefits. (We get the same result no matter how we measure candidate ideology.)

Political scientists have developed approaches that get beyond the correlations that plague popularist models. We have used these approaches to test whether moderation wins elections. We built a comprehensive measure of ideology by combining over a dozen distinct metrics—from congressional voting records and campaign finance data to the policy positions on candidate websites to voter evaluations. And we employed research designs that can distinguish causal effects from correlations. For example, we looked at very closely contested Democratic primaries that included both a moderate and a progressive candidate. Because the primaries are so closely contested, the winner is essentially random, as if someone flipped a coin to determine which candidate would get the nomination. (This research design ensures that we’re studying the effect of moderation independent of all other factors, such as whether the district was red, blue, or purple.) We then considered whether the moderates who won these coin-flip primaries did better in the general election than the progressives who won them. They didn’t.

Putting all this together, the results are clear. Moderation used to help candidates in past decades, but in the Trump era, a candidate’s ideological moderation has no consistent, measurable effect on their vote share. Moderation as a strategy hit its ceiling in 2024. With moderates already dominating the battlefield in every competitive district, the potential gains from moving further to the center were exhausted. Still, Democrats made their wager: from the presidential nominee down to local candidates, they fully deployed the popularist strategy of moderate, “kitchen table” messaging. The result? A Republican trifecta. The party bet on moderation, but the strategy, on its own, proved unable to meet the moment. The record of who actually wins and loses reinforces the point. Of the twenty-two Democratic incumbents who lost competitive seats between 2016 and 2024, twenty-one were moderates—only one was a progressive.

We are not alone in these findings. A number of recent scholarly studies converge on a theoretically interesting and practically important conclusion: yes, moderation was once a powerful force in U.S. politics, but its effects have shrunk to the point of (at most) being small and highly context-dependent. A small moderation bonus may exist in some contexts, but it has nowhere near the campaign-defining effect its boosters claim. In the Trump era, moderation has no consistent, measurable effect on vote share. Whatever small advantage once existed has been exhausted, and with moderates already running in nearly every competitive district, there are no more gains to be had. The strategy has been fully deployed. It has not delivered.

This raises a pair of key questions. What has changed? And what has taken the place of the old rules?

To begin to answer, it helps to step back from election statistics and think about the substance of modern American politics.

One of political science’s most elegant theories offers a starting point: the “median voter theorem.” It was given expression in Anthony Downs’s 1957 book, Economic Theory of Democracy. Imagine voters arranged along a line from left to right. The median voter sits in the middle, with half the electorate to the left, half to the right. If candidate Jones moves away from the middle to the left, they will lose the voters who are now closer to candidate Smith, and not gain anything because the voters further to the left were already going to vote for Jones. Like ice cream vendors setting up in the center of a beach, candidates should rationally move toward that midpoint to win the most votes.

The model is very simple and, for decades, it worked. In the 1980s and 1990s, moderates often outperformed their parties. But that advantage has eroded for structural reasons.

First, who the median voter is depends on who shows up to vote. When turnout among younger and more diverse voters surges, the median shifts left; when it’s low, older and more conservative voters pull it right. The strategic question isn’t just how to appeal to the median voter, but how to build an electorate that generates a median voter favorable to your goals. So the popularist focus on persuading a small number of swing voters is misplaced: that focus may dampen turnout and shift the median voter to the right.

Playing defense forfeits the moral clarity and collective purpose that have powered successful anti-authoritarian movements worldwide.

In contemporary elections, what really matters is which party gets more of its voters to the polls. The turnout gap between Democrats and Republicans is a powerful predictor of election outcomes. As the data show, Democratic national vote share rises and falls with their ability to close this gap. When Democrats matched GOP turnout rates (as in 2008 and 2018), they performed exceptionally well. When the gap widened (as in 2010, 2014, and 2024), they struggled. Moreover, our own research finds that progressive candidates are slightly more effective at mobilizing their base and increasing turnout among registered Democrats. This finding comes from tracking the voting behavior of over 120 million individuals across five election cycles, a design that isolates the effect of candidate ideology by comparing the same voters across different elections.

Second, local politics has been swamped by national tides. In our hyperpolarized system, voters increasingly choose Team Blue or Team Red based on national sentiment, not the ideology of their local candidate. Every Democrat—from the progressive Alexandria Ocasio-Cortez to the moderate Jared Golden—now rises or falls with the party brand. When inflation hit 8 percent, voters didn’t parse ideological differences; they punished the party in power. When Donald Trump was unpopular in 2018, Democrats of all stripes saw sizable gains.

This reality is well-documented in political science. One of us—Grumbach—has written a whole book showing how voters have become more attentive to national news and how state parties have increasingly become agents of their national counterparts. The result is that congressional elections now function more like national referendums, where a candidate’s individual positioning is often overwhelmed by these larger forces. To be sure, candidate quality still matters some, but it matters less, and moderation itself does not seem to be an especially important quality.

