The March Jobs Report and Time

Let me be straight: the March jobs report was much better than I had expected. I always give my wife my predictions just before the report comes out, because that’s when I have all the data I’m going to have. My pre-release numbers were 40k jobs and 4.5% unemployment. The actual numbers were 178k jobs and 4.3% unemployment. That’s a big miss.

Okay, so what did I get wrong? Starting with the unemployment rate, fans of the household survey know that the monthly data are erratic. A slide of 0.1 percentage point could just be sampling error. In fairness, carrying it to the next decimal, the drop from February to March was almost 0.2 percentage points. That’s not huge, but large enough to be real.

The other data in the survey also was mostly positive, there was a drop of 0.6 percentage points in the unemployment rate for Black workers to 7.1% and 1.0 percentage points in the unemployment rate for young workers, between the ages of 20-24, to 6.4%. I and others had pointed to the sharp rise in unemployment in 2025 for these disadvantaged groups as an early warning sign for a deterioration in the labor market.

This turnaround, which together with improvements in the prior three months, largely reversed the 2025 rise. That is unambiguously good news.

There was also a rise in the share of unemployment due to quits from 11.4% to 12.4%. That’s still low for an economy with 4.3% unemployment, but this measure of workers’ confidence in their employment prospects was looking healthier in March than through most of the last two years.

I could go through more items, but most of the data in the household survey tell a decent story of the labor market. It’s not great, we were down at 3.4% unemployment two years ago, with an unemployment rate for Black workers of 4.8%, but the March household survey does not support a story of a continuing deterioration.

The Jobs Numbers

While the household survey wins a clear “better than expected” seal of approval, most of the focus was on the 178k job growth in the establishment survey. All of us nerds pointed out that 30k of this growth was due to the return of workers in the healthcare sector, who were on strike in February. That’s true, but I knew about the return of striking workers when I told my wife 40k. So, what’s the deal with the other 138k?

Much of the story continues to be the category healthcare and social assistance, which accounted for 89,900 of the new jobs in March, more than half the total. But construction added 26,000 jobs in March, a number almost equal to its total in the prior year. Restaurants added 21,500 jobs, but this was after losing 26,200 jobs in February.

And manufacturing added 15,000 jobs (airplanes and metals, for those wondering), the largest monthly gain in over two years. Local governments added 14,000 jobs and retail added 9,700, so we had decent growth across sectors.

But there is another part of this story that got less attention. The index of aggregate hours worked fell by 0.2% in March, as a reduction in the length of the average workweek more than offset the increase in the number of people working.

To be clear, this is just one month’s data, and the hours data are highly erratic, but there does seem to be an interesting picture here going back over the last two years. Taking the aggregate data first, from January of 2023 to March of 2024, hours were rising at a decent pace, increasing by just under 0.9 percent, the equivalent of 1.5 million jobs or 100,000 a month. After March of 2024 aggregate hours essentially flatlined increasing by less than 0.3% over the last two years. In fact, it has literally flatlined over the last 11 months; the index for March of 2026 was the same as the index for April of 2025.

This is also the story with most major sectors. Education and healthcare is the major exception, but even here there has been some slowing. From January 2023 to April 2025, hours worked increased by 7.1%, a 2.9% annual rate. Since then, they have increased at less than a 1.0% annual rate.

There is a similar story with the leisure and hospitality sector, the other big gainer. Hours worked had risen by 2.2% as of April 2025, a 0.9% annual rate. In March, hours worked were actually down slightly from the April level.

Hours worked in the huge professional and business services sector, which employs more than 22 million workers, has been trending downward over this whole period. The March 2025 measure was 1.7% lower than the January 2023 level.

This is the same story with manufacturing, although the rate of decline has slowed in the last year. Hours worked is now 3.4% lower than its level 3 years and 3 months ago.

What Do Hours Worked Tell Us?

I see three takeaways from the stagnation of hours worked over the last two years, and especially the last year. First, the demand for labor is weaker than is generally recognized. Job growth has been slow, but the growth in hours has been even slower. Employers have less need for labor.

The decline in hours worked matters also in workers’ paychecks. The average weekly wage actually fell slightly in March. This was due to both weak hourly wage growth and a decline in hours. This is bad news for workers struggling to keep pace with inflation.

The second is a positive story. Insofar as we continue to see GDP growth even as hours worked stagnates, it is driven by productivity growth. If GDP rises 2.0% year-over-year, with no change in hours worked, that means productivity has risen by 2.0%. That’s not earth-shaking; we had 3.0% annual growth from 1947 to 1973, but it’s much better than the 1.0% rate we saw in the decade following the collapse of the housing bubble. AI may be part of the story here, but this is not the mass collapse of employment that some have warned us about.

The third point is that employers often cut back hours before laying off workers. That story is not always true, but we should see the weakness in hours as a warning.

One last point, there is nothing wrong with people working fewer hours insofar as this is their choice. Workers in West Europe put in 20-25% fewer hours a year on average than workers in the United States. This is because they have far more vacation, paid family leave, and paid sick days. Taking the benefits of productivity growth in more leisure is a perfectly reasonable thing to do and was a major demand of the labor movement for more than a century.

But We Have a War

Anyhow, I may be twisting too much to analyze data that are likely becoming irrelevant due to the Iran War. The survey period only covers the first half of March, a point at which the rise in gas prices was still limited, and many hoped the war would soon be over.

As the damage from the war increases, prices rise further, and there is still no end in sight, economic calculations will change. The jobs picture for April is likely to look very different.

This first appeared on Dean Baker’s Beat the Press blog.

The post The March Jobs Report and Time appeared first on CounterPunch.org.