Prediction Markets

framework · economics · structured-empirical

Aggregating dispersed information about future events through trading mechanisms that produce price-based probability estimates.

Prediction Markets are speculative markets in which contracts pay out based on the outcome of future events, with the equilibrium price interpretable as a market-aggregated probability estimate of the event's occurrence. The framework draws on Friedrich Hayek's 'The Use of Knowledge in Society' (American Economic Review, 1945) which argued prices aggregate dispersed information that no central planner could possess, and was operationalized in academic prediction markets including the Iowa Electronic Markets (founded at the University of Iowa in 1988 under Robert Forsythe, George Neumann, and colleagues) which has run continuous election-prediction markets since. Robin Hanson's 1990s work on information markets and the controversial 2003 DARPA Policy Analysis Market proposal (which produced congressional shutdown over the appearance of trading on terrorism predictions) shaped subsequent development. Contemporary prediction markets include Polymarket, Kalshi, and Manifold Markets, alongside academic-research operations and intra-organizational markets (HP, Google, and others operated internal markets in the 2000s-2010s with mixed sustained adoption). The framework has substantial empirical support for accurate price-based probability estimates in markets with sufficient liquidity, though regulatory constraints in the United States have shaped which markets are legally accessible.

Originators

Robert Forsythe, George Neumann, Forrest Nelson, and collaborators (foundational Iowa Electronic Markets at University of Iowa, founded 1988 originally as Iowa Political Stock Market); Robin Hanson (foundational theorist of information markets and prediction-market mechanism design, GMU economist, 1990s onward); intellectual antecedents in Friedrich Hayek's 1945 'The Use of Knowledge in Society', Eugene Fama's efficient-market hypothesis (1965 onward), futures-market traditions (Chicago Board of Trade and commodities futures stretching back centuries), James Surowiecki's broader Wisdom of Crowds (2004) framing; Subsequent development through Justin Wolfers and Eric Zitzewitz academic research, Kenneth Arrow, Robert Shiller, and other economists' advocacy for broader prediction-market adoption, regulatory frameworks (CFTC no-action letter for IEM, Kalshi DCM designation 2020), commercial prediction-market platforms (Intrade until 2013 shutdown, Polymarket, Kalshi, Manifold), DARPA Policy Analysis Market controversy 2003, internal corporate prediction markets at HP (Charles Plott research), Google, Intel, Microsoft, and others high

Year / Decade

1945 (Hayek information-aggregation foundation); 1988 (Iowa Electronic Markets founding); 1990s (Hanson information-market theoretical development); 2003 (DARPA PAM controversy); 2020 (Kalshi DCM regulatory milestone); ongoing development high

Primary sources

Hayek, F.A. (1945). 'The Use of Knowledge in Society', American Economic Review (information-aggregation foundation), Forsythe, R., Nelson, F., Neumann, G.R. & Wright, J. (1992). 'Anatomy of an Experimental Political Stock Market', American Economic Review (Iowa Electronic Markets foundational paper), Hanson, R. (2003). 'Combinatorial Information Market Design', Information Systems Frontiers, Wolfers, J. & Zitzewitz, E. (2004). 'Prediction Markets', Journal of Economic Perspectives, Arrow, K.J., Forsythe, R., Gorham, M., Hahn, R., Hanson, R., Ledyard, J.O., Levmore, S., Litan, R., Milgrom, P., Nelson, F.D., Neumann, G.R., Ottaviani, M., Schelling, T.C., Shiller, R.J., Smith, V.L., Snowberg, E., Sunstein, C.R., Tetlock, P.C., Tetlock, P.E., Varian, H.R., Wolfers, J. & Zitzewitz, E. (2008). 'The Promise of Prediction Markets', Science high

