CFTC bans White House staffer over Trump speech bet manipulation

Gabriel Perez did not need a model. He needed a calendar and a script.

The former White House teleprompter operator placed trades on prediction markets tied to whether President Trump would speak specific phrases during public appearances, then collected on those trades using knowledge no outside bettor could have. The CFTC banned him from prediction markets for three years. Then, on September 22nd, the agency issued a staff advisory warning that "mention markets" — contracts that settle on whether a named individual utters a specific phrase — carry structural manipulation risk that most platforms are not equipped to stop.

The advisory is careful about what it is not. It creates no binding rules, establishes no enforcement precedent, and provides no no-action relief. What it does is name the mechanism: contracts whose settlement "turns on the discrete conduct" of a single individual, where outcomes are neither independently generated nor externally verifiable, fall outside the basic discipline that makes a prediction market a market. You cannot aggregate dispersed information about what one person will say at a scheduled event when one trader already has the script.

The CFTC's framing points to Core Principle 3 of the Commodity Exchange Act, which prohibits Designated Contract Markets from listing contracts readily susceptible to manipulation. The advisory stops short of banning mention markets outright, instead calling on platforms to implement what it terms "prophylactic trading rules." That phrase is doing a great deal of work for a memo that acknowledges it cannot be enforced.

Here is where I break from the consensus reading. Most commentary on the advisory treats it as a warning to platforms — tighten your rules or face scrutiny. I think that misses where the pressure actually lands. The Perez case was not primarily a failure of platform surveillance. It was a failure of access controls. Perez had information that was structurally unavailable to the market. No trading rule catches that. The only intervention that works is refusing to list contracts where a single insider can determine the outcome before the event begins, and doing so requires a definitional line the CFTC has declined to draw.

The advisory's non-binding character is therefore the story, not the background. The agency identified a real manipulation vector, named the precise legal principle at stake, and then explicitly declined to write the rule that would close it. Platforms are left holding a document that tells them what they are doing wrong and grants them no clear guidance on how to stop.

That gap matters beyond mention markets. The Perez case is the most legible version of a problem that scales across any contract where settlement depends on the conduct of a named individual with fewer than a few hundred people in a position to observe it in real time. Speech markets are the obvious case. They will not be the last one.