From Iran to Taylor Swift: Informed Trading in Prediction Markets

The New Frontier of Insider Trading

In the hours before the February 28, 2026 U.S.-Israeli strike on Iran—one of the most closely guarded military operations in recent history—six newly created Polymarket wallets collectively earned approximately $1.2 million by purchasing ‘Yes’ shares in the ‘US strikes Iran by February 28?’ contract at prices as low as $0.10. One account, operating under the handle ‘Magamyman,’ placed its first trade seventy-one minutes before the news broke, when markets implied only a 17% probability of a strike. When those markets resolved in the affirmative, the account’s profits totalled approximately $553,000.

This episode is striking because it is hardly the first time this has occurred. Two months earlier, a pseudonymous Polymarket account called ‘Burdensome-Mix’ earned roughly $485,000 from a $38,500 investment in contracts tied to the capture of Venezuelan President Nicolás Maduro—placing its largest trades just hours before a covert military operation was publicly announced. Israeli authorities separately indicted a civilian and an IDF reservist for allegedly using classified wartime information to profit on Polymarket. A trader earned over $1 million by predicting with uncanny precision the results of Google’s proprietary Year in Search rankings. Another appeared to have advance knowledge of OpenAI’s browser launch. And a user named ‘romanticpaul’ purchased Taylor Swift engagement contracts aggressively in the days before Swift publicly announced her engagement to Travis Kelce.

These cases are not merely colorful anecdotes. They represent a systematic challenge to the legal and regulatory frameworks that govern the use of inside information in connection with trading in traditional instruments like stocks, bonds and futures. Our paper, From Iran to Taylor Swift: Informed Trading in Prediction Markets, presents the first systematic empirical and legal study of this phenomenon.

A Systematic Analysis of Insider Trading in Prediction Markets

To examine whether there is a broader pattern, we developed a systematic statistical screening of all Polymarket markets from February 2024 through February 2026—a universe of over 93,000 distinct markets and nearly 50,000 unique wallet addresses.  The screening combines five signals—cross-sectional bet size, within-trader bet size, profitability, pre-event timing, and directional concentration—into a composite score measuring the probability that a given (wallet, market) pair reflects informed trading rather than lucky speculation. The unit of analysis is the wallet-market pair rather than the wallet itself, as an insider may have advance knowledge about one event but not others.

We find that across 210,718 suspicious wallet-market pairs, flagged traders achieved a 69.9% win rate—a result that exceeds the null distribution of random chance by more than 60 standard deviations under a permutation test. We estimate approximately $143 million in aggregate anomalous profit across the study period. This figure is almost certainly a lower bound: the screening is buy-side only, excludes positions below $500, and cannot detect traders who deliberately limit their bets to avoid detection.

To be sure, we cannot observe the full portfolio of any given trader, meaning that a profitable trade may have been part of a strategy with other unprofitable legs or a hedge against another position. And in any study of trading data lacking direct evidence of the information available to a given trader, we cannot rule out the possibility that we are identifying unlikely coincidental patterns rather than genuine exploitation of material nonpublic information. We believe the patterns are nonetheless sufficiently unusual to warrant serious scrutiny.

Why Existing Law Falls Short

Our paper’s legal analysis identifies three structural reasons why existing law has failed to address informed trading in prediction markets.

First, the two principal theories of insider trading liability under U.S. securities law—the classical theory, rooted in the fiduciary duty a corporate insider owes shareholders, and the misappropriation theory, recognized in United States v. O’Hagan (1997)—both require that the relevant trading involves securities. Most prediction market contracts whose payoffs depend on geopolitical or macroeconomic events are unlikely to qualify as securities and may instead be commodities. Securities law’s anti-fraud provisions do not apply to commodity contracts.

Second, the CFTC’s principal anti-fraud vehicle for commodity derivatives—Rule 180.1, modelled on SEC Rule 10b-5—is narrower than its securities law analogue in certain respects. For one, it is unclear that Rule 180.1 yields a Cady Roberts-style duty to disclose, and trading on ‘lawfully obtained’ commercial information, without deception or fraud, remains legal. The CFTC has never applied Rule 180.1 to prediction markets in a contested enforcement action, though it issued an advisory in February 2026 confirming its authority to do so and announced its first prediction market enforcement actions—modest fines against a MrBeast video editor and a political candidate who traded on platform insider information.

Third, federal wire fraud statutes—which the Second Circuit considered in the context of NFT insider trading in United States v. Chastain (2025)—faces certain difficulties in the prediction market setting. In particular, wire fraud requires that the exploited information had commercial value to the party from whom it was misappropriated. For information concerning military operations, geopolitical developments, or personal social knowledge, that commercial value may be minimal or non-existent for the original source, even if the information is enormously valuable to a prediction market trader.

