MNI POLICY: US Jobless Figures Face More Labeling Errors: BLS

--Millions Have Been Counted As Employed After Covid-19 Layoffs

By Brooke Migdon

     WASHINGTON (MNI) - U.S. unemployment figures may continue to show errors in

classifying workers laid off amid Covid-19, and a Bureau of Labor Statistics

official told MNI the agency is reviewing how to further improve interviews to

fix the problem.

     Millions of workers have been misclassified as employed but absent from

work due to 'other reasons.' The May unemployment rate was 3 percentage points

higher than reported, not including seasonal adjustments, the BLS said last

week.

     Errors continue to skew the jobless rate even though misclassification has

declined each month, says Julie Hatch Maxfield, the BLS Associate Commissioner

for Employment and Unemployment Statistics. Data collectors have been given

additional training since the Covid-19 pandemic was declared in mid-March and

in-person interviews were suspended, she said.

     "The training did help the issue," she said. "On an interviewer basis,

those instances only occurred a small fraction of the time, but unfortunately

because there are so many Census Bureau interviewers, that small fraction added

up across the nation."

     The BLS took another step to improve the next report for June by altering

the tools used by interviewers to reduce misclassification, she said. "Because

it appears this will be a long-going issue and training does not seem to be

doing the job by itself, it makes sense to actually change the survey

instrument," she said. 

     --MISSING QUESTIONS

     The BLS and Census this week launched an investigation into employment data

collected through May to identify whether the misclassification of workers

occurred more frequently in certain Census zones or among particular

interviewers, she said. While the investigation may show the extent of past

errors, the BLS won't revise published data based on the findings.

     The investigation will allow BLS and Census to refine the number of

misclassified workers versus the "crude" method used to date, she said.

     The exact total will never be known because some people weren't asked

questions of respondents who were out of a job, such as if they are available

for work or actively looking for a job, she said.

     The errors add to the Fed's struggle to figure out if and when the job

market will fully recover. Weekly jobless claims figures are also being skewed

by workers who under federal stimulus legislation can draw state benefits

without looking for a new job.

--MNI Washington Bureau; +1 202 371 2121; email: brooke.migdon@marketnews.com

[TOPICS: MAUDS$,M$U$$$,MT$$$$]