The (Poly)Market Nobody Knows How to Govern
Prediction markets or gambling websites? Inside the regulatory wild west
A platform whose CEO calls it “the future of news” opened a market on the question of nuclear detonation, let nearly a million dollars in bets accumulate, and then quietly archived it. They didn’t just close it, they archived it so no record of the bet remains.
The platform is Polymarket. Its closest competitor, Kalshi, operates on the same basic premise: users bet real money on real-world outcomes, from sports results to economic indicators to elections. Together, the two platforms have become the dominant players in an industry that didn’t meaningfully exist for ordinary Americans five years ago.
Increasingly, these prediction sites seem to be everywhere (including on Substack). So who is regulating them? In theory, the answer is the Commodity Futures Trading Commission (CFTC). It is also, at the moment, a commission-in-transition with only one member. While a fully functioning CFTC normally has five commissioners, the agency asserting exclusive jurisdiction over a multi-billion dollar industry is, currently, a single person.
Polymarket launched in June 2020. Less than two years later, the CFTC fined it for operating without CFTC registration and blocked it from operating in the United States. The platform spent the next three years operating offshore, officially closed to American users. Then, in October 2024, a federal court ruled against the CFTC in Kalshi v. CFTC, finding that the agency had improperly blocked election betting contracts. Polymarket moved quickly and in July 2025 it acquired QCX, a CFTC-registered derivatives exchange, accelerating its return to the U.S. market by leveraging an existing registration, with official re-entry secured by late 2025. A U.S. app soft-launched via waitlist in December 2025.
Around the same time, Donald Trump Jr. joined its advisory board. (He is also a strategic advisor with Kalshi.) Under such circumstances, few believe the new entities will be subject to close regulatory attention from Trump’s government, creating an incentive for states to jump in.
The logic behind prediction markets is unlike a poll or a pundit. Instead, the bettor has to publicly put money on the line about their expectations and will lose hard cash for being wrong. This should incentivize accuracy, but only if one assumes the money on the line reflects genuine belief rather than privileged access.
Evidence that it may reward privileged access to information is accumulating.
On Polymarket, eight recently created accounts, all made around March 21, collectively wagered close to $70,000 on a ceasefire occurring. If a deal is struck before March 31, those accounts could walk away with over $800,000 in winnings. A separate account, opened just before the US attacked Iran on February 28, placed a successful bet predicting those strikes as well. That same account has made no other wagers.
Kalshi also wants you to know it won’t engage in disquieting, illegal or corrupt behavior because of federal regulation!
But death markets are occurring, and become all the more likely in times of conflict.
One concern is whether bets involving the use of military force might encourage bets by well-informed insiders. A user pocketed over $400,000 on bets placed before Nicolás Maduro was taken from Venezuela, another wagered on Ayatollah Khamenei’s removal from power before his death in Tehran. Here, it appears the market did not aggregate crowd wisdom. Rather, it rewarded someone who appears to have known something the crowd didn’t.
Legislators have been approaching the problem from two directions simultaneously. The End Prediction Market Corruption Act, introduced in early March, targets what can be traded. It would ban prediction market contracts on sports events entirely, essentially siding with the states that have argued these are unlicensed gambling operations. The PREDICT Act, announced only a couple weeks later, by a bipartisan group of lawmakers, targets who can trade: it would prohibit members of Congress, the president, vice president, and others in the executive branch from participating in prediction markets involving political events.
In response to these government actions, both Kalshi and Polymarket announced new rules aimed at banning insider trading on their platforms. For instance, Kalshi announced that regulations would include banning political candidates from trading on their own campaigns and prohibiting athletes from trading contracts related to their sport.
But the category of people with relevant non-public information extends well beyond elected officials and professional athletes. Coaches, military officers, intelligence analysts, diplomatic staff, campaign workers, bankers, regulatory officials – anyone with advance knowledge of the outcomes being traded on is a potential insider. How either platform intends to identify, monitor, or enforce rules against that universe of people remains unclear.
The question of who shouldn’t be trading on these platforms is easier to answer than the question of whether anyone can trust what the platforms are actually showing. When prediction markets are cited as evidence of what the public believes, and the odds are being moved by actors with non-public information, the “future of news” starts to look less like a forecasting tool and more like an information asymmetry problem with a financial instrument attached. For example, Polymarket calls itself “News 2.0” despite pumping out misleading content.
Before stepping down in early 2025, outgoing CFTC chairman Rostin Behnam declared the agency was well positioned to be the regulator for “digital commodity assets.”
