Kalshi vs Polymarket vs Fed Funds Futures: Live FOMC Divergence Tracker
The same Fed meeting, priced on three separate books. The spread on every outcome, why the venues disagree, and the running record of which one has been right.
Quick Answer
Kalshi, Polymarket and CME fed funds futures currently disagree by 6.5 percentage points on the most likely outcome of the Sep 15–16, 2026 FOMC meeting — that is the live cross-market divergence, measured only where the three venues price the same question. Which venue has been right: no meeting has been graded yet; the first is September 16, 2026.
Sep 15–16, 2026 meeting — the daily series →Cross-Market Divergence
Sep 15–16, 2026 · 13 days outThe same meeting, priced on 3 venues at once. On the outcome the market considers most likely, they disagree by 6.5pp.
| Outcome | Kalshi | Polymarket | Futures | Spread |
|---|---|---|---|---|
| 50bp+ cut | 0.5% | 0.1% | — | 0.4pp |
| 25bp cut | 0.0% | 0.5% | 0.0% | 0.6pp |
| Hold | 44.0% | 44.5% | 40.0% | 4.5pp |
| 25bp hike | 54.0% | 53.5% | 60.0% | 6.5pp |
| 50bp+ hike | 1.5% | 0.8% | — | 0.8pp |
Widest gap: 6.5pp on 25bp hike, between Polymarket and Futures.
Which venue has been right
0 meetings graded. Cross-venue capture began August 30, 2026, so the first meeting that can be scored on all three is September 16, 2026. Nothing is estimated or backfilled here — the record starts empty and fills one meeting at a time, because a pre-decision price cannot be reconstructed after the fact.
Polymarket prices are from the international book. Polymarket US is a separate exchange with its own order book — verify the price there before taking a position.
CME fed funds futures implied is derived from the CME fed funds futures curve using the published FedWatch methodology, anchored on the daily effective fed funds rate. It yields a single expected rate, so it is mapped onto a cut / hold / hike triple — it cannot express a 50bp outcome at all, and those cells read “—” rather than 0%.
Updated Sep 2, 8:00 PM ET · refreshes every 30 minutes.
The day-by-day version of this board — every venue’s daily close on the outcome the market considers most likely — lives on the Sep 15–16, 2026 meeting page.
Why Kalshi, Polymarket and futures disagree
Most “Kalshi and Polymarket disagree about the Fed” takes are wrong, and the reason is not liquidity or fees. It is that, for every meeting except the next one, the two contracts are asking different questions.
Kalshi’s KXFED contract asks whether the Fed’s target range will be above a given level following a meeting. That is a rate level, measured against today. If the Fed hikes a quarter point in September and another quarter point in October, Kalshi’s October contract lands in the “50 basis points higher” bucket even though October itself was a single quarter-point move. Polymarket’s per-meeting ladder, and the probability we derive from CME fed funds futures, ask what the Fed does at that meeting. One move, at one meeting.
The two questions coincide only for the next meeting, because no decision intervenes between today and it. Measured on the first day we captured all three venues, August 30, 2026:
- September 2026 (the next meeting): Kalshi and Polymarket agreed to within 2.5 percentage points on all five outcomes. Same question, nearly the same answer.
- October 2026: hold read 35.0% on Kalshi versus 70.5% on Polymarket.
- December 2026: a 50 basis point hike read 33.5% versus 3.7% — a 29.8-point gap.
Up to 35.5 points of “divergence” that is entirely basis and zero edge. It is large, stable and directionally consistent, which is exactly why it reads as a persistent venue lean instead of the accounting difference it is — and why every volume, liquidity and book-quality check is blind to it.
So this tracker subtracts only same-basis venues. Kalshi is compared with Polymarket and futures for the next meeting only; for October and December it is shown as context, never in the spread column. The moment September settles, October becomes the next meeting and its Kalshi contract becomes comparable. The full contract language is on the KXFED settlement page.
The math, briefly
Kalshi’s ladder is a set of threshold contracts (“above 3.50%?”, “above 3.75%?” …), so a per-outcome probability is the difference between adjacent thresholds — a CDF difference — and the outcomes need not sum to exactly 100%. Polymarket lists five independent binaries per meeting; they need not sum to 100% either (0.9965 on August 30) and we deliberately do not renormalize them, because forcing agreement between venues would erase the thing being measured. The futures figure inverts the CME curve with the published FedWatch method, anchored on the daily effective fed funds rate rather than the monthly average, because the inversion amplifies anchor error by roughly 8.6 points of probability per basis point; it yields one expected rate, mapped onto cut / hold / hike, and cannot express a 50 basis point outcome — those cells read “—”, never 0%. Two ceilings guard the board: a book-vs-book divergence above 25 points, or a futures-vs-book divergence above 40, is treated as a data fault and withheld rather than published.
