You should start looking at the state level household employment series; (the unemployment rate is calculated using this variable). Here is Wisconsin’s, as of today’s release.
Of course, there are many reasons not to focus on the household series at the state level, but several readers are very dense (you know who you are) insofar as this point is concerned. For more discussion, see posts on Wisconsin and California.
Both household and establishment series will be benchmarked come the January release.
Is “employment” defined the same in both series? For example, if somebody is working as a private contractor (e.g., in construction, as a hairdresser, or as an uber driver), but is not an employee, would he or she be included as employed in the establishment series? Would he or she be included as employed in the household series?
The idea that any but very minor differences could arise from cross-state employment just doesn’t seem plausible (if that is the gist of of Justin’s NYT article, which I was unable to read).
don: See here.
Thanks. So even without statistical error, i would expect the household series to be more volatile and sensitive to changes in the economy.
“more volatile and sensitive to changes in the economy”
You may volatile sound like a good thing as if it were more reliable. I doubt many labor economists would agree with you. Keep in mind both series are surveys. There is a long literature on the differences between the two series. Neither is necessarily perfect which is why BLS provides both surveys.
don: I think you could expect the household series to be more volatile — the m/m 90 CI is ±500,000 for household, ±115,000 for establishment at the national level. Doesn’t mean it’s more sensitive to economic conditions, though.