Predicting markets cannot do without insider trading, but insider trading is killing it.

Author: Nic Carter

Translation: Deep Tide TechFlow

Deep Tide Guide: U.S. military special forces personnel used confidential information to make $400k on Polymarket, and this is just the latest scandal. Nic Carter points out that prediction markets are caught in a vicious cycle: relying on insider trading to produce accurate prices, but this causes retail traders to feel the market is manipulated and exit. This contradiction determines whether prediction markets can survive long-term.

As I wrote in February this year, prediction markets have serious insider trading issues, and this is no accident. It leads to a major failure mode:

The social value of prediction markets comes from incentivizing insiders to leak confidential information with money, but over time, this destroys retail traders’ confidence in the market.

Two days ago, the biggest scandal so far broke: the U.S. Department of Justice accused a Green Beret sergeant, Gannon Ken Van Dyke, of misconduct involving confidential information. Before the Maduro raid mission, he made $400k on Polymarket. He is not an ordinary soldier but a senior Green Beret member responsible for special operations planning and execution.

To put it simply, although many people call for leniency because many congressmen engage in (legal) insider trading, he still should go to jail. His actions may have leaked details of the raid to Venezuelans through trading activities, which is morally and legally problematic. While Venezuelans may not have noticed, the government cannot accept such a precedent: elite operators leaking details of upcoming operations through market activities for personal gain. I sympathize with Van Dyke, but he indeed violated the law and the confidentiality he swore to uphold.

This is just the latest in a series of real or suspected insider trading scandals in prediction markets. Previously, two reservists in Israel were arrested for trading using military intelligence. Markets about the start time of the Iran war, ceasefire agreements, the killing of Khamenei, and Biden’s pardons have also been suspected, but no arrests have been made yet. Kalshi and Polymarket have also flagged and suspended accounts involved in trading on markets that involve their own interests, such as three congressional candidates betting on their own campaigns.

You might think that as more people realize that trading with confidential information is illegal not only in securities markets but also in prediction markets, these issues would disappear. But I believe the problem is deeper.

The premise of prediction markets is that they are informationally efficient because they reward informed insiders.

In other words, prediction markets are “good” because they gather a large number of uninformed retail traders, who create economic incentives for insiders to disclose private information. (This concept—retailers creating incentives for informed insiders to participate—has been well established in financial literature, and a recent paper further extends it to prediction markets.) Then, prediction markets can reliably claim to have social utility because they provide signals that are better and more timely than other platforms (experts, polls, etc.). Both Kalshi and Polymarket know this but are reluctant to admit it explicitly. But they do imply it in their marketing!

Kalshi’s CEO Tarek Mansour explicitly stated on the Sourcery podcast, “Commodities markets have no insider trading. In fact, it’s all insider trading,” which is… an extremely creative legal interpretation. He added:

I think some non-public information (traders) shouldn’t be tradable, but I think we’re restricting it a bit too much now.

Kalshi has used slogans like “trade anything” and “everyone is an expert in some area,” both implying that ordinary people, if they happen to have some privileged information, can monetize it on the platform.

Last year, Polymarket’s CEO Shayne Coplan had this exchange with CBS:

Anderson Cooper: But prediction markets do rely on some people having insider information.

Shayne Coplan: Hmm. Yeah. I think it’s good that people have an advantage in the market. Obviously, you need to manage it, define very clearly and strictly the boundaries, like ethically, and we’ve spent a lot of time on that. But to some extent, it’s unavoidable, and it can bring a lot of benefits. You know, people will adapt.

Shayne also said prediction markets are “the most accurate thing humans have right now, until someone creates some super crystal ball.” Some of that accuracy comes from insiders.

Robinhood CEO Vlad Tenev (who collaborates with Kalshi) said:

Prediction markets can actually get you news faster, sometimes even before it happens. I think it has enormous economic value.

Economist Robin Hanson, often regarded as the godfather of prediction markets, directly embraces this view and has written lengthy defenses of insider trading in prediction markets. In 2024, he stated:

If the purpose of (prediction) markets is to get accurate price information, then you definitely want insiders to trade, even if it makes others feel it’s unfair and they don’t want to bet, because it makes prices more accurate. That’s the priority.

I must point out that both Kalshi and Polymarket have policies against insider trading. Kalshi is regulated by the CFTC and has explicitly banned trading based on material nonpublic information (MNPI), with market surveillance in place. When I wrote a blog post in February, I noted that Polymarket did not have explicit sanctions against insider trading, but in March they updated their rules manual with detailed prohibitions, including:

Trading based on stolen confidential information (if you are a soldier, operational plans do not belong to you but to the government)

Trading based on information illegally passed to you by insiders

Trading on any contracts where you can influence the outcome

The focus of this section is not to blame Kalshi or Polymarket or their leadership for implying traders have an informational advantage. I believe their policies (since the March 2026 update) are sufficiently clear. Instead, I want to highlight the fundamental contradiction troubling these markets:

Prediction markets rely on informed traders to produce accurate prices, but also depend on uninformed traders to create economic incentives to attract informed trading flow. This creates a tension:

If insider trading is too tolerated, uninformed traders may exit because they feel it’s unfair.

If insider trading is too restricted, the market may exclude its most valuable information sources.

Thus, there is a trade-off between informational efficiency and perceived fairness. A visual version of this idea:

Chart: The trade-off curve between informational efficiency and perceived fairness

So we ultimately face several different failure modes:

Sharks are too many, eating all the fish

Insider trading standards are too lax, making the market highly efficient but retail traders clearly feel it’s “manipulated,” always betting against insiders. As a result, retail traders leave, and market liquidity drops. This is the failure mode I discussed earlier. This is where we are now, but I believe we will bounce back in another direction.

No sharks, no advantage

This is the other end of the spectrum. Insider trading is strictly regulated on the platform, with real-time market monitoring and strong regulatory reporting, so informed trading flow is kept far away. The social value of these markets is thus less, merely aggregating sentiment rather than generating “news before the news.” As a result, the platform cannot market itself effectively.

The existential question is whether there is a golden middle ground: maximizing liquidity, making retail traders feel the market is “fair enough,” and still allowing informed trading to be rewarded for information gathering. The chart suggests it might exist, but reality is messier.

My February prediction still holds. As I said at the time:

Serious risks remain—insider trading scandals could make retail traders feel the market is manipulated, leading them to abandon the platform. I predict a series of insider trading incidents this year, prompting platforms to significantly tighten market surveillance, especially pushing Polymarket away from anonymous mode.

I expect Polymarket will completely eliminate the ability to trade without KYC (which is the case for non-U.S. platforms now), and will strengthen flags on suspicious trades. There will be many criminal cases related to stolen insider information, but the temptation will still exist. Although platforms won’t admit it, there is indeed a “social optimum” insider trading volume. But can they calibrate it optimally? Will regulators allow them to do so?

It’s worth noting that not all informed traders are insiders. You can inform yourself by collecting public information and trading based on it. But some informed traders are indeed insiders who have stolen information.