>You didn't say "compared to the UI universe of employers", so those who
>heard you talking of comparing to the UI records are not crazy to think you
>were referring to comparisons to those filing for unemployment insurance.
>
>But frankly, saying that 97% of US workers are registered with the UI
>system seems inherently unlikely. The GAO estimates that 6.3% of the
>population are independent contractors and another 4.8% are self-employed.
Independent contractors and the self-employed are not covered by the establishment survey. (And isn't there overlap between independent contractors and the self-employed? <ftp://ftp.bls.gov/pub/news.release/conemp.txt>) You claimed some expertise in this area, so I didn't think it was necessary to spell it out in detail. I also provided a URL that offered some detail on this:
At 4:03 PM -0400 8/6/04, Doug Henwood wrote:
><http://stats.bls.gov/ces/cesbmkqa.htm>
>
>>II. Background Information
>>
>>(1) What is the establishment payroll survey?
etc.
The self-employed are covered by the HH survey. Scrupulous analysts look at both. But the employment/population ratio has tracked the movement in the establishment survey quite closely over the last couple of years.
Also, lots of undoc'd workers are covered by the establishment survey. Employers are asked how many workers they have, how many hours they work, and how much they're paid. Whether they're citizens or legal immigrants isn't in the questionnaire.
>As I said, the fact that much of my work deals with these kinds of
>quasi-workers may make me take them too seriously as components of the
>economy, but I don't think I am. Add in the more general questions about
>methodology raised by others about the BLS establishments survey and I just
>don't feel a great need to accept its pronouncements as the Delphi Oracle.
I thought the Oracle was famous for her ambiguity.
>Not that we shouldn't take trends shown in the numbers seriously, but
>taking them seriously is not the same as taking them as gospel.
Who takes them as gospel? Really? I spend a good bit of my life looking at this stuff very closely. I know the data's flaws and virtues better than people who make a lot more money than I do.
Doug