Author:Scott Kominers
Compiled by: Jiahuan, ChainCatcher
Prediction markets allow people to trade on the outcome of events. Last year, they entered the U.S. on a large scale and are now used to track everything from geopolitics to entertainment awards. But what are they exactly?
As an economist who has long studied markets and incentive mechanisms, my answer is simple: prediction markets are, at their core, markets. Markets are fundamental tools for allocating resources, ensuring that goods and services flow to those who value them most.
In this process, markets also aggregate information: market clearing (i.e., reaching supply-demand equilibrium) is essentially a mechanism that gathers the knowledge of all participants and distills it into a price signal.
Prediction market platforms and products directly harness this information-aggregating capability to forecast specific future events: they design an asset tied to an event that pays out if a particular outcome occurs, and people then trade this asset based on their judgment of whether the outcome will happen.
This application is not new.
Companies have long used prediction markets to gather tacit information from employees, such as predicting whether a key product will launch on time; scientists use them to assess which experiments are likely to be successfully replicated; and today, multiple media outlets partner with prediction markets, using the "wisdom of the crowd" to supplement their own sources and reporters' coverage.
Prediction markets collect information directly from participants—each person's judgment about the future—and aggregate it within a market to answer how likely an event is to occur.
People can "bet" on these events, much like they "bet" on a company's future value in the stock market or "bet" on the future price of a commodity like oil. The difference is that demand for an asset like oil is influenced by many factors simultaneously, whereas the assets designed by prediction markets generate returns only if a specific event occurs.
If oil prices rise, we know that demand has increased relative to supply, but we may not know the reason behind it: it could be an expected escalation of conflict in the Middle East, or someone finding a new use for oil.
With prediction markets, you can isolate and forecast each possibility individually.
For example, a prediction market on "whether the Strait of Hormuz will remain open at a specific time" could revolve around a contract that pays one dollar per unit if the event occurs.
As people repeatedly buy and sell this asset, its market price becomes a "probability barometer," reflecting the collective judgment of traders on the likelihood of that event.
How does it work specifically? Suppose the unit market price for an outcome is $0.50, equivalent to a 50% probability. If you believe the likelihood of the strait remaining open is higher than 50%, say 67%, you would buy; if your judgment is correct, you get a total return of $0.67 for a cost of $0.50.
This purchase then pushes up the market price and the corresponding probability estimate, essentially saying "someone thinks the market is underestimating it." The reverse is also true: when someone thinks the price is too high, they will sell (or short) at a lower price, thereby pulling down the market's overall probability estimate.
When prediction markets function well, they have several distinct advantages over other forecasting methods.
First, they directly provide a probability estimate, which in itself is a "superpower."
Opinion polls and questionnaires only give a "share of opinions"; to translate this into a probability, you must perform statistical inference to determine the relationship between the measured proportion and the population. Moreover, polls are often snapshots at a single point in time, while prediction markets can update in real-time as new participants and information emerge.
More crucially, prediction markets have built-in incentives: buyers and sellers put real money on the line and lose if they are wrong. This encourages participants to carefully weigh their information and put their money on questions they are most confident about.
Conversely, the ability to profit from information and professional judgment in prediction markets incentivizes people to actively conduct research and gain clarity.
(A well-known example: before the 2024 U.S. presidential election, a prediction market participant even conducted their own poll, using unconventional methods to gather information not captured by standard polling agencies.)
Finally, prediction markets also have a significant advantage in coverage. Someone who knows which events might affect oil demand can, in principle, go long or short on oil; but for many outcomes we want to predict, there is no corresponding commodity or stock market to bet on. In such cases, prediction markets become an ideal choice.
For instance, a batch of prediction markets has recently emerged specifically to aggregate judgments on questions like "which AI model performs best on various tasks"—these are too granular to be reflected in traditional commodity markets. And anyone can create and fund a prediction market for such specific questions.
These ideas are not new. Similar practices existed as early as 16th-century Europe, when people used them to predict the next pope.
The foundations of modern prediction markets lie in economics, statistics, market design, and computer science. Charles Plott and Shyam Sunder proposed the earliest formal academic framework in the 1980s, followed shortly by the first modern prediction market, the Iowa Electronic Markets.
With the internet, this model can now aggregate dispersed information from around the globe. However, several preconditions must be met for prediction markets to truly realize their potential.
One set involves infrastructure issues: how to verify and reach consensus on "whether an event has occurred," how to ensure market operation is transparent and auditable, and how to handle, at scale, the settlement of contracts that may be controversial or even manipulated.
The other set consists of market design challenges. First, people who actually possess relevant information must be willing to participate. If participants are uninformed, the price signal says nothing; conversely, only by getting people with diverse information to participate can the prediction market's estimates avoid distortion.
I pointed out in 2016 that prediction markets might have underestimated the probability of Brexit and Trump's first election win because participants at the time did not sufficiently understand the rise of populism.
Another issue is when someone has "perfect" information, such as knowing the actual result in advance, which is also problematic, especially if they can influence the outcome.
Imagine: what if an insider from the papal conclave placed bets in a "next pope" prediction market, trading before the official announcement of Leo's election, or even tried to sway the election to ensure their chosen candidate wins?
Precisely because of this, once potential participants expect insiders to be trading in the market, the rational choice is to simply stay away, and the market collapses.
Finally, there is also the possibility of people deliberately distorting prediction market prices to influence public perception of the probability of a certain outcome, turning it from a tool for "aggregating beliefs" into a tool for "manipulating beliefs."
For example, a candidate's PR team, wanting to make the outside world believe they are a sure winner, might use part of their campaign funds to influence the relevant market.
However, prediction markets have some self-correcting ability in this regard: once the probability of a contract is pushed to an unreasonable height, someone will always be willing to take the other side of the trade.
All of this indicates that prediction markets require greater transparency and clarity in participation management, contract design, and operation. But as long as designers can solve these challenges, prediction markets have the potential to become one of our core tools for anticipating the future.
