The stubborn inflation the Fed can't shake

Core PCE inflation ran at 4.6 per cent year-on-year in May, and Federal Reserve officials can think of little else. Headline figures have drifted lower as energy prices fell and supply chains returned to normal. Core services have barely budged. Goods prices have done most of their falling, which leaves a core driven by what service businesses pay their staff.

A skip, not a pivot

The Fed raised its policy rate at ten consecutive meetings between March 2022 and May of this year, taking the target range from near zero to 5 to 5.25 per cent. In June it held steady for the first time in more than a year, yet the committee's own projections pencilled in two further increases before the year is out. The combination is telling. Officials want time to watch the delayed effect of the fastest tightening since the early 1980s, and they are plainly not persuaded that the job is finished.

The May numbers explain the hesitation. Headline PCE inflation fell to 3.8 per cent, well below last summer's peak, as petrol and food prices cooled. The core measure, which strips out both and is the one the Fed's 2 per cent target is judged against in practice, has hardly moved since January. A central bank that looked only at the headline would be tempted to declare victory. One that looks at the core sees a problem that has not yet started to shrink.

Three kinds of price

Jerome Powell has taken to splitting core inflation into three parts: goods, housing services, and other services. Goods inflation surged when locked-down households bought furniture and electronics and ports jammed, then subsided as shipping costs collapsed and inventories piled up. Housing inflation is still high in the official figures, but it follows leases that reset only once a year, so the cooling already visible in new-lease rents will show up in the index only gradually. That leaves the third bucket, services outside housing, as the part of the basket that will decide how long rates stay high.

Wages set the pace

Services outside housing are labour-heavy, and their prices ignore commodity cycles and container freight rates. They move with annual pay reviews and with the pricing power of the firms selling them. Demand for services is not very sensitive to price, so businesses have passed higher wage bills on to customers without much trouble. Price rises feed pay claims, and a modest policy tweak will not break the cycle.

The labour market gives that cycle plenty of fuel. Unemployment stood at 3.4 per cent in January, the lowest since 1969, and has stayed not far above that level since. Restaurants, hotels, care homes and airlines spent much of the past two years struggling to fill vacancies, and firms in that position raise pay first and ask about productivity later. A haircut or a hospital visit cannot be imported from a cheaper country or produced by a machine, so the extra cost goes straight into the price.

Why the decline is slow

In statistical terms, services inflation behaves like a series with high persistence: each month's reading inherits most of the previous one, and shocks fade over quarters rather than weeks. Goods prices behave differently. Once a supply bottleneck clears, they can fall back almost as fast as they rose, which is exactly what happened to used cars and freight. A model fitted to the goods experience of the past year would therefore badly overstate how quickly the rest of the core will come down.

The persistence has a simple cause. Wage contracts, rental agreements and insurance premiums are set for a year or more, and each round of increases is benchmarked against the inflation of the round before. Expectations do the rest. If a firm believes its competitors will raise prices by 4 per cent next year, raising its own by less looks like leaving money on the table. Breaking that chain requires enough slack in the economy that firms lose customers when they push prices up.

The 1970s lesson

The Fed's reluctance to call time owes a good deal to history. In the mid-1970s the central bank under Arthur Burns eased once inflation dropped from its first peak, only to watch it return higher later in the decade. It fell to Paul Volcker to crush it with rates near 20 per cent and a deep recession. Powell has said repeatedly that history warns against loosening policy too early, and the stickiness of services prices is precisely the kind of evidence that makes that warning feel current.

The comparison has limits. Inflation expectations in bond markets and household surveys have stayed far better anchored than they were in the 1970s, and the Fed has moved much faster this time. Even so, the institutional memory matters. Officials would rather hold rates high for a few quarters too long than repeat the Burns error and have to start again from a worse position.

Who pays for patience

A policy rate of 5 to 5.25 per cent against core inflation of 4.6 per cent leaves real rates only modestly positive, which is why the Fed is inclined to keep tightening or at least keep rates where they are. Borrowers on floating rates bear the cost first: small businesses with bank credit lines, private-equity-owned companies with leveraged loans, and owners of commercial property whose loans need refinancing. Regional banks, already rattled by the failure of Silicon Valley Bank in March, face both higher funding costs and a weaker commercial property book.

Savers are the obvious beneficiaries, since money-market funds now pay more than inflation for the first time in years. Labour-intensive service companies sit awkwardly in the middle. They can raise prices while demand holds, but they are the firms most exposed if the Fed eventually engineers the slowdown in hiring it wants.

The case for faster relief

The strongest objection to this view is that policy works with long lags, and the full weight of last year's increases has not yet reached hiring decisions. Job openings have fallen from their peak, fewer workers are quitting to chase higher pay, and wage growth has eased somewhat from last year's pace. If those trends continue, services inflation could slow faster than persistence models suggest, and the Fed would risk overtightening into a recession it did not need.

That argument deserves weight, but it describes a direction rather than a destination. A labour market with unemployment still below 4 per cent is not yet the slack the Fed needs, and each monthly core reading so far has shown little of the turn that the optimists expect.

The indicators to watch are the monthly jobs report, the quarterly employment cost index, and the services-excluding-housing component of the PCE data. Until the labour market cools enough to slow pay growth, core inflation is likely to sit well above the Fed's target, whatever the headline figures say. The most probable result is at least one more rate rise this year and no cut before well into next year, with floating-rate borrowers and commercial property owners carrying most of the cost.