Third, being moderate doesn’t necessarily mean having a more popular platform. As political scientist David Broockman has shown, voters who appear moderate often hold a mix of extreme positions rather than consistently centrist ones: confiscate billionaire wealth, for example, but also ban abortion. There is no coherent “center” to triangulate toward, and a candidate who tries may end up pleasing no one.

Fourth, the median voter theorem may underestimate the role of political leadership in the current environment. The Trump era is full of examples of voters seeking a leader who “tells it like it is.” These voters want authenticity and disruption, not careful ideological positioning. An appeal to moderation may completely miss the mark with voters who want a candidate who projects strength and a willingness to challenge the system.

This desire for authenticity is closely related to the idea that voters “follow the leader”: politicians influence voters’ policy attitudes, not the other way around. Michael Barber and Jeremy C. Pope have a creative analysis of this phenomenon focused on Donald Trump’s contradictory policy statements. (Evidence for this theory appears in Gabriel Lenz’s 2012 book, Follow the Leader, as well as in much of the work of Achen and Bartels.) Trump has garnered the support of the most anti-vaccine voters in the United States, despite having overseen the creation of Operation Warp Speed in 2020 to rapidly develop COVID-19 vaccines. And he has taken flatly contradictory positions on abortion and the minimum wage, sometimes in the same interview. When researchers show voters one of these statements, many adopt that position. The leader shapes the followers, not the other way around.

But this dynamic has an important precondition: voters must already identify with the leader. Trump can shift positions freely because his supporters are following him, not his platform. The policy positions are downstream of the relationship. For candidates still trying to build that relationship, the calculus is different. Strategic repositioning can signal inauthenticity, which is precisely what prevents the bond from forming in the first place.

The recent history of high-profile Democratic moderates confirms this. Consider Jared Golden, who since 2019 has represented Maine’s 2nd congressional district, one of the few districts in the country that has split its ticket for Trump and a Democratic House member. Golden is the archetypal moderate: a Marine veteran who broke with his party on key votes, opposed Nancy Pelosi for House Speaker, and carefully cultivated an independent brand. If the moderation thesis were correct, Golden should have been thriving. Instead, by October 2025 his favorability had collapsed to 16 percent, and 57 percent of his constituents said he did not deserve reelection. He chose not to run again this year. And he is not alone. Kyrsten Sinema, the most prominent Senate moderate of the Trump era, offers perhaps the starkest warning. By late 2022, her strategy of triangulation had achieved a rare feat: she was unpopular with Democrats, Republicans, and independents alike. Having alienated her base without securing the opposition, she, too, declined to seek reelection.

Finally, strategic moderation often undermines the very credibility it seeks to build. When candidates without the leadership clout of people like Trump or Manchin shift positions to chase the political center, voters see them as unprincipled, not pragmatic. The strategic calculation is often transparent, validating an opponent’s charge that the candidate lacks core beliefs. When center-left parties in Europe adopt anti-immigration positions to counter the far-right, it brings them no electoral benefit; if anything, researchers have found, strategies that “accommodate” right-wing positions “lead to more voters defecting to the radical right.” Democrats who chase moderate Republicans by adopting their frames risk the same outcome.

The median voter theorem made sense in an era of stable coalitions, low polarization, and localized elections. That era is over. Democrats need strategies built for this new reality.

If moderation isn’t sure to deliver victory, and if popularism leads to paralysis, what should Democrats do? The honest answer is that nobody knows for certain. Our issue with popularism is its air of certainty. There is reasonable room for debate about the virtues of particular kinds of moderation in particular contexts. But political elites should not interpret the results of biased statistical analyses as a form of scientific justification for any political strategy, something we see frequently in recent discourse.

The reality is that electoral politics has entered an era of profound volatility, one when yesterday’s certainties become today’s mistakes. We are not making a general case for running to the left instead of to the center but for dispensing with outdated conventional wisdom. Instead, we favor experimentation and exploration. Embracing these requires expanding our sense of possibility and the range of our explorations, partly by paying close attention to what has worked in other countries that have faced democratic backsliding. Taking inspiration from their experiences, we draw three important lessons.

Reframe the Battlefield

The most effective opposition in a polarized society doesn’t try to win a few converts on an entrenched battlefield; it redraws the map entirely. Successful anti-authoritarian movements do not win by softening their positions. They win by building unlikely coalitions around a broadly resonant grievance.

One promising possibility is to focus on corruption. Anti-corruption has been a powerful axis of political mobilization throughout history and across democracies. It is not a poll-tested “issue” like health care or immigration; it is a frame that reorganizes politics around a fundamental question of legitimacy. And the conditions for such a frame are present. A May 2025 poll from Yale and George Mason found more Americans said they are “very worried” about government corruption than about the cost of living or the economy. This frustration is rooted in a deep-seated belief that the system is rigged; a recent YouGov survey, for instance, found that a staggering 73 percent of Americans believe a member of Congress would be likely to accept a bribe if offered one. The revelations about Clarence Thomas accepting undisclosed gifts from billionaire donors, or members of Congress trading stocks on nonpublic information, confirm what many already suspect: the rules don’t apply to the powerful.

