Feburary 18, 2023
In the latest Bureau of Labor Statistics (BLS) Employment Situation for January, the official unemployment rate fell a tenth of a percent to 3.4%. If we had no hidden unemployment, this would be a very positive number. And even with uncounted joblessness, it is a very low rate. The rate for black workers fell .3 to 5.4% and for black teens from 17.1 to 15.8–both moving in the right direction but still very high. The unemployment rate for disabled people jumped by 2.1%. And the hidden unemployment rate–which adds to the official estimates part-timers who want but cannot find full-time work and job-wanters who have not actively searched for work– increased by 300,000. That lifts real unemployment by a tenth from 8.7% last month to 8.8% this month.
On the plus side, a separate government report showed that initial weekly unemployment claims were the lowest since April 22 of 2022. And from the Employment Situation, a shocker. The economy added 517,000 jobs in January–this in an economy which features headlines about tech-sector layoffs every other day.  This jobs increase was such a surprise that some commentators are calling foul and attributing the good news to the annual update of BLS statistical methods.
Statistical Methods Can Be Difficult: I Know
I am not a statistician and cannot judge the validity of BLS sampling methods, annual revisions of population bases, and the construction of seasonal adjustment factors. But I will make a couple of general points.
First, job additions after seasonal and other adjustments in January are always high. Reality is a dicey area for large statistical estimates. This January, prior to the annual revisions of population bases and seasonal factors, total non-farm jobs were actually down by 2,505,000. Sounds terrible, right? But this happens every January (and to some extent in December also). I checked January for the last 6 years and sampled 5 more going back to 1991. Before adjustments, total actual jobs fell every January. Seems that some people cannot go to work because of the weather and other things, some get laid off from temporary holiday jobs, and, perhaps, some employers decide to enter the new year with a trimmed labor force. Anyway, the actual count for January 2023 was, as always, way down, but it showed fewer layoffs than normal, so after seasonal and population adjustments, the Bureau determined that it was as though the economy added 517,000 jobs. That 517,000 is supposed to be a better indicator of where the job market is trending than the actual minus 2,505,000 jobs. One benefit of the BLS seasonal adjustment is that we have fewer politicians and economic pundits warning us that the economy is going to hell every January when the problem is, in part, a seasonal one. 
Is the Tech-Sector a Harbinger of Recession?
Tech-sector employers seem to be leading the way with highly publicized layoffs. The sector purge has been underway for about four months. Twitter, Meta, Google, Dell, Zoom, and others have announced layoffs totaling between 200,000 and 300,000 jobs. Is that the lead-up to a bad job recession? Possibly not. For one thing the number of jobs we are talking about is still pretty small in an economy with 160,000,000 jobs. Also, as Don Lee of the Los Angeles Times (2/10/23) pointed out, some fired tech workers are getting re-hired pretty quickly, and the unemployment rates in two tech centers, the San Francisco Bay Area and Austin, Texas, were still very low at 2.4% and 2.7%.
Another point: Data from the BLS’s Job Openings and Labor Turnover publication through December do show that in the Information sector monthly layoff totals rose, but job openings and hires fell a bit. But for the whole non-farm labor force, job openings, which have been quite high for months, actually increased in December. Also, hiring rates have not fallen in recent months, nor have worker quit-rates. Falling quit-rates might indicate that employees are not confident about finding new, better jobs.
Perhaps just now the tech sector is different from other sectors. For one thing, there is that bad actor leading the way toward nasty policy: Elon Musk aggressively firing people at Twitter, opened the door for copy-cats. More fundamentally, during the low points of the pandemic, the tech sector may have overspent more than some other sectors as people stayed home to work online and required an array of new machines and services to do so. That market has shrunk as many people have more of the tools they need or have gone back to the office. And finally, as Brian Merchant laid out in a well-researched piece in the Los Angeles Times (1/30/23), tech executives must be angry and worried about rising class-consciousness and worker organizing. So while layoffs are partly a matter of the markets, they are also a matter of class-control hardball. Managers want their workers to feel more insecure and less rebellious. There may be other explanations for the way the layoffs have been carried out, but the anonymous, seemingly random selection of victims may have aimed at increasing workers’ sense of precarity. The “clinical cruelty” of the way layoffs were administered at Google was very 1984. No human contact: some Google employees showed up at work to find that their keycards had been deactivated and that they were locked out of their email accounts. No going-away parties.
But managers in another sector are doing things differently. Manufacturers are not increasing layoffs just yet, even though output is slowing a little. The rank-and-file work force has actually been growing in recent months. In fact, it’s now larger than in any month since 2008. We won’t get back to the much higher factory work forces of the 70s, 80s, 90s, but the work force is growing rather than declining, even as business is not. Why? Several commentators think some employers are hoarding labor. Seems bosses remember that there is still a kind of “labor shortage”; they want to keep employees and especially employees who know the business and know the people on their teams. A real recession could end employer worries about labor shortages, but for now, labor hoarding is an interesting phenomenon.
 The household sample from which we get the unemployment rate showed that after adjustments for seasonality and new population estimates and so on, the number of employed Americans increased by 894,000. Another shocker, but one that observers don’t discuss much because the sample is on the small side.
 The methodological issues are complex, and I do not understand most of them. They are the kinds of things that make life difficult for the statistically non-inclined. Years ago, as a professor, I was awarded university support to audit a colleague’s statistics course. His name was Herman Loether and he was well-known for a textbook he co-wrote. That did not help me. I did not finish the course, did not learn much, and would have flunked the exam, had I stayed around for it. But I did not blame the teacher.
Frank Stricker is on the board of the National Jobs for All Network and a member of Democratic Socialists of America. He wrote American Unemployment: Past, Present, and Future.