Kalshi and Polymarket have rolled out new insider-trading bans this week, with both platforms confirming policies that bar politicians from trading contracts on their own campaigns, athletes from trading on their own leagues, and employees from trading on their own employers.

The rules drop into a category that has, until now, operated without the kind of material-non-public-information rules every regulated commodities or securities exchange treats as table stakes. The new policies are a direct response to CFTC pressure and the federal insider-trading prosecution of Master Sgt. Gannon Van Dyke earlier this year over the 2024 Maduro election market on Polymarket.

Both platforms framed the rules as voluntary code-of-conduct updates rather than CFTC-required compliance changes. The substance, however, lines up with what the CFTC’s enforcement division has been pushing for since the Van Dyke charges landed.

What the New Rules Cover

The three core prohibitions are tightly drawn. Politicians and their staff cannot trade contracts that resolve on the politician’s own electoral outcome or on legislation they sponsor. Athletes and team employees cannot trade contracts that resolve on games involving their own league. Employees of any company referenced in a contract cannot trade that contract.

Family members fall under the same restriction in a “household” definition that mirrors federal securities-law conventions. Both platforms also added record-keeping obligations for high-volume traders and a self-certification requirement at account creation that the user is not a covered person for any active market.

Enforcement is platform-side: account closure, position liquidation at trade-time prices, and referral to the CFTC for any pattern that looks like deliberate trading on material non-public information. Neither platform claims the rules are perfectly enforceable. Both have made clear they will refer enforcement gaps to federal investigators rather than litigate them in-house.

The CFTC Pressure That Forced This

The Van Dyke case did most of the work. Federal prosecutors charged the Air National Guard sergeant earlier this year with commodities fraud over a $409,881 winning position on the Maduro election market, alleging he had traded on classified information he encountered through his military intelligence role. The case is the first criminal commodities-fraud prosecution tied directly to a prediction-market platform.

For the CFTC, the prosecution opened a regulatory window. The agency had already been signalling that prediction markets needed insider-trading rules consistent with other commodities products. The Van Dyke charges gave the agency a live fact pattern to point at, and the platforms moved before the CFTC could move first. Whether the new policies satisfy the agency’s expectations, or whether the CFTC tightens further, remains the open question.

Hanson’s Contrarian Pushback

Not everyone agrees the new rules are good policy. Economist Robin Hanson, who pioneered prediction-market theory through the Foresight Exchange and Iowa Electronic Markets work in the 1990s, told Fortune that insider trading “is the whole point” of prediction markets. Hanson’s argument is that prediction markets exist to aggregate information, and the most informed traders are the ones with material non-public information. Banning them from the market, the argument goes, removes the very signal the market is built to surface.

The Hanson position is the academic minority view, but it is not a fringe one. The Iowa Electronic Markets explicitly allowed insider trading in its early academic-research configuration, on the same theory: better information in, better forecast out. The applied-economics literature on prediction markets has consistently found that insider-trading bans reduce forecast accuracy in lab settings.

The platforms’ counter-argument is regulatory survival. Whatever the academic case for letting insiders trade, the political case is unsustainable. A CFTC that watched a Pentagon employee make $409,881 on classified information cannot leave the rules where they were. Kalshi and Polymarket have chosen forecast accuracy as the cost of staying open.

What Counts as Material Non-Public Information

The grey area is not in the named categories but in the surrounding informational economy. Reporters with embargoed information, financial-sector analysts with privileged operator data, and lobbyists with advance read of legislative whip counts are not formally covered. The platforms’ rules require self-disclosure for those edge categories, but do not actively police them.

The first test cases will likely come in election-related markets and corporate-event markets, where the line between general industry knowledge and material non-public information is harder to draw than in a sports league. Regulated US operators watching the prediction-market category, particularly those running election-adjacent product internally, are watching the enforcement pipeline closely.

What the rules establish, regardless of how cleanly they get enforced, is that prediction markets now have an insider-trading regime. Of sorts.