A U.S. servicemember used classified information about the planned capture of Venezuelan President Nicolás Maduro to make more than $400,000 trading on Polymarket. A Google engineer with the online moniker “AlphaRaccoon” converted confidential company information about search traffic into more than $1 million in prediction market profits. In France, authorities are investigating whether someone used a hair dryer to manipulate an airport temperature sensor and then cashed in on weather contracts.

It is easy to look at stories like these and conclude that popular prediction markets like Polymarket and Kalshi have an insider trading problem—and that lawmakers need to do something about it.

But consider another trader. Before Super Bowl LX, a 21-year-old TikToker flew from Cincinnati to San Francisco, spent hours listening to national anthem rehearsals outside the stadium, timed them with a stopwatch, and then reportedly made more than $50,000 trading on the length of the anthem at halftime.

He knew something most other traders did not. And that is precisely the kind of informational advantage prediction markets are designed to reward.

In a new article, we argue that this distinction between the source and quality of information should anchor the debate over regulating prediction markets. There is no doubt these markets need protection from manipulation and the misuse of confidential information. But policymakers and regulators should be careful not to turn “insider trading” into an overly broad label for anyone who turns knowledge others don’t have into profit. Doing so could undermine the information-producing function that makes these markets uniquely useful. The situation creates a bit of a regulatory paradox: Some forms of informational advantage can corrupt a prediction market, while others are the reason to have the market at all.

What’s in a Prediction Market

Prediction markets have existed for decades, though until recently they were largely confined to academic experiments and specialized uses. One early example was the University of Iowa’s election market, launched in 1988 with fewer than 200 traders. Despite its modest size, the market predicted that year’s presidential popular vote within 0.2 percentage points of the actual result. In the decades that followed, academics, corporations, and government agencies experimented with prediction markets as tools for forecasting elections, business outcomes, economic conditions, and even geopolitical events.

What has changed is scale and access. Platforms such as Kalshi and Polymarket have transformed what was once a niche forecasting mechanism into a retail product. The underlying premise, however, remains the same—that markets can aggregate dispersed information and use the wisdom of crowds to make useful predictions that beat experts deciding alone.

Prediction markets work as follows. In a typical market, contracts trade between $.01 and $1.00, with the price roughly reflecting the market’s estimate of the probability that an event will occur. A recent Kalshi contract, for example, asked whether there would be more tech layoffs in 2026 than in 2025. A “yes” contract trading at $.91, effectively signaled a roughly 91 percent market-implied probability while offering a $.09 profit if the prediction proved correct. As traders bring new information to the market, those prices move.

The magic comes from putting money behind those predictions, giving participants incentives to find and reveal information. The better the information entering the market, the more useful the resulting forecast can be—a forecast that’s available to and benefits everyone, not just the winning trader.

The Law Is Less Empty Than It May Seem

With the risks and benefits of prediction markets in mind, we conducted a comprehensive survey of existing public and private law tools to prohibit improper trading on prediction markets. What we found helps advance the discussion of regulating these markets in a more nuanced manner.

Our first finding is that prediction markets are not the legal “Wild West” they are sometimes made out to be. The familiar rules of securities insider trading do not map neatly onto prediction contracts because most of these contracts are not securities. But that does not mean nothing bars suspicious trading.

The Commodity Futures Trading Commission (CFTC) has taken the position that §6(c)(1) of the Commodity Exchange Act and CFTC Rule 180.1 prohibits trading based on misappropriated material nonpublic information. That connection is doctrinally important. Rule 180.1 was expressly modeled on §10(b) of the Securities Exchange Act and SEC Rule 10b-5, but the relevant analogue is not the “classical” theory of insider trading, under which a corporate insider owes duties to the corporation and its shareholders. Prediction markets ordinarily have no corporation, shareholder, or comparable fiduciary relationship at the center of the trade.

Instead, the better fit is the misappropriation theory recognized by the Supreme Court in United States v. O’Hagan. Under that theory, the deception occurs when a person entrusted with confidential information secretly uses it, in breach of a duty owed to the source, for an unauthorized trading purpose. The CFTC has imported that anti-fraud principle into the commodities context: The problem is not simply possessing material nonpublic information, but obtaining or using it in violation of a duty of trust or confidence and then trading on it.

