Prediction markets are platforms that allow users to speculate on the outcome of future events. Their intellectual pedigree stretches back to Kenneth Arrow and Gerard Debreu, who envisioned a world where every uncertain future could be priced, hedged and insured against.
Polymarket is the largest prediction-market platform. In May 2025 traders wagered over $1bn on the platform, up from $63m a year earlier. Polymarket is off-limits to American users, having been accused of running an unregistered derivatives-trading platform. Kalshi, a rival, is approved to operate in America but remains mired in regulatory disputes over which contracts constitute legitimate futures and which amount to gambling. Trading volume on Kalshi rose 12-fold to $24bn in 2025.
Prediction markets outperformed opinion polls in forecasting the November 2024 presidential election, correctly pricing in a victory for Donald Trump when polls called it a toss-up. They also did a better job than other sources at predicting Israel's strikes on Iran and the result of New York City's Democratic mayoral primary.
In a 2026 paper, Anthony Diercks of the Federal Reserve and two colleagues found that Kalshi outperforms futures and options contracts at forecasting the Fed's benchmark interest rate. For GDP growth and unemployment—where market-based alternatives do not exist—Kalshi probabilities appeared as good as analysts' consensus forecasts for unemployment and core inflation, and better for headline inflation. Surveys that try to capture expectations take weeks to compile; prediction-market probabilities are always available and constantly updated. Trading volumes on Kalshi and Polymarket together surpassed $50bn in 2025, up from $16bn the year before, though economic contracts remain just 1-2% of trading.
Despite their forecasting successes, prediction markets have yet to attract serious institutional capital. Goldman Sachs cites them in its research, but active trading by large financial institutions remains virtually non-existent on the platforms. Low liquidity outside headline events makes markets vulnerable to price manipulation. Traditional financial instruments — inflation swaps, federal funds futures, commodity futures — already cover many of the same risks. The most active financial market on Polymarket in mid-2025 had attracted wagers worth just $22m, compared with nearly $2trn in the market for Treasury inflation-protected securities.
In 2018 the Supreme Court let states permit sports betting; 39 of the 50 have done so. Since then, Kalshi has expanded from event contracts into sports, and traditional gambling firms have entered prediction markets. In August 2025 FanDuel, America's biggest sports-betting site, announced plans to launch a prediction market with the Chicago Mercantile Exchange, the world's largest futures exchange. Interactive Brokers opened its own prediction market, ForecastEx. Robinhood offered ForecastEx's presidential-election contracts and then Kalshi's broader markets, including sports, to its users. Keynes warned in 1936 of "the capital development of a country becom[ing] a by-product of the activities of a casino"; Warren Buffett wrote in 2023 that "markets now exhibit far more casino-like behaviour than...when I was young".
With around $2bn of wagers placed each week on Kalshi and Polymarket combined by late November 2025—four times bigger than in summer 2025—prediction markets have made insider trading much more attractive. Unlike stock markets, where the SEC has decades of enforcement experience, prediction wagers are regulated as derivative contracts by the Commodity Futures Trading Commission (CFTC), which has much less experience enforcing insider-trading rules. In July 2025 an appeals court overturned a conviction for trading NFTs on inside information. A case alleging that a group including a current NBA player profited from inside information in gambling markets could provide a crucial precedent for prediction platforms.
In February 2026 Israel arrested two men, one a former army reservist, for betting on the precise timing of Israel's attacks on Iran using classified information. They were charged with bribery, obstruction of justice and serious security offences. The rules on insider trading for "commodities" (a category which includes prediction-market bets) are much less stringent than those for "securities" such as shares or bonds. Banning informed traders outright would harm forecast accuracy—a positive externality—but there may be a case for prohibiting government officials from betting on information they obtain in the course of their jobs, as the Stock Act of 2012 does for securities.
Unlike stock markets, informed trading on prediction markets is not inherently illegal: futures law permits it, and the resulting price discovery benefits all participants. Regulators should worry chiefly about contracts where disclosure is harmful (such as military action) or where traders can affect outcomes. Polymarket withdrew a market on whether the Artemis II spacecraft would explode. Platforms should verify identities, sources of funds and employers.
Under the Trump administration, regulators became more open to prediction markets, though it remains uncertain whether future administrations will follow suit. Some platforms have lobbied to put sports betting and event contracts on an equal legal basis, which risks tying economically meaningful markets to more controversial retail offerings.
The generation of random numbers is too important to be left to chance.