Part one of “The price of tomorrow,” a series on prediction and futures markets.
In November 2024, professional journalists covering the U.S. presidential election began citing Polymarket alongside Reuters and the Associated Press. Not as a curiosity. As a signal. Polymarket is not a polling organisation, not a modelling outfit, not a firm with political scientists on staff. It is a decentralised, crypto-native prediction market built on a blockchain, operated by no single authority, accessible to anyone in the world willing to trade in cryptocurrency. Its prices are not produced by analysts. They emerge from anonymous participants placing bets on binary outcomes.
That a platform of this description was being treated as a credible signal by professional journalists says something about how far prediction markets have come. But it also raises a question that matters more than the coverage itself: what is a prediction market actually doing, and how does it differ from the futures markets that have priced uncertainty in global finance for nearly two centuries?
The two instruments look superficially similar. Both produce a number representing a claim about the future. Both attract participants who believe they know something the market has not yet priced. But beneath that surface similarity lies a divergence that concerns not just mechanics but the assumptions each market makes about the nature of uncertainty itself. That distinction is not academic. Reading either instrument without understanding which question it is answering produces wrong conclusions from the answer.
Chicago, 1848
The futures market was not born in a trading room. It was born on the shores of Lake Michigan, out of a practical problem that was rooted in soil and seasons.
In the mid-nineteenth century, American grain markets were chaotic in a specifically seasonal way. Harvests arrived in bulk, prices collapsed, and farmers bore ruinous losses. Months later, supply dried up, prices spiked, and millers absorbed the cost. The Chicago Board of Trade, founded in 1848, introduced a standardised contract committing buyer and seller to exchange a fixed quantity of grain at a predetermined price on a specified future date. Risk was not eliminated. It was transferred. The farmer who could not afford to gamble on December prices locked them in. A speculator who believed prices would rise took the other side.
The premise of this arrangement is not prediction. It is management. The futures market was designed for a world in which the future is uncertain but not unknowable in its broad contours, and in which the primary problem is not figuring out what will happen but deciding who should bear the consequences when it does. That logic still governs the vast machinery of modern futures markets. When an airline buys jet fuel futures, it is not forecasting crude oil prices with any particular confidence. It is removing a variable it cannot control from its operating model. When a portfolio manager shorts equity index futures, she may be expressing a bearish view, but she may equally be hedging an existing long position for reasons entirely internal to her fund’s risk architecture. The price that emerges from these interactions contains information. But it is information entangled with hedging pressures, institutional flows, and the structural needs of participants who are not primarily in the business of forecasting.
The futures market treats the future as something to be managed: parcelled out, hedged against, absorbed by those best positioned to carry it.
A different claim entirely
Prediction markets begin from a different premise. They are not designed to transfer risk. They are designed to aggregate information and forecast outcomes.
The structure is direct. A contract pays out one unit of currency if a specified event occurs and nothing if it does not. If you believe the probability of the event is higher than the current price implies, you buy. If you think it lower, you sell. The price that emerges is, in principle, a direct expression of the market’s collective probability estimate. It is not a hedging instrument. It is an instrument of collective belief.
Friedrich Hayek’s 1945 essay “The Use of Knowledge in Society” remains the most precise statement of the intellectual tradition behind this: no central authority can possess all the information dispersed across millions of individual actors, but a price mechanism can distil that knowledge into a single number that guides behaviour. Prediction markets take that Hayekian logic and apply it to a specific question with a verifiable answer. The result, when conditions are right, is a probability estimate that reflects not any single expert’s view but the aggregated judgment of everyone who has bothered to place money behind their beliefs.
The empirical record is credible enough to take seriously. Berg, Forsythe, Nelson, and Rietz (2008), analysing the Iowa Electronic Markets across five U.S. presidential elections from 1988 to 2004, found that IEM vote-share predictions were closer to the eventual outcome than contemporaneous polls 74% of the time, significantly outperforming polls when forecasting more than 100 days in advance. Wolfers and Zitzewitz, in a 2004 survey published in the Journal of Economic Perspectives, drew on a wider range of prediction contexts to conclude that market-generated forecasts typically outperform most moderately sophisticated benchmarks. The underlying mechanism is Darwinian: traders who are systematically wrong lose money and eventually exit, while those who are well-calibrated profit and remain.