Core components

Contract structure: binary contracts pay $1 if event occurs (and zero otherwise), with equilibrium price interpretable as probability estimate
continuous-outcome contracts pay based on numerical realization
Information-aggregation mechanism: traders with private information trade until prices reflect collective information, with profit-motivated traders incentivized to incorporate accurate information rather than express opinions
Liquidity-and-accuracy relationship: market accuracy increases with liquidity (number and informativeness of participants), with thin markets producing noisy estimates
Calibration assessment: comparing market prices to realized outcomes across many events tests whether prices correspond to probabilities — empirical research generally finds reasonable calibration in liquid markets
Mechanism design: market scoring rules (Hanson's logarithmic market scoring rule, market-maker subsidies), pari-mutuel structures, and other mechanism-design variants
Combinatorial markets: markets covering conditional and joint probabilities, attempted but operationally complex
Regulatory frameworks: US Commodity Futures Trading Commission (CFTC) regulation, Iowa Electronic Markets' no-action letter for academic research, Kalshi's 2020 designation as designated contract market, gambling-versus-financial-product legal distinctions
Internal-organizational applications: companies including HP, Google, Intel, Microsoft have operated internal markets for product launch dates, sales forecasts, and other questions with mixed sustained adoption
Public-policy applications: hypothesis markets for science policy, terrorism prediction (DARPA PAM controversy), election prediction
Strengths over alternative aggregation methods: incentive-compatible mechanism that rewards accurate forecasting and penalizes inaccurate forecasting in real time
Weaknesses: regulatory constraints in US substantially limit market liquidity, low-probability event prediction is particularly susceptible to noise, manipulation risk in thinly-traded markets

Primary use case

Information-aggregation framework for probabilistic prediction of future events; applied principally in: election prediction (Iowa Electronic Markets continuous since 1988, PredictIt 2014-2022, Polymarket 2020 onward), policy and macroeconomic prediction (Kalshi markets on inflation, Fed decisions, GDP releases), sports-outcome prediction (substantial overlap with sports betting industry), entertainment-event prediction (Hollywood Stock Exchange), corporate internal forecasting (HP, Google internal markets), public-health prediction (Metaculus, Manifold's COVID-19 markets), AI-development and technology forecasting (Metaculus AI markets), academic research on information aggregation; standard reference in market-design economics, public-policy debate about institutional forecasting, intelligence-community discussions of analytic methods.

Common criticisms

Prediction Markets have substantial empirical foundation but specific critiques exist — Nassim Taleb has argued prediction markets perform poorly for rare-high-impact events (the domain he has emphasized) because thin markets for low-probability outcomes produce noisy probability estimates substantially affected by individual large trades
critics including Hilary Putnam and political theorists have raised normative concerns about prediction markets on certain topics (terrorism, deaths of public figures, military operations) where market trading would create incentives the broader society finds repugnant — the DARPA Policy Analysis Market controversy in 2003 produced congressional shutdown over precisely these concerns
the regulatory framework in the United States substantially constrains prediction-market liquidity — most US persons cannot legally trade on offshore markets like Polymarket, with CFTC's Kalshi exemption and ban on event contracts producing ongoing litigation
manipulation risk in thinly-traded markets has been demonstrated empirically — coordinated trading can move prices in ways disconnected from underlying probability, with limited liquidity preventing arbitrage correction
the relationship between long-run accuracy and short-run usefulness has been argued to be misleading — markets may produce accurate probabilities across many events while individual prices at any moment may be substantially noisy, complicating real-time decision use
corporate internal prediction markets have a mixed sustained-adoption record — pilots at HP, Google, Microsoft, and others produced encouraging research results but limited sustained operational adoption, suggesting organizational implementation barriers
the long-tail-question accuracy is methodologically harder to evaluate because few resolved long-tail questions exist
specific high-profile prediction-market price-deviations from final outcomes (2016 US presidential election, 2024 election cycle particularly) have generated debate about whether markets failed or correctly assigned non-zero probability to outcomes that occurred
the relationship between prediction markets and adjacent traditions (sports betting markets, financial-derivatives markets) raises classification and regulatory questions
Robin Hanson's broader 'futarchy' proposal (governance through prediction markets on policy outcomes) has been criticized as politically infeasible and conceptually contested — even prediction-market advocates often resist the more expansive governance applications.

Lineage

Siblings: Superforecasting, Efficient Market Hypothesis