Prediction market bettors are the new political pundits. Why are we listening to them?

Sites like Kalshi and Polymarket offer faux populism as elites reap the benefits – and voters risk being misled

Online betting is a bit like Donald Trump.

The prediction markets – platforms where users bet on whether events will occur – pretend to be populist, pushing the nabobs aside and representing the “wisdom of the crowd”. “This data is bottom-up, not elites-driven,” a Kalshi ​spokesperson told Reuters. But in fact, the markets are controlled by a few, skillful traders.

Their users are led to believe they can strike it rich. But like gamblers everywhere, most of them lose most of the time, while enriching the casino and bankrolling the winner’s takings.

The markets pretend honesty and transparency, but they open the door to fraud and insider trading.

Like opinion polls, they affect voters’ expectations of the outcomes of elections but make them no smarter about whom to vote for.

Fundamentally, the assumptions projected by the prediction markets are Trumpian. The companies suggest that their results are improved by the financial stake customers have in them.

“Prediction markets are a way to see what people are saying when there’s money on the line,” said the Kalshi spokesperson. When money talks, in other words, its message is more authentic and valuable than mere political speech.

The prediction markets made a better collective call on the 2024 elections than conventional polls and pundits, particularly in the swing states. But why?

The authors of a much-cited 2025 study from Vanderbilt University infer that their results – that Polymarket was more accurate than polls – were “in alignment with research on the ‘Wisdom of Crowds’ theory, which suggests a large group of people are often accurate in predicting outcomes, even if they are not necessarily experts or closely aligned with the issue at hand”.

But more recent analyses suggest that the betting masses are not so wise after all. A new working paper from scholars at the London Business School and Yale University looked at publicly available trading data from Polymarket. Whereas almost every wagerer’s outcomes were no better than chance, they discovered, the “accuracy comes from a minority of persistently skilled traders, around 3% of accounts”. The crowd “generates most of the volume but little of the information”.

The crowd also furnishes most of the cash but reaps little of it. Analyzing 1.6m Polymarket accounts, or “wallets”, traded since November 2022, the Wall Street Journal found: “Casual traders are bleeding cash while a small number of sophisticated pros – including trading firms with access to vast streams of data – eat their lunch.”

What skills do these pros employ? They react quickly to public and financial news, say the London and Yale authors. They gather poll data and buy access to big data streams, which can cost $200,000 annually. They analyze the data using “algorithms to predict price movements and manage risk”, says the Journal, “faster than any human”.

Operating at scale and speed, with commensurate information and discipline, the pros can make tens of thousands of trades a day. And the more they trade, the more they earn.

In the Journal’s analysis, three-quarters of the profits went to 0.1% of the accounts. Of at least 2.3 million accounts on Polymarket’s site, just 2,000 netted a total of almost half a billion dollars.

Meanwhile, the median Polymarket user placed 46 trades across 10 active trading days in 2026, according to Pew Research. And because the few have the skills to trade against the mistakes of the many, the losses of the many bankroll the profits of the few.

The traders steering the prediction markets are not PhD pollsters and PBS talking heads. But they are elites nonetheless – the same elites who dominate Wall Street and make economic policy, whichever party is in office.

If it were a simply a matter of some people’s money squandered or increased, the problems of the traders would remain with the traders. But the press and the public are listening to them more and more. “Prediction market odds shift voters’ expectations substantially,” a study funded by the Brennan Center found.

The same features that rig the markets in favor of the few can distort what we think we know about coming elections. Explained CoinMarketCal, a newsletter serving crypto traders: “A market price reflects trades, liquidity, and available capital, not a representative poll. A well-funded trader can move a thin contract without any change in voter sentiment.”

And it’s not just expectations that prediction markets influence. They can also change voter behavior. “Elections have the property that what one believes is going to happen can actually affect what happens if a candidate is thought to be losing momentum or the campaign is flagging,” the Barnard College economics professor Rajiv Sethi told NPR.

Gaming on votes is not new, and its history is thick with corruption. Tammany Hall, New York’s Progressive Era Democratic party machine, kept a war chest for electoral wagers and publicized them vigorously to boost political momentum. In the 1904 presidential elections, the New York Times accused the Republicans of finagling the betting odds to help Teddy Roosevelt.

The corruption has simply moved online. There’s two-bit graft like the trick the expelled Republican ​representative George Santos pulled in 2023: he bought Kalshi contracts on whether he would attend the State of the Union address, posted on social media that he would, then stayed home and earned $18,000 on the bet. “There was ​absolutely no intent to deceive any person,” the compulsive liar’s lawyer protested.

While prediction market participation by elected officials and candidates constitutes illegal insider trading, the law did not stop aides this spring from betting on their own candidates. Polymarket has investigated dozens of cases of potential fraud related to the midterms. But with millions of trades moving both in the country and offshore, it may be impossible for companies to keep up with the fraudsters.

Especially not under an administration stinking of corruption and a presidential family deeply engaged in the crypto and prediction markets.

Donald Trump Jr’s venture firm 1789 Capital has invested half a billion dollars in Polymarket. On the day after Daddy’s 2025 inauguration, Kalshi announced it was “beyond excited” that Junior had joined the exchange as a paid adviser.

Trump Sr, who made $1.4bn on crypto in 2025 and has vowed to make the US the “crypto capital of the world”, declared in May that prediction markets also “will thrive” under his presidency.

There is overwhelming popular, bipartisan opposition to betting on politics and elections. Half the states ban it to some degree. But federal regulation is unlikely any time soon. Before Michael Selig was appointed to chair of the Commodity Futures Trading Commission, which regulates online prediction and crypto markets, he was a corporate lawyer for prediction markets and crypto.

Without government oversight or publicly traded stocks, it is even hard to evaluate election betting sites as businesses.

Estimates of trading volume on the midterms range widely. The Anti-Corruption Data Collective, the New York Times’s source, says it has surpassed $220m in 2026, up from $92.4m in the 2024 election cycle. Meanwhile, CNN quotes spokespeople for the two biggest platforms, Kalshi and Polymarket, who claim a total of $1.4bn for the upcoming midterms.

But the house doesn’t profit directly from the bets; users determine the prices and the odds. Revenue comes from trading fees and money earned on the funds sitting in the companies’ coffers between buys and payouts. Increasingly, media, financial and tech companies, including CNN, Fox, Dow Jones and Palantir, are striking deals to use the data.

Kalshi was recently valued at $22bn and Polymarket at $9bn. This could be a bubble that will burst, as artificial intelligence’s appears to be doing.

Some condemn the prediction markets as inherently morally and psychologically harmful. On this, I’m agnostic. A casino on your phone makes betting easier, but it does not turn every user into a fiend. And while the psychological profile of these online bettors – young, male, impulsive, irrationally optimistic, overly self-confident – makes them unnerving as pundits, if the question is whether Zimbabwe’s cricketers will beat Afghanistan’s or the US will confirm the existence of aliens, my feeling is, let the people bet.

But democracy is not cricket. These markets commodify a politics already gravely corrupted by money and gamify a discourse ever thinner on content or nuance. They should be banned.

Judith Levine is a frequent contributor to the Guardian. Her Substack is Today in Fascism