A bad jobs report

Unlock the Editor’s Digest for free

Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.

Good morning. The war in the Middle East is entering its second week with few signs of abating, and oil has reached $100 a barrel. The higher energy prices won’t please consumers — or the Federal Reserve, whose two mandates are increasingly in tension. Send us your thoughts: [email protected] 

A soft but stable job market  

The February jobs report was bad. I’m especially sorry to write that, because less than a month ago I was happy to write that “the jobs report from January was good”. The fact that we can go from one end of the spectrum to the other in 24 days should tell us something about taking any one month too seriously. Tracking the level of employment in a huge economy is hard, and it’s the trend, not the point estimate, that matters. 

Still, what an absolute clunker of a report this was. The unemployment rate nudged up to 4.4 per cent, there was a net decline of 92,000 jobs and almost every sector was weak. Healthcare, which has been a job-creation mainstay, was dragged down by a strike that accounted for about a third of the job losses, but even adjusting for that numbers look terrible. There has now been no net job creation since May. 

The six-month moving average (the blue line below) shows that job creation is stable — at around zero.

As a measure of cyclical activity in the job market we like to look at private employment without healthcare and social assistance. On that front, you can see an improving trend — towards less job losses — in the six-month rolling average, but you have to squint a bit.

Finding something to be hopeful about in this report is possible, just. First, blame the big recent swings on the eccentricities of the Bureau of Labor Statistics’ birth/death model, which is designed to account for the opening and closing of new businesses. That allows you to take the average of January and February, excluding the January strike, and conclude that the economy added 33,000 jobs a month in the first two months of the year, mostly driven by healthcare. 

Finally, have a look at the growth rate of hourly wages, which has averaged 4 per cent annualised over the past six months, and was steady in February. That, paired with low lay-offs and a stable unemployment rate, suggests that labour demand, while low, is stable. That, and the threat of war-driven inflation, will probably keep the Fed policy on hold, at least until the composition of the Federal Open Market Committee changes in May. 

AI and productivity

Well, this looks pretty good:

Some people look at the recent acceleration in productivity growth and get positively giddy about our future as animal-machine work Minotaurs. Here is the information economist Erik Brynjolfsson writing about AI in the FT a few weeks ago:

This [acceleration in productivity] aligns with the productivity “J-curve” that my colleagues and I have explored in earlier research. General-purpose technologies, from the steam engine to the computer, do not deliver immediate gains. Instead, they require a period of massive, often unmeasured investment in intangible capital — reorganising business processes, retraining the workforce and developing new business models. During this phase, measured productivity is suppressed as resources are diverted to investments. The updated 2025 US data suggests we are now transitioning out of this investment phase into a harvest phase

We’re all reading the same anecdotes and case studies about AI making work happen faster, but this seems a little early to me. And as Jason Furman — who believes that AI is probably adding something to productivity now — points out, annual productivity growth in this cycle (which he dates to the end of 2019) is the second best we have seen in half a century (after the 2001-07 cycle), but is not extraordinary. Indeed, the current trend looks about like what we were seeing in 2018-19. Certainly it makes sense to me that we would have a good cycle after the workforce went through a massive reshuffle in Covid, at the same time as consumer demand and investment have remained strong.

Others are less cautious in their scepticism. Here is Dario Perkins, head of macro at TS Lombard:

There is no evidence that AI deployment is either boosting productivity or damaging US employment . . . While US productivity has been strong and hiring weak, our analysis finds that cyclical forces — not automation — are to blame. Since Liberation Day, US companies have had to endure a significant tariff-induced margin squeeze. They have also faced extreme policy uncertainty. Their response was predictable: they hired fewer people and pushed their existing workforce harder . . . That outcome was possible primarily thanks to the inefficiencies that appeared after the pandemic, when all the focus was on headcount (owing to severe staff shortages) and following a period of exceptionally high worker turnover . . . 

I’m in the Perkins camp. We have a pretty good understanding of why we are in a sluggish job market (Covid hangover, bad trade policy, immigration control) and why the economy is growing (high business investment, strong household balance sheets supporting consumption). Combine those things, and you get higher productivity. AI may well change everything about the economy, but it’s probably not doing it quite yet. 

One good read

Spring cleaning.

FT Unhedged podcast

Can’t get enough of Unhedged? Listen to our new podcast, for a 15-minute dive into the latest markets news and financial headlines, twice a week. Catch up on past editions of the newsletter here.