Kalshi's Fed-cut ladder is mispricing the tails

The September contract ladder implies a tidy, unimodal path for rates. The macro data says the wings are fatter than that — and that's where the contracts are cheap.

Kalshi lists the September FOMC outcome as a ladder of strikes — 25bp cut, hold, 50bp — and the market prices each rung independently. Stitch them together and you get an implied distribution over the policy rate. Right now that distribution is suspiciously smooth: a single hump centered on a 25bp cut, thin shoulders, almost nothing in the tails.

That shape is convenient. It's also wrong. The data-dependence the committee keeps signaling means September is a function of two more CPI prints and a jobs report — and those don't resolve to a single modal path. They resolve to a fork.

Run the same data through a regime model and the hump splits. Sticky services inflation keeps a "hold" outcome alive; a soft jobs print pulls a 50bp cut into play. The body of the distribution thins and the wings fatten. The ladder, priced rung-by-rung, never sees this — each strike is quoted by a different marginal trader with a directional view, not a distributional one.

If the ladder is unimodal and the world is bimodal, you don't need a view on direction. You need a view on shape.

That reframes the trade. Selling the body and buying the wings is a textbook long-convexity structure — and on Kalshi it's executable in size, because the tail contracts trade at a few cents with real depth behind them. The carry is negative while you wait, but the payoff if September forks is asymmetric.