A Google Engineer Made $1 Million on Polymarket, and a White House Aide Made $100K on Kalshi—You're the Exit Liquidity 

Prediction markets are marketed as democratic tools harnessing the "wisdom of the crowd." Recent scandals involving a Google engineer allegedly betting $1 million and a White House aide wagering $100,000 with information obtained outside of market data have exposed the truth: the mechanisms we call prediction markets are places to exploit the retail trader's access to information available to all as exit liquidity for those with nonpublic information. 

When you bet on a market that someone else is designed to win, you are always on the wrong side.

Microsoft Just Rewrote Its Insider Trading Policy—and Named Prediction Markets

New regulatory peril is not the surprising insight here. What is more surprising is the manner in which Microsoft has begun amending its insider trading policy to explicitly include prediction markets. The change was revealed in an updated filing with the SEC this month, as Microsoft expands its corporate insider trading prohibition to include prediction markets in which employees, vendors, partners and competitors (yes, competitors) engage. 

"Throughout the vast sectors of artificial intelligence and cloud computing in which we participate, our corporate technologists and innovators work with each other, as well as researchers and technologists throughout the technology community to build an ecosystem that advances the state of the art." 

Microsoft spends tens of billions of dollars annually to build the "ecosystem that advances the state of the art." Now it is additionally spending millions of dollars to ensure their employees do not enable those outside to exploit their "innovators" when they get bit by the betting bug.

The $165,000 Bet on a Microsoft Executive's Vocabulary

An analysis of the event-based wagering market in a specific instance illustrates exactly how it works. Looking at the second most recent trade on the "Microsoft Earnings Call Will Feature At Least One Executive Using Unusual Vocabulary," noted that the $165,000 wager "is way bigger than any I've seen on Intrade." 

With the ticker symbol MSFEXPVOC, the market is designed as an "opportunity to express your opinion on the Microsoft Earnings Call." Microsoft, which has a gigantic stake in the markets in which its employees participate, is candid in its Annual Report about the fact that only "people with knowledge about how Microsoft operates who are expected to be in the meeting room" truly know whether or not the "information will be public or private." 

Additionally, those who have the scripts for the earnings call enjoy an even bigger edge. In other words, the winner of this prediction market will almost certainly be someone in the room or with one of the scripts. Those with the knowledge to produce a better forecast, such as those who follow Microsoft closely and consume rival analyst reports, have a mathematical disadvantage. 

The mechanism that delivers these forecasts to attentive buyers, selling the other side, does so to the benefit of those with privileged information. The market construction itself effectively auctions the information rather than aggregating dispersed information. Exactly the same mechanism used for price discovery functions as the engine for insider trading.

No Flagging System Can Flag What Actually Matters

The Microsoft example is revealing when it comes to understanding the words "flagging systems" and "terms-of-service" on betting site disclosures. They are inadequate. Volume detection will help an institution identify users behaving erratically. Once alerted, Microsoft would have the tools to investigate further and put the user on notice or permanently bar her. 

That said, these automated systems can't distinguish between smart public analysis and insider leaks, nor between a gambler reacting to a real story or reacting to a story generated by someone else inside the company. Goldman Sachs chose not to run the risk of a hole in its omnibus insider trading policy, banning participation on prediction markets entirely. Microsoft is smart enough to know that a proper insider trading policy must be broad enough to prohibit even those companies to which it sells products or does business. 

Its response is to prohibit only internal betting, which is significantly less effective, but effectively reveals that Microsoft now considers its prediction market engagements illicit as well. When Microsoft, a giant of the cloud computing industry, is making these moves, this speaks volumes. In an environment in which billions of dollars in venture capital investment will likely go to the next prediction market, a small mind is fighting the basic structure of the product itself rather than external bad actors.

The Existential Threat Isn't Regulation. It's the Business Model.

This liability cements the existential threat these platforms pose. The industry's legal defense is that, in the aggregate, they are sophisticated financial infrastructure rather than gambling. They circumvent state prohibitions on betting markets on the grounds that their activities are more akin to buying and selling insurance contracts and that they provide insights and price discovery better than crude voting mechanisms. 

Insiders who leak facts to prediction market traders by telling them what they know will threaten this entire legal argument. Prediction markets operate by promising the average crowd a "winnings" tax against it. The existence of insiders for whom prediction markets serve as an exit liquidity mechanism means that these platforms operate not as instruments of the wisdom of crowds, but as instruments of insider trading. This runs headlong into a state and federal regulatory framework built precisely to prevent such practices.