When September goes from “maybe” to “base case,” the first losses show up in duration, not in jobs

The key change isn’t whether the Fed hikes in September—it’s that the market’s probability mass moved decisively enough that positioning can treat a hike as the starting assumption rather than a tail risk. In late August, coverage tied the September odds to CME FedWatch, describing a move from about a ~56% hike probability to a market view “now a coin flip,” with odds later rising above 80%.

Two implications follow immediately: 1) After the probability flips, new data has to be “wrong the other way” to change the consensus. 2) The discount rate embeds not just a single meeting outcome, but the probability-weighted interest-rate path that affects duration assets, credit pricing, and global funding spreads.

To understand what “80% probability” actually means for trading, you need the mapping from derivatives to the expected short-rate distribution.

The Atlanta Fed’s Market Probability Tracker explains the framework: it estimates probability distributions implied by options tied to SOFR futures, referencing the three-month compounded average SOFR. The model uses option premiums to infer the probability that realized SOFR over the reference quarter ends up above or below specific strikes.

So when the probability crosses a threshold (like ~80%), it’s not just a headline number: it’s a shift in the probability distribution of SOFR over the following quarter(s), which then feeds into bond yields (especially the front end and the part of the curve most sensitive to the next policy steps).

A probability regime shift is rarely random. In this cycle, coverage linked the rise in hike odds to the Fed’s narrative tone—especially after Warsh’s Jackson Hole speech.

In his keynote remarks, Warsh emphasized a price-stability focus, argued that progress over the past two years has been “modest,” and stressed that policy changes should follow a reaction-function standard: if the Fed is not confident that underlying inflation is moving to 2% “clearly and at sufficient speed,” the Fed has “work to do.”

That kind of conditional language changes expectations even before new inflation prints arrive, because markets infer what the Fed will do under a wide range of scenarios.

Why this ordering? Because duration and discounting are mechanical: higher expected policy rates move discount factors immediately. Credit and housing require behavioral/financial transmission—refinancing calendars, underwriting standards, mortgage-rate pass-through, and demand shifts.

That’s why the most useful investor question after the probability flip is not “Is a hike bullish or bearish?” but “Which cash-flow profile reprices first under a higher-for-longer distribution—and which balance sheets have the greatest sensitivity to front-end rate changes?”

To anchor the mechanism in fundamentals, consider large US banks where net interest income is a direct function of rate levels and repricing dynamics.

Using income statement line items, JPMorgan Chase's net interest income was $95.443B in FY2025 (reported Feb 13, 2026) vs. $92.583B in FY2024 (reported Feb 14, 2025). Bank of America's net interest income was $60.096B in FY2025 (reported Feb 25, 2026) vs. $56.060B in FY2024 (reported Feb 25, 2025).

This matters because a base-case hike raises the expected path of short rates and can lift the interest-income side quickly. The risk is that higher rates also raise funding costs and can eventually pressure credit quality—typically with a lag. In other words: bank stocks can rally on “NIM math” before the market fears the credit cycle.

Equities with durable, long-dated cash flows are the cleanest expression of the duration channel.

Consider Microsoft and Apple. Their income statements show continued scale even as interest-rate expectations move; for FY2025 (for the latest available fiscal year in the dataset), Microsoft reported revenue of $281.724B and FY2025 for Apple shows revenue of $279.745B. These are not “rate-sensitive operating numbers” in the quarter you get a probability flip.

But the market can still reprice them quickly because discounting changes with expected policy rates. That’s why a probability crossing tends to precede corporate earnings revisions: the market adjusts discount factors first, not revenues or margins.