(The Center Square) – Federal Reserve Chair Kevin Warsh used his first Jackson Hole speech to make his position plain: the economy is strong, the labor market is stable, and inflation is still the problem that deserves the Fed’s attention.
On the substance, he is half right – but the missing half is the part that matters.
Start with what he gets right. Unemployment is 4.1% and has barely moved in two years. Business investment is growing at its fastest pace since 2021, profits at the largest firms are up more than 20%, and consumers are still spending. Warsh said he would be hard pressed to call financial conditions restrictive, and on the average he has a point.
But an average can hide as much as it reveals.
More than half of this year’s investment growth, Warsh noted, comes from the buildout of artificial intelligence – the data centers, chips and power needed to run them. He also acknowledged that housing and agriculture are under strain.
That is not a broad expansion. It is a narrow one.
Homebuilding is having its weakest summer since 2020: July starts fell 14% from a year earlier and single-family starts were down nearly 17%. Real GDP grew 1.5% annualized in Q2, and equipment plus intellectual-property investment accounted for roughly 1.2 percentage points of that growth. Everything else combined contributed just 0.3 percentage points. On a year-over-year basis, real GDP growth slowed from 2.7% in Q1 to 2.1% in Q2.
There is another reason not to lean too heavily on the AI boom: more of it now runs on borrowed money. Goldman Sachs estimates roughly a third of hyperscaler capital spending will be debt-financed this year, alongside a fast-growing layer of project finance, private credit and off-balance-sheet deals.
That debt has to be serviced whether or not the bet pays off.
And that makes productivity the central question of this expansion.
Population growth has slowed and labor-force growth is slowing with it. Without stronger productivity, slower growth in the number of workers eventually means slower growth in the economy’s productive capacity.
We are adding capital at extraordinary speed. The productivity payoff has yet to show up.
Total factor productivity – the “doing more with the same workers and machines” part of growth – slowed from roughly 1.5% in 2024 to 0.8% in 2025.
That is the bet. With population and labor-force growth slowing, stronger productivity is no longer just upside. It is what we need to bring inflation down without crushing demand, raise real incomes and grow the economy fast enough to make today’s debt burden easier to carry.
As I argued in "Resilient, but Narrow", we are already financing the investment. What remains uncertain is how large the payoff will be, how quickly it will arrive and who will capture the gains.
The labor market tells the same story of a strong headline resting on a softer foundation.
Warsh explained today’s unusually low hiring, quitting and firing partly as the aftermath of the “rematching” that happened when workers changed jobs in droves after the pandemic. There may be something to that.
But job openings have fallen sharply from their post-pandemic highs. Labor-force participation has fallen to 61.4%, the labor force itself has contracted over the past year, and real hourly earnings have slipped as nominal wage growth failed to keep pace with prices. Recent graduates are taking longer to find work.
This may be a comfortable labor market if you already have a job. It is much less comfortable if you need one.
On inflation, Warsh is right to resist declaring victory. More than half of the PCE basket is still running above 3%, and inflation has been above the Fed’s target for more than five years.
But his own evidence cuts both ways: medium-term inflation expectations remain stable. The case for vigilance is strong. The case for urgency is weaker.
And some of today’s inflation is coming from sources higher interest rates cannot directly fix. Energy shocks raise prices while reducing households’ purchasing power. Tariffs raise import costs while also weakening demand. The Fed can stop those shocks from spreading, but it cannot produce more oil or remove a tariff.
The one place Warsh has a real point is the AI boom.
Credit spreads remain narrow, issuance is strong and financing continues to flow toward the firms at the center of the investment surge. There is little evidence that today’s borrowing costs are restraining the part of the economy doing the most investing.
The restraint is showing up elsewhere: housing, entry-level hiring and household budgets.
And this is where the Fed’s tool shows its limits.
It has a single, blunt instrument, and it cannot lean against the AI credit boom without leaning on everyone else.
Rates high enough to discipline exuberant borrowing around AI also fall on the priced-out homebuyer, the graduate who cannot find work and the worker whose paycheck is slipping behind – none of whom shared equally in the boom.
Warsh has the right instinct that bringing down inflation should be the Fed’s top priority. But the Fed is setting policy for the average American in an economy where the average has stopped describing almost anyone.
The strength is real. So is the weakness.
They are just in different places – and only one of them shows up when you look at the mean.
