This article is an on-site version of our Chris Giles on Central Banks newsletter. Premium subscribers can sign up here to get the newsletter delivered every Tuesday. Standard subscribers can upgrade to Premium here, or explore all FT newsletters
Shock greeted the latest US labour market data on Friday, as non-farm payrolls in August grew by only 22,000 jobs, far below expectations. Worse, June’s figures were revised further downwards to a 13,000 drop. Financial markets think the Federal Reserve is certain to cut rates in its meeting next week. And recessionary talk is back, for example from my colleague Tej Parikh in his Free Lunch newsletter.
What really matters for the Fed is not whether the monthly change in payrolls figure is high or low, but what the broader sweep of labour market data tells us about the US economy. When you have a weak payrolls number, there are four broad possibilities about what is going on: weak demand, a freak month, weak labour supply, or cooling in supply and demand.
Given the volatility under President Donald Trump, it is a brave person who would make a definitive call about where the American economy will go next. But what we are seeing has all the hallmarks of bad policy hitting labour supply, reducing jobs growth and even raising the level of unemployment consistent with stable and low inflation.
It certainly does not scream “loosen monetary policy now”, and the Fed will need to be careful when it cuts interest rates next week not to cut its credibility at the same time.
What is happening with jobs?
There is no doubt jobs growth has slowed. Although Trump on Friday held out hope that “corrections” are often made to economic data, saying “so many elements aren’t included yet”, the data points to falling employment growth in 2025. This rules out the freak monthly data option.
The trouble with non-farm payrolls is that, as an indicator, they fail spectacularly to distinguish between demand and supply. Payroll growth might be low because companies and households are holding back spending, or because there are too few candidates to fill job openings.
With spending holding up, there is little sign of severe demand weakness. (No prizes for guessing what the supply restrictions might be.) Tariffs, curbs on migration and government job cuts are highly plausible explanations for low jobs growth.
Some content could not load. Check your internet connection or browser settings.
Not only has jobs growth slowed sharply, the Bureau of Labor Statistics on Tuesday revised down the number of jobs in the economy in March 2025 by 911,000. This revision was based on its preliminary benchmarking of jobs data to its Quarterly Census of Employment and Wages, which uses administrative unemployment insurance tax records.
With the downward revision for March 2025, payroll growth each month between April 2024 and March 2025 was on average 76,000 lower than previously thought, shown in the chart below.
This time, however, we can be pretty sure this is an effect of supply, not demand. A big one. Other labour market data for the same period, including unemployment figures, has not been revised. So the best way to interpret the numbers is that in the year to March 2025, the US economy needed to create 76,000 fewer jobs per month to keep unemployment stable.
Some content could not load. Check your internet connection or browser settings.
US unemployment rose a tenth of a percentage point in August from July. But it is low on a historical basis, and has only edged higher by the same amount over the past year.
This is further powerful evidence to suggest we are not seeing demand weakness. At 4.3 per cent, the unemployment rate is within the central tendency of FOMC members’ estimates of the long-run stable level of unemployment, and bang on the Congressional Budget Office’s estimate of “non-cyclical unemployment”.
Both of these official estimates of equilibrium unemployment are probably out of date and should be higher to take account of the distortions Trump’s tariff and migration policies have imposed on the US economy.
Some content could not load. Check your internet connection or browser settings.
Remember a year ago? Then, the Fed and everyone else justified looser monetary policy on the basis of a sharp rise in the unemployment rate over the previous 12 months. Everyone started talking about the “Sahm rule” recession indicator which, after being triggered in July and August 2024, led to the Fed cutting rates by a half point in September.
The Sahm rule is designed to capture a sudden drop in demand, and is triggered when the three-month moving average of the unemployment rate is 0.5 percentage points higher than its minimum over the previous 12 months. The value is now 0.13, suggesting the labour market has cooled a bit over the past year. But that is nothing to be alarmed about, as the chart below shows.
