In June I wrote that prediction markets were becoming real financial infrastructure with a manipulation problem. Congress, the states, and the Supreme Court are about to spend two years proving both halves.
Hi Everyone,
Last month I wrote about prediction markets for the first time — the fake-liquidity problem, the strange new machine that had correctly called the 2024 election while the pollsters shrugged. I ended that letter saying the interesting part would be what happens when the machine collides with the law. The collision is here, and it happened in Wisconsin.
Bet on the election, lose your ballot.
The Wisconsin Elections Commission warned last week that a state law disqualifying anyone who has placed “a bet or wager” on an election may apply to prediction-market users. Kalshi — which has hundreds of thousands of users in Wisconsin alone — fired back within a day, accused the state of voter suppression, and then did something wonderfully American: it launched a midterm-elections hub, a live odds board for the races, built on how its users are betting. The CFTC, run by a Trump appointee, is suing states on the platforms’ behalf. Multiple states call it illegal gambling. This is going to the Supreme Court, and both sides seem to want it there.
With control of both the Senate and the House in play, this November will be the first American election where a regulated betting market publishes a real-time forecast next to the polls. That is simply new. It has never happened before.
My honest, two-handed opinion.
So let me give you my honest, two-handed opinion, because I hold both of these views at full strength. These markets are the best free forecast in America. When someone must put money behind an opinion, the opinion improves; that is the entire theory of every market I work in. And: putting a wager next to a ballot box is genuinely corrosive in a way the enthusiasts wave off too quickly. A voter with a position is a voter with a conflict, however small. The software developer the Journal quoted said it plainly — elections should not carry a direct financial incentive, no matter how modest. Both things are true. The law will spend two years deciding which one it cares about more, and I do not know the answer.
What I actually use these markets for, professionally: calibration, never action. There is a market right now on whether the S&P 500, gold, or bitcoin finishes 2026 with the best return. The crowd says S&P 64%, gold 20%, bitcoin 16%. After a week in which the Dow fell four figures in a day and oil swung ten dollars, the money still prefers the boring, diversified index over the shiny alternatives — which happens to be how your accounts are built. When that number flips, I will not trade on it, but I will absolutely want to know why it flipped.
What I will never do with your money.
What I will never do: bet on outcomes, or position client money based on what an odds board says about November. We trade appropriations, not polls. When Washington actually writes a check — like the $120 billion defense restock this week — that is real. The odds of a check are not a check.
They can see where this is going.
Robinhood, meanwhile, is in talks to deepen its prediction-market business. The brokers can see perfectly well where the retail order flow is going. Toll roads get built where the traffic already is.
Part III of this letter will probably be written the week the Supreme Court hears the case. Between now and then, watch the odds boards if you enjoy them — read them like a poll with money on it, not an oracle.
Bring the question, and we’ll run the numbers together.
A short call is usually enough to know whether anything in your plan needs to change before the next quarter. No prep required.
— Sean Anees Saifi