How to fix insider trading on platforms like Polymarket and Kalshi

Ever since Polymarket correctly tipped a Donald Trump victory in 2024, months before the polls and pundits did, prediction markets have gone from strength to strength. 

Monthly trading volumes grew $4.3 billion in the lead up to the election, and have continued sharply upwards to $10.6 billion in March this year.

But allegations of shady insider bets, on everything from the content of MrBeast videos to the timing of the Iran war, began to dog both Polymarket and Kalshi as they became more popular.

Karoline Thomsen, international law and relations researcher at the University of New South Wales, says there is evidence that prediction markets can surface and aggregate useful information — but they can also incentivize insider trading.

“I think with Polymarket, what you also get are bets that lend themselves to people misusing insider information,” she says. “At that point, you’re not predicting what is going to happen, you are telling other people, or signaling to other people, what is going to happen. ” 

What prediction markets do in theory

The promise of prediction markets, as laid out by economist Robin Hanson, is that wagering on an outcome will reveal people’s actual beliefs — rather than saying the “right” things to pollsters — aggregate dispersed information and provide the crowd’s informed best guess about a future outcome expressed as a price. 

Unlike traditional bookies, who set the odds and the house takes the profits, prediction markets essentially trade in shares of Yes or No, allowing the odds to rise and fall along with sentiment.  

This works pretty well, with Polymarket claiming an average of 90% correct predictions, from a month out, on resolved markets.

Elisabeth Diana, communications head for Kalshi, explains prediction markets are a valuable public service.  

“Around 70% of people on Kalshi actually just come to see the odds, to see the forecasts. They don’t trade,” she tells Magazine.”I mean, the value is enormous. They see it as a source of truth.”

In an ideal world, prediction markets would be based on informed traders using public, but not widely known, knowledge. The Pentagon Pizza Report is a good example — it gives an indication of whether everyone at the Pentagon is working late, which raises the chance that something big is about to happen.

Diana gives another example. “If you parked outside a Walmart and watched everyone come in and out every day for three weeks, you’d have a better sense of how Walmarts are doing. That’s not illegal information,” she says.

When does information cross the line to become ‘inside information’

But it becomes insider trading when someone misuses “material nonpublic information.” 

“A lot of times, for example, if you’re permitted to share information publicly, you can trade on it. If you’re not permitted to disclose it, you can’t trade on it,” Diana says.

“For example…. if you’re a dancer employed by Bad Bunny at the Super Bowl and you traded on information about his first song, that is insider trading because you were employed.”

However, if you were standing outside the stadium and heard Bad Bunny rehearsing a song, betting on that information is fine.

“That is legitimate, because that is public. Anyone can do that.”

Insider trading on the war(s)

The trouble is that prediction markets have vastly expanded the pool of people that possess inside information. Insider trading used to be the preserve of CEOs and board members, but now everyone, from Google search developers to members of the Nobel Committee, face financial temptation. These aren’t random examples; there were suspicious last-minute bets on the 2025 Nobel Peace Prize and Google’s 2025 Year in Search rankings.

Prediction markets also create scenarios where a single actor can affect the outcome. Coinbase CEO Brian Armstrong famously read out the entire list of words that pundits had bet he might say during an earnings call. 

But it’s been the misuse of classified information in wartime that’s drawn the most concern.

On Jan. 3, in the hours before US troops were deployed to Venezuela, a newly created Polymarket account bet on that eventuality and picked up $400,000 in…