Further Thoughts on the January Jobs Report

Like most analysts, I was surprised by the January jobs report. It was definitely stronger than I expected. Now that I have had a day to reflect on it, I thought I would share some thoughts. Briefly, I will deal with four issues:

+ Was it fake? (quick answer, no)

+ What does the establishment data tell us about the economy?

+ What does the household data tell us about the labor market?

+ Are there reasons to discount the January data?

Is it Fake?

I’ve had many people ask me this one. We know the Trump administration would have no moral qualms about faking jobs data. Why not, they lie about everything. But as a practical matter it would not be easy.

They couldn’t just fake a single number, like the unemployment rate. They would have to fake a whole set of numbers so that they lined up without any one of them calling attention as being obviously absurd. This would be made harder by the fact that the underlying data in the household survey are made public, so researchers around the country would be able to quickly check the numbers BLS reported.

This doesn’t mean that the data couldn’t be faked, but it would almost certainly require a large number of people to be in on the process. The people who work BLS are serious professionals with integrity. If they got orders from Trump to cook the data, we would have heard about it.

There may be some point down the road where Trump has MAGArized BLS and the data are no longer trustworthy, but we are not there yet.

130,000 Jobs Used to be a Bad Month

Many folks have lost their bearings in looking at the January jobs number. It was much better than expected (my monthly prediction to my wife, just before the report, was 40k), so in that sense it was good, but we created an average of 103,000 jobs a month in 2024 and 188,000 in 2023. The January figure was hardly something to brag about, which of course the Trump administration did.

In fairness, with immigration largely shut down, the labor force is growing far more slowly. This means we need many fewer jobs, likely 30,000-50,000 a month, to keep pace with the growth of the labor force.

Moving below the aggregate, it is striking how concentrated job growth was. The category “healthcare and social assistance” accounted for 123,500 of the job growth, 95 percent of the total. If we add in the 27,800 jobs in restaurants, we’re up to 151,300 jobs. That means on net, everything else lost jobs.

There is nothing in principle wrong with jobs in health care and social assistance, but this is a very narrow base for the economy. It certainly is not the manufacturing renaissance Donald Trump has promised.

There were some positives in traditional blue-collar jobs. Manufacturing added 7k jobs, its first monthly gain since Trump took office. Construction added 33,000 jobs in January, but that makes the average gain in the sector less than 4k per month since Trump took office.

The oil industry lost just under 1,000 jobs in the month, bringing the loss since Trump took office to just under 14k, 3.5 percent of employment in the sector. Apparently, Trump has not realized that low oil prices reduce incentives to drill. The trucking industry also lost jobs in January, bringing the loss since Trump took office to 30,000, 2.0 percent of employment in the industry.

Wage growth was 3.7 percent year-over-year. That’s still a healthy pace, but down from the 4.0 percent pace in 2023-2024.

Bad Employment Trends Were Largely Reversed in January

In addition to the small downtick in overall unemployment, many of the disturbing rises for specific demographic groups were at least partially reversed in January. At the top of this list is the drop in the unemployment rate for Black workers from 7.5 percent in December to 7.2 percent. The unemployment rate for Black workers had been 8.2 percent in November, so this is a substantial decline, although it still leaves the unemployment rate for Black workers a full percentage point above its year-ago level.

In the same vein, the unemployment rate for young workers, ages 20-24, fell by 1.1 p.p. in January to 7.1 percent and is now 0.9 p.p. below its year-ago level. It was 9.2 percent in September. The share of unemployment due to voluntary quits rose to 13.7 percent from 11.1 percent in December. Involuntary part-time employment fell by more than 400k and the share of multiple jobholders was back to its year-ago level.

Data fans know the household survey is highly erratic, so when we see a big turnaround in these numbers in January, as we did for youth unemployment, there are three possible explanations. One is that the prior months’ numbers were the result of measurement error. A second is that the January numbers are the result of error. The third is that there was an actual change in the economy that is responsible for the shift.

Put me down as a skeptic for number three. I don’t believe that the unemployment rate for Black workers actually fell by 1.0 p.p. in two months. Of course, it could be a mix of the three. Maybe the fall numbers were driven in part by measurement error, as was the improvement in January, and the labor market could have improved somewhat. Anyhow, the data in the household survey show a markedly better labor market than what the data showed in the fall.

January Weather May Have Made the Employment Picture Look Better

We all know about the bad storms that hit much of the East in January, but these storms all hit in the second half of the month, after the reference period for the surveys. The weather in the first half of the month was actually pretty good in most of the country.

This matters because our seasonal adjustment factors are looking for bad January weather. This is easily seen by comparing the unadjusted data with the adjusted data. The unadjusted data show that instead of falling from 4.4 percent to 4.3 percent, the unemployment rate rose from 4.1 percent in December to 4.6 percent in January. (Part of this story is the holiday shopping season.) To be clear, I’m not advocating using unadjusted data, I’m just pointing out that the adjustment factors, especially in January are large. Just as especially bad weather would make the employment picture look worse than it is, unusually good weather can make it look better.

To see this story with the establishment survey, instead of the 130k job gain we’re all discussing, the unadjusted data show a loss of more than 2.6 million jobs. Instead of the 30k job gain reported for construction, the unadjusted data show a loss of 213k jobs. Manufacturing lost 86k jobs in the unadjusted data. And the 27.8k job gain reported for restaurants is a loss of 246k jobs in the unadjusted data.

Again, there is nothing illicit in using seasonally adjusted data. If we didn’t adjust the data, it would look like we’re going into a recession every fall and seeing a boom in the spring. The point is simply that the seasonal effects are large, and better or worse than normal weather will have an impact on the data we see.

The Jobs Report Matters, But it is Just One Month and One Report

We can think of our efforts to understand the economy as being like putting together a jigsaw puzzle. Each data report gives us a piece, but none of them can tell the whole picture. Many of us saw a story of a weak labor market based on more than half a year of weak job growth, rising unemployment, and slowing wage growth. We also have data on job turnover showing workers are very reluctant to leave their jobs. This also fits with the stories from the consumer confidence indexes showing unhappy customers.

In addition, we have independent data sources like Indeed, showing very weak job listings data and ADP, which show very weak job growth. The January jobs report is an important piece of data pointing in the other direction, but it is just one piece of the puzzle.

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

The post Further Thoughts on the January Jobs Report appeared first on CounterPunch.org.