Data quirk inflates coronavirus jobless figures by 3.7M

Economic data that has predictable seasonality, like holiday retail hiring, is usually adjusted to provide a more accurate long-term perspective

Figures for weekly job losses caused by COVID-19 are being inflated by adjustments the government uses in a normal year to smooth out seasonal factors and show broader growth patterns, according to one investment adviser.

While Labor Department reports show more than 54 million Americans have filed for first-time unemployment benefits since stay-at-home orders were issued in mid-March, that number is inflated by 3.7 million due to seasonal abnormalities related to COVID-19, meaning total first-time filings should be 50.3 million, David Eichhorn, president and head of investment strategies at St. Louis-based NISA Investment Advisors, told FOX Business.

“Ignore what is typically the headline number and look at actual [continuing] claims until we get back to more normal flows in employment,” Eichhorn said. The U.S. Department of Labor did not respond to FOX Business’ multiple requests for comment on the numbers.


Typically, economic data that has predictable seasonality is subject to adjustments to provide a smoother figure. A good example of this would be retail sales, which see a boost ahead of the holidays but disappear soon thereafter.

Growth during the season could be determined using unadjusted numbers, but since the gain is short-lived, that would give a misleading picture of overall labor market expansion.

The adjustments that enable a more accurate long-term assessment, however, can skew short-term results, particularly in an atypical year such as 2020.

For instance, the summertime is usually when automakers retool their production lines, but that process occurred earlier this year as factories were shuttered to slow the spread of COVID-19.

Because of such quirks, a single week in March saw 850,000 job losses created statistically when “nothing was going on” due to COVID-19, Eichhorn said.

While the number of initial jobless claims are overstated because of the seasonal adjustments, continuing claims, which were more than 17 million last week, are coincidentally “close enough” to their real number due to the way things have netted out, said Eichhorn, who believes that's the number to watch to “until we get back to normal business-cycle job gains and losses.”

Seasonal-adjustment distortions are likely to abate in November, when they typically return to zero, but the number could be misleading, either better or worse, until weekly first-time filings fall back to the 300,000 to 400,000 range.


As for how the big claims numbers from 2020 might distort the data going forward, Eichhorn wasn’t sure, noting that we could “have some weird numbers next year” unless the Department of Labor’s larger data set smooths them out.