The US midterm elections are unfolding in a strange new atmosphere, one where the familiar drama of campaign rallies, political ads, and razor-thin polling margins is now accompanied by something that feels more like a Wall Street trading floor than a civic ritual. For the first time in a major national election cycle, prediction markets have gone fully mainstream, and they are impossible to ignore. Sites like Polymarket and Kalshi allow anyone with a credit card to buy and sell contracts on thousands of political outcomes, from which party will control the House or Senate to who will win a specific governor’s race and by how much. Hundreds of millions of dollars are now riding on these digital wagers, and the flashing probabilities these markets generate have become a regular feature of news coverage, social media feeds, and casual conversation. It is a controversial new scoreboard, and experts are increasingly worried that it is changing how Americans understand democracy itself. The concern is not just that people are betting on elections, although surveys consistently show that most Americans oppose the idea. The deeper fear is that live odds, presented with the clean precision of a stock ticker, are being mistaken for actual election outcomes, and that this confusion is creating fertile ground for misinformation, conspiracy theories, and baseless accusations of fraud.
The two companies at the center of this shift, Polymarket and Kalshi, have spent years fighting legal battles over whether they should be regulated as federal commodities markets or as gambling operations subject to state laws. They have largely prevailed in their push to operate, and they have aggressively expanded their partnerships with prominent news organizations, election forecasters, and social media influencers. Their pitch is straightforward: prediction markets aggregate the knowledge and money of thousands of participants, and the resulting probabilities are often more dynamic and responsive than traditional polls. Kalshi, in particular, has tried to position itself as a responsible data provider rather than a betting parlor. Its communications director, Elisabeth Diana, told reporters that the company sees its product as a complement to polling, not a replacement. On a prediction market, she said, traders express what they think will happen, whereas a poll captures what voters say they want to happen. That distinction sounds reasonable in theory, but in practice it has become dangerously blurred. Markets do not predict the future with certainty; they reflect the current collective judgment of people who have money at stake. A probability of 85 percent does not mean an outcome is guaranteed, and a probability of 15 percent does not mean an outcome is impossible. But in the fast-moving, emotionally charged world of election night, those nuances are easily lost. When a favorite loses, the natural human reaction is not to accept the inherent uncertainty of probability; it is to conclude that something must have gone wrong, that the fix was in, or that the votes were not counted fairly.
This is precisely what many researchers and election officials fear. The proliferation of prediction market odds into mainstream coverage has created a new way for voters to be misled, and for bad actors to manufacture chaos. Davina Hurt, the government ethics director at Santa Clara University’s Markkula Center for Applied Ethics, put it plainly: elections are not games, and the financial markets’ expectations should never carry more weight in the public’s mind than the actual votes. High probability is not a promise, she argued. When a prediction fails, that is not evidence that the election failed, but people are going to read it that way. The danger is compounded by the fact that President Donald Trump has already made unsubstantiated allegations of fraud central to his political movement before the November vote. Any discrepancy between prediction market favorites and final results is ripe to be twisted into proof of a rigged system. Even before Election Day, there are concerns that lopsided odds could depress voter turnout, as supporters of a candidate who is widely expected to lose may stay home, convinced their vote does not matter. And when results do come in, losing candidates and their allies can point to the confident online projections that favored them for weeks and demand to know how such an upset could possibly be legitimate. The markets, in other words, are not just passively reflecting the horse race; they are actively shaping the narrative around it.
A striking preview of this danger occurred in Los Angeles in June, during the mayoral primary. Spencer Pratt, a reality television personality known for his appearances on “The Hills,” ran as a Republican candidate and was by no means a serious contender by traditional political standards. Yet on some prediction markets, his odds of advancing to the general election were surprisingly high, enough that right-wing influencers seized on them as evidence of a vast conspiracy when he ultimately lost. The odds, they claimed, had proven that the election was stolen from Pratt, and they spread this message across social media with confidence. Some of the posts repeating the misinformation carried “paid partnership” tags, revealing that they were part of sponsored content arrangements with the very prediction market platforms that had issued the favorable projections. The Los Angeles County registrar’s office was forced to respond to a wave of public confusion. Michael Sanchez, a spokesman for the office, told reporters that the June primary had reinforced how quickly expectations about election outcomes can become confused with actual election results. Even without any evidence of fraud, the mere existence of financial markets that favored a losing candidate was enough to convince many people that something sinister had occurred. The episode became a cautionary tale for election administrators across the country, who suddenly realized that they were now competing not just with pollsters and pundits, but with billion-dollar trading platforms that have every incentive to keep engagement high.
In the wake of that controversy, Kalshi moved to impose new restrictions on its paid partners, though critics note that these rules are only as effective as their enforcement. In a memo shared exclusively with reporters, the company told its partners that they “shall not call into question the integrity or accuracy of any official vote tally result, legal ruling, or other official or formal determination in connection with the election.” Partnerships would be terminated, the memo warned, if anyone posted without evidence that an election result had been corrupted, rigged, compromised, or otherwise improperly administered. It was a direct acknowledgment that the platforms themselves had become vectors for misinformation, even if unintentionally. Polymarket, for its part, did not respond to requests for comment on the issue. For researchers who study online misinformation, this is an alarming development. At the University of Washington’s Center for an Informed Public, scientists who have spent years tracking election falsehoods are now adding prediction market distortions to their monitoring efforts. Senior research scientist Rachel Moran said the financial element attached to these rumors makes them more powerful and more likely to spread. People are not just making political arguments; they are making money, or losing money, and that creates an emotional investment that goes far beyond ordinary partisanship. The result is that an already fragmented and toxic information ecosystem now has another powerful engine of confusion, one that dresses up speculation as objective market wisdom.
For election officials on the front lines, the emergence of prediction markets as a major force in political discourse is an unwelcome burden. Jared DeMarinis, the elections chief of Maryland, has asked state authorities to investigate whether election betting through these platforms is even legal under local law. He has banned poll workers in his jurisdiction from placing any election-related bets, an unprecedented step that reflects how seriously he takes the threat. But he is deeply worried that the damage may already be done. Lopsided odds, he argues, could discourage voters from turning out, and the inevitable spread of false fraud claims will wear down his team during what is already an intensely stressful period. Election administrators across the country are already under attack, facing pressure from the Trump administration and from citizens who have been convinced that the system cannot be trusted. Adding money to that volatile mix, DeMarinis said, is like pouring gasoline on a fire. The legitimacy of the outcome is now in question even before the votes are cast, and that is a profoundly dangerous place for a democracy to be. The markets are not going away, and their influence over how Americans understand elections is only likely to grow. But the responsibility for ensuring that they do not become a tool for undermining faith in democracy rests not just with the platforms, or with election officials, or with the media. It rests with every citizen who sees a flashing probability online and remembers that an election is not a game of chance, but a collective act of self-governance. It is a reminder that the numbers on a screen are never as important as the quiet act of casting a ballot, and that the winner of an election is determined by actual people, not by the trading floor.