The thing about economic reports from government agencies — local, state and, especially, federal — is that the first publication of the data is basically a first draft.
It will get revised again and again and again.
Still, when you hear that the government's payroll employment data have been overreported by 818,000 between April 2023 and March 2024, you wonder what happened.
Related: Analyst resets Nvidia stock price target before earnings
You shouldn't be surprised — but try not to be outraged. This is just one more thing you can blame on the Covid-19 pandemic.
“What you’re seeing is an echo of the large shocks that we’re just working our way through,” Joe Brusuelas, chief economist at RSM, told The Washington Post.
How the labor numbers are put together
The U.S. Bureau of Labor Statistics releases its monthly jobs report the first Friday of every month. The numbers in their report will be revised the next month and the month following. And it will be revised again at year-end.
The jobs report has two pieces: the data that produce the unemployment rate; and payrolls employment, i.e., how many people are employed in whichever jobs.
The payroll employment data are studied just as intensely as the unemployment rate and can and do move markets.
The payrolls report depends first on BLS staffers calling up a specific number of employers once a month and asking how many employees they have. From that data the agency projects an estimate for the entire country.
The surveys have become less reliable in recent years in part because the pandemic first caused massive losses. Then, the recovery meant labor shortages. And the Labor Department gets hit, too: It hasn't had enough people to make the calls, and many employers say they don't have time to answer the questions.
This month, the BLS started to match its data with the payroll data that states collect for their unemployment-insurance programs, using March as the anchor date. And that can refine numbers quickly because now we're dealing with names and Social Security numbers.
The big in the big overcount
That said, there was a bit of excitement Wednesday when the BLS announced its Current Employment Statistics Preliminary Benchmark report.
This is the BLS's first crack at matching payroll numbers with state data. The preliminary report suggested the jobs were overcounted — by 818,000 over the last 12 months.
At first blush, it was not a pretty number. But two points are in order:
- The percentage is still pretty small — only 0.5% out of a total payroll employment number in March 2024 of 158.1 million
- But yes, it's a bit of an outlier. Normally, the revisions come to about 0.1% up or down.
More Economic Analysis:
- American Express stock analyst flags concerning shift in consumer behavior
- Revised Jobs Data Could Be the Bear That Knocks on Goldilocks' Door
- Struggling whiskey brand files for bankruptcy, closes
But remember this is a preliminary announcement of what the benchmarking will show when a final tally is announced in March. Indeed, some analysts believe the 818,000 number ultimately will be cut in half.
One last point, courtesy of The Wall Street Journal. It is believed that employers offer all their job counts when the Labor Department workers do their surveys. That might include undocumented workers. But undocumented workers don't qualify for unemployment insurance and, thus, won't appear in state unemployment tallies.
Here's where the biggest numbers in the overcount occurred:
- Professional and business services (everything from lawyers, accountants and architects to trash-pickup staff), off 358,000.
- Leisure and hospitality, off 150,000.
- Retail trade, off 129,000.
- Manufacturing, off 115,000.
- Transportation and warehousing, 56,000.
Related: Veteran fund manager sees world of pain coming for stocks
