One good inflation report does not make a trend

Headline US inflation came in at 7.7 per cent for October, down from 8.2 per cent in September, and core inflation eased to 6.3 per cent. The figures, released on 10 November, set off the fiercest global rally of the year. The S&P 500 jumped 5.5 per cent, its best day since the worst of the 2020 pandemic, and two-year Treasury yields fell by nearly thirty basis points. Traders celebrated as if price stability had returned overnight. They have mistaken a single month of slower inflation for the end of an inflationary regime.

Markets have spent the past eighteen months spotting inflation peaks that turned out not to exist. Each softer monthly print is greeted by macro traders as the signal that the Federal Reserve will stop tightening and go back to expanding its balance sheet. That misreads how the Fed works. Central bankers do not set policy on one data point. They watch the persistent momentum of core service prices, and on that measure the fight is far from over.

Down from the peak, far from target

Context matters. Headline inflation peaked at 9.1 per cent in June, the highest in more than forty years, and core inflation reached 6.6 per cent as recently as September. The Fed raised its policy rate by 75 basis points on 2 November, the fourth such increase in a row, to a range of 3.75 to 4.00 per cent. Even after October's relief, core prices rose 0.3 per cent in the month, a pace that would still run at more than three per cent a year, well above the Fed's 2 per cent target.

Officials were quick to say so. Within days Christopher Waller, a Fed governor, said that markets had run well ahead of a single report and that the central bank still had a long way to go. That is the message a committee sends when it wants to take back an easing of financial conditions it did not sanction.

Used cars did the work

A cool look at the October report shows that most of the relief came from a few volatile goods categories. The biggest contributor to the downside surprise was a 2.4 per cent monthly fall in used car and truck prices. There was also a sharp slowdown in medical care services, but that came entirely from an annual accounting adjustment to health insurance, not from cheaper doctors' bills.

Used car prices soared to absurd premiums during the chip shortages of 2021 and are now falling back towards normal depreciation as dealer inventories recover. That is supply chains healing, and it is welcome. It is also a one-off. Once used vehicle prices are back near their pre-pandemic level, the drag they exert on the headline index will disappear.

The health-insurance quirk is also temporary in its own way. The statistical agency measures insurance prices indirectly, from insurers' retained earnings, and refreshes that estimate once a year. The October update will subtract from the index each month for a year, flattering core readings without any change in what patients pay for care, and then drop out.

Rents and wages keep rising

Core services excluding energy, the real engine of underlying inflation, kept compounding at an uncomfortably steady pace. Shelter, nearly one-third of the consumer price basket, rose 0.8 per cent on the month, the biggest monthly increase since 1990. Market measures of new apartment leases have started to cool, but the official shelter index is collected on a rolling twelve-month basis, which all but guarantees high readings well into 2023.

Service prices outside shelter still move with wages. Unemployment is 3.7 per cent and average hourly earnings are growing at an annual rate of 4.7 per cent, so service businesses have both the reason and the pricing power to pass higher labour costs on to customers. A central bank that loosens financial conditions too early risks letting that wage-price loop settle into permanent inflation expectations.

The 1970s false dawn

The historical warning is familiar to every central banker. After peaking above 12 per cent in 1974, American inflation fell below 5 per cent by the end of 1976, and the Fed eased. It climbed back into double digits before the decade was out, and it took the far harsher tightening under Paul Volcker to break it. The Fed's current leadership has cited that sequence repeatedly as the mistake it most wants to avoid.

The lesson is about persistence, not levels. In the mid-1970s the fall in inflation came mostly from energy and food, while wage settlements kept running hot, much as goods prices are doing the work today while service costs keep climbing. A committee that remembers that episode will want several months of slower core services inflation before it treats the problem as solved.

Where the optimists have a point

The bullish case deserves a fair hearing. Goods prices are now falling across a broader range of items than used cars, shipping costs have collapsed from their pandemic highs, and private measures of new rents point to shelter inflation turning down during 2023. If those trends feed through, core inflation could fall faster next year than the Fed's own projections assume.

That is plausible, and it is also a forecast about next year, not a description of this one. Markets priced it as settled on the strength of one release. The November figures, which arrive on the first day of the Fed's December meeting, will show whether October's easing was the start of a trend or another false peak.

The rally after the October CPI print was an unhedged bout of collective self-indulgence. By lifting equity multiples, narrowing corporate bond spreads and knocking twenty-five basis points off expected peak rates, traders eased financial conditions against the Fed's wishes, and the Fed is likely to answer by talking the peak rate back up at its next meetings. Policymakers will not let a market party undo their progress on inflation, and they will probably keep rate expectations higher for longer until the services trend clearly breaks, which on the evidence of rents and wages is unlikely before the middle of next year.