I like big data, and I cannot lie, you other economists can’t deny . . . In America, no data release is bigger than the monthly jobs report. But size isn’t everything, right? Right? I also like my data to stir me by helping me to generate penetrating insights. Accordingly, it would also be nice to have a jobs report that was easy to interpret—one where an economist or other observer could simply check the numbers and say, “We’re adding a lot of jobs” or conversely, “We’re not adding a lot of jobs.” Alas, today’s data render it unclear how well we are (or are not) doing.
Let’s get into the weeds, shall we? There are two monthly jobs report components: the Current Employment Statistics Survey (CES, also known as the establishment survey), which is sent to about 650,000 establishments each month and is the source of the payroll employment data that we spend much of our time discussing, and the Current Population Survey (CPS, also known as the household survey), which is sent to 60,000 households and is the source of unemployment and labor force data.
There are three reasons to think the CES data are wack. First, the survey’s response rate has plummeted, dipping from 60% in January 2020 to about 48% by August. There are many explanatory factors, including the fact that pandemics are stressful for businesses. During times of stress, people are less likely to fill out voluntary surveys. Also, some portion of the surveyed businesses have likely failed and are no longer operational.
The second reason to—I don’t want to say distrust—question the CES data takes the form of wacky seasonal adjustments, which are based on history, but that history isn’t as reliable during a pandemic that has utterly altered behavior. Today’s seasonal adjustments, which are supposed to help us compare labor market performance from one month to the next as seasons change, are based on the hiring patterns of the past. I wrote about this at length back in the September jobs post (scroll down to the fifth paragraph if you are so inclined).
Third, a lot of people have likely become self-employed, and those workers are primarily captured in the household data. The response rate to the CPS survey has also declined, but from 81.7% to a still respectable 75.9% in October. According to the household data, U.S. employment increased by 651,000 in December. The ADP Employment Report had U.S. payrolls up by 807,000 in December.
Many economists still use the pre-pandemic rule of thumb that the actual employment change number is to be computed as 80% CES + 20% CPS. Using this formula, we come to about 290,000 jobs added in December, which is a bit closer to the pre-release consensus of about 400,000 jobs added in December. It’s also likely that the 199,000-headline estimate will be revised higher. Indeed, October and November job totals were revised upwards by a combined 141,000.
Here’s what we know for sure. The labor market is tight as a drum. Unemployment dipped to 3.9% in December according to the household data, and the labor force participation rate was unchanged at 61.9% even as the labor force expanded by 168,000 for the month. Labor force participation is 1.5 percentage points below its pre-pandemic level.
As always, you can read my in-depth thoughts regarding the construction industry’s labor market at Associated Builders and Contractors.
Three (Somewhat) Key Takeaways
The unemployment rate for those with a high school diploma or less fell sharply in December (from 5.5% to 5.2% for less than a high school diploma and from 5.2% to 4.6% for high school graduates). Because the unemployment rate for college graduates is currently 2.1%, future job gains will by definition have to come largely from those with less than a college degree.
The labor force participation rate for Americans with a disability fell in December (from 23.3% to 22.3%) but remains significantly above the pre-pandemic level of 20.7%. Hopefully the advent of remote work leads to a more inclusive labor force and employers look more aggressively to hire from this population in the context of ferocious competition for talent.
The share of workers who teleworked because of the pandemic inched down to 11.1% in December (from 11.3% in November). This is probably due more to the fact that non-teleworking jobs are getting filled rather than teleworking employees being called back into the office in large numbers. By the way, omicron is not built into the December numbers, which is why we look forward to more clarity in January.
What to Watch
We need more people to rejoin the American workforce. If you could go back in time and have more children circa 1989, that would be most helpful.
Want to Hear me Speak Live?
My 2022 presentation is called No Time to Buy, and yes, the theme is James Bond. If you want to book a presentation (in person or virtual), please contact my assistant Julia (jcomer@sagepolicy.com).
Econ by Anirban has Expanded!
Reminder: this newsletter is expanding, and those who become paid subscribers will get a variety of additional content, including a Week in Review column every Friday. We’ll be sending out the first weekly post later today to all subscribers, but starting next week it will only be available to those who pay.