Record Gap Between U.S. Employment Surveys

As noted several times in the past, unlike most countries, the U.S. two different employment surveys, one based on households and one based on employers ("the payroll survey"). Since they're supposed to describe the same real-world phenomenon they should generate the same numbers, but they often don't since they have different data sources, and both are somewhat unreliable.

The latest month and year, the divergence has been record large, with the payroll survey coming with far stronger numbers than the household survey. Here's how they changed the latest month:

Payroll: +204,000 jobs
Household: -735,000 jobs

And the latest year:
Payroll: +2.33 million jobs
Household: +0.24 million jobs

Part of the most recent monthly divergence could be related to furloughed federal workers during the partial shutdown counting as employed in the payroll survey but not in the household survey, but even adjusting for that the household survey still suggests a much weaker labor market than the payroll survey.