Despite the occasional headline about big payouts from prediction market gamblers, hedge funds have mostly avoided the nascent space.
Trading on platforms like Kalshi and Polymarket specifically is often not worth it for many funds, which require deeper markets to make bigger bets on macro developments and would likely struggle to get their compliance team’s sign-off on using these platforms.
While some proprietary trading firms are now dabbling in the space, including Susquehanna, which posted job openings for prediction market traders, the so-called smart money’s interest is focused more on the data being generated by these platforms to help inform their bets.
Similar to the rush to track retail traders discussing stocks on Reddit forums after the GameStop phenomenon in early 2021, funds are, at the very least, ingesting data on activity on platforms like Polymarket and Kalshi. And these platforms make it easy, with a free data feed on trading volumes. Polymarket has also signed partnerships with exchange and clearing house Intercontinental Exchange and Dow Jones that will likely produce more data products for funds to use one day.
Data companies and forecasting firms have also started building products using the exhaust of prediction markets. Dysrupt Labs, an Australian data and forecasting company that works with hedge funds and family offices, pulls prediction market data into its internal algorithms to decide if the “informed minority” is in line or drifting away from the consensus expectations, said Karl Mattingly, the firm’s CEO.
Mattingly said in an interview that the signal they can generate from recurring economic releases, like inflation or jobs data, can give users an “early view on if the prevailing view is going to change in the next two to four days.”
He said that Dysrupt’s research has found that 95% of the time, the consensus from traditional sources like economists and consultants aligns with that of prediction markets. But the other 5% is a chance for traders to make money on breaking away from the pack.
Prediction markets are “the fastest way to model a known unknown,” he said, adding that the average drift from the consensus generates up to 12 basis points of uncorrelated gains.
“Financial markets need better and faster information,” he said, and “this is a really fast way of looking at things.”
Still, the newness of these platforms means hedge funds aren’t entirely sure what the data is useful for just yet. Macro managers aren’t yet building in Kalshi’s odds on a Chinese invasion of Taiwan into their models, said Daryl Smith, head of research at Neudata, a data consulting firm.
“The wider interest has been at the intersection with sports betting,” Smith said in an email. A data executive at a smaller hedge fund said they mainly use prediction market data to track the general interest in gambling, as that’s a helpful signal for the success of stocks like DraftKings and FanDuel parent, Flutter Entertainment.
“We have not seen compelling evidence of demand for macro event-related prediction markets data,” Smith said.
