Prediction markets need clearer rules to attract big investors

Prediction markets are often called “truth machines”. By letting people bet on anything from the outbreak of war to what celebrities will wear, their prices reveal the closest thing to humanity’s best guess about the future — and that information can help people and firms make better decisions, and even hedge risks.

One American, for example, bet big against the government forgiving his student debt, allowing him to pay some down whatever happened. The promise is both entertainment and economic utility.

A boom built on fragile settlement

That promise is helping volumes surge. Volumes on Kalshi, the biggest such financial exchange, climbed to $33bn in June, a 40-fold year-on-year increase. The firm is reportedly seeking to raise funds at a $40bn valuation; Polymarket, its main rival, at $15bn.

Yet both markets face growing pains. Neither has devised a reliable way to settle contracts.

Bookmakers have long faced such problems. They dealt with them by building up a body of rules and precedent. Prediction markets have no such infrastructure to call on.

Two flawed ways to decide winners

Polymarket’s gameable crypto oracle

The process at Polymarket’s main exchange, which is crypto-based, verges on the absurd. Its contracts are loosely worded and often rely on “a consensus of credible reporting” to resolve them, inviting disputes.

When markets are contested, the exchange turns to holders of a little-known cryptocurrency, UMA, to vote on which way a bet should close. A democratic resolution mechanism may sound attractive, but the “oracle”, as it is known, is gameable. An investigation by The Economist has revealed that a handful of wallets control a majority of the ballot and that big Polymarket traders routinely vote on how to define the outcome of their own bets.

Kalshi’s overstretched rulebook

Kalshi, which unlike Panama-domiciled Polymarket operates under the beady eye of America’s financial regulator, has a saner approach. It has hired a small team, led by former university debating champions, to write precise rules for every trade. They settle contracts based on those rulebooks. No one involved in the process is allowed to place bets.

But the team is too small to cope with the more than 10,000 contracts trading on any given day, spanning topics from congressional salaries to the price of compute. In the rush to grow, it is listing ambiguously worded bets and publishing typo-strewn rulebooks. Resolution has been inconsistent. Kalshi did not pay out on the “Khamenei out” trade in full, but did settle a contract on Jimmy Carter attending Mr Trump’s inauguration as a “no” after the former president died.

What credible prediction markets need next

Prediction markets must become more predictable, if they are to keep growing fast — and certainly if they are to attract big trades by institutional investors.

Contracts should be clear in advance; rules should be set by subject-matter experts with adequate resources; and decisions should set precedents. Only if it is known what contracts mean can prediction markets become important in global finance — and realise the vision of the truth machine. ■

Report a correction