When the Theory Does Not Fit: Prediction Markets and the Limits of Insider Trading Enforcement

Prediction markets have exploded in popularity over the last few years, with billions of dollars now wagered each month on future events ranging from sports and pop culture to federal elections. As the money has grown, so too has the temptation for people with “inside” knowledge to cash in. And this has captured the attention of federal regulators and prosecutors. The question, however, is whether the government’s existing insider-trading tools can even reach this new conduct. That uncertainty is precisely why companies should pay attention now, before an employee’s bet becomes a company’s problem.

This alert proceeds in four parts:

1. An explanation of prediction markets and why regulators are now closely watching.
2. A review of the recent wave of enforcement.
3. An overview of the insider trading framework, as currently applied.
4. An argument to why that framework likely does not fit prediction markets and the use of non-public corporate information. 

1. Prediction Markets Have Captured the Attention of Users and Regulators

Prediction markets are online platforms that allow participants to bet on the outcome of real-world events, such as election results, interest-rate decisions, sporting outcomes, or geopolitical developments. These “event contracts” often use a binary “yes/no” structure: either the event happened or it did not. A participant who correctly predicts the event outcome receives a fixed payout, usually reflecting the odds of the outcome, while the losing side receives nothing. In practice, the price of a “yes” contract works like a probability: A contract trading at 60 cents that pays $1 if the event occurs reflects a roughly 60% market-implied chance that it will.

While registered US exchanges have listed event contracts since 1992, the number and variety of these contracts increased sharply beginning in 2021, expanding well beyond traditional commodities and financial indicators. Platforms such as Kalshi and Polymarket have been designed with these non-traditional contracts in mind. Instead of trading on the prices of commodities, users can bet on who wins the presidential election, whether the United States goes to war, or who wins the Oscar for Best Actor. Only recently have several platforms registered with the Commodity Futures Trading Commission (CFTC), the federal agency that oversees futures and derivatives trading. The same growth that has fueled these new markets has also increased the opportunity for users to exploit nonpublic information to win big. And this has caught the attention of federal regulators and the US Department of Justice (DOJ).

The CFTC and DOJ have, in recent months, attempted to regulate prediction markets in ways that established securities markets have been historically regulated. In many respects, the markets are similar: prices respond to information, liquidity, and shifting public expectations, and the value of a position moves as traders reassess the probability of an event before it resolves. But the markets are quite different. Betting on the length of a White House press briefing is fundamentally different from trading corporate stock.

2. Corporate Information, Military Secrets, and State of the Union Attendance: Recent Enforcement Actions Have Focused on Headline-Grabbing Conduct 

Regardless of whether the analogy sticks, regulators and prosecutors have already begun to act. Three recent matters illustrate the range of conduct now under scrutiny. 

CFTC v. Spagnuolo

On May 27, the DOJ and the CFTC initiated parallel criminal and civil proceedings, filing criminal and civil complaints against a former Google software engineer, Michele Spagnuolo. The complaints alleged that Spagnuolo engaged in insider trading on Polymarket using the handle “AlphaRaccoon.” According to the complaints, Spagnuolo used confidential, nonpublic data concerning Google’s 2025 “Year in Search” rankings — accessed through an internal tool bearing a “Google Confidential” banner — to make roughly $1.2 million. Spagnuolo allegedly used the internal tool to trade event contracts relating to Google’s year-end search rankings, including contracts such as “Will Pope Leo XIV be the #1 searched person?” and “Will Donald Trump rank in the Top 5 most searched?”

United States v. Van Dyke

On April 23, the DOJ unsealed an indictment, alleging that Master Sergeant Gannon Ken Van Dyke used classified information to make more than $400,000 on Polymarket by betting on the timing and outcome of the operation that led to the capture of Nicolas Maduro. The indictment alleges Van Dyke signed nondisclosure agreements (NDAs) covering operational information, used a VPN to disguise his location, and later tried to delete his account as scrutiny grew.

George Santos and the State of the Union

Former US Congressman George Santos is reportedly under scrutiny in connection with possible insider trading on Kalshi. Leading up to the February State of the Union address, Kalshi users were placing bets on who would be attending the event. Shortly before the address, Santos tweeted that he would be attending, significantly increasing the odds that he would be appearing. According to reports, Santos allegedly bet that he would not appear. Santos, in fact, did not attend the speech. Kalshi reportedly detected that Santos traded against his own attendance, froze his account, and referred the matter to the CFTC and DOJ.

