Nonfarm payroll employment falls by 701,000 in March; unemployment rate rises to 4.4% https://t.co/1Y9cSWJUIB #JobsReport #BLSdata
— BLS-Labor Statistics (@BLS_gov) April 3, 2020
How does this report jive with the unemployment claims?
Unemployment Insurance Weekly Claims
Initial claims were 6,648,000 for the week ending 3/28 (+3,341,000).
Insured unemployment was 3,029,000 for the week ending 3/21 (+1,245,000).https://t.co/ys7Eg5LKAW
— US Labor Department (@USDOL) April 2, 2020
We have two options. First we could claim conspiracy or we can understand how the data is collected.
The job total data is collected from a monthly survey. The survey is fielded for a week. The week it was fielded is the middle of the month. This is standard procedure.
However, we are not in standard times. The survey was in the field before most of the state led social distancing began. We are seeing the pre large scale social distancing universe with a twenty day delay.
The unemployment claims are derived from weekly, state level, administrative data. The first week of March was a totally typical week. And then things went to shit.
Understanding how data is collected is important when weird things happen. The March survey is effectively useless as a meteor hit the US economy right after the survey was brought back in from the field. April data will show a better appreciation of reality but that fuzziness is an artifact of data collection methods and not anything nefarious.