For example, a military officer who uses classified operational plans to trade has not merely done good research. Nor has an employee who monetizes confidential information about an unreleased product, a campaign aide who trades on private strategy, or a team trainer who bets on a player’s undisclosed injury. In each case, the problem is not superior knowledge; it’s the misuse of information obtained through a relationship of trust.

Second, existing law also reaches a different category of trading: manipulative trades. If a trader takes a position and then deliberately rigs the event that determines whether the contract pays, the trader is not improving the market’s forecast. The trader is corrupting the “ground truth” the market is supposed to predict.

The French hair dryer episode captures the problem. If someone trades on the temperature in Paris and then artificially heats the device that determines the official reading, that is not an informational edge. The same is true of an athlete who trades on a proposition tied to his own performance and then intentionally alters that performance to make the contract pay. Existing market manipulation and fraud laws already provide substantial tools for addressing conduct of this kind.

Public law is also only part of the picture. Prediction platforms can use their terms-of-service agreements to prohibit categories of trading, identify restricted participants, impose “know your customer” and surveillance requirements, and suspend or ban traders who violate those rules. Those contractual restrictions can go beyond what public law independently prohibits and can also help surface suspicious conduct for regulators.

Employers have a parallel role. Employment agreements, confidentiality provisions, and internal policies can identify categories of nonpublic information that employees may not use for prediction market trading and make clear when an employee owes a duty not to monetize information. Properly drafted, these private law rules reinforce the same distinction that should guide public enforcement: Genuinely confidential information can be protected without converting every employee’s knowledge or expertise into non-tradable information.

A difficult question remains, however. What should happen when a trader has unusually good information but neither stole it nor breached a duty to obtain it?

Regulate the Source of the Advantage

Here, it’s important to note that prediction markets differ importantly from securities markets. Capital markets serve a crucial function by allocating capital and facilitating investment, making broad participation and investor confidence important regulatory objectives. Prediction markets principally produce information. Their social value lies in generating accurate forecasts.

That means regulation should be especially cautious about suppressing what we call “high-information trading.”

A researcher who develops a superior forecasting model, a citizen who notices facts others have overlooked, or an expert who draws a novel inference from lawful sources should be allowed to trade on that knowledge. So should the TikTok creator who spent hours timing Super Bowl rehearsals. He did not steal information, violate a confidentiality obligation, or deceive anyone. He simply worked harder to find the answer.

Prohibiting such trading because other participants were less informed would amount to a general rule against informational advantage. That would be a strange rule for a market whose principal purpose is to aggregate information.

The most difficult cases arise when a trader is both unusually well-informed and capable of influencing the event itself. The allegations surrounding former Congressman George Santos illustrate this boundary problem. Santos reportedly traded on whether he would attend the State of the Union. No one is better positioned to know Santos’ plans than himself. His informational advantage, standing alone, might make the market more accurate.

But if a trader first takes a position and then acts to change the event—or spreads misleading information designed to move the market—that begins to look very different. The key question is not simply, “Was this trader an insider?” Rather, it’s, “How did the trader acquire the information, and did the trader manipulate the outcome or misuse information entrusted to him?”

All this leads, we argue, to a relatively simple regulatory framework.

Market manipulation—changing the underlying event to make a contract pay—should be aggressively prohibited. Trading on misappropriated confidential information should be prohibited as well, because it converts duties owed to employers, governments, customers, and other information sources into private profits. But lawfully acquired informational advantages should generally remain tradable, even when those advantages are substantial.

The broader point is that policymakers should seek optimal, not maximal, restrictions on information in prediction markets. There are already ample public and private law tools available to police the most troubling conduct. The challenge is to use them without flattening meaningful distinctions among manipulation, misappropriation, and superior information.

Todd Haugh is a professor, and John Holden and Matthew Turk are associate professors, of business law at Indiana University’s Kelley School of Business. This post is based on their recent article, “Optimal Insider Trading in Prediction Markets,” available here.