Jobs Report: When the Numbers Get This Jumpy, Ya Gotta Look At The Trend

Plus, real concerns about real wages.

Well, it’s jobs day again. Funny how that keeps happening.

Payrolls were up a much-higher-than expected 178,000 in March, more than twice expectations for around 65,000. The unemployment rate ticked down to 4.3%, though, as I’ll explain in a moment, for “bad” reasons. Revisions took Feb’s payroll loss down even further than the initial report, from -92K to -133K. January was revised up, so on net, revisions led to a small decline of 7K in Jan/Feb payrolls.

Part of that big swing from a big Feb negative to a big March positive was due to a strike: nurses were out in Feb; back in March.

But look, folks: in reality, labor markets tend to run a lot less jumpy than what you see below in ‘25 and ‘26 so far, moving from losses to gains like that. Some of this is sampling noise as the payroll survey grapples with measurement issues generated by immigration changes, changes to the way they measure business births and deaths, and so on. These estimates are by no means “wrong.” They’re just noisy, monthly, sample-based estimates.

In such cases, you know my methods: Extract a smooth trend. I do so below and find the underlying pace of payroll gains to be around 30,000. That’s at least, if not ahead, of the breakeven number: the number of jobs needed to employ enough labor market entrants to keep the unemployment rate from rising.

If you’re thinking, “aha—that must be why the unemployment rate ticked down from 4.44% to 4.26% in March,” I must annoyingly tell you: good intuition, but nope (sorry). The jobless rate and the payroll counts come from different surveys and in any given month, they can disagree for all sorts of reasons (see sampling points above). In March, the unemployment rate ticked down because the labor market contracted, not because more people left unemployment for jobs. Consider this more noise; the point is unemployment remains relatively low and pretty stable in the 4%-4.5% range.

It is however, about a point above its low of 3.4% three years ago in April ‘23. And between this and the very low hiring rate we see from other BLS surveys, we can assume that workers have less bargaining clout than was the case back then. The figure below shows nominal yearly wage growth for the 80% of workers that are blue-collar production workers or non-managers in services—think mid/low wage growth. It’s been sliding down for a while now, first due to post-pandemic normalization and more recently due to the weaker labor market which dampens both workers bargaining clout and their ability to upgrade their jobs (quit rates boomed for this reason in the very strong post-pandemic job market).

For awhile, this wage was growing at ~4% and inflation was in the 2.5% range, so real wage gains were >1%. At this point, war-induced inflation forecasts are in the 3-3.5% range, so the high-end of that forecast is uncomfortably close to the nominal wage pace. What I’m saying here is that the combination of the war pushing up headline inflation and the already weakened labor market hitting nominal wage growth looks to me like it’s leading to a kind of real wage stagflation. That could be short-lived, but it’s worth watching. If the buying power of people’s paychecks takes a hit, that’s going to be bad for them and their families, as well as for aggregate consumer spending in the overall economy.

A few other notes:

—Monthly noise not withstanding, the March payroll report showed the 57% of private industries added jobs, compared to 49% in Feb.

—The end of the strike contributed to 76,000 jobs added in healthcare. The sector continues to be a reliable driver of job gains.

—The federal government continues to shed jobs, and this one is a trend, not a blip. Employment in the sector was down 18,000 last month, and, according to the Bureau, “since reaching a peak in October 2024, federal government employment is down by 355,000, or 11.8 percent.”

—Factory employment popped 15,000 in March but here again, the trend is not our friend.

—With low hiring, it’s harder to get out of unemployment. The share of the unemployed who have been so for at least half a year was 25% last month, up from 21% a year ago.

BL not UF, the job market is even harder to read than usual but it’s at least okay in a snapshot-from-last-month sense, and the underlying pace of payroll gains, while low in historical terms, is fine given the decline in labor supply. Yet there are more than the usual number of things to worry about. First, this read is all prewar so it’s a rear-view-mirror report. As the war’s impact on energy prices lowers families’ disposable income, that could hurt future demand for goods and services, which in turn could boost layoffs, which, importantly, have been and remain low. Yes, this scenario is importantly a function of war duration; btw, here’s a deck from a presentation I gave yesterday on war econ.

Second, yes, low-fire; but also low-hire. Employers were already largely in wait-and-see mode before the war, and they’re even more so now. Third, the Fed is unlikely to help juice demand with any rates cuts, at least near term, as they too are in wait-and-see mode. Fourth, there are the new real wage concerns I presented above.

I know—this take has even more on-the-one-hand-on-the-other than usual. But that’s the world we’re in right now, and in such a world, with such a data flow, you’ve gotta be a multi-handed economist.