3. Insider Trading: An Established Framework in Securities Markets

Within the securities markets, insider trading occurs when someone buys or sells a corporation’s securities based on material, nonpublic information in violation of a duty of trust or confidence. “Material” information is information a reasonable investor would consider important in making an investment and “nonpublic” means it has not yet been released to the market. The textbook example is a corporate executive who learns of a coming merger, an unannounced product, or negative earnings, and then trades the company’s stock before the news goes public.

Critically, the law does not punish merely having and trading secret information. In Chiarella v. United States, 445 U.S. 222 (1980), the US Supreme Court rejected the idea that anyone who possesses material nonpublic information and trades is liable. Instead, the Court held that liability arises out of a duty to “disclose or abstain;” an insider must have an obligation to either disclose relevant confidential information to shareholders or to abstain from trading. The Court emphasized that neither Congress nor regulators have ever adopted a “parity of information” rule — the principle that all traders must have equal information. That single point — no duty, no liability — is the hinge on which the prediction-market problem turns.

The Court expanded on the “classical” theory of insider trading, established in Chiarella, in United States v. O’Hagan, 521 U.S. 642 (1997). Under the “classical” theory, a company insider (an officer, director, or “temporary insider” such as outside counsel) who trades the company’s stock breaches a duty owed to the company’s shareholders. In O’Hagan, the Court held that a corporate “outsider” can engage in insider trading by misappropriating confidential information in breach of a duty owed to the source of that information, and trading on that information. O’Hagan was a lawyer who obtained and traded information given to attorneys in his firm who were representing a client in a planned tender offer. O’Hagan did not have a fiduciary duty to the company whose stock he bought, but he did have a fiduciary duty to his law firm and its client. Two related decisions provided further guidance on liability for insider trading. In Dirks v. SEC, 463 U.S. 646 (1983), the Court held that, in order to establish liability for insider trading, a “tipper” who shares inside information must receive a personal benefit from the disclosure of that inside information. And in Salman v. United States, 580 U.S. 39 (2016), the Court held that providing inside information to a trading relative is a personal benefit which would establish liability for insider trading.

The defining feature of the insider trading framework is that the universe of potential insiders is limited and identifiable. The information relates to a corporate issuer; the duty arises from recognizable sources such as corporate fiduciary obligations, employment agreements, and confidentiality contracts; and the information flows through traceable channels. When suspicious trading appears before a corporate announcement, regulators have well-developed tools — US Securities and Exchange Commission (SEC) filings, trading records, and relationship mapping — to identify who had access to inside information.

4. But Does It Fit? The Insider Trading Framework Likely Does Not Apply to All Conduct

Recent enforcement actions have demonstrated that the federal government is attempting to police “insider trading” in prediction markets — largely through use of the misappropriation theory. But the misappropriation theory depends on a duty owed to the source of information. In the stock market, the underlying asset is a corporate issuer with defined relationships. In a prediction market, however, the underlying “asset” is a real-world event — a military operation, an election, a central-bank decision, a media company’s announcement. The people who possess nonpublic information about those events are often soldiers, campaign staffers, government employees, contractors, and their families, not people whose roles are defined by corporate law. As one commentator put it, the duty relationships that anchor the misappropriation theory become progressively harder to identify the further you move from formal corporate employment, and in many cases may not exist at all.

The recent cases illustrate this difficulty.

Some of these matters fit the existing framework comfortably because a clear duty exists. Spagnuolo easily fits within the misappropriation theory of liability. As a Google employee bound to keep the “Year in Search” data confidential, he had the precise duty the misappropriation theory requires, and the “Google Confidential” banner and access restrictions help prove it. 

Van Dyke is similar. The sergeant signed NDAs and had access to classified operational information; he was an identifiable official with a contractual obligation not to use the information for personal gain. The US Attorney for the Southern District of New York Jay Clayton described the breach of Van Dyke’s obligation and the personal benefit he derived from the breach: “the defendant allegedly violated the trust placed in him by the United States Government by using classified information about a sensitive military operation to place bets on the timing and outcome of that very operation, all to turn a profit.” In both cases, an identifiable employment, contractual, or official duty supplies the missing piece the misappropriation theory requires. 

