The midterm elections are tomorrow and, of course, everyone wants to know who is going to win what. In the old days, the main way to gauge the likely outcome was the results of the polling industry. Polling was hardly infallible, so observers greeted the arrival of prediction markets (for example, Kalshi or Polymarket) with interest. Maybe these markets were a better way to forecast election results?

Building on his previous publication about prediction markets, AAF’s Thomas Kingsley takes this on in his timely piece: Prediction Markets and the Problem of Calling Elections. To cut to the chase, why should you choose one or the other? Polls and prediction markets are two sources of information, so it does not make sense to toss away information. Instead, the key question is whether having prediction market information in addition to the polls improves forecast performance.

Polling and prediction markets both convey information, but they are not the same thing. As Kingsley puts it:

Pollsters must recruit respondents, construct representative samples, estimate turnout, and translate responses into an assessment of voting behavior, and each step introduces error. Prediction markets by contrast, let participants fold polling, campaign reporting, and economic data into a common price.

There are three tests of a prediction market’s results:

Outcome accuracy asks whether a forecast identifies the winner; calibration asks whether events assigned a given probability occur at roughly that frequency across enough forecasts; and incremental predictive value asks whether market information improves forecasts that already incorporate polling. A market can be accurate without being well calibrated, or without telling analysts anything they did not already know.

There is now a growing literature on the performance of prediction markets, and evidence from the 2024 campaigns and the 2026 primaries suggests they are generally well calibrated. But the reliability of a prediction market depends on the regulatory environment, and thus the information and incentives participants face. The rules for prediction markets are in their infancy and will depend on the ongoing decisions of the Commodity Futures Trading Commission and the courts.

Thus, at this juncture, prediction markets appear to belong on the menu of options, although “what they add beyond polling remains an open empirical question.” Of course, the conclusion may look different on Wednesday after evaluating the data from tomorrow’s elections.