Prediction Markets? The Accuracy and Efficiency of Political Prediction Markets in the 2024 Presidential Election

Nearly 2,200 political prediction markets involving more than $3.2 billion in platform-reported trading volume (Confino, 2024) were traded across the Iowa Electronic Markets, Kalshi, PredictIt, and Polymarket during the final nine weeks of the 2024 U.S. presidential campaign. Overall, 86% of these actively traded political markets performed better than a coin-flip, and they were also generally well-calibrated. Even so, there was modest evidence of longshot bias, and the impressive topline accuracy partly reflects the composition of the contracts being traded. 66% of all markets were priced below $0.20 or above $0.80 on the eve of Election Day, and 96% of those contracts resolved as priced. Among markets priced between $0.40 and $0.60 (12% of the sample), 55% resolved in favor of the priced favorite. Market efficiency was more mixed. Prices involving related political events generally covaried, but daily price changes were considerably less synchronized than price levels — suggesting that short-run movements were often market-specific. Arbitrage opportunities in presidential winner-take-all markets were also present on nearly every day. The prices of political prediction markets may provide useful signals about expected outcomes overall, but aggregate evidence of past accuracy and calibration neither guarantees future market performance nor implies that the price of any particular contract provides a definitive or fully efficient assessment of political reality.