But is it?
The problems outlined above, insider trading, manipulation, and macabre markets, are not simply enforcement failures. They are symptoms of a federalism problem: a new industry that arrived faster than the laws written to govern it is now caught between a federal regulator claiming exclusive authority and states insisting it is theirs to police. Neither framework was designed with this in mind.
The Lineage of Prediction Markets
In a recent Wall Street Journal piece, current CFTC Chairman Michael Selig wrote:
The Commodity Futures Trading Commission for decades has overseen regulation of prediction markets — or event contracts, as we refer to them — that help market participants hedge risk, aggregate information and test hypotheses about future outcomes.
Prediction markets have existed since the late 1980s. The Iowa Electronic Markets (IEM) began in 1988. It was founded by three University of Iowa professors who wanted to know if they could predict polling data better than pollsters. Known as the "Iowa Political Stock Market," it was originally created for the 1988 Presidential election. From the start, the platform ran into legal friction with gambling laws and securities regulations prompting it to restrict participation to the University of Iowa community.
Ordinarily, futures contracts fall under CFTC jurisdiction via the Commodity Exchange Act (CEA). The IEM was no exception in theory. In practice, the CFTC chose not to regulate it due to its academic focus and the small scope. Instead, it issued a series of no-action letters, a formal mechanism by which the agency signals it will not pursue enforcement action against a particular entity.
In 1992 and 1993, those letters permitted the IEM to expand beyond the University of Iowa to other academic institutions, but under strict conditions: participation was capped at 2,000 traders per market, with a maximum investment of $500 per participant. The CFTC wasn’t overseeing the IEM. It was formally looking the other way, on the condition that the platform stayed small and academic.
It still operates today, still capped at $500 per participant, still run as a nonprofit academic research project by the University of Iowa, and still operating under a CFTC no-action letter rather than normal registration. Its director, University of Iowa professor Thomas Gruca, describes its purpose in terms that would be unrecognizable to Polymarket’s investors:
The purpose is not to make money, we’re nonprofit, we’re even non-revenue. Our purpose is to see if different groups of people can put their heads together and solve really hard forecasting problems.
The professors affiliated with the IEM are aware that prediction markets like Kalshi and Polymarket use them as a justification for existing. Finance professor and Iowa Electronic Markets board member Thomas Rietz has noted that while Kalshi and Polymarket prioritize opening many markets to increase site traffic and betting for revenue, “The difference is we are designing our markets and our contracts specifically to make the most accurate forecast possible.”
In March 2026, Yahoo Financereported that Polymarket recorded a single-day notional trading volume of $478 million, with their politics category accounting for over $200 million of that activity. The IEM has no equivalent figure because there is no equivalent: a $500-capped nonprofit research tool operating under a regulatory carve-out and a billion-dollar commercial platform are not the same thing and were never designed to be governed the same way.
When Michael Selig says the CFTC has overseen prediction markets “for decades,” he is technically correct about the genealogy of these platforms. But he is describing something radically different in scale, scope, and purpose. The original version was a university research project explicitly exempted from normal regulation because it was small and academic. What exists now is a multi-billion dollar commercial platform trading on nuclear war, assassinations, and elections.
The Category Problem: States vs the CFTC
One of the fundamental problems with regulating something like Polymarket is that there is no agreement on what it actually is. The answer determines everything: who has jurisdiction, what rules apply, how users are protected, and whether the platform should exist in its current form at all.
Platforms like Polymarket brand themselves as a kind of digital commodity market offering contracts not gambling. Michael Selig says that prediction markets function the same way as other futures contracts with standardized derivatives obligating parties to exchange assets at prices set on future dates. Hedging against something like weather, the argument goes, is not betting against the house in the way that sportsbook companies do.
The CFTC’s legal authority here is genuinely broad. As per the CFTC:
Under the plain language of the Commodity Exchange Act, event contracts are “swaps.” They are derivative instruments that allow two parties to speculate on future market conditions without owning the underlying asset. In the wake of the 2008 financial crisis, Congress expressly granted the CFTC comprehensive authority over any such contract based on a commodity. The statutory definition of “commodity” is extraordinarily broad and includes practically all goods, articles, services, rights and interests except for onions (due to a history of market manipulation) and movie box-office receipts (because of Hollywood lobbying).
