Skip to main content

Fed Funds Futures & CME FedWatch (Rate Odds)

Updated Jun 24, 2026 at 2:35pm

Research Draft High 1,246 words

Fed funds futures are exchange-traded contracts whose price embeds the market's collective bet on the average overnight policy rate over a given calendar month, and CME FedWatch is the tool that translates those prices into headline-friendly probabilities like "78% chance of a hold at the next meeting." Together they are the single most-cited real-time read on what the market expects the Federal Reserve to do — a market-implied forecast that often moves stocks, bonds, and the dollar the instant it shifts. The core tension is that this "probability" is a price-derived inference, not a poll: it is contaminated by a risk premium and rests on simplifying assumptions, so it is a very good directional gauge and a noisier point forecast.

How it's calculated / formed

The underlying instrument is the CME's 30-Day Federal Funds futures contract (ticker ZQ), traded on CBOT. It is quoted in "IMM index" terms: price = 100 − implied rate. A contract pricing at 95.50 implies a 4.50% average rate for that month. At expiry the contract is cash-settled against the simple average of the daily effective federal funds rate (EFFR) for the delivery month — the sum of each day's EFFR (published by the New York Fed) divided by the number of days in the month (CME rulebook Ch. 22; kisfutures fact card). Contract size is $5 million notional, so one basis point of monthly average rate is worth $41.67.

Because settlement is a monthly average, a meeting that changes the rate mid-month splits the month into a pre-meeting and post-meeting period. FedWatch handles this with arithmetic, not magic. Its methodology (CME):

1. Extract the average rate implied by the ZQ contract for the meeting month (100 − price). 2. Knowing the rate that prevailed before the meeting and the number of days on each side, solve for the implied end-of-month rate — the rate the market expects after the decision. 3. Compare that implied end rate to the current target. Under the assumptions that moves come only in 25 bp multiples and EFFR is bounded at zero, distribute the implied change across the discrete outcomes (e.g., a hold vs. a 25 bp cut) to get probabilities. 4. Chain meetings into a conditional probability tree: each later meeting's odds are computed assuming the outcomes of prior meetings, then unconditional probabilities are read off the tree.

For months without a meeting, the contract directly pins the prevailing rate, which anchors the start rate for the next month.

How it's used in practice

Traders and analysts use FedWatch and ZQ futures as the market's live consensus on policy. The quick-and-dirty version needs no tool: subtract the front-month ZQ price from 100 to read the implied rate, and compare nearby contract months to see how much easing or tightening is priced over the coming year (the "strip"). FedWatch packages this into per-meeting hike/hold/cut probabilities.

Common applications:

  • Event positioning: gauge whether an FOMC decision is "priced in." A decision matching ~95% odds is usually a non-event; the surprise — and the volatility — lives in the gap between expectation and outcome, and in the dot-plot/press-conference guidance.
  • Cross-asset context: a sudden repricing toward cuts typically supports rate-sensitive equities (growth, small caps, REITs), pressures the dollar, and steepens or rallies the front end of the curve. Equity and macro desks watch FedWatch shifts intraday around CPI, jobs, and Fed-speak.
  • Narrative tracking: the change in odds across a week is often more informative than the level — it shows how new data is reshaping the policy path.

Adoption, debate & evidence

FedWatch is ubiquitous in financial media and is effectively the default rate-expectation reference. The underlying ZQ market has been studied for decades and is genuinely informative, with two important caveats.

First, the futures are biased by a risk premium. Piazzesi and Swanson's influential work (NBER w10547) shows excess returns on fed funds futures are positive on average and countercyclical — predictable from employment growth and credit spreads — so the raw futures rate systematically over-predicts the future funds rate, with the bias growing at longer horizons. The expectations hypothesis fails even at short horizons. FedWatch's probabilities do not strip out this premium, so they are best read as "market-implied" rather than "true" odds. The bias is small for the very next meeting and material several meetings out.

Second, on accuracy: at short horizons the futures are quite good. The Richmond Fed's work (Owens & Webb, 2001) and subsequent literature find near-term futures forecast policy well, with errors and premium rising with horizon. A recent peer-reviewed evaluation, Bonini, Huang & Simaan's "Watching the FedWatch" (Journal of Futures Markets, 2026), reports the FedWatch model predicting FOMC decisions with roughly 88% accuracy about 30 days ahead (versus ~75% for the outright fed funds futures baseline) — a figure that names its source and should not be treated as a guarantee. Note the survivorship caveat: most meetings are holds, so a high hit-rate partly reflects the base rate of "no change." The honest framing is calibration, not infallibility: when FedWatch says 70%+, the favored outcome tends to occur, but the tool is a snapshot that can reprice violently on a single data surprise or an inter-meeting shock (e.g., March 2020).

Strengths & limitations

Strengths: real-time, liquid, transparent, and forward-looking across many meetings at once; far timelier than surveys; directionally reliable for the near term. The #1 practical value is detecting changes in the policy path the instant data lands.

Limitations and the #1 misuse: treating the headline percentage as a literal, premium-free probability. It is risk-adjusted-price-derived. Further pitfalls: the 25 bp-multiple assumption breaks for unconventional moves (50 bp hikes/cuts, emergency actions); EFFR can drift within the target band, adding basis-point noise; thin or distorted front-end liquidity (e.g., funding stress) can warp implied rates; and the tool says nothing about why — the dot plot and forward guidance often matter more for markets than the binary decision itself.

Sources

Dispute flagged: The ~88% accuracy (Bonini 2026) and the risk-premium bias (Piazzesi–Swanson) are not contradictory but are often conflated — FedWatch is well-calibrated near-term yet systematically premium-biased, especially at longer horizons. The 88% figure is also inflated by the high base rate of "hold" decisions.