Federal Reserve Research: Kalshi Prediction Markets Are Just as Good, if Not Better, Than Traditional Forecasting Methods

The rise of prediction markets offered by companies like Kalshi and Polymarket has captivated a wide swath of risk-takers, from investors in traditional markets to sports gamblers. The new markets allow people to purchase contracts on the outcomes of future events, ranging from elections to sporting events to economic indicators to other events that were previously more difficult to wager on.

The contracts are priced based on the broader market's perceived likelihood of an event and are time-sensitive, meaning there is usually a specific date by which something must occur. Prediction markets have also become a key research tool for looking at the likelihood of certain events. For instance, the prediction markets were among the first to suggest that President Donald Trump would win the 2024 presidential election.

In fact, a new study by the Federal Reserve found that Kalshi prediction markets performed just as well, if not better, than traditional forecasting methods at predicting certain economic outcomes.

Image source: Getty Images.

Kalshi is remarkably accurate

The Fed's study focused on Kalshi's ability to predict economic data releases, including the Consumer Price Index (CPI), a key gauge of inflation; gross domestic product (GDP); the unemployment rate; and the federal funds rate. The Fed specifically examined how Kalshi's prediction markets performed relative to more traditional forecasting methods, such as the Fed's Survey of Market Expectations, Bloomberg consensus estimates, and federal funds futures.

The Fed conducted its study by scraping trade-level data from Kalshi and treating the price of a "yes" contract as the market's baseline expectation of an event occurring. Then the Fed compared the results to more traditional forecasting tools.

The Fed's research found that Kalshi expectations for headline CPI (year over year) were a "statistically significant improvement" over the Bloomberg consensus. Kalshi expectations for core CPI (excluding food and energy prices) and unemployment were statistically similar.

The study also found that Kalshi was better at predicting the federal funds rate than fed funds futures. "The mode of the Kalshi distribution, for example, has
perfectly matched the realized federal funds rate by the day of each meeting since 2022, a feat not achieved by either surveys or futures," the Fed concluded.

More than just outcomes, the Fed also found that Kalshi filled several gaps that more traditional forecasting methods could not. For instance, Kalshi incorporates the views of retail investors, offering a perspective that has not historically been represented in complex financial derivatives or Fed surveys.

The real-time nature of Kalshi's markets is also an advantage over surveys, which can quickly become stale, or derivatives, which can be thinly traded. Kalshi also allows bets on economic indicators such as unemployment, GDP, and core CPI, which historically were difficult to wager on through derivatives.

The breadth and real-time nature of Kalshi's markets also allowed the Fed to study how certain data points and surprises impacted other data points and market reactions. Overall, the Fed was impressed by how Kalshi could provide real-time, broad-scale outcomes, including information on market uncertainty and tail risks:

... these markets yield well-calibrated, rapidly updating density forecasts on important economic variables, including several for which alternatives are not available. Our study highlights the promise of prediction markets as a new benchmark for measuring expectations and informing monetary policy decisions.

Caveats and considerations

While the paper seems overwhelmingly bullish about prediction markets, at least when it comes to predicting economic indicators, the Fed did mention several caveats, including that contracts representing events less likely to occur often had low trading volume, similar to derivatives. The Fed also found that probabilities could be distorted by a risk premium, or the return expected for taking on additional risks.

It's certainly hard to argue that prediction markets aren't useful for assessing the probabilities of certain events and economic indicators. However, one issue that these platforms face is the risk of insider information, similar to insider trading. Now, this seems more prevalent in random events than in economic data releases. But the more events there are to bet on, the higher the likelihood of insider trading, especially with so many stakeholders now able to participate in these markets.

Who's to say that insider information doesn't naturally make the prediction markets more accurate? Overall, I do find these platforms quite useful and expect them to grow in popularity, but I also believe certain regulatory issues remain.

The Motley Fool has a disclosure policy.