Which venue has been right
After each decision, every venue is scored on its last close before the announcement on the outcome that actually happened; the venue with the highest probability on the realized outcome is the closest. A grade needs at least two venues, and a pre-decision price cannot be reconstructed after the fact, so the record starts on the first meeting captured on all three — September 16, 2026 — and fills one meeting at a time. It is never estimated, never backfilled, and never renders a 0-of-0 as 0%.
0 meetings graded — first grade Sep 16, 2026. The Kalshi-only accuracy record back to July 2026 is on the historical accuracy page.
Every 2026 FOMC meeting
Use this data
The daily three-venue series is free to use with attribution — link to this page. JSON by default, CSV with ?format=csv; meeting=YYYY-MM-DD selects a meeting (default: the next one), outcome=HIKE-25 filters one rung. Cached five minutes, CORS open, only derived probabilities.
GET https://predictionmarketspicks.com/api/public/fed-divergence
GET https://predictionmarketspicks.com/api/public/fed-divergence?meeting=2026-09-16&outcome=HIKE-25
GET https://predictionmarketspicks.com/api/public/fed-divergence?format=csvEmbed the live board — one line, no signup, updates itself:
<iframe src="https://predictionmarketspicks.com/embed/fed-rate-tracker/article?theme=indigo" width="100%" height="400" style="border:0;max-width:728px" loading="lazy" title="Live Fed rate decision odds from prediction markets"></iframe>Series today: CME fed funds futures implied, Kalshi, Polymarket. Polymarket prices are from the international book; Polymarket US is a separate exchange with its own order book. The futures figure is a probability derived from the CME curve, not a contract price.
Trade the Sep 15–16, 2026 Fed decision
The spread above is where the venues disagree. Take a position on cut, hold or hike on Kalshi’s KXFED ladder — new accounts get a sign-up bonus through our referral.
Get the cross-market Fed read before each FOMC decision — free
No spam. Unsubscribe anytime.
Related
Frequently Asked Questions
Why do Kalshi and Polymarket show different Fed odds?
Three reasons, in order of size. First, for any meeting past the next one they are answering different questions: Kalshi’s KXFED contract asks whether the target range will be above a level FOLLOWING the meeting (a cumulative level versus today), while Polymarket’s per-meeting ladder asks what the Fed does AT that meeting. Measured on August 30, 2026, that basis difference alone was 35.5 percentage points on October’s hold outcome. Second, they are different crowds on different books — Kalshi is a CFTC-regulated US exchange, the Polymarket prices here are from its international book — with different fees and different liquidity. Third, genuine disagreement: on the next meeting, where the question is identical, the two books typically sit within a few points and that gap is a real difference of opinion.
Is CME FedWatch more accurate than prediction markets?
Nobody has published a graded comparison, so we are building one. The CME fed funds futures figure on this page is derived from the futures curve using the published FedWatch method, anchored on the daily effective fed funds rate. It reflects banks and hedge funds hedging rate risk, so it is the deepest of the three books, but it yields a single expected rate and cannot express a 50 basis point outcome at all. The record below scores each venue on its last pre-decision close, one meeting at a time, starting September 16, 2026. Until then the honest answer is that it reads zero meetings graded.
Which market has been right about the Fed?
None yet, because nothing has been graded yet. Three-venue capture began August 30, 2026, and a pre-decision price cannot be reconstructed after the fact, so the first meeting that can be scored on all three venues is September 16, 2026. The record fills one meeting at a time and is never estimated or backfilled.
Can you compare Kalshi and Polymarket odds for October or December?
Not directly, and this page refuses to. For any meeting past the next one, Kalshi’s contract prices the rate LEVEL after that meeting relative to today — two separate 25 basis point hikes at earlier meetings land in the same 50 basis point bucket — while Polymarket and the futures curve price the move at that specific meeting. Subtracting one from the other measures the difference in the question, not a mispricing: 35.0% versus 70.5% on October’s hold and 33.5% versus 3.7% on December’s 50 basis point hike, as of August 30, 2026. We compare only same-basis venues and show the rest as context. The two questions become the same the moment the intervening meeting settles, so October becomes comparable on the evening of September 16.
Where does the data come from and can I use it?
Every 30 minutes we capture Kalshi’s KXFED ladder, Polymarket’s per-meeting Fed ladder (international book) and the CME fed funds futures curve, store each venue’s implied probability per outcome, and roll them into a daily close. The daily series is free to use with attribution as JSON or CSV at predictionmarketspicks.com/api/public/fed-divergence, and the live board can be embedded. Only derived probabilities are published — never a raw futures contract price.
Latest board: 5 comparable outcomes, headline spread 6.5pp · CUT-50 0.4pp · CUT-25 0.6pp · HOLD 4.5pp · HIKE-25 6.5pp · HIKE-50 0.8pp · venues CME fed funds futures implied 40.0% hold, Kalshi 44.0% hold, Polymarket 44.5% hold.