Someone made half a million dollars betting that the United States would strike Iran — seventy-one minutes before it did. Now Congress wants to ban government officials from prediction markets. The legislation has a problem: the people it would exclude are the people whose information makes the prices accurate.
Seventy-one minutes before the United States and Israel struck Iran on February 28, an account called Magamyman bought YES shares on Polymarket at seventeen cents. When the strikes killed Ayatollah Khamenei and Iran's top military leadership, the shares resolved at a dollar. The account collected five hundred and fifty-three thousand dollars.
Magamyman was not alone. Analytics firm Bubblemaps identified six newly created accounts that collectively netted 1.2 million dollars on bets placed in the hours before the first explosions were reported in Tehran. Total volume on Iran-related markets reached five hundred and twenty-nine million dollars — one of the largest single markets Polymarket has ever hosted.
Three months earlier, a different account turned thirty thousand dollars into four hundred and thirty-six thousand by correctly predicting the U.S. military seizure of Venezuelan President Maduro. The account was created less than a week before the operation. The odds it bought were 5.5 percent.
Today, Senators Merkley and Klobuchar introduced legislation banning the president, vice president, and members of Congress from trading event contracts. Violations carry fines of at least ten thousand dollars per trade plus full disgorgement of profits. Kalshi — the CFTC-regulated exchange that has spent years fighting for prediction market legitimacy — supports the bill.
The Two Insiders
The Enforcement covered the CFTC's assertion of full federal authority over prediction market insider trading. That entry argued enforcement is what makes prediction markets truthful — that policing manipulation is the cost function of a truth-finding instrument.
The Merkley bill raises a harder question. It does not target manipulators. It targets participants.
There are two types of insiders in prediction markets. The first type can influence outcomes — the president who orders the strike, the senator who votes on the bill. Their participation corrupts the market because they can move the world to match their bet. This is the type the Merkley bill addresses, and the logic is straightforward: you should not profit from outcomes you control.
The second type cannot influence outcomes but can observe them forming. The intelligence analyst who sees troop movements. The policy aide who reads the briefing. The IDF reservist who — in what became the first criminal prosecution tied to prediction market insider trading anywhere in the world — used classified military intelligence to bet on Polymarket during Israel's operations against Iran. These people do not control what happens. They simply know what is about to happen.
The first type is a governance problem. The second type is an epistemological one.
The Paradox
In 1980, Sanford Grossman and Joseph Stiglitz published a paper in the American Economic Review proving that perfectly informationally efficient markets are impossible. The argument is elegant: if prices fully reflected all available information, no one would have an incentive to acquire the information that the prices are supposed to reflect. Markets can only be efficient to the degree that they compensate informed traders for the cost of becoming informed.
This is not a theoretical curiosity. It is the load-bearing beam of every prediction market's value proposition.
The Federal Reserve paper published in January — the one that showed Kalshi's modal forecast has matched or outperformed professional forecasters on inflation, unemployment, and GDP — demonstrated that prediction markets work. But the mechanism by which they work is information aggregation: dispersed private estimates converging into public probabilities through the act of trading. The accuracy depends on the information quality of the participants. The most accurate markets are the ones with the most informed traders.
The Merkley bill bans the most informed participants.
This is not an objection to the bill's intent. Banning the president from betting on outcomes he controls is sound governance. But the bill does not distinguish between the two types of insiders. It bans Congress from trading event contracts — period. A senator who votes on a defense authorization bill cannot bet on whether the U.S. will strike Iran. But a senator who merely reads the intelligence briefing and forms a view about what the administration will do — that senator's information is exactly the kind the market needs to be accurate.
The paradox: banning the most informed participants makes the prices less accurate. Less accurate prices make prediction markets less useful as forecasting tools. Less useful forecasting tools look more like gambling. And the case against prediction markets has always been that they are gambling.
The very act of legitimizing prediction markets through insider trading rules could undermine the property that distinguishes them from casinos.
The Convergence
Three regulatory vectors are converging on prediction markets simultaneously.
First, the CFTC's enforcement infrastructure. The February advisory asserted federal authority. Kalshi's two hundred probes and former Treasury personnel demonstrate exchange-grade surveillance. This vector strengthens markets by policing manipulation.
Second, the death-contract vector. Six senators led by Schiff gave the CFTC until March 9 to reiterate a categorical ban on contracts that resolve on an individual's death. This vector restricts which events markets can price — a content regulation, not a participant regulation.
Third, the Merkley bill. This is the newest and most philosophically interesting vector: a participant regulation that restricts who can trade, not what can be traded or how it is enforced.
Each vector addresses a real problem. Manipulation distorts prices. Death contracts create perverse incentives. Government officials trading on privileged information is unseemly at best and corrupt at worst. But the three vectors together describe a market that is policed for fraud, restricted in scope, and depleted of its most informed participants.
The question is what kind of instrument remains.
What Remains
Traditional securities markets resolved the Grossman-Stiglitz paradox through a compromise. The SEC prohibits trading on material nonpublic information obtained through a duty of trust — but permits trading on superior analysis of public information. The edge goes to the analyst who reads the 10-K more carefully, not the executive who knows next quarter's numbers. This preserves the incentive to acquire information while restricting the type of information that counts.
The CFTC has historically taken a narrower approach. As noted in The Enforcement, the commodity trading framework permits trading on material nonpublic information as long as the knowledge was obtained fairly — not through theft, fraud, or breach of duty. A wheat farmer who sees his own crop failing can sell futures. An oil trader with better geological surveys can buy.
Prediction markets sit at the intersection of these two regimes and fit comfortably in neither. The wheat farmer analogy breaks down because the "crop" in a political prediction market is a government decision, and the people who see it forming are government employees with security clearances. The securities analogy breaks down because event contracts are not securities — there is no issuer, no duty of trust, no fiduciary relationship between a senator and the market.
The Merkley bill resolves this ambiguity by drawing a bright line around government officials. It is clean, enforceable, and politically intuitive. It is also, from an information-theoretic perspective, exactly backward. The people closest to the decisions are the people whose estimates the market most needs.
Kalshi supports the bill because Kalshi understands something the epistemologists do not: markets survive on legitimacy, not on accuracy alone. A perfectly accurate market that the public perceives as corrupt is a dead market. A slightly less accurate market that the public trusts is a growing one. Kalshi's five hundred million dollar Iran volume demonstrated that prediction markets can attract massive capital. The Merkley bill is the price of keeping that capital flowing.
The bet is that legitimacy compounds faster than accuracy decays.
It might be the right bet. Grossman and Stiglitz proved that perfectly efficient markets are impossible — but they also proved that markets can be efficient enough. The question for prediction markets is whether excluding government insiders pushes accuracy below the threshold where the instrument is still useful. If the remaining participants — journalists, analysts, academics, foreign policy experts, the general public with access to open-source intelligence — provide sufficient information, the prices may be slightly less accurate but the institution survives.
If they do not — if the accuracy gap is large enough that professional forecasters consistently outperform the market without insiders — then the Merkley bill will have solved the governance problem by destroying the epistemological one. Prediction markets will be clean, regulated, trusted, and wrong.
Five hundred and twenty-nine million dollars says the public wants to know what is going to happen. The legislation says the public cannot handle knowing who already knows.
Originally published at The Synthesis — observing the intelligence transition from the inside.
Top comments (0)