经济周刊：劳动力市场可能并不像人们想象的那么疲软

16 January 2026 Richard de Chazal, CFArdechazal@williamblair.com+44 20 7868 4489 Economics Weekly Louis Mukamalmukama@williamblair.com+1 312 364 8867 The Labor Market May Not Be Quite asWeak as Believed William Blair Over the last two quarters, the Fed has shifted its empha-sis away from the inflation half of its dual mandate andback toward the employment side. It is taking the viewthat inflation will continue to decelerate, but employmentgrowth is starting to more tangibly slow and further rate and stable. The data are showing rates that are far fromthose that might normally raise eyebrows. If we run a single factor regression of annual rates ofchange for jobless claims and payrolls from 1967-2019(just before the pandemic) with a correlation coefficientof 64%, it shows that payroll growth would have in- cuts were needed to shore up growth and prevent further deterioration. In our view, while there is a clear decelera- tion in job growth taking place, it looks to be driven asmuch, if not more, by supply-side dynamics, as opposed to The U.S. economy has gone from exceptionally strongrates of payroll growth over the last few years to a sharpdeceleration seemingly overnight. In 2023 nonfarm pay-rolls grew by 2.5 million, in 2024 by 2.0 million, and in2025 by just 584,000, with the average monthly changedropping to 69,000 from 168,000 in 2024. Meanwhile, Moreover, employment growth in the household surveyof employment has similarly shown no signs of modera- But has the labor market actually deteriorated as much asthe establishment survey payroll data seem to be telling us? Historically, investors have tended to view the establish-ment (nonfarm payroll) survey as the better gauge of em-ployment growth. Not only is the sample size much largerthan the household survey, but it also involves samplingcompanies, not “knocking on doors” as the household The fact is that the weekly initial jobless claims data—which were not impacted by the shutdown and do not suf- William Blair to 50,000. Brookingsestimatesthat in 2026 this will de-cline further, reaching a lower range estimate of -20,000to 20,000 and an upper range of 10,000 to 50,000. (e.g., changes in immigration) or misperceptions aroundwhat constitutes being employed by the general public. Companies also normally count undocumented or illegalworkers as employed on their books. This is becausewhile workers are required to present their work permits,employers are not obligated to authenticate their docu-ments beyond a “reasonable” inspection. It is also moredetrimental for the employer to be caught hiring illegal This would imply that even if we start to see some nega-tive nonfarm payroll prints in the coming months, theymight still be compatible with a growing economy and Due to these substantial and rapid supply-side changes,financial market participants have quite clearly decidedthat the best way to cut through all of these measurementissues is to effectively ignore the payroll numbers and However, this survey comes with its own set of measure-ment issues; it is an incomplete sample taken each month,with an imputed births-deaths model to account fornew or closing businesses. This is then reconciled with Unfortunately, the monthly mismeasurement from thismodel in recent years has been substantial; establish-ment survey growth data were revised down by 818,000jobs in the year to March 2024 and by 911,000 in theyear to March 2025. In his last post-FOMC press con-ference, Chair Powell stated that the Fed believes thatpayrolls in the most recent few months are likely being According to the Sahm Rule, such an increase has justabout always been associated with the onset of a reces-sion. But that clearly is not happening today. GDP growthin the third quarter increased 4.3% (including a 3.5%increase in consumer spending), and the Atlanta Fed’s We think a couple factors may help to explain this newscenario. First, the core members of the labor market—i.e., prime-aged workers 25-54 years old—are still fullyemployed. Their participation rate has been stable at ahigh level, as depicted in exhibit 4. Such a high level of On the supply-side, the two biggest factors contributingto the growth slowdown are the sharp reversal in net Brookings estimates that the U.S. experienced a nega-tive net migration flow in 2025 for the first time in manydecades. This follows very substantial annual net inflowsof 2 million-3 million in the few years up to 2024. Thisreduction in the supply of labor needs to be taken into ac- Whereas between 2022 and 2024 roughly 150,000 to200,000 jobs were required to stabilize the unemploy-ment rate (the number of unemployed/labor force), in the William Blair to unemployment (exhibit 6). This is encouraging and con-sistent with at least some rising demand for labor—likely Second, while baby boomers are retiring and pushingdown labor force growth and participation rates (exhibit5), they are also moving out of employment by choice.T