Forecasts, not puppet strings: The myth that prediction markets rig reality (Part II)

Prediction markets have become an increasingly prominent tool for forecasting elections, economic indicators, geopolitical developments, and other uncertain events. Their growing popularity has also fueled concerns that allowing people to profit from future outcomes will create incentives to manipulate those outcomes.

Critics often argue these markets provide a shortcut from prediction to control: identify an outcome, place a bet, and then manipulate reality to collect the payoff. But that assumption overlooks the basic economics of these markets.

As discussed previously, a common criticism of prediction markets is that they encourage participants to manipulate future events for financial gain. The Roosevelt Institute’s analysis is a prominent example of this argument. The examples it relies on involve attempts to influence the information used to settle prediction market contracts, not the events themselves.

If prediction markets were truly tools for routinely manipulating elections, wars, and economic conditions, we would expect to see a well-established pattern of successful event manipulation. Instead, the basic economics of prediction markets point in a different direction: changing reality is usually far more difficult than forecasting it.

That is because prediction markets are not another form of gambling. Unlike a casino, where the house takes sides and tilts the odds toward itself, prediction markets make money regardless of the outcome. Prediction market venues charge the same fees to participants who take opposing views about future events, so they are neutral on the outcome.

Traders profit on prediction markets by identifying information that others have missed. A participant who correctly anticipates an election result, economic trend, or geopolitical development can benefit simply by being more accurate than other traders. There is no need to influence the underlying event itself.

In fact, attempting to do so would put the trader in a much more precarious position that traders can avoid simply by identifying mispriced information. That is because the events tracked by prediction markets are rarely the product of a single decision-maker. They emerge from complex systems that are difficult to influence and even harder to control.

Elections depend on millions of voters and institutional processes. Macroeconomic conditions emerge from the actions of consumers, businesses, and policymakers. Wars involve governments, militaries, and adversaries.

A trader who wants to profit from a presidential election market cannot simply change millions of individual preferences or bypass the institutions that administer elections. Someone seeking to profit from a market on economic growth cannot easily alter consumer behavior, interest rates, supply chains, and government policy. The gap between predicting these outcomes and controlling them is gargantuan.

Critics may respond that prediction markets do not require participants to completely control events in order to create problematic incentives. But influence and control are different concepts. Many actors attempt to influence elections, markets, and public debates every day without having the ability to determine the outcome. For prediction-market manipulation to become a rational strategy, the potential payoff would have to justify the substantial costs, uncertainty, and legal risks involved. Prediction markets also contain a built-in corrective mechanism: traders who believe prices are wrong have an incentive to trade against them.

For example, in 2004, a trader attempted to influence the Ireland-based TradeSports prediction market for the US presidential election by making large trades that moved the price toward John Kerry. He did not permanently distort the market. Instead, other market participants took the opposing side, and the market ultimately continued to reflect broader expectations.

The same dynamic appeared in 2012 when a trader known as the “Romney whale” placed enormous bets on Mitt Romney on the Ireland-based Intrade prediction market in an apparent attempt to influence market perceptions by moving the odds. These traders briefly changed prices, but did not affect the overall market trajectory. These examples illustrate how the incentive to identify mispricing also encourages traders to correct it. The incentive structure therefore rewards information gathering, not market manipulation.

If prediction markets truly engineered outcomes, there would be a pattern of successful event manipulation, not isolated examples of trading misconduct, benchmark disputes, or hypothetical concerns. This pattern does not exist.

That’s because prediction markets are not puppet strings attached to the real world. They are information systems that aggregate knowledge, allow participants to weigh evidence, and update expectations about uncertain events. There is much speculation that takes place there, just as it does in traditional financial markets. But describing prediction markets as either gambling or reality manipulation is what’s at odds with reality.

Prediction markets should be evaluated and regulated based on what they actually do, not on the unsupported assumption that they persistently provide a means to manipulate reality. Overregulating them on that basis would make it harder for voters, policymakers, and businesses to access market-generated information about events that affect their decisions.