Midterm trading volumes on prediction markets alarm election officials

Jared DeMarinis delivered the clearest statement of what election administrators are actually afraid of. It isn't fraud, exactly. It isn't manipulation in the crude sense. It's the feedback loop: Americans watching odds move in real time, concluding the market knows something they don't, and losing confidence in a result before a single ballot is counted.

DeMarinis runs Maryland's State Board of Elections. When someone in that position calls something a troubling trend that administrators across the nation must deal with, the language is careful and the alarm is genuine. Election officials do not use the word "troubling" for things they expect to resolve themselves.

The concern the platforms have answered — insider trading — is not the concern DeMarinis is raising. Kalshi and Polymarket both point to their federal compliance frameworks: required disclosures, trading suspensions, the Buckhout case in August as evidence the rules have teeth. Those are real controls on a real and narrow problem. They say nothing about what happens when a contract on a Senate race moves twenty points in a week and a local television station runs the graphic.

Here is where I part from the consensus read on this. Most of the commentary treats election officials' concerns as institutional self-protection — bureaucrats uncomfortable with markets they can't control. I don't think that's where this lands. The stock market comparison that Columbia's Joshua Mitts offers is technically correct and practically beside the point. When equity prices move on election expectations, the feedback into voter behavior is diffuse and slow. When a prediction market contract with a price between one and ninety-nine cents is packaged as a real-time probability and broadcast into an information environment already saturated with contested claims about voting fraud, the mechanism is different. The signal is explicit, it is simple, and it arrives at the same moment as the vote.

Kalshi's own calibration argument — that sixty-percent contracts resolve correctly at roughly sixty percent — is a defense of accuracy, not a defense of influence. A market can be well-calibrated and still distort the thing it is measuring. This is not a hypothetical problem. It is a known problem in financial markets, where price discovery and price impact coexist, and where researchers have spent decades trying to separate them. The prediction market industry has not seriously engaged with the literature on that question, because engaging with it would require acknowledging that the question is open.

The state-level crackdowns spreading from Missouri to Maryland are messy, legally inconsistent, and in several cases constitutionally dubious. None of that changes what DeMarinis is pointing at. The industry's best answer to the election integrity concern remains the insider trading framework — a response to a different question. The gap between what officials are asking and what platforms are answering is not closing.