Representative Bryan Steil's bill to bar members of Congress from trading on prediction markets, advanced with the White House's endorsement, lands as a clean political artifact and a messy regulatory one. We have read it the way we read every gambling-adjacent statute that crosses this desk — against the public record of how the gambling sector is actually policed. The contrast is instructive. The UK Gambling Commission, by way of comparison, levied a £17m regulatory settlement against Ladbrokes and Coral in August 2022 for social-responsibility and AML failings that touched millions of customers, a figure documented on its public register. The Steil bill's universe of covered persons is 535.
The Ban Confuses the Symptom for the Sickness
Read the bill at the level of incentive design and a particular shape emerges. It treats the lawmaker-trader as the locus of risk. That is a defensible political instinct. It is also, by the standards of how regulated gambling-adjacent markets are policed elsewhere, a remarkably narrow lens. A serving member of Congress who buys a contract on a contested vote does, of course, present a textbook conflict — informational asymmetry, fiduciary duty to constituents, the trade itself a kind of self-dealing. The framing of the bill is that removing the trade removes the conflict. We are skeptical that the equation is that clean.
The reason we are skeptical is forensic. When the UK Gambling Commission writes an enforcement notice — and we have read dozens — the operative language is almost never about the identity of the bettor. It is about the controls of the operator. Look at the Hillside (Bet365) settlement of £582,120 in December 2022. The notice does not say that the customers behaved badly. It says that the operator failed to interact with high-risk players in the prescribed manner, failed to identify deposit patterns that should have triggered review, failed at the level of system rather than at the level of individual user. The regulator's instinct, refined over a decade of enforcement, is that you do not police a derivatives-adjacent market by chasing individual bad actors. You police it by binding the venue.
The Steil bill does the opposite. It selects 535 individual actors and tells them not to trade. That is a meaningful gesture. It does not bind the venue. Nothing in the bill — at least nothing in the public record of the version we have seen — requires Kalshi, Polymarket, or any prediction-market operator to implement the kind of customer-interaction protocol that the UKGC has, for years, required of every licensed operator on its public register. Nothing requires a regulator equivalent to the Glücksspielbehörde's cross-operator deposit-cap enforcement, the system by which Germany's GGL tracks combined monthly deposits across every licensed operator and prevents a user from exceeding €1,000 in aggregate regardless of how many accounts they hold. The contrast is striking. Germany's regulator treats the user as someone the system must protect from itself across operator boundaries. The Steil bill treats one narrow class of users as people the system must protect itself from. Those are different theories of risk, and the second one is the smaller one.
What would a sickness-level intervention look like instead? It would look like the architecture of a tier-one gambling regulator. Mandatory operator interaction with users who exhibit high-velocity trading on a single political contract. Audit trails on the order book inspectable by a Commodity Futures Trading Commission-style supervisor with subpoena power. Real reporting obligations on aggregated exposure. The piece of legislation that does that has not been introduced. The piece of legislation that has been introduced bans 535 people from logging in. The ratio of headline to substance is the part of this we keep coming back to.
On the Public Record, Lawmaker Exposure Is the Smaller Story
Here is the proportion problem. The global online gambling market — a useful comparator because prediction markets are functionally a subset of binary betting on real-world outcomes — generated roughly $94bn in gross gaming revenue in 2024, per H2 Gambling Capital. Flutter Entertainment alone reported $14.0bn in revenue for the same year, on a customer base of 14.1m registered users, as set out in its results centre disclosures. Entain reported £4,833m in revenue for FY2024 against 28.0m active customers, the number disclosed on page-level basis in its 2024 Annual Report. These are the numbers that show how concentrated the actual policy surface is. You are looking at tens of millions of retail accounts at a single operator and, in aggregate, hundreds of millions across the licensed online sector globally. The Steil bill, by contrast, addresses a covered population of 535 people who, even in the most aggressive theoretical model of lawmaker trading, cannot move a Kalshi political contract's volume in any month by a fraction of what one well-distributed Polymarket retail cohort can.
We do not mean this as a defense of lawmaker trading. We mean it as a calibration exercise. The public-policy weight assigned to a prohibition should bear some relationship to the share of the harm it addresses. Apply the same calibration the UKGC applies in its enforcement priorities and the proportion looks different. When the regulator went after Sky Betting and Gaming for £1.17m in March 2023 — a fine documented in the Commission's enforcement publication — it did so because the operator's social-responsibility framework was failing across a customer base in the millions. The £1.17m number is small relative to Flutter's revenue. The customer harm number is not. That is the analytic gap the gambling regulators have learned to close: they go where the volume is, not where the cleanest political optics live. The Steil bill inverts the priority. The optics are clean. The volume isn't there.
