Wall Street Draws a Line Against the Prediction Markets

Goldman Sachs and other major banks are banning employees from using prediction markets. They see them less as wisdom-of-the-crowd and more as a compliance disaster waiting to happen.

Prediction markets were supposed to be the wisdom of the crowd, crystallized into a price. A decentralized truth machine. Wall Street, it turns out, sees them as a liability engine. Banks including Goldman Sachs and Morgan Stanley are now updating employee codes of conduct to explicitly ban their use for bets on financial markets or political events. The message is clear: you can make money for the house, but you cannot play at the casino down the street using what you learn on the job. The game has simply gotten too real, too fast.

At their core, these markets are simple. Platforms like Kalshi, which is regulated in the US, allow users to buy and sell contracts on the outcome of future events. "Will unemployment claims be above 220,000 this week?" A 'Yes' contract might trade at $0.60; if the event happens, it pays out $1.00. The problem isn't the mechanism, it's the information asymmetry. These markets are pure signal, and they react violently to non-public information. The recent case where The Commodity Futures Trading Commission and Department of Justice charged a Google employee with using internal knowledge to profit on Polymarket contracts is the platonic ideal of this risk. One trader with an edge can look a lot like a criminal.

The crackdown isn't academic. According to Reuters, Goldman Sachs has told staffers they cannot use platforms for event-based contracts tied to finance or politics that could create even a perceived conflict of interest. This isn't a gentle reminder about ethics. According to Bloomberg, repeat violations could lead to termination and forfeiture of any gains. This is about managing liability at institutions like JPMorgan Chase and Bank of America, which are rolling out similar restrictions. The fear is that a banker, knowing about an unannounced merger from their day job, could make a perfectly legal bet on a prediction market that looks an awful lot like insider trading to a prosecutor.

This is just the start. The banks cannot un-invent these markets, but they can build a wall around their own operations, isolating their employees' information advantage from public betting pools. In the next few years, expect prediction platforms to face immense pressure to police themselves or be policed by regulators looking to apply old laws in this new context. The arms race will be between decentralized information discovery and centralized compliance enforcement. The platforms will either become surveillance partners for the institutions they threaten, or they will be regulated into a different shape. The real question is not whether you can bet on the future. It's whose information is considered fair game when you place that bet.