This should be the final nail in the coffin for both the simple demand and freak data explanations of low payroll jobs growth. The benchmark revisions show a weakness in supply was dominant.
Some content could not load. Check your internet connection or browser settings.
Limiting analysis to just payrolls and the unemployment rate can give a misleadingly narrow view of the labour market. The good news is, when you look at all the data, a remarkably consistent picture emerges.
In the chart below, I have shown trends in the unemployment rate, the quits rate, job openings per person unemployed, and wage growth, comparing each series with its 2001 to 2019 average to assess how hot or cold it is compared with the past.
All of these measures are stronger now. And while they have cooled over the past year, most still lie outside one standard deviation from the mean. This suggests they have been weaker 84 per cent of the time this century. This is not a struggling labour market.
But it is a cooling labour market, hit by supply shocks and overheating less than in 2022 and 2023. Since wage growth is decelerating, it suggests one which is still warm.
Some content could not load. Check your internet connection or browser settings.
The Atlanta Fed goes one step further with its helpful, 16-variable spider diagram of labour market strength. That the orange line — August 2025 — is generally inside the green line — August 2024 — shows the data has cooled. The orange line is also generally outside the grey median values, underlining that the labour market is still hotter than normal, even with weak payrolls.
Put this all together and the best conclusion is that the US labour market has been hit by a big supply shock at a time when it has also cooled from a year earlier. This explains why payroll growth has deteriorated rapidly and other indicators remain in the warm zone. As Fed chair Jay Powell said in Jackson Hole last month, the labour market is showing “a curious kind of balance that results from a marked slowing in both the supply of and demand for workers”.
The Fed and the labour market
The Fed cannot justify an interest rate cut on September 17 on a divergence from its maximum employment mandate, since the unemployment rate is exactly at its estimate of the long-term equilibrium.
Instead, the justification will need to be that a cooling labour market with some risks of acceleration demands a slightly less restrictive monetary policy stance than now, when rates are in the 4.25 to 4.5 per cent range.
It is an argument the Fed will have no difficulty justifying now, but will be harder to repeat if it wants to start a series of rate cuts this autumn while the labour market remains warm. Payrolls should not feature in the central bank’s reasoning.
What I’ve been reading and watching
I spent a fascinating half-hour talking with Peter Conti-Brown of the Wharton School at the University of Pennsylvania about Fed history for the Economics Show podcast. Much of the conversation was spent discussing when threats to the Fed would cross a line.
On that subject, US Treasury secretary Scott Bessent said that Fed independence was under threat in an opinion article in the Wall Street Journal. But he blamed the central bank for its own predicament, saying it “jeopardised” its status by expanding beyond its remit. He called for it to scale back distortions, such as its enlarged balance sheet. With the US president wanting cheaper government borrowing, that was a bold call.
In his hearing before the Senate Banking Committee for the vacant Fed governor seat, Stephen Miran said he would not stand down from his White House role as chair of the Council of Economic Advisers.
Katie Martin made a wise distinction between genuine fiscal concerns in bond markets and silly talk of IMF bailouts. I suggested the UK was a relative saint, not sinner when it came to fiscal policy.
A chart that matters
Financial markets took a brief glance at the labour market figures on Friday and decided “payrolls bad, Fed cut rates”. The chart below shows the market-implied probability of different monetary policy settings for the next three FOMC meetings. A cut is fully priced into next week’s Fed meeting, with a small probability of a jumbo half-point reduction.
In October, markets think it more likely than not that there will be a further quarter-point easing of policy and rates will end the year in the 3.5 to 3.75 per cent range. You can click on the chart to see how the probabilities change, and visit Monetary Policy Radar for this data updated daily.
Some content could not load. Check your internet connection or browser settings.
Central Banks is edited by Harvey Nriapia
Recommended newsletters for you
Free Lunch — Your guide to the global economic policy debate. Sign up here
The Lex Newsletter — Lex, our investment column, breaks down the week’s key themes, with analysis by award-winning writers. Sign up here