The FOMC meets on September 15 and 16, with the decision and a fresh dot plot landing at 2pm ET on the 16th. A month ago the only live question was when the cuts would start. It is now a serious argument about whether the next move is a hike.
Here is the odd part. Three different venues are pricing the same binary event and arriving at three different answers. Fed funds futures, via the CME’s FedWatch tool, put the odds of a quarter-point hike at roughly 56%. Kalshi has it at 48%. Polymarket, 49%. Goldman Sachs, meanwhile, has told clients a September hike has become “very unlikely,” pointing to softer retail sales, a cooling labour market and inflation prints that are decelerating rather than accelerating.
The move itself is not in dispute. Before Fed Chair Kevin Warsh’s Jackson Hole speech, the odds of no change in September sat near 70%. His commitment to fighting inflation repriced the front end inside a session: on Friday the two-year yield rose about nine basis points to roughly 4.32% and the ten-year added four to about 4.72%, while equities closed lower.
Why the numbers disagree
They are not measuring the same thing. FedWatch probabilities are derived from fed funds futures, and that is a hedging market before it is a forecasting one. Institutions with duration exposure buy protection against the hawkish tail whether or not they believe in it, and that demand lifts implied odds above what the median participant actually expects. Kalshi and Polymarket are direct binary wagers with no hedging demand attached and much thinner order books — cleaner as opinion, noisier as data.
So the eight-point gap between futures and prediction markets is not an arbitrage sitting on the table. It is roughly the cost of insurance against a Fed that has just told everyone it is worried about inflation. Core PCE ran at 3.3% year over year in July. Headline is at 3.7%. Neither of those is a number a hawkish chair ignores.
The week that decides it
Markets are closed Monday for Labor Day. Then the data arrives in a compressed stack:
- Tuesday, September 2 — ISM manufacturing, final S&P Global manufacturing PMI, construction spending.
- Wednesday, September 3 — JOLTS job openings, factory orders, the Fed’s Beige Book.
- Thursday, September 4 — jobless claims, Challenger job cuts, ADP employment, trade balance, ISM services.
- Friday, September 5 — nonfarm payrolls and the unemployment rate.
Four of those five readings are labour-market readings. That is not a coincidence: a hike this cycle requires a jobs market strong enough to justify one, and the labour data is the only thing standing between a hawkish chair and a hawkish decision.
Our take: The dispersion is the information. When futures, two prediction markets and a major dealer’s economics desk are spread across a 56-to-“very unlikely” range, nobody has an edge on the outcome — they have different exposures to being wrong about it. What that spread does tell you is that the September meeting is currently underpriced as an event, not overpriced as a hike. Both tails are live, the dot plot lands the same afternoon, and positioning built on “cuts are coming, timing unclear” has not been rebuilt for a world where the next move might go the other way. Friday’s payrolls report is the single number that collapses the range.
What to watch
- Whether the three venues converge after payrolls. If futures stay materially above Kalshi and Polymarket even once the jobs number is out, the gap really is hedging demand rather than disagreement.
- The two-year yield. It moved nine basis points on a speech. It is the cleanest read on how seriously the front end is taking the hike scenario.
- The Beige Book on Wednesday. Anecdotal, unloved, and the one release that describes labour conditions in words rather than a seasonally adjusted headline.
- The dot plot, not just the decision. September is a projection meeting. A hold accompanied by a materially higher path would be more hawkish than a hike with a flat one.