CFTC speech market test leaves a gap no transcript can close

On the morning of September 24, Kalshi listed a contract on whether Donald Trump would say "China" at least five times during Xi Jinping's state arrival ceremony. The contract had roughly $125,000 in volume. The resolution method was straightforward: a live video feed, with official transcripts as backup, and a bar on trading by anyone holding material nonpublic information about the event.

That last clause is where the September 22 advisory from the CFTC's Division of Market Oversight becomes instructive — and where the Kalshi listing's public terms fall short of what staff is now asking for.

The advisory applies to what the industry calls mention markets: contracts tied to whether a named person says a specific word, makes a specific appearance, or takes a specific action. Staff may treat these as presumptively "readily susceptible to manipulation" — a phrase that carries weight because regulated exchanges have an affirmative duty to list only contracts that are not readily susceptible. That duty is not new. What is new is the four-factor framework staff wants applied to this specific category.

The framework asks whether the person who determines the outcome faces independent obligations that deter gaming it; whether they could be pressured; whether the result can be verified under substantial public scrutiny; and whether the exchange's surveillance is specifically tailored to that contract's risk profile. A video of what was said addresses the third question. It does not touch the first, second, or fourth.

The Kalshi listing's trading restriction names the class of people who should not participate — employees of source agencies, holders of material nonpublic information. What it cannot show, on its face, is how the exchange would detect a script holder trading through a third account, or identify coordinated activity between someone with access to a prepared remark and a market participant with a meaningful position. Those are surveillance problems, not verification problems, and the advisory treats them as distinct.

The Bessent and BlackBerry listings on the same platform confirm that the staff's concern extends well beyond political speech. A Treasury Secretary's television interview and a corporate earnings call present the same structural vulnerability: a small number of people know what words will be spoken before the market resolves, and those people may or may not be subject to the specific trading restrictions the exchange has written into its rules.

The advisory is staff guidance. It creates no binding rule and does not prohibit the category. Kalshi continued listing mention markets after it was issued. But the guidance is also not costless. Exchanges that list contracts staff considers presumptively susceptible to manipulation without an adequate contract-specific filing take on the risk that a future enforcement action finds they failed their listing duty under the Commodity Exchange Act. The duty runs to the contract, not the category — which means a generalized surveillance program, however sophisticated, is not the same as a tailored explanation of how a specific contract's specific risks are controlled.

The ESMA's July conclusion that certain binary contracts qualify as financial instruments under MiFID II is quietly relevant here. European regulators are reaching the same structural conclusion the CFTC reached years ago — these instruments are derivatives — while arriving at different answers about which ones are listable and under what conditions. The FCA's division of the perimeter, placing non-financial event contracts under the Gambling Commission rather than itself, suggests that even within the financial-instrument framework, the mention market category may not resolve cleanly in any jurisdiction.