The rise – and risks – of prediction markets

What are prediction markets?

Prediction markets are online peer-to-peer exchanges that let users wager money not just on sports and politics, but on everything from the chances of a full US ground invasion in Iran to what colour tie Donald Trump will wear on a particular day. The two biggest sites, Polymarket and Kalshi, have exploded in popularity over the past two years, as US federal regulators have taken a far more relaxed approach.

Trading volumes on the two sites were $50 billion in 2025, up from $16 billion the year before, and are set to be many times that this year. According to analytics firm The Block, cited by The Wall Street Journal, trading volumes on the two leading platforms rocketed to $24.2 billion last month, up from $1.8 billion a year earlier. Donald Trump Jr is an adviser and investor in Polymarket, and Trump Sr’s media company is planning its own site, Truth Predicts. Meanwhile, several US states are suing the firms on the grounds that they facilitate illegal gambling.

Are prediction markets risky?

Betting on events that might or might not happen is obviously wide open to people taking advantage of inside information and/or seeking to influence events. Last month, a US soldier was the first person to be charged with insider trading on prediction markets. Gannon Ken Van Dyke is charged with using classified information to place roughly 13 bets worth $33,034 on positions including “US Forces in Venezuela” and “Maduro out”.

The sites claim they serve an important wider function in terms of price and information discovery – indeed, the Federal Reserve, no less, recently published a paper finding that Kalshi’s market participants did a usefully good job of predicting changes in interest rates and GDP. Polymarket says it “aggregates wisdom from informed users, often outperforming experts”. But if those users are too “informed”, it stacks the deck against the rest of the players. Exactly how “informed” a user is permitted to be before they are doing something illegal is a question that’s certain to play out in the courts many times in forthcoming months and years.

Are prediction markets allowed in the UK?

In theory, both Polymarket and Kalshi are regarded as unlicensed gambling sites by UK authorities and are blocked to UK-based users. In practice, there are routes (involving VPNs and cryptowallets) available to those who wish to bet. This week, the site was offering 20 different markets related to the local and devolved government elections in the UK, including mayoral elections. Unless an awful lot of Americans have developed an unlikely expertise in Britain’s local politics, it seems probable that the “informed users” in question are UK-based.

How do prediction markets work in practice?

The principle is similar to sports-betting exchanges, in that prediction markets involve peer-to-peer betting (or “trading” for those who fancy themselves as pros) rather than betting against a bookmaker. On Polymarket and its competitors, unlike a conventional bookmaker, the house doesn’t set the odds: it facilitates a peer-to-peer exchange – all bets are binary Yes/ No – and takes a fee.

At the time of writing, Polymarket is offering a market on what the highest temperature will be in London on 7 May. A high temperature of 16˚C is priced at 39 cents, meaning the market of interested punters thinks that there’s a 39% chance of that event happening. If you bet Yes – and get it right – you get a dollar, and have made 61 cents profit on each 39 cents you stumped up.

Can you trade your position?

Yes, you can sell out of either winning or losing positions before the event is resolved – and most users do just that. Say you buy Yes at 45%, and it moves to 55% as people think it’s becoming more likely. You can cash out early and make a smaller profit of 10 cents. This is an important point in terms of market manipulation and abuse by insiders – you don’t need to win your prediction to make money, you simply need to move market sentiment and cash out. On 9 April, for example, the official temperature recorded at Charles de Gaulle airport jumped suddenly before falling back. The spikes were due to suspected tampering with sensors and coincided with suspicious bets on Polymarket, with hundreds of thousands of dollars in play.

Do prediction markets facilitate corruption?

It appears so. According to analysis by the Anti-Corruption Data Collective, a non-profit research and advocacy group, more than half of “long-shot” bets on military action made on Polymarket – defined as wagers of $2,500 or more at odds of 35% or less – are successful. That suspiciously high average win rate (of 52%) compares with just 25% in politics and 14% for all long-shot markets. The day before the US attack on Iran on 28 February, 150 people made trades of at least $1,000 predicting an imminent strike.

The risks go beyond insider trading, says Sam Freedman on Substack. “The more lucrative these markets become, the more predictions about the future will affect decision-making and behaviour” – and the more public trust will erode as misinformation aimed at manipulating markets becomes widespread.

Should you have a punt?

For a bit of fun, maybe. As a way of making money, probably not. The Wall Street Journal finds that 70% of users lose money and that 67% of all profits on Polymarket go to just 0.1% of accounts. “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.”

Separate analysis by Charles Martineau, a professor at Toronto, came up with similar results. His paper concluded that 69% of Polymarket customers lose money, while the top 1% captured three-quarters of the profits. On Kalshi, too, the large majority of users lose money, with 74% of accounts unprofitable over the past month (on the firm’s own figures). Good luck beating those odds.

This article was first published in MoneyWeek’s magazine. Enjoy exclusive early access to news, opinion and analysis from our team of financial experts with a MoneyWeek subscription.