For most of 2025 the question in American economic policy was how fast interest rates would come down. By September 2026 the Federal Reserve was raising them again. The reason sits in the official statistics on prices, pay and spending, and in the Fed’s own account of what worried it.
The energy shock reached the pump
The disruption to shipping through the Strait of Hormuz since the end of February 2026 cut oil and gas exports from the Gulf and pushed up world fuel prices. American households felt it most directly in petrol.
The Bureau of Labor Statistics (BLS) consumer price index shows how concentrated the rise was:
| 12 months to September 2026 | Change |
|---|---|
| All items | 3.7% |
| Energy | 20.4% |
| Petrol (all types) | 34.2% |
| Electricity | 4.4% |
| Food | 2.5% |
| Shelter | 3.2% |
| All items excluding food and energy (“core”) | 2.6% |
The headline rate had been 2.7% in December 2025. It reached 4.2% in May 2026, eased, then climbed again to 3.7% in September as petrol rose 9.1% in that month alone. October brought a smaller monthly rise in prices, 0.4%, but no 12-month figure: BLS did not collect price data in October 2025 because of the lapse in federal funding, so it cannot compare the two months for most series.
The core rate is the striking part of the table. Prices outside food and energy were rising at a pace not far above the Fed’s 2% goal. This is what an energy shock looks like in its first stage: large moves in a few prices, rather than a broad rise in everything.
The Fed’s preferred gauge tells a less comfortable story. The Bureau of Economic Analysis (BEA) price index for personal consumption expenditures (PCE) was up 3.6% in the year to September, and its core version 3.1%. The two families of measures differ in what they cover and how they weight it, and the gap between them in 2026 is unusually wide. Which one better captures underlying inflation is not settled by the data.
Jobs held up; pay did not keep pace
The labour market has not cracked. In October 2026 payrolls rose by 165,000, the unemployment rate was 4.1%, and the two previous months were revised up. Over the preceding year, though, job growth averaged only 48,000 a month, so October was strong against a modest trend.
Pay is where households have lost ground. Average hourly earnings were $38.00 in October, 3.1% higher than a year earlier. With prices rising faster, BLS calculates that real hourly earnings fell 0.7% in the year to September. A worker whose pay rose in line with the average could buy slightly less with it than a year before.
Spending kept going, savings took the strain
Despite falling real pay, consumers did not retrench. BEA’s advance estimate shows the economy growing at an annual rate of 3.4% in the third quarter, with consumer spending among the contributors.
The monthly figures show how that was financed. In September, real disposable income fell 0.3% while real spending was flat; in August, spending rose 0.5% while real income did not grow. The personal saving rate was 3.8% in September. When spending holds up while real income falls, the difference has to come from saving less or borrowing more.
The New York Fed’s household debt report points the same way. Total household debt rose by $334 billion in the third quarter to $19.1 trillion, and credit card balances reached a record $1.29 trillion. The share of debt in some stage of delinquency rose to 5.0%. The picture is mixed rather than alarming: the rate at which card balances moved into serious delinquency edged down from a year earlier, while the equivalent rate for mortgages rose.
Why the Fed turned
The Fed cut three times between September and December 2025, taking its target range to 3.50–3.75%. It then held through the first half of 2026 as the energy shock unfolded. On 16 September 2026 the committee, now chaired by Kevin Warsh, voted unanimously to raise the range by a quarter point to 3.75–4.00%. Its statement was brief: “Inflation remains elevated”, activity was expanding at “a solid pace”, and the move would “support a timelier return” to 2% inflation.
The minutes of that meeting give the fuller reasoning. Participants said they had not seen enough progress in lowering inflation, and some worried that more than five years above 2% could start to affect expectations and price- and wage-setting. Measures of short-term inflation expectations were elevated. The worry was that price rises in particular sectors, energy above all, could broaden into “more persistent inflation dynamics.” At the same time, they judged the labour market close to maximum employment, which removed the main argument for caution.
The mechanism matters here. Interest rates cannot make oil cheaper. What they can do is slow demand and signal resolve, so that businesses and workers do not build today’s energy prices into tomorrow’s prices and wage demands. The September rise reads as insurance against that second-round effect rather than a response to the oil price itself; this is our interpretation of the minutes.
In October the committee held the range at 3.75–4.00%, but two members, Lisa Cook and Beth Hammack, voted for another rise. The statement upgraded its description of growth from “solid” to “strong”.
What it means, as mechanisms
For households. Energy takes a larger share of the budget of lower-income families, and the Fed’s minutes record that low- and moderate-income households faced strains, while spending was supported by stock market gains concentrated among higher earners. Higher policy rates pass through to credit card and other variable-rate borrowing, adding cost for households already leaning on credit. Mortgage rates were already high, and Fed participants noted that financial conditions were not supporting housing.
For companies. Firms that sell to consumers face customers with less real income but still spending. Firms with heavy fuel costs, such as transport, see margins squeezed unless they can pass the cost on, and passing it on is exactly what the Fed is watching. Fed participants also described financial conditions overall as supportive, with narrow corporate bond spreads, so the effect of a single quarter-point rise on borrowing costs is modest so far.
For markets. When a central bank moves from cutting to raising, the expected path of short-term rates shifts, which feeds into bond yields and into the rate at which investors discount future profits. Equity valuations that depended on falling rates face a different backdrop. How much of that has already been priced in is not something official data can tell us.
What would change the picture
Three readings will say most. First, whether core inflation, on either measure, starts to rise as energy costs feed through to other prices. Second, whether wage growth catches up with prices, which would ease the squeeze on households but add to the Fed’s concerns. Third, whether the saving rate and delinquency figures show households running out of room. The Fed next meets on 8–9 December 2026.

