September 1, 2026 · Decision research
Rates Fed Regime
Finding
The weakest assumption is false: FedWatch is not merely a probability-weighted average. CME uses futures-implied monthly EFFR to construct a discrete distribution of target-rate scenarios, while the SEP reports policymakers' individual assessments. Correcting that category error improves the analysis, but it does not create a directional trading edge.
What changed
The freshest defensible market evidence remains close to even: a timestamped August 28 report put CME FedWatch at nearly 56% for a 25 bp September hike, while a separate OIS dashboard as of 00:00 August 31 showed 57% hike pricing. Official Treasury data show the 2-year yield at 4.34% on both August 28 and August 31, and the 10-year at 4.73% and 4.75%, respectively.
Decision
Do not change the posture today. Do not take an unhedged directional September Fed bet on the basis of a modest, instrument-dependent hike lean. Change the rationale from "FedWatch and SEP are incomparable" to "the scenario is tradable only after the market and front-end rates confirm a one-sided move."
Watch next
The next decision-grade catalysts are the official July JOLTS release scheduled for September 1 at 14:00 UTC and the August Employment Situation scheduled for September 4 at 12:30 UTC. Require those data, the September probability distribution, and the 2-year yield to move together before changing posture.