The harder cases are those in which the trader’s informational advantage is real, but the duty is uncertain. As other practitioners have warned, those are the cases that matter and that will become “significant battlegrounds” for the developing enforcement landscape. Consider a campaign staffer who bets before internal polling is released. NPR has reported that such bets are commonplace; staffers compare a leaked internal poll to the market odds and place bets on prediction markets before the poll goes public. Whether that is unlawful may turn entirely on whether an employment agreement, campaign handbook, NDA, or platform user agreement created a duty not to use the information. As you move further away from a defined duty of confidentiality — a volunteer with no NDA, a congressional staffer who overhears a discussion between members of Congress, or a contractor’s spouse who learns something over dinner — the duty becomes harder to identify and harder to distinguish from a lawful analytical edge.

The Court’s rejection of a parity-of-information rule in Chiarella makes it more difficult to establish insider trading liability. The CFTC has also previously stated that “unlike securities markets, derivatives markets have long operated in a way that allows for market participants to trade on the basis of lawfully obtained material nonpublic information.” Derivatives markets depend on participants with superior, lawfully obtained information: the grain trader scouting crop conditions, the analyst reading satellite imagery, and the specialist monitoring shipping lanes. And in securities markets, the corporate-disclosure framework draws a workable line between lawful research and insider trading. But in prediction markets, no such line yet exists.

Returning to the Santos matter, if the “inside information” was simply Santos’ own knowledge of whether he intended to attend, it is not obvious that he misappropriated information from another source or breached a confidentiality duty to anyone. Public reports have not identified a specific confidentiality agreement, fiduciary duty, employment duty, or official-duty source he allegedly breached. Trading on one’s own intentions is not the same as stealing someone else’s secret. 

The Congressional Research Service notes that Kalshi’s rules bar trading by any “insider” with access to material nonpublic information about the underlying event, by anyone able to influence the subject of the contract, and by any decision maker or person with influence over the outcome. Because those prohibitions do not require breach of a duty, they appear to extend beyond the misappropriation theory — and whether the same conduct would also trigger liability under CFTC Rule 180.1, the Commission’s primary anti-fraud rule, remains unclear. In the absence of developed doctrine, platforms reportedly flag consistent winners — an outcome-based method borrowed from sportsbooks that cannot reliably distinguish a genuine insider from a skilled analyst or a lucky bettor.

The bottom line for companies is twofold. First, in clear-duty cases — an employee trading on confidential employer information — the legal exposure looks much like traditional insider trading and the government’s path is relatively smooth. Second, in the many cases where the duty is murky, federal liability is genuinely uncertain, but platform-level consequences are not, a trade may still trigger account freezes, a forced return of ill-gotten profits, and referral to regulators even where a federal case would be hard to prove.

Key Takeaways for Companies and Corporate Leaders 

• Review and clearly define corporate codes of conduct and confidentiality policies. The proliferation of prediction markets is a reminder for companies to update their codes of conduct and confidentiality policies, so that they can cover the unique aspects of the event contracts traded on prediction markets. Companies should also be alert to the danger that insiders might not merely trade on, but actively influence, or time, events to profit on prediction-market positions. Internal policies should ensure that such behavior is prohibited. 

• Treat platform rulebooks as the compliance baseline. Exchange rules (such as Kalshi’s) sweep more broadly than federal law and can trigger account freezes, disgorgement, suspension, and referral to the CFTC and DOJ even where liability under Rule 180.1 is uncertain. Companies should expressly bar employees from trading event contracts tied to the company’s information, products, or events, and write that prohibition into codes of conduct, trading policies, and confidentiality agreements.

• Monitor legislative and regulatory developments. Numerous bills in Congress would restrict prediction-market trading by federal officials and employees or ban certain categories of event contracts. The US Senate has already barred senators and their staff from trading on these markets, and the SEC has begun to signal interest alongside the CFTC. The boundaries of who regulates what and what conduct is prohibited are likely to keep shifting.

Additional research and writing from Fuyi Kuang, a 2026 summer associate in ArentFox Schiff’s San Francisco office and a law student at UC Berkeley.