Why core inflation stopped falling

Headline consumer prices rose 4.9 per cent in the year to April, the first reading below 5 per cent in two years. Bond traders and central bankers were not cheered. Core CPI, which strips out food and energy, came in at 5.5 per cent, and its three-month and six-month annualised rates have gone nowhere since January, stuck in a band between 5.0 and 5.5 per cent. The easy part of disinflation is over. What remains is services inflation, which is much harder to shift.

Anyone who looked at the components of the index could have seen the stall coming. The fast disinflation of late 2022 came almost entirely from a sharp fall in core goods prices, as shipping costs returned to normal and retailers cleared surplus stock with sales.

Goods have done their bit

By April 2023 that effect had largely run out. Goods prices flattened, and the monthly drag they had exerted stopped hiding the strength of domestic services.

April even brought a reversal. Used-car prices, which had fallen for much of last year and did more than any other single item to pull core inflation down, jumped again as wholesale auction prices recovered earlier this spring. Core prices rose 0.4 per cent in the month, the same pace as in March, which works out at close to 5 per cent a year.

Services carry the load

Core services excluding housing kept rising at more than 0.4 per cent a month, pushed along by steady wage growth in healthcare, professional services, auto insurance and hospitality. Shelter inflation may be peaking, but it feeds into the official indices very slowly.

The labour market behind those services prices has barely loosened. Employers added 253,000 jobs in April, and unemployment fell to 3.4 per cent, matching its lowest since 1969. Average hourly earnings grew 0.5 per cent in the month. For a hospital, a hotel or an insurer's claims department, where pay is most of the cost, that is a steady upward push on prices that no improvement in shipping can offset.

Car insurance, a small item with a large recent contribution, shows how the lags work. Premiums are rising at double-digit annual rates because the cost of claims jumped first: used cars became dearer to replace, spare parts and repairs cost more, and mechanics' wages rose. In many states insurers need regulatory approval before raising rates, so premiums catch up with claims costs only a year or more after the event. Medical services follow a similar pattern, with many prices set in contracts that are renegotiated annually and so reflect last year's wage and cost increases. Inflation of this kind looks backwards by construction, and it will keep rising for some time after the shocks that caused it have faded, whatever the Fed does to demand in the meantime.

Shelter will help, slowly

The optimists' best argument is housing. Rents on newly signed leases slowed sharply over the past year, and because the official index averages rents across all existing leases, that slowdown will reach the index gradually through the rest of 2023. Shelter is the largest single item in core CPI, so its decline alone will take a meaningful slice off the core rate.

That is a reasonable forecast, and it explains why the Fed expects core inflation to ease. It does not touch services outside housing, which are the part of the index most closely tied to wages. A core rate that falls because rents slow, while labour-intensive services keep rising at an annualised 5 per cent, would leave the Fed's problem unsolved.

The hard yards

That leaves the Federal Reserve in a corner. Getting inflation from 9 per cent to 5 per cent was comparatively easy, since supply-chain bottlenecks cleared and global commodity prices fell.

The Fed raised rates again this month, to a range of 5 to 5.25 per cent, and dropped the language promising more increases. The risk in that pause is that it is read as the end of the fight. Tighter bank lending after the regional banking failures may do some of the work, but credit conditions take months to affect hiring.

A lesson from the 1970s

The period the Fed most wants to avoid is the mid-1970s. Inflation fell sharply after the 1974 recession, the Fed eased, and prices accelerated again later in the decade to rates higher than before. By the time Paul Volcker arrived in 1979, the Fed had to push rates far higher than would have been needed had it held its nerve earlier.

That history explains why officials worry more about a stall at 5 per cent than about the level itself. Inflation that stops falling well above target tends to become embedded in wage settlements and price-setting, and it then takes a recession to shift it.

Getting core inflation from 5.5 per cent to the 2 per cent target will need something far more painful: weaker demand for services at home and a looser labour market. The next test comes with the May consumer price figures, released on the day the Fed's June meeting begins, and the committee's new projections at the end of it. Goods disinflation has run its course, and core inflation will probably stay above 5 per cent until the Fed accepts the slowdown it has so far avoided. Getting to two per cent means cooling the jobs market, and the Fed has yet to show it is willing to do that.