Record consumer gloom, and the inflation reading that ties the Fed's hands

American consumers have rarely been so miserable. The University of Michigan's Index of Consumer Sentiment fell to 47.8 in the first half of September, its second-lowest reading in 74 years of the survey, down from 51.7 a month earlier and well short of the 51 that forecasters had expected. Only the record 44.8 of May 2026, set when the war with Iran sent oil prices shooting up, was ever lower. The survey's authors gave a terse explanation for the slide: a resurgence in fuel prices, trade tensions and the lingering conflict.

The gloom is worth taking seriously, but it is not the most important number in the release. That honour belongs to inflation expectations, which jumped from 4.0% in August to 4.6% for the year ahead, the highest since June and far above the 3.4% recorded before the conflict began. A household that feels poorer and expects its money to buy less is a cross-pressure on spending, and both forces trace to the same source: the petrol pump.

Two inflations, one mood

The pump explains the divide between what households feel and what economists measure. Consumer prices overall rose 3.4% in the year to August. But that average hides a split that is now doing the real work in the economy. Strip out food and energy, and the "core" measure the Federal Reserve watches sits at 2.4%—close to target and drifting down. The energy index, by contrast, is up 16.3% on the year; petrol alone jumped 3.9% in August and 27.4% over twelve months. Households do not read core inflation. They read the price on the forecourt, and it tells them prices are running away.

This is the textbook definition of a supply shock, and it poses a textbook problem for the central bank. The Fed's settled doctrine is to "look through" such shocks—to tolerate a temporary burst in headline inflation rather than respond to a jump that tightening cannot reverse, because it reflects a shortage of a commodity, not an overheating economy. That looks wise when core inflation sits below the annual price rises of the past two years.

The doctrine carries a precondition, however. It works only so long as longer-term expectations stay anchored. And the number the Fed actually polices—the survey's five-year expectation—ticked up from 3.3% to 3.4% in September, its first rise in four months after holding steady through the summer. The move is small. The direction matters more than the size. A 3.4% expectation is well above the 2.8–3.2% band that prevailed through 2024, a reminder of how far the anchor has already drifted from the 2% the Fed claims as its goal. One-month moves in the one-year reading are noisy, driven by the price on the forecourt; a drift in the five-year reading is a signal that the shock is beginning to rename people's baseline. That is precisely the contamination the doctrine forbids.

A central bank in a bind

The consequence is a central bank squeezed from both sides. The Fed has held its key rate at 3.50–3.75% since March. In July it chose patience, but only narrowly: the decision went 9–3, with three members voting to raise rates to contain price pressures. That internal split reflects the new problem. The political and electoral pressure to cut is obvious—sentiment near a record low is not a backdrop in which tight money wins friends—yet the inflation expectation argues just as forcefully against easing. Indeed the median official projection for year-end inflation has already been revised up to 2.7% from 2.4%. The squeeze is visible at the household level too: only 8% of consumers now expect their incomes to grow faster than prices, down from 18% in late 2024.

For investors this collision matters because it sets the price of money, and therefore the multiple that equities command. Future earnings are discounted at whatever rate the Fed's path implies; a dollar earned next decade is worth less if today's policy makes that dollar's purchasing power doubtful. The two errors the Fed could make are not symmetric in their consequences. If it holds and the oil shock proves temporary, the damage is confined to a quarter of pinched spending and a headline number that recedes. The worse outcome is the opposite one: if the energy rise keeps feeding the five-year expectation, the Fed is eventually forced to tighten into a weakening economy to defend its credibility—the error everyone fears, because it turns a supply shock into a demand collapse.

The coming days answer the live question. The next policy decision lands within days of this survey, and it will be judged by whether officials treat the 4.6% one-year reading as the transient pump noise the see-through doctrine says it is, or the first sign that a hard-won anchor is slipping. Consumers cannot be talked out of the price on the forecourt; that is not the Fed's job. Its job is to keep that price from rewriting what people think five years from now. The sentiment print is the economy's complaint. The five-year number, and the committee's answer to it, will decide whether the Fed must choose which side of its remit to betray.