No, a million more men than women didn’t enter the labor force in January – Grid News

Introducing Grid Health, our new weekly health and policy newsletter

No, a million more men than women didn’t enter the labor force in January

“Pay strict attention to what I say because I choose my words carefully,” the Manhattan bank thief Dalton Russell says at the beginning of Spike Lee’s “Inside Man.” For everyone’s sake, you should treat the Bureau of Labor Statistics the same way.

Within this January’s jobs report, there was something strange. It seemed like more than 1 million men joined the labor force, meaning they were working or looking for work, compared with women.


Hear more from Matthew Zeitlin about this story:


Axios blasted out a chart-based story (the outlet later corrected the story) working off of a number in the monthly jobs report, linking it to the widespread concerns that the covid-19 pandemic had set back women’s participation in the work force because women tend to face disproportionate burdens at home, especially with child care.

But it wasn’t the case. Or, if gender-based domestic burdens from omicron were affecting the entry of men and women into the labor force, it wasn’t going to show up in those figures.

ADVERTISEMENT

What really happened is that the Census Bureau, which helps conduct the household survey used by the BLS for its labor force estimates, developed what are known as new “population controls” based on updated population data.

If you look at the month-over-month change in the size of the labor force on its own, maybe using a service like the Federal Reserve Bank of St. Louis’ data repository, FRED, you would see that the male labor force grew by 1.2 million while the female labor force grew by 175,000, a gap of about 1 million.

It’s worth remembering that in a typical month, the jobs report is open to sizable revisions, especially to the much-paid-attention-to payroll number, which reports the total number of “nonfarm” jobs in the country. Much of this is due simply to the businesses that answer the survey taking extra time to report the information to the surveyors. And then at the end of the year, the numbers can be revised even further as BLS updates its data.

These two sets of revisions, along with the “seasonal adjustments” they use to iron out predictable variations in employment (like retailers taking on new employees in November and December and letting them go in January), can produce large disjunctions, especially in the January report, where the data is updated through the annual adjustment process and there’s one of the more sizable seasonal adjustments, as well as the typical revisions.

While many economists and perhaps even the White House expected meager or even negative jobs gains because of omicron peaking across the country, the headline number was far stronger than expected.

ADVERTISEMENT

Population estimates, like the ones that caused that change in men versus women employed, according to the BLS, “reflect updated birth and death statistics, which include the effects of the coronavirus (COVID-19) pandemic; new estimates of net international migration, which reflect reduced international migration due to the pandemic and associated travel restrictions; and methodological improvements to the estimation process.”

What this means is that the household survey responses are re-weighted according to new estimates of how large the “civilian noninstitutional population age 16 and older” is.

“The only thing that a survey can ever measure is a proportion … a public opinion survey can tell you the proportion of the population that plans to vote for Biden, just as it can tell you the proportion of the population that’s in the labor force. To turn this into an aggregate number — the total number of votes for Biden or the total number of people in the labor force — you have to multiply by a population estimate,” Justin Wolfers, a University of Michigan economist, told Grid.

The new population estimates found that the overall 16-plus population not in prison, nursing homes or the military was about 1 million people larger than the previous estimate, but the labor force was about 1.5 million people larger.

The explanation for this disjunction is that the survey found more men. Since the male labor force participation rate is higher than the female labor force participation rate, this mechanically made the male labor force larger.

ADVERTISEMENT

“The BLS uses a population estimate that comes from the Census Bureau. Through 2021, it was using one set of population estimates. In January 2022, it switched to a new set of population estimates. That new set of estimates has a lot more men, so a (roughly) unchanged male labor force participation rate multiplied by a larger number of men yields a larger estimate of the male labor force,” Wolfers explained in an email to Grid.

“Virtually all (or possibly even more than all) of the rise in male labor force participation in January was due to changing population controls,” he said.

The BLS helpfully also publishes changes in labor force statistics from December to January that didn’t use the updated population figures. Without the new population figures, the civilian noninstitutional population would have grown by 93,000, the labor force would have shrunk by 137,000 and the employment level as measured by the household survey would have decreased by 272,000. But it’s important to remember that changes in measurements like this aren’t changes in the underlying reality. They instead show how the underlying reality was obscured by older, less accurate measurements.

“The December 2021 data was systematically underestimating employment and that the 2021 employment numbers were better than first published. However, the headline change in employment, labor force participation, and unemployment between December and January is distorted by the incorporation of these new estimates and therefore not as good as it initially seems,” Joseph Politano, a Bureau of Labor Statistics economist who publishes an independent newsletter, wrote late last week.

Another annual revision process caught the attention of economic observers: the so-called benchmark revisions to the establishment survey, which, combined with updating seasonal adjustments, can lead to big swings in reported data that aren’t necessarily month-over-month changes.


In this case, the big headline changes were that the total number of jobs gained over the last year was higher, with some decreases in the summer months and increases toward the end of the year. But, unlike the population adjustments to the household survey producing incorrect inferences about covid and the labor force, these adjustments really did tell us something new.

For example, the new data shows that employment in the retail sector had suffered more than previously thought and employment in warehousing and storage has been hotter than we knew before.

From February 2020 to December 2021, according to the older data, employment in transportation and warehousing rose about 3.5 percent, while according to the revised data, the increase was more like 8.5 percent.

Nicholas Bunker, an economist at the job-search website Indeed, told Grid the warehousing and storage subsector had “really done very, very well,” and went from “not really showing up in the top 10 industries for job growth,” to showing the most employment growth since before the pandemic, rising over 30 percent, while the previous data showed a 13.5 percent increase.

In “Inside Man,” Russell delivers his monologue at the beginning of the movie, from what he describes as what “could most readily be described as a prison cell,” but, spoiler alert, he is really in the bank itself, waiting to walk right out, unnoticed by the police. Sometimes all you need to know is right in front of you, but only if you know where to look.

  • Matthew Zeitlin
    Matthew Zeitlin

    Domestic Economics Reporter

    Matthew Zeitlin is an economics reporter at Grid focused on the domestic impact of major stories such as coronavirus, the supply chain and economic volatility.