Author: nic_carter
Translation:白话区块链
Just two days ago, the biggest scandal broke: the U.S. Department of Justice indicted Army Staff Sergeant Gannon Ken Van Dyke, accusing him of using classified information about the Maduro raid to make illegal trades. He traded on Polymarket before the mission and made $400,000. He wasn't a junior soldier, but a seasoned Green Beret special forces member with significant planning and operational responsibilities in special operations missions.
Incidentally, although many have called for leniency, citing rampant (legal) insider trading in Congress, he still deserves prison time. What he did risked leaking the imminent raid to Venezuelans through his trading activity, which is both morally and legally problematic. While the Venezuelans didn't seem to notice, the government cannot accept a precedent where elite combat operators leak details of upcoming operations through market activity. I sympathize with Van Dyke, but he broke the law and violated his oath of secrecy for personal gain.
This is just the latest in a series of real or suspected insider trading scandals on prediction markets. Previously, Israel arrested two reservists for allegedly using classified military intelligence to trade on Polymarket. Trades on markets for the start of a war with Iran, a ceasefire, the assassination of Khamenei, and a Biden pardon have also been questioned, though no one has been arrested yet. Kalshi and Polymarket have also flagged and frozen accounts trading on markets about themselves, such as three congressmen running for reelection betting on their own race.
Now you might think these problems would go away as more people realize that trading on confidential information in prediction markets (not just securities markets) is equally illegal. But I think the root cause runs much deeper.
The premise of prediction markets is this: they are informationally efficient because they reward informed insiders.
In other words, prediction markets are "good" because they pool large numbers of ignorant noise traders, creating economic incentives for insiders to convert private information into public signals. (This concept of noise traders creating participation incentives for informed insiders is well established in financial literature, and a recent paper further extends it to prediction markets.) Prediction markets can then reliably market themselves as socially valuable, as they indeed provide better, more timely signals than other platforms (experts, polls, etc.). Kalshi and Polymarket know this well, but are reluctant to admit it outright. Still, they hint at it in their marketing!
Kalshi CEO Tarek Mansour said clearly on the Sourcery podcast: "There is no insider trading in commodities. Actually, it's all insider trading." That... is a highly creative interpretation of the law. He added:
I think there is a small set of non-public information that [traders] cannot trade on, but I think our current restrictions are a bit too strict.
Kalshi uses marketing language like "trade anything" and "everyone is an expert in something," both suggesting that ordinary people with certain privileged information can monetize it on the platform.
Polymarket CEO Shayne Coplan had this exchange with CBS last year:
Anderson Cooper:
But prediction markets do rely on some people having inside information.
Shayne Coplan:
Um. Yes. I think it's a good thing that people have an edge in the market. Obviously, you have to manage them, you have to be very clear and strict about boundaries and ethics, and we spend a lot of time on that. But it's somewhat inevitable, and there are a lot of benefits from it. People will adapt.
Shayne also said prediction markets are "the most accurate tool humans currently have, until someone invents some kind of super crystal ball." And part of that accuracy comes from insiders.
Robinhood CEO Vlad Tenev (Kalshi's partner) once said:
Prediction markets actually let you get news faster, in some cases even before the event happens. I think this undoubtedly has huge economic value.
Economist Robin Hanson, widely regarded as the godfather of prediction markets, directly acknowledged this and defended insider trading in prediction markets. He said in 2024:
If the point of [prediction] markets is to get accurate price information, then you absolutely want to allow insiders to trade, even if it discourages others from betting because prices become more accurate. That's the priority.
I must point out that Kalshi and Polymarket both have anti-insider trading policies. Kalshi, regulated by the CFTC, has always clearly prohibited trading on MNPI (Material Non-Public Information) and conducts market surveillance. When I wrote my previous blog post in February, I noted that Polymarket didn't explicitly sanction insider trading, but in March they updated their rulebook and detailed the following types of prohibited trading:
Trading on stolen confidential information (if you're a soldier, the operational plan doesn't belong to you, but to the government).
Trading on illegal tips from insiders.
Trading in any contract where you can influence the outcome.
The point of this section is not to blame Kalshi, Polymarket, or their leadership on the insider trading issue. I think their policies (after the March 2026 update) are clear enough. Instead, I am pointing out the fundamental contradiction plaguing these markets:
Prediction markets rely on informed traders to generate accurate prices, but also rely on ignorant traders to provide economic incentives for informed capital flows. This creates a conflict:
If insider trading is tolerated too much, ignorant traders may withdraw due to perceived unfairness.
If insider trading is restricted too tightly, the market may shut out its most valuable source of information.
The result is a trade-off between information efficiency and perceived fairness.
Therefore, we end up with several different failure modes:
Too many sharks, eating all the small fish. Insider trading standards are too loose, making the market highly informationally efficient, but noise traders get a clear sense that the market is "rigged" and that they're always betting against insiders. Thus, noise traders leave, and the market ultimately lacks liquidity. This is the failure mode I mentioned earlier. This is where we are now, but I think we will bounce in the other direction.
No sharks, no edge. This is the other end of the spectrum. Platforms crack down hard on insider trading with real-time market surveillance and strong regulatory reporting, driving informed capital flows away. As a result, these markets produce less socially valuable information, becoming mere sentiment aggregators rather than generating "news before it happens." Thus, platforms cannot effectively market themselves.
The core question is whether a happy medium exists, one that maximizes liquidity, makes noise traders feel the market is "fair enough," while still allowing informed capital flows to be compensated for gathering information. The chart suggests it might, but reality is more complex.
My prediction from February still stands. As I said then:
There remains a serious risk that insider trading scandals will make retail investors perceive the market as manipulated, leading them to leave the platform. I predict that a string of insider trading incidents this year will force platforms to significantly strengthen market surveillance and push Polymarket, in particular, to move away from anonymity.
I expect Polymarket to completely remove the ability to trade without KYC (its non-U.S. platform is already like this) and increase the flagging of suspicious trades on the platform. There will be many criminal cases about stolen insider information, but the temptation will remain. Though platforms won't admit it, there is a "socially optimal" amount of insider trading. But can they calibrate it optimally? Will regulators allow them to?
