As the monthly nonfarm payroll employment numbers repeatedly blew past economists’ forecasts over the past couple of years, one small sector in particular stood out. In 2023, according to the US Bureau of Labor Statistics, sports teams and clubs accounted for just 0.1% of US payroll jobs but 2.6% of all job growth. As of April, they employed 75% more people, on a seasonally adjusted basis, than before the pandemic.
When I first stumbled across this apparent sports-employment boom a couple of months ago, I wondered what could be driving it. Was it the rise of sports analytics? The proliferation of weird new sports leagues? Something else? (It’s not the professionalization of college sports — those jobs appear all to be counted under private and public education.)
Happily, I didn’t write anything about this. Happily because now I suspect that the sudden sharp increase in sports-team employment since early 2023 is mostly a statistical fluke.
The monthly payroll numbers, aka the Current Employment Statistics, are derived from a survey of 119,000 businesses and government agencies covering 629,000 distinct worksites, aka establishments. The survey data are adjusted by estimates, made with something called a “net birth death model,” of the number of jobs created by new business establishments not represented in the survey minus jobs lost at establishments that closed down. The BLS later benchmarks these numbers against the Quarterly Census of Employment and Wages, which is compiled from state unemployment insurance records that cover more than 95% of US jobs. And right now, the QCEW, available through December in not-seasonally-adjusted form, is painting a wildly different picture of sports team employment since early 2023 than the monthly payroll survey:
Which of these surveys is right is currently a question of interest far beyond the sports world. The recently released fourth-quarter QCEW numbers showed employment growth slowing to less than 70,000 jobs a month in the second half of 2023, while in the payroll survey it continued to average more than 200,000. This has led to widespread speculation — detailed in a Bloomberg News article earlier this month — that the payroll numbers will eventually be revised downward, revealing the job market to be less robust than currently portrayed.
How much less robust is the subject of some debate. Anna Wong, chief US economist for Bloomberg Economics, estimated that actual payroll gains have averaged less than 100,000 a month and may have been negative last fall. Ernie Tedeschi, until recently chief economist at the White House Council of Economic Advisers and now director of economics at the Yale Budget Lab, forecast a much smaller adjustment that would reduce monthly jobs gains by about 40,000 (thus leaving them above 160,000).
The actual revisions will be made on the basis of the March 2024 QCEW numbers. The first official word on how big they’ll be comes in August, when the BLS releases those numbers and then makes a “preliminary benchmark announcement” of the likely implications for nonfarm payrolls overall and for the main industry sectors such as manufacturing, financial activities and leisure and hospitality (which sports teams and clubs fall under). The final, detailed revisions aren’t announced until the next February.
This process takes so long because the QCEW statistics get revised too as more-complete data become available. Since 2017, the vast majority of revisions to QCEW employment numbers have been upward, a good reason to think the payroll employment revisions will be smaller than a glance at the fourth-quarter QCEW statistics would indicate. They’ll probably still be negative, though. Over the past 11 years, the nonfarm payroll number has been revised downward six times and left effectively unchanged three times, leaving only two upward revisions. All these revisions have been quite small on an economy-wide basis, averaging positive 0.1% since 2017 for the QCEW and negative 0.1% since 2013 for nonfarm payrolls.
For individual industries, the revisions can be much larger — which brings us back to those sports teams and clubs. The most recent QCEW-based benchmark revision to their monthly payroll number was a whopper, raising (not seasonally adjusted) employment in March 2023 to 99,000 from 68,100, a 31% increase. The net birth-death model used to estimate payroll job gains from new businesses had apparently undershot in a big way as spectator sports bounced back from pandemic shutdowns.
As new data from the QCEW were fed into the sports teams and clubs birth-death model after that, it may have overshot in a big way last year. The models are updated quarterly using five years of data on actual employment changes related to births and deaths of business establishments, with the newest numbers a year old at the time of the update and more recent data given more weight. Here’s what those data look like for the broader category of performing arts and spectator sports (numbers for just sports teams aren’t available).
In 2020, employment losses from business deaths vastly exceeded the gains from births. In 2021 and 2022, employment gains from births outstripped deaths. It’s hard to make accurate predictions based on such volatile data, and the sector probably hasn’t really added 63,400 (seasonally adjusted) jobs since March 2023 and didn’t really account for 2.6% of US employment gains in 2023. It has also probably been growing at a healthier pace than the fourth-quarter QCEW numbers indicate, but again I would guess that those numbers are closer to correct than the payroll numbers — in part because those payroll gains seem too big to be plausible.
So I’m putting that think piece on the sports-jobs boom on hold until August at the earliest and have acquired a new appreciation for how noisy and just plain wrong single-industry jobs numbers can be, at least at first. That these errors tend to mostly cancel out in the nonfarm payroll totals is part of the magic of statistics. But it still means there are a lot of errors out there.
A message from Advisor Perspectives and VettaFi: To learn more about this and other topics, check out our podcasts.
Bloomberg News provided this article. For more articles like this please visit
bloomberg.com.
Read more articles by Justin Fox
