Did inflation's progress survive the revisions?

On 9 February the Bureau of Labor Statistics published its annual benchmark revisions to the consumer price index, and bond traders watched nervously. A year earlier the revisions had wiped out much of the reported disinflation and set off a sharp sell-off. This time they brought relief.

Why a routine release mattered

Every February the BLS recalculates the seasonal factors it uses to strip out regular calendar patterns from monthly price changes, such as the start-of-year price resets and the seasonal swings in clothing and travel. The new factors are applied to the previous five years of data, so monthly inflation readings that markets have already traded on can change after the fact. Usually the adjustments are small and attract little notice. Last year they did not, and that made this year's release a test of whether the progress on inflation was real.

In February 2023 the revisions raised inflation in the final months of 2022, turning what had looked like a clear slowdown into something more stubborn. December's headline reading, first reported as a small fall, became a small rise. Coming alongside a strong January jobs report, the revision helped push Treasury yields sharply higher and convinced markets that the Federal Reserve would have to keep raising rates. It did so until July.

Revisions leave the trend intact

The revised data showed core CPI slowing in late 2023 by almost exactly as much as first reported, so the disinflation was real and owed nothing to faulty seasonal adjustment. Some monthly core readings were nudged up a little. The broader path towards price stability was unchanged. Central bankers and bond investors can take the disinflation story at face value.

How seasonal adjustment moves inflation

Seasonal adjustment does not change the total rise in prices over a year; it only redistributes it between months. If the new factors attribute more of the typical January increase to seasonal patterns, the adjusted January reading falls, and some other month's reading rises to compensate. That is why revisions matter most for short-term measures such as three- and six-month annualised rates, which the Fed and markets have been watching closely to judge whether inflation is converging on 2 per cent. The annual rate is unaffected.

The worry this year was residual seasonality, the tendency for inflation to look higher early in the year even after adjustment, perhaps because companies raise prices in January more than they used to. If the new factors had shifted inflation from the second half of 2023 into the first, the recent run of soft readings would have looked less impressive. They did not, which means the slowdown in the second half reflects genuine easing in price pressures rather than a statistical artefact.

A second routine change arrives with the January data. The BLS now updates the expenditure weights in the index every year, using a single year of household spending data, where it used to update them every two years. The new weights will slightly change how much each category, from shelter to used cars, contributes to the headline figure. Such shifts are usually small, but in a year when the Fed is parsing every tenth of a point they add another source of noise.

What the Fed was watching

Fed officials had flagged the release. At its meeting on 31 January the committee held rates at 5.25 to 5.50 per cent and said it wanted greater confidence that inflation was moving sustainably towards target before cutting. Jerome Powell said a March cut was unlikely, and a strong January jobs report with 353,000 new positions reinforced that. At least one Fed governor had said publicly that he would watch the revisions closely, given what happened a year earlier. Had the data been rewritten upward, the case for waiting until the summer or later would have strengthened.

The Fed's preferred measure, the personal consumption expenditures price index, tells a similar story. Core PCE inflation fell to 2.9 per cent in December, and over the past six months it ran below 2 per cent at an annualised rate. The CPI revisions give no reason to doubt that picture, since the two indices share much of their underlying price data.

Winners from the clean result

Holders of Treasuries were the immediate beneficiaries, as one source of upside risk to yields disappeared. Borrowers waiting for lower mortgage rates gain from anything that keeps the path to rate cuts intact. Equity investors, whose rally since October has rested on the expectation of easier policy, avoided a setback. The Fed itself benefits from a cleaner signal, since it can respond to incoming data without worrying that its judgements rest on numbers that will later be changed.

Where doubt remains

The counterargument is that the revisions settle only one question. January inflation is due next week, and it has run hot in several recent years as firms reset prices. Shelter costs are slowing more gradually than rents on new leases suggest they should, and service prices outside housing remain firm. A strong labour market, with wage growth near 4.5 per cent, could keep services inflation elevated. The revisions show that the progress in late 2023 was real, not that it will continue.

The indicators to watch are the January consumer-price report on 13 February, the January PCE figures at the end of the month, and the Fed's updated projections in March.

The revisions confirm that late-2023 disinflation was genuine. One source of doubt has gone, and the Fed can keep edging towards cuts without fearing that the data will be rewritten beneath it, leaving next week's January figures as the real test of whether a cut comes by June.