Reason Foundation study finds prediction markets and polls roughly even in 2026 midterm accuracy
A Reason Foundation analysis compared Polymarket prediction-market forecasts against traditional polling for the 2026 midterm elections. Researchers checked which forecasts failed to predict the election winner. The results were roughly even between the two methods. The study offers a counterpoint to claims that prediction markets hold a systematic edge over conventional polling. Researchers found that Kalshi and Polymarket forecast the 2024 election more accurately than other platforms, according to a separate referenced study.
The roughly even result weakens the forecasting-utility claim that Kalshi and Polymarket have leaned on in court and Congress. State attorneys general in Ohio and Tennessee are already suing to classify event contracts as gambling, and they have cited accuracy gaps as evidence that markets amplify sentiment rather than predict outcomes. The Reason finding gives them a published source to cite directly.
Senators Warren and Curtis have demanded internal trading data under bipartisan deadline, so a mainstream foundation casting doubt on prediction-market superiority enters the congressional record at a vulnerable moment. Platforms now face the risk that legislators and judges treat their core defense as unproven marketing rather than established fact.