Looking at the data behind prediction markets
2 days ago
- #information markets
- #prediction markets
- #forecasting
- Prediction markets were initially promoted by scholars to improve decision-making by aggregating dispersed knowledge through prices.
- Current public markets like Polymarket and Kalshi handle billions monthly, but most volume is for sports betting, crypto, and elections, not useful information.
- Useful prediction markets can provide risk monitoring, interpret news, inform policy outcomes, create accountability, and generate novel information.
- Risk monitoring markets, such as those for geopolitical conflicts, show clear demand and are cited by mainstream media.
- Markets for interpreting news and policy outcomes have high volume but often duplicate existing forecasts or focus on entertaining topics like Trump.
- Accuracy improves with higher trading volume for long-duration markets, but short-term markets show no significant link between volume and accuracy.
- AI chatbots like Claude are emerging as competitive forecasters and may overshadow prediction markets by offering narratives and follow-up questions.
- The future of prediction markets may depend on media coverage normalizing their use, but current evidence suggests they primarily serve bettors, not truth-seekers.