It is also worth noting where the comparable bill is not aimed. We are not aware, on the public record, of a parallel piece of legislation that would require prediction-market operators to implement self-exclusion architecture equivalent to GAMSTOP in the United Kingdom, the single-registration system that automatically binds every UKGC-licensed online operator and now covers 420,000 self-excluded users — a population that grew by 35% year over year, per Gamstop's own disclosures. Self-exclusion, as a mechanism, is the cleanest example of a policy intervention that addresses a real and measurable harm. It is also strikingly absent from the prediction-market discussion in the US, where the legal architecture of the venues themselves remains contested between the CFTC, state gambling regulators, and the operators' own preferred framing of themselves as event-derivatives venues rather than betting platforms. That ambiguity is precisely the kind of thing the Steil bill could have addressed and did not.
The proportion problem extends to the language of the bill itself. Reading it against, say, the Entain Deferred Prosecution Agreement of December 2023 — a £585m settlement with the UK Crown Prosecution Service relating to the former Turkey-facing business of a subsidiary sold in 2017 — you can see what an enforcement document looks like when it grapples with structural risk. Entities. Subsidiaries. Beneficial ownership. Time horizons of nearly a decade. The Steil bill is, by comparison, an instrument of personal abstention. It tells a person not to do a thing. It says nothing about the structure within which the thing is done.
What This Bill Quietly Leaves Untouched
Here is the part we find most clarifying. Run the bill against four checklist items that any serious gambling regulator would address and see what survives the test. First: does the bill require operators to implement responsible-trading interaction protocols modelled on the kind of customer-interaction obligations that the UKGC fined Sky Betting and Gaming for failing — the failures documented in that 2023 enforcement notice? It does not. Second: does the bill mandate the kind of cross-operator deposit-cap architecture that Germany's GGL operates under the State Treaty on Gambling, the €1,000 monthly aggregate that binds operators rather than relying on the user to police themselves? It does not. Third: does the bill condition operator licensing on participation in a unified self-exclusion register — the architecture that makes GAMSTOP functionally binding across the UK licensed sector and that is mirrored in Portugal's RSA and Germany's OASIS? It does not. Fourth: does the bill require the kind of operator-level disclosure that the New Jersey Division of Gaming Enforcement publishes monthly on its public reporting portal, the granular reporting that makes US sports-betting market shares — FanDuel at 28.5% of New Jersey sportsbook revenue, DraftKings at 27.0% — knowable on the public record? It does not.
Four checklist items, four gaps. We mean the gaps in a clinical sense. The bill is what it is. Its scope is narrow on purpose. The political theory underlying it appears to be that the lawmaker conflict-of-interest problem is the rate-limiting harm and that fixing it produces a meaningful share of the social benefit. We do not see how the underlying math supports that theory. The customer-protection harms documented in the UKGC enforcement record are orders of magnitude larger in affected-population terms than the conflict-of-interest harms from lawmaker trading could be, even under aggressive modeling. A serious incentive-misalignment analysis would have started with the larger pool and worked toward the smaller.
There is a real argument that some kinds of policy address symbolic concerns and that symbolic concerns matter. We accept that. We have spent enough time reading enforcement notices to know that public legitimacy of a market depends on more than the strict utilitarian sum of its harms. A ban on lawmaker trading sends a signal about the seat of public trust. The signal is not nothing. What we are flagging is the divergence between signal weight and substance weight. The signal weight here is substantial. The substance weight, measured against the regulatory infrastructure that the gambling sector has built over twenty years and that prediction markets functionally lack, is light.
The cleanest way to see this is to look at what a tier-one regulator's annual register looks like and compare it to the prediction-market environment the Steil bill leaves in place. The UKGC's public license register currently lists 268 licensed online operators, each subject to a published license condition framework, each capable of being placed on a published enforcement action, each obligated to integrate with the GAMSTOP self-exclusion architecture by virtue of holding the license. The US prediction-market environment does not have the equivalent register, the equivalent enforcement publication cadence, or the equivalent self-exclusion infrastructure. A bill that prohibits 535 people from trading on the venues operating in that environment is doing one thing. It is not doing the other thing. The other thing is the larger one.
What we would have written instead, had we been asked, would not have looked like a ban. It would have looked like a licensing condition framework. It would have specified the customer-interaction protocols. It would have specified the disclosure obligations. It would have specified the operator's duty to integrate with a self-exclusion architecture that did not yet exist and would have had to be built. It would have placed the cost of compliance on the venue. That cost is where the policy weight should sit, and the Steil bill does not place it there.
