
Official U.S. jobs data are sorely needed at a time when a slowdown in immigration and AI-related job cuts are making it difficult to assess the health of the labor market, and private data are no substitute, former Bureau of Labor Statistics Commissioner Katharine Abraham told MNI.
The October household survey from which the unemployment rate is calculated is unlikely to be collected, she said as the record federal government shutdown forced a second straight missed publication of the BLS's closely-watched monthly employment report Friday.
"We were starting to see upticks in long-term unemployment. Employers are not hiring as much, and it seems like there’s a lot of caution in bringing people on board," she said in an interview. "AI is also a factor. You’re starting to see companies talking about not hiring or laying people off because they can be more efficient."
On the other hand, with a large decline in immigration, the number of new jobs needed to hold the unemployment rate steady has fallen to something like 50,000 a month from some 120,000 a year ago, she said. "That would be a natural explanation of why payrolls has slowed, not necessarily an indication we’re not keeping pace with new workers becoming available."
"It’s hard to parse out all of these conflicting signals, so it’s a time when having good current data seems particularly important."
HOLE IN THE DATA
Various sources of private sector data can enrich an understanding of the labor market, but they rely on official benchmarks and offer only a partial view of the U.S. economy, said Abraham, economist at University of Maryland.
A payroll service provider like ADP adjusts its customer-generated data to the Labor Department's quarterly census of employment and wages, so it can upweight industry cells where there's less representation among their customers to come up with a more representative measure, she said. "If they didn’t have benchmark data, they couldn’t do that."
It would be difficult for the BLS to ask about the original reference week for October belatedly, as it did following the 1995-1996 and 2013 shutdowns, she said. In those episodes, surveys were conducted just a week late. Now, much more time has passed and respondents are likely to make errors in recalling what they were doing weeks earlier.
Operationally, changing the November questions and reprogramming the computerized survey instruments would also be difficult, she said. "There will probably will be a hole in the data." (See MNI INTERVIEW: US BLS On Track To Miss Nov CPI Report -Beach)
CPI data that's typically collected physically from stores and doctors offices may be lost as well for as long as the shutdown continues, Abraham said. If the BLS manages to collect November price data, it may put out an estimate of the one-month change for October, linearly interpolating from the November price index levels, she said.
"It would be better if they did so there’s a monthly number people could work with," she said.