But states argue that prediction markets are a loophole for state gambling regulations, particularly on sports. Several states including Tennessee, New York, New Jersey, Nevada, and Massachusetts have taken Polymarket and Kalshi to court, arguing that event contracts on sports are effectively gambling and are therefore subject to state jurisdiction and taxed differently than financial markets. User eligibility requirements differ as well: prediction markets generally allow users 18 and older, while state gambling is restricted to those 21 and older.
Massachusetts won an injunction against Kalshi, temporarily banning the firm from operating sports-related contracts in the state. In response, Kalshi argued that the state was “trying to block [its] innovations by relying on outdated laws and ideas.” Polymarket filed its own lawsuit against the state, seeking to avoid what it called “imminent and irreparable harm” from the enforcement of state gambling laws against a federally regulated derivatives exchange.
From the CFTC’s perspective, this is not a gambling question. It is a jurisdictional one that states have no authority to be involved in. It has referred to state governments as “overzealous” in their attempts to “undermine the agency’s exclusive jurisdiction over these markets.” The agency has pushed this position in court, filing an amicus brief asserting exclusive jurisdiction over event contract markets. Multiple rulings on the issue have been split, with a federal court in Tennessee leaning toward CFTC jurisdiction while courts in Nevada side with the state.
But the jurisdictional fight obscures a deeper problem that neither side is asking.
In a standard commodity futures market, the underlying asset, such as corn, oil, an interest rate, exists independently of the contract written against it. Political prediction markets are different. When enough money moves on an election outcome, the market doesn’t just reflect political reality, it potentially shapes and reshapes it. Campaigns track them. Donors read them as signals. News outlets and pundits refer to prediction markets as reasonable measures of public opinion and political expectations. That feedback loop is not something the Commodity Exchange Act was written to address. The question of whether a Polymarket or Kalshi-type prediction market should exist at all is still a question to be reckoned with.
Let’s say the answer is yes – how can one be sure they are operating fairly?
These platforms rest on a core premise: that markets aggregate real information through real financial stakes. However, there’s evidence that challenges this.
Researchers at Columbia University found that roughly 25% of all trading on Polymarket over the past three years consisted of “wash trading.” This is an illegal form of market manipulation where users rapidly buy and sell the same contracts to artificially inflate activity. The figure varied by category: 17% of election-related trading, 45% of sports-related trading. It was made possible by Polymarket’s decision not to charge transaction fees, its use of pseudonymous crypto wallets, and its operation on a public blockchain. These are structural choices that, intentionally or not, created the conditions for manipulation at scale. Kalshi does not operate on a blockchain, so its trading data is not available to outside researchers. Whether the same problem exists there is unknown.
Regulatory Momentum, Sort Of
In early February 2026, the CFTC assembled an Innovation Advisory Committee, describing its goal as ensuring “the CFTC’s decisions reflect market realities so the agency can future-proof its markets and develop clear rules of the road for the Golden Age of American Financial Markets.” Alongside executives from traditional financial institutions like Nasdaq, the committee includes executives from FanDuel, DraftKings, and Gemini, as well as executives from Polymarket and Kalshi.
The platforms the committee is meant to help regulate are represented on the committee itself.
On February 23rd, six Senate Democrats issued a letter to Michael Selig urging the CFTC to prohibit prediction contracts that correlate with or incentivize geopolitical outcomes including assassination, death, terrorism, and war. Two days later, the CFTC’s Enforcement Division issued an advisory emphasizing that all trading practices prohibited under the Commodity Exchange Act apply equally to prediction market contracts. On March 9th, Selig announced that he had instructed CFTC staff to prepare an advance notice of proposed rulemaking on prediction markets.
These actions have the appearance of regulatory momentum. But it is worth being precise about what each one does.
The Senate letter asks a narrow question about the most obviously objectionable markets. The enforcement advisory reminds the industry that existing fraud rules apply, which was already true. The advance notice of proposed rulemaking is the earliest and most tentative step in a regulatory process that typically takes years. Meanwhile, the CFTC’s most concrete action to date has been battling in several court cases not to scrutinize the platforms but to ensure that the states trying to constrain them cannot.
That is not incidental. The rulemaking process is not simply reactive enforcement. It is an attempt to lock in federal jurisdiction over prediction markets by defining them as federally regulated financial instruments, with the downstream effect of foreclosing state regulatory claims entirely.
Neither the Senate letter nor the rulemaking process is asking the question this piece has been raising: what does it mean for democratic life that a private platform, backed by venture capital and advised by political insiders, has become the infrastructure through which Americans form probabilistic beliefs about their own political future?
The institution claiming sole authority to govern it has, so far, spent most of its energy making sure no one else can.
 | A guest post by
|