The energy generated by disclosure of the Epstein files points in the same direction: large numbers of people saw a club of powerful insiders acting with impunity. Focusing on corruption shifts the debate from divisive ideological questions to a simple moral choice: Are you on the side of the people, or of a rigged system?

Build Anti-Establishment Credibility

An anti-corruption platform is useless without anti-establishment credibility, which many Democratic leaders now lack. They need to position themselves as the party that will take on the real elites—billionaires buying Supreme Court justices, corporations price-gouging families, insiders trading stocks on classified briefings.

This will require genuine, costly reform—and self-criticism—that goes beyond rhetoric: ending deceptive fundraising practices, rejecting corporate PAC money, and championing popular measures like banning congressional stock trading. Recruiting authentic, working-class candidates—nurses, teachers, veterans—is also crucial, not as a matter of ideological positioning, but as a demonstration that the party is not beholden to the same professional class. Let a thousand flowers bloom; these candidates should put forward platforms and campaigns based on what they authentically think would help the country, not what might attract donors, please crypto lobbyists, or satisfy a popularist’s survey samples. Action, not messaging, is the only path to earning the credibility needed to lead this fight.

Boost Turnout

Most critically, Democrats must fix their turnout crisis.

While much of the punditry has focused on swings toward Trump among young men and Latinos in 2024, the turnout gap between registered Democrats and Republicans was much more devastating. This turnout gap appeared across all voter categories—among reliable voters from 2020, sporadic voters who voted in 2020 but sat out in 2024, and newly registered voters. The gap between Black and white Americans reached 10.9 percentage points—the widest in over three decades, as Black participation dropped more sharply than white. Youth turnout was also down from 2020. The Republican turnout advantage in 2024 exceeded both 2020 and 2016. It’s true that nonvoters likely leaned slightly toward Trump over Harris in November 2024, but large segments of these nonvoters were progressive young people and most others might have been persuaded by a more credible Democratic Party.

The choice now is to transform the party’s strategy to meet the scale of the threat, or to risk the end of democratic self-governance itself.

The popularist strategy pours resources into chasing a dwindling number of swing voters while neglecting the millions of disengaged citizens who power its victories. A policy-based appeal that proposes incremental changes to a status quo that many see as fundamentally unfair and corrupt is not enough to overcome the deep-seated cynicism that keeps many voters home. More promising is to mobilize people who believe the system is rigged with a credible promise to un-rig it. A genuine anti-corruption fight provides the moral clarity and purpose needed to energize the base, young voters, and even disillusioned conservatives disgusted by a system they believe has been captured by special interests. It answers the fundamental strategic question not of how to appeal to the median voter, but of how to create a new, mobilized electorate by giving them a cause larger than any single candidate.

The false choice between “moderate” and “progressive” has trapped Democrats in an irrelevant debate while democracy hangs in the balance. The real divide isn’t left versus center—it’s whether Democrats view Trump as an aberration to be waited out or as an authoritarian threat that requires an extraordinary response. Journalist Ronald Brownstein recently identified this fault line: those who view authoritarianism as a “distraction” from kitchen-table issues versus those who see it as the existential crisis defining all other questions. The “distraction” camp still operates by 1990s rules, counseling moderation and patience, trusting failed institutions.

This brings us back to a crucial point: successful anti-authoritarian movements don’t win by moderating their positions on a traditional left-right axis but by creating an entirely new one. They mobilize previously disengaged citizens by framing the struggle not as a contest over policy, but as a fight for the fundamental fairness of the system itself.

Democrats can keep debating whether to stand five degrees or fifteen degrees left of center, fine-tuning messages drowned out by the latest Trump spectacle. They can chase moderate voters our data shows won’t materialize while their base stays home. They can maintain faith that reasonableness will eventually triumph.

Or they can accept the evidence: the old rules are dead. In a nationalized media environment with an authoritarian movement capturing one major party, electoral politics has become existential conflict. This requires not moderation but mobilization, not positioning but purpose, not just messaging but genuine reform to prove that Democrats will fight for democracy itself by first attacking the corruption that rots it from within.

Scholars of democratic breakdown know that moments like this demand institutional coordination, civil society mobilization, and the political courage to name and confront the authoritarian threat on its weakest flank. Every democracy facing this challenge has learned you don’t defeat authoritarians by being more reasonable. You defeat them by being more determined and by uniting the country against their most visible vulnerability: their corruption.

History won’t judge Democrats on whether they were moderate enough. It will judge them on whether they fought hard enough, and smart enough, when democracy was threatened. The politics of careful positioning and poll-tested moderation have been tested, and on their own, they have failed to deliver the victories needed to protect democracy. The choice now is to transform the party’s strategy to meet the scale of the threat, or to fail. In this contest, failure is not just an electoral defeat; it may be the end of democratic self-governance.

Independent and nonprofit, Boston Review relies on reader funding. To support work like this, please donate here.