The calibration case is not, however, without friction. Manski (2006) cautioned that prediction market prices confound beliefs with risk preferences, which complicates the interpretation of a price as a clean probability estimate. And the Hayekian mechanism depends on conditions that are not always met: broad and diverse participation, no single actor with a position large enough to move prices, and no structural incentive to manufacture a signal rather than respond to one. In well-functioning markets with sufficient liquidity, these conditions broadly hold. Whether they hold in thinner, higher-stakes markets is a question the empirical record has not yet fully resolved.
Where futures markets treat the future as manageable, prediction markets treat it as knowable. That distinction carries more weight than it first appears, and more qualifications than its advocates sometimes acknowledge.
The decentralisation question
Polymarket sharpens this distinction considerably, because it adds a dimension that conventional prediction exchanges do not carry. There is no operator setting prices, no market maker with a structural interest in a particular outcome, no regulator mandating the terms of participation. Prices emerge from a global pool of anonymous participants trading in cryptocurrency, subject to no jurisdiction and answerable to no authority beyond the logic of the market itself.
That is a fundamentally different claim about who gets to set the price. A regulated futures exchange operates within an institutional framework that constrains behaviour, enforces transparency, and provides legal recourse. Its prices are produced by actors who can be identified, audited, and held accountable. Polymarket’s prices are produced by actors who may be anywhere, trading under any identity, with any motivation
That is not, in itself, an argument against it. Ungoverned does not mean uninformed. Its record on the 2024 presidential result attracted attention that was difficult to dismiss as coincidence. But the anonymity that enables broad participation also enables concentrated positions that can move prices independently of any informational signal. In October 2024, a small cluster of accounts (subsequently identified through Wall Street Journal reporting as linked to a single French trader calling himself Théo) collectively held approximately 25% of all Trump-wins-Electoral-College contracts on the platform and over 40% of the contracts on Trump winning the popular vote. The trader’s stated intent was profit, not manipulation. But the Journal was unable to rule out links between the accounts and political organisations, and the market’s own structure made it impossible to distinguish superior information from a deliberate attempt to manufacture a signal. That is not a hypothetical governance concern. It is a structural one, inscribed in the design.
Cousins, not twins
There is a temptation to treat this contrast as settled: futures markets are serious, institutional, and trustworthy; prediction markets are novel, speculative, and unproven. That hierarchy deserves scrutiny.
The Chicago Board of Trade was not born legitimate. In its early decades, it was regarded by many as a venue for gambling on grain prices, which, in a certain light, it was. The futures contracts that now underpin global commodity markets were once derided as divorced from productive economic activity. The institutional legitimacy that futures markets now possess was not granted at founding. It was accumulated over generations, through demonstrated utility, regulatory incorporation, and the simple passage of time.
Prediction markets are earlier in that arc. They have demonstrated utility in specific domains. Kalshi, after a period of regulatory litigation with the Commodity Futures Trading Commission, received approval to offer event contracts including election markets in 2024, a meaningful shift from the ambiguity that had characterised the sector. Polymarket itself was fined by the CFTC in 2022 for operating without registration and subsequently shut down its U.S. operations. Whether prediction markets will travel the full distance from novelty to institutional infrastructure is not yet clear. The honest answer is that it depends on whether the accuracy record survives contact with higher-stakes, higher-liquidity conditions, and on whether regulators conclude that the social value of aggregated probability estimates justifies the governance challenges that open, anonymous markets create.
The gap between the two
What Polymarket’s moment in the 2024 election coverage actually revealed is not that prediction markets have surpassed futures markets as forecasting tools. After all, they are built on different premises and asking different questions.
Futures markets ask who should bear the risk, and at what price. Prediction markets ask what the collective intelligence of the market believes will happen. These are genuinely different questions, and the prices they generate reflect genuinely different things. When the two instruments diverge on the same underlying event, as they sometimes do materially, the gap between them is itself analytically significant. It is not noise to be averaged away. It is a signal about which participants, under which institutional conditions, with which motivations, are pricing the same uncertainty differently.
That gap is the more interesting question. And it does not yet have a clean answer.
The “Price of tomorrow” continues next month with part two: “The Ethics of Knowing,” on whether pricing the future carries a moral charge, and what the 2024 election’s most controversial trader reveals about the responsibilities that prediction markets create.
Disclaimer: This article is for informational purposes only and does not constitute investment advice. Investors should conduct their own due diligence or consult with a financial advisor before making any investment decisions.
Photo by Anatolii Grytsenko on Pexels.