The regulatory picture is further complicated by a structural asymmetry between platforms. Kalshi operates as a CFTC-designated contract market, subjecting its contracts to anti-fraud and surveillance obligations. Polymarket, which has hosted the majority of documented suspicious trading, operates on a decentralized blockchain infrastructure and exists in a legal gray area. This asymmetry creates a perverse incentive: informed traders who wish to exploit material non-public information face fewer legal constraints by trading on Polymarket than on Kalshi, precisely because Polymarket’s regulatory status is unresolved.

Legislative Responses and Their Limits

Given the abundance of insider trading in prediction markets, lawmakers have begun to consider whether legislative intervention is warranted.  Representative Ritchie Torres introduced the Public Integrity in Financial Prediction Markets Act of 2026 (H.R. 7004), which prohibits federal elected officials, political appointees, and congressional staff from trading on prediction market contracts where they possess material non-public information.

But H.R. 7004 illustrates the limitations of contract-level regulation. It is unclear that the bill would have deterred the cases we document. The Maduro and Iran strike trades appeared to involved non-public information possessed by military and national security personnel who are already covered by separate statutes—or, in the case of the Iranian operation, possibly foreign nationals. The IDF case involved a non-U.S. (Israeli) reservist. The OpenAI browser case involved corporate trade secrets. And the Taylor Swift case involved personal social knowledge entirely outside government information channels. It is unclear whether H.R. 7004 would have any measurable effect on insider trading in prediction markets.

A Regulatory Response

We identify three possible elements to a regulatory response.  The first is platform-level regulation: mandatory registration, surveillance, and reporting obligations for any prediction market operator that offers contracts to U.S. persons, regardless of where the operator is incorporated or on what technology it operates. Once a platform is regulated, it is straightforward to apply anti-fraud rules without requiring advance identification of which specific contracts create insider trading risks. The principal disadvantage is that platform regulation reaches only operators who voluntarily submit to regulation or can be reached through CFTC extraterritorial jurisdiction. Polymarket’s planned U.S. re-entry would, if consummated under a DCM framework, resolve this gap for the platform that has hosted most documented suspicious trading.

The second is contract-level regulation: targeted rules for specific categories of event contract identified as high-risk for insider trading, particularly contracts tied to government data releases, military and diplomatic operations, and corporate material events where the insider population is identifiable and the informational asymmetry is most severe. Contract-level regulation offers precision that platform-level regulation lacks. But it faces an identification problem: the universe of potential event contracts is effectively unlimited, and sophisticated actors will route their trades through uncovered contract types.

Finally, we consider a liability theory directed at informed traders themselves on decentralized platforms that resist operator-level regulation. We argue for extending the misappropriation doctrine’s breach-of-duty framework, through CFTC rulemaking, to cover the non-securities duties of confidentiality that government employees, military personnel, and corporate insiders owe to their respective principals. Under this extended theory, a soldier who uses classified operational plans to trade on prediction markets breaches the duty of confidentiality owed to the government, even if the resulting contracts are not securities. The IDF prosecution in Israel approximates this theory under Israeli law; we argue that Rule 180.1 can be interpreted to reach the same conduct under U.S. law, with clarifying rulemaking to confirm its scope.

Broader Implications

Our paper joins a literature that has long recognized prediction markets as valuable instruments for aggregating dispersed information. The foundational Hayekian insight—that decentralized markets harness private knowledge more efficiently than any centralized mechanism—has been empirically validated in contexts ranging from presidential elections to economic forecasting. A February 2026 Federal Reserve Board study finds that Kalshi’s macroeconomic prediction markets achieve accuracy on CPI and GDP releases that rivals or exceeds professional forecasts, with rich intraday dynamics that daily data entirely miss.

But there is a tension at the heart of the prediction market enterprise that this literature has not adequately addressed. The same features that make prediction markets epistemically powerful—their incentive to reward private information with financial returns—also make them uniquely attractive venues for the exploitation of material non-public information. When the information in question concerns classified military operations, the profits come not merely from superior analysis but from theft of government secrets. When it concerns corporate product launches, the profits represent misappropriated trade secrets. When it concerns personal social knowledge, the profits represent monetization of relationships and confidences.

The informed trading we document is not, in the aggregate, trivial. $143 million is a conservative lower-bound estimate of anomalous profits extracted over two years from a single platform. These profits represent transfers from uninformed retail participants to those with access to material non-public information—a regressive outcome that undermines the democratic appeal of prediction markets as venues where ordinary forecasters can profit from accurate beliefs. Whether this is normatively desirable depends on one’s views as to whether the Hayekian benefits of centralized pricing justify the social costs (if any) of gambling of this sort.

The regulatory and enforcement community is beginning to respond. The CFTC’s February 2026 advisory, U.S. Attorney Clayton’s February 2026 speech emphasizing that antifraud law applies to prediction markets, the Israeli indictment in the IDF case, and Representative Torres’s legislation all reflect a recognition that prediction markets have outpaced the legal frameworks designed to govern them. Our paper aims to provide the empirical grounding and legal analysis necessary to close that gap.