This piece does not cover the constitutional question of whether the federal government can validly prohibit a sitting member of Congress from holding a position in a regulated derivatives market while permitting them to hold equities — that is a securities-law-and-separation-of-powers argument we are not qualified to make. It does not cover the CFTC's evolving position on event contracts, which is a moving target and deserves its own treatment when the public record stabilizes. And it does not cover the tax treatment of prediction-market winnings under existing IRS guidance, which is materially different from gambling tax treatment and would warrant a separate piece. Each of those is its own argument and would need its own grounding.
This started as a note on a single bill and turned into a longer argument about how policy chooses its targets. We did not set out to defend lawmaker trading on prediction markets. We set out to read the bill against the public record of how the broader sector is regulated. The reading, in the end, is the editorial. The Steil bill is a clean political artifact. The harder regulatory question — who supervises the venue, what the customer-protection obligations look like, where the self-exclusion architecture lives — survives this bill entirely intact. It is the question the desk will return to as the rest of the legislative cycle plays out.
FAQ
What does the Steil bill actually prohibit?
The bill, advanced with White House endorsement, would bar members of Congress from holding or trading positions on prediction markets such as Kalshi and Polymarket. The covered population is 535 lawmakers. The bill does not address staff, family-office vehicles, or beneficial ownership through trust structures in the version of the text we have read, and it does not impose obligations on the venues themselves. It is a personal-conduct rule rather than an operator-licensing framework.
Why compare a US lawmaker-trading bill to UK Gambling Commission enforcement actions?
Because prediction markets are functionally adjacent to binary betting markets, and the UK Gambling Commission has spent two decades building the operator-level regulatory infrastructure — license conditions, self-exclusion integration, customer-interaction obligations — that the US prediction-market environment lacks. UKGC fines like the £17m Ladbrokes and Coral settlement show what a mature operator-supervision regime looks like, which is the comparator the Steil bill conspicuously does not invoke.
Does the bill require prediction-market operators to implement self-exclusion?
No. On the public record, the bill imposes no obligation on Kalshi, Polymarket, or any other prediction-market operator to integrate with a self-exclusion register equivalent to GAMSTOP, which automatically binds every UKGC-licensed online operator. GAMSTOP currently covers 420,000 self-excluded UK users, a population that grew 35% year on year. No comparable US infrastructure exists for prediction markets, and the Steil bill does not create it.
How big is the prediction-market harm relative to the broader gambling sector?
The global online gambling market generated approximately $94bn in gross gaming revenue in 2024 across hundreds of millions of registered customers, per H2 Gambling Capital. Flutter alone reported $14.0bn in revenue and 14.1m registered users. The lawmaker-trading subset is, by any measure, a tiny share of the harm surface. Calibrating policy weight to harm size suggests operator-level supervision of prediction markets would deliver more public benefit per unit of legislative effort.
Does the bill address staff or family-member trading?
On the public record we have reviewed, the bill's covered population is members of Congress themselves. We have not seen language extending the prohibition to congressional staff, immediate family members, or trust structures in which a lawmaker is a beneficial owner. The narrowness of the covered population is one of the structural features that makes the bill closer to a symbolic instrument than to a substantive market-supervision rule.
What would a substantive operator-licensing framework look like?
It would mirror the UKGC license-condition architecture: customer-interaction obligations that produced enforcement actions like the £1.17m Sky Betting and Gaming fine in March 2023, cross-operator deposit caps along the lines of Germany's €1,000 monthly aggregate enforced by the GGL, mandatory self-exclusion register integration, and monthly operator-level reporting comparable to what New Jersey publishes on its DGE portal. The Steil bill does none of this.
Is the lawmaker-trading prohibition entirely without merit?
Not entirely. We accept that conflict-of-interest rules carry signal weight, and signal weight matters for public legitimacy. The point of this analysis is the divergence between signal weight and substance weight. The bill produces a high-visibility prohibition that addresses a small share of the underlying market harm. A more analytically honest framework would have paired the lawmaker rule with the operator-licensing architecture, and the legislative record does not yet show that pairing.
Where can a reader verify the regulatory comparators cited in this piece?
The UK Gambling Commission publishes its enforcement actions and licensed-operator register at gamblingcommission.gov.uk, including the public register of 268 currently licensed online operators. Germany's regulator publishes at gluecksspiel-behoerde.de. New Jersey's DGE publishes monthly market data at nj.gov/oag/ge. These are the primary documents this desk reads against any gambling-adjacent statute that lands on the public record.