267
Pages, pre-publication
Every prior cut of this paper closed on the same open question — whether the CFTC's promised Q2 rulemaking would arrive or hand the courts the vacuum to fill. On June 10 it arrived: 267 pages, and it is the most permissive federal posture toward prediction markets in the agency's history.
The Notice of Proposed Rulemaking — RIN 3038-AF65, "Prediction Markets; Public Interest Determinations" — does two things. It stops treating event contracts as presumptively suspect, and it draws the first set of bright lines around the contracts that will never be allowed. Both moves run through a single reinterpreted word: involve. Start there.
For fourteen years the CFTC read the word "involve" the way a worried parent reads a permission slip — broadly, suspiciously, with the benefit of the doubt running against the market. If a contract was even related to war, gaming, or terrorism, it was in scope and probably dead. That is the reading the agency used to kill Nadex's election contracts in 2012 and Kalshi's in 2023. It is also the reading a federal court told the CFTC was wrong.
This proposal concedes the point in regulatory prose. The Commission now says a contract "involves" one of the five prohibited activities only when its settlement is determined by an occurrence — or the extent of an occurrence, or a contingency — inside that activity. Not when it relates to it. Not when it equates to it. Only when the thing that decides who gets paid happens within the banned activity itself.
It's easy to file this under housekeeping. Don't. Every major prediction-market dispute of the last decade turned on that one word — involve — and rewriting it resets everything downstream: which contracts are even eligible for prohibition, and which never enter the rule at all. The cleanest illustration is the one the proposal works through itself.
The proposal's sharpest worked example runs through the same chokepoint we modeled in "The Hormuz Actuarial Blockade" — armed conflict in the Strait of Hormuz. Take that one real-world event and write two contracts that both depend on it. Under the old reading, both die. Under the new reading, only the settlement trigger matters.
This is why oil, freight, insurance, and macro-hedging contracts that key off geopolitical stress survive while the geopolitics-as-settlement contracts do not. The agency built the doctrine to let the commercially useful instrument live next door to the prohibited one. For anyone building risk product on event contracts, that is the sentence to underline.
The Hormuz pair is the clean two-contract version. Generalize it and the proposal is a classification engine: it sorts contracts by what determines settlement, not by what influences the underlying event. Three gates, three bins. The engine below runs that logic on the proposal's own worked examples — pick a contract and trace where it falls.
NOTE The five Enumerated Activities are fixed by statute (CEA §5c(c)(5)(C)): activity unlawful under federal or state law, terrorism, assassination, war, gaming, and any "other similar activity" the Commission later names. Gate 02 is a filter, not a verdict — involving one of these makes a contract eligible for prohibition, never prohibited automatically. The agency went out of its way to say "may," not "shall."
The proposal is permissive in the aggregate, with a tightly bounded set of prohibitions. The contracts on the right are not "sports" or "politics" as categories — they are specific structures the agency views as manipulable, violence-adjacent, or informationally empty. Everything else now starts the review presumed allowed.
The Nadex and Kalshi orders both rested on a single move: gaming equals gambling, gambling is wagering on an uncertain outcome, and every event contract is a wager — therefore every event contract is gaming. The proposal calls that reasoning a tautology that would swallow the whole rule, and replaces it with a definition built around the activity itself.
The proposal reads almost like a post-mortem on why the agency lost to Kalshi — it concedes the "involve" standard was misapplied, that elections were never games, and that not every wager-shaped contract is gaming.
Strip the citations away and the document is the CFTC adopting its opponents' brief. In measured agency voice it calls the 2012 and 2023 orders "incorrect," concedes it read "involve" too broadly, and grants that elections and award shows were never gaming. The framework is the small news. The big news is that the proposal rejects the analytical foundation of the Commission's own Nadex and Kalshi orders. Federal agencies rarely propose rules that concede the central legal argument advanced by the parties they were, until recently, opposing in court. The debate has moved from "should prediction markets exist?" to "which specific contracts cross a line?"
The constructive read is real, but a careful compliance officer will reach for the other column first. Five things the NPRM leaves exactly where they were.
None of that is a reason to wave the proposal off; it's a reason to state the conclusion carefully. The right way to put it is not that prediction markets were legalized, but that the Commission is proposing to treat many event contracts as presumptively permissible rather than presumptively suspect — a posture it can be held to and built against.
The proposal rewards two things: the right contract set and the capacity to administer it. That is exactly where an infrastructure provider sits, and it takes the risk out of the macro-and-sports-outcome bet just as the third venue goes live.
Weekly jobless claims and core PCE — the self-certified macro contracts — are on the agency's own list of instruments that fall outside the Special Rule. Baseball game outcomes sit squarely in the favored sports-outcome bucket. The June-30 launch lands into a friendlier framework than it filed under.
The public-interest factors reward venues that can run surveillance, settlement integrity, and league-grade integrity data-sharing. That is a capacity argument, not a contract argument — and it favors the clearing-and-risk layer over anyone who just lists tickets. The compliance burden is the durable advantage.
Injury, officiating, single-play micro, altercation, and any national-security settlement are now documented dead ends. That leaves a short, obvious product map: macro, financial, political, award, and sports-outcome contracts on top, with the integrity infrastructure to defend the public-interest case underneath.
For two years the question was whether prediction markets would be allowed to exist. This proposal moves past it: for the first time, the Commission is proposing to treat many event contracts — macroeconomic, financial, political, and sports-outcome — as presumptively permissible rather than presumptively suspect, judged one at a time against a public-interest standard rather than a categorical ban.
It creates the clearest federal pathway yet for that core book, with prohibitions bounded around violence, national security, athlete safety, and trivially manipulable micro-events. It is still a proposal, and the discretion remains the Commission's, so the move for anyone building here is to be on the record during the comment window, not to declare victory.
The most important sentence in the proposal is also the simplest: a contract only "involves" a prohibited activity when the occurrence that determines settlement happens within that activity itself. That interpretation does not merely resolve the current fight — it sets the frame through which the next generation of event contracts will be judged. The argument is no longer whether prediction markets should exist. It is where the red lines belong.