A Coming Contradiction
How we could see job elimination but low unemployment
For econometricians and macroeconomists, today’s era of shrinking gross domestic product, a stubbornly tight labor market, the strong dollar, and rising interest rates has made for a flurry of apparent contradictions. It has led to debates about the technical definition of recession. A further quirk might be coming: it is possible that the Bureau of Labor Statistics (BLS) may report elimination of jobs in the economy, while labor-force participation remains steady and unemployment low. If the total number of jobs were to decrease, how could unemployment remain low if labor-force participation remains constant? As it turns out, this possibility is a quirk of BLS survey techniques and the changing nature of the American workforce.
The two most common measures of U.S. employment are the monthly unemployment rate (officially “U-3 unemployment” – jobless, but actively seeking employment) and the total number of jobs in the economy (officially “Total Nonfarm Payroll Employment”). These so-called headline numbers summarize the BLS’s relatively detailed monthly survey results.
The headline unemployment rate comes from the Consumer Population Survey (CPS), while the headline payroll number comes from the Current Employment Survey (CES). The CPS surveys 60,000 households monthly, inquiring about their employment status. The CES collects payroll numbers from 160,000 businesses or government agencies covered by unemployment insurance. Conceptually, the CPS surveys households, and the CES a specific (but broad) kind of employer.
The CES’s “Total Nonfarm Payroll Employment” excludes agricultural workers and -- more significant -- domestic workers, private household workers, and the self-employed, whose “employers” would not appear in its databases. The last category (self-employed) covers all unincorporated individuals, including sole proprietors and independent contractors, with the latter group consisting of anyone earning income from the so-called gig economy: rideshare drivers, food delivery, on-demand handymen, and a range of other digitally enabled businesses that use independent contractors, as well as anyone receiving income from the relatively new pursuits of selling cryptocurrency projects, video streaming, or a range of other “digital” professions without registering as a LLC. This category of self-employed was among the fastest to recover following the employment shock of the 2020 pandemic, perhaps because many people hit by lockdowns turned to these digital professions in the interim. All these workers are excluded from the CES survey pool because they have no formal employment with an institution that could be sampled from CES’ reference database. Therefore, even if this sector picks up (with a need perhaps for more Uber drivers), its payroll numbers would not figure in the CES survey.
In contrast, the CPS survey (and corresponding U-3 unemployment calculation) covers independent contractors (and other self-employed individuals) and counts them as employed. Therefore, a shift from traditional forms of work to gig-style jobs might well drag down figures for job creation but not alter the unemployment rate.
The situation is further muddled by ongoing legislative and judicial debate about how to classify such gig-economy workers. In March 2021, a New York State Supreme Court judge ruled Uber drivers eligible for unemployment insurance – historically available only to traditional employees and a core distinction between CES and CPS respondents – despite their status as independent contractors. It remains unclear how this will affect the CES and CPS surveys: if Uber were to be sampled in the CES survey, it seems unlikely their independent contractors would count on the payroll, even if Uber is mandated to pay into unemployment insurance programs. What is certain, however, is that employment numbers will be more difficult to interpret if gig-economy-style labor becomes more common.
Thanks for reading Magis! Subscribe for free to receive new posts and support my work.
A fully comprehensive methodology can be found here: https://www.bls.gov/opub/mlr/2006/02/art2full.pdf
Anyone who is unincorporated, that is. If an individual forms a single member LLC and employs themselves, they would be counted. I have not found a good source estimating the percentage of contractors that do so.
The March 2020 pandemic period has seen a rise in the unincorporated self-employed, reaching levels not seen since the Global Financial Crisis but far from historical peaks: https://fred.stlouisfed.org/series/LNS12027714
This contradiction is surprising at a constant Labour Force Participation rate. This rate is also covered by the CPS survey. Obviously, unemployment can also remain low alongside job eliminations if the unemployed immediately exit the workforce.
Metrics are breaking down all over the place: the number of measured COVID cases has less and less to do with the number of actual COVID cases, and the number of measured abortions (I argue) is going to deviate more and more from the number of actual abortions. Alternative data can help--if you can find the actual ground truth to calibrate it against...