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

Prediction markets have become an increasingly prominent tool for forecasting elections, economic indicators, geopolitical events, and other uncertain outcomes. As their popularity has grown, so too has criticism that allowing people to profit from future events will inevitably encourage efforts to manipulate those events.

A recent Roosevelt Institute article advances this familiar claim by arguing that prediction markets are shaping real-world events and eroding public trust. It points to examples such as traders allegedly attempting to influence a weather sensor used to settle a contract and efforts to manipulate the information used to resolve a prediction market about the war in Ukraine.

Critics use these episodes to argue that prediction markets create incentives to manipulate real-world events. But the more fundamental question is whether these incidents actually show that prediction markets uniquely create or commonly encourage the manipulation of the underlying events they forecast.

Too often, manipulation is treated as a problem that prediction markets themselves create. That assumption is incorrect. The desire to influence valuable outcomes in order to profit from them is a longstanding feature of human behavior, not a novel consequence of prediction markets.

Nearly every market faces the possibility that someone will attempt to cheat, manipulate, or gain an unfair advantage. Investors have attempted to move prices through pump-and-dump schemes, as demonstrated by the Stratton Oakmont case. Companies have manipulated information, as Enron and WorldCom did when executives used accounting fraud to mislead investors. Competitors have also attempted to alter outcomes for financial gain, as shown by match-fixing scandals in sports.

The existence of incentives for misconduct, however, has never been sufficient reason to eliminate or heavily regulate an entire market. If that standard were applied consistently, there would be little room for financial markets to exist. The appropriate response is the same one developed across financial markets: improve contract design, strengthen settlement procedures, and punish fraud where it occurs. Prediction markets should be viewed through this same lens.

Like other financial institutions, prediction markets can face multiple forms of misconduct. But those forms should not be treated as interchangeable. Doing so risks turning a narrow problem with a specific solution into a sweeping indictment of an entire market.

Discussions of prediction-market manipulation often blur together three distinct concerns: manipulation of market prices, manipulation of the sources used to settle contracts, and manipulation of the underlying events those contracts measure. These categories can overlap in some cases, but they involve different mechanisms and raise different policy questions.

Market manipulation involves distorting the price signal generated by trading activity. Manipulating a settlement source involves influencing the benchmark or metric used to determine a contract’s outcome. Event manipulation involves changing the real-world outcome being forecast.

The distinction matters because only the third category directly supports the claim that prediction markets are reshaping reality. Yet many examples highlighted by the Roosevelt Institute involve alleged efforts to manipulate the information used to resolve particular contracts as opposed to the underlying events being forecast.

For example, the Roosevelt Institute article discusses a Ukraine-related prediction market in which a trader allegedly attempted to influence a map used to resolve a contract concerning Russia’s capture of the Ukrainian city of Myrnohrad. It also cites an alleged effort to affect weather sensor data used to settle a weather contract, as well as traders pressuring journalist Emanuel Fabian over reporting that affected a contract outcome.

Even assuming these allegations are accurate, they have a common feature: the alleged misconduct involved attempts to influence the information used to determine a contract’s resolution, not attempts to change the underlying events themselves.

Altering a map used to track Russian territorial gains would not change the course of the war in Ukraine. Influencing a weather sensor would not change the temperature or weather conditions being measured. And attempting to influence a journalist’s reporting would not change the events that the reporting describes.

This is important because manipulating a settlement mechanism is a challenge familiar across financial markets, with solutions focused on contract design, reliable benchmarks, and enforcement.

A broken scoreboard does not mean the game has been fixed. Likewise, evidence that traders may attempt to influence how a contract is resolved does not show that prediction markets systematically encourage people to manipulate the events those contracts measure.

This highlights the central flaw in many arguments about prediction-market manipulation: they often confuse the ability to exploit a market’s rules with the ability to control the world outside the market. That issue deserves separate attention, which is why an upcoming piece will examine the economics of event manipulation and whether the incentives to change reality are as strong as critics suggest.