Why Now Daily.

Published

Prediction Markets: Prices, Probabilities and Limits

Prediction markets let participants trade contracts whose payoff depends on a defined future event.

Timeline

  1. 1988-01-01: The University of Iowa created the Iowa Presidential Stock Market, a modern real-money prediction market for research and teaching.
  2. 2006-05-01: NBER published formal analysis of conditions under which binary prediction-market prices can approximate traders’ mean beliefs.
  3. 2026-03-12: CFTC staff issued an advisory reminding designated contract markets of obligations when listing and supervising event contracts.
  4. 2026-10-12: Current CFTC, NBER and University of Iowa materials were reviewed for this guide.

A prediction market is a market for contracts whose payoff depends on a specified future event. A common binary contract pays $1 if an event occurs and $0 if it does not. Someone paying 70 cents for “yes” risks that amount to receive $1 at settlement if the official outcome is yes; before fees, the maximum gain is 30 cents and the maximum loss is 70 cents. Other contracts can pay according to a numerical range, vote share or several mutually exclusive outcomes, so the payout formula must be read before interpreting the price. [1][3]

For a $1 binary contract, a price of $0.70 is commonly described as an implied 70 percent probability. That interpretation comes from expected value: under simplified assumptions, a risk-neutral trader who believes the event has a 70 percent chance would value the payoff at 70 cents. The number is the market’s current price, not a certified statistical probability. NBER research finds prices often provide useful estimates of average beliefs, while also showing that the relationship depends on conditions including traders’ risk aversion and the distribution of their beliefs. [1][2]

Prices change as participants submit bids and offers based on public information, private analysis, hedging needs and expectations about other traders. People who think “yes” is underpriced buy it; people who think it is overpriced sell or take the opposing position, subject to the platform’s design. The Iowa Electronic Markets use price-and-time ordered queues with visible best bids and asks, illustrating how an order book aggregates competing views. A last trade, midpoint, best bid and best ask are different numbers, especially when the spread is wide. [3]

Liquidity determines how much confidence to place in the signal. A market with many independent participants, tight spreads and meaningful depth can absorb new information with less price disruption. A thin market may move sharply after one trade or show a stale last price. Position limits, fees, restricted access and difficulty moving money between venues can prevent arbitrage from correcting inconsistencies. Participant selection also matters: a market can confidently reflect the beliefs of a narrow or similarly informed crowd and still miss information held elsewhere. [1][2][3]

The contract’s wording is part of the forecast. Traders need an exact question, deadline, eligible outcome, authoritative source and process for cancellations, corrections and disputes. “Will it rain?” is incomplete without a location, measurement threshold, time window and data source. A surprising resolution can result from precise rules rather than the event as casually described in headlines. CFTC consumer guidance says customers should receive transparent terms explaining payouts, prices, who decides settlement and how the determination will be made. [1]

Prediction markets can aggregate information, but they can also face manipulation, insider advantages and conflicts around events that a participant may influence. A trader can temporarily push a thin price without changing the underlying likelihood. In regulated U.S. markets, designated contract markets have surveillance and rule-enforcement duties aimed at fraud, manipulation, abusive trading and misuse of inside information. Those controls reduce risk but do not make every price accurate or every trade profitable, and protections can be limited on unregistered or overseas platforms. [1][4]

Forecasting and wagering are different uses of the same price. A business might monitor an event contract as one input to planning or use a permitted contract to offset a related economic exposure; another trader may simply speculate. Neither purpose turns the quote into a fact. Compare the market with polling, base rates, model forecasts and direct evidence, and examine how quickly it updates after news. For multiple mutually exclusive outcomes, check whether prices add to roughly the total payout after accounting for spreads, fees and the platform’s contract structure. [1][2]

Before relying on a prediction-market number, verify the contract and venue, read the resolution rules, note the quote type and time, and inspect volume, depth and bid-ask spread. Ask who can participate, whether positions are capped, which costs affect returns and what happens during a dispute or market cancellation. Legal availability and regulatory treatment vary by contract, venue and jurisdiction, so use the relevant regulator’s current information. Treat the price as a continuously updated, incentive-weighted opinion with measurable limits, not as certainty about the future. [1][2][3][4]

Sources

  1. Commodity Futures Trading Commission — Understanding Prediction Markets and Event Contracts
  2. NBER — Interpreting Prediction Market Prices as Probabilities
  3. University of Iowa — Iowa Electronic Markets Forecasting the 2024 Presidential Election
  4. Commodity Futures Trading Commission — Staff Issues Prediction Markets Advisory

Related stories