>The second reason pertains to the numerator of the unemployment rate. BLS
>uses headcount rather than full-time equivalent (FTE) to account for
>employment. That is, a person who is employed only 10 hours a week counts
>the same as the one who is employed 40 hours a week. This, BTW, is a
>well-known weakness of the employment stats. Thus, significant numbers of
>the under-employed can skew the employment rates upward, as compared to the
The BLS publishes reams of data on part-time and multiple jobholding as well as hours worked.
>I suspect that underemployment (part-tiem, temporary arrangements) are more
>prevalent in the US than in Europe, and that again skews the US stats
>upward as compared to Europe.
In 1996, according to the OECD, 18.3% of U.S. employment was part-time, compared with 22.1% for the U.K., 18.9% for Canada, 16.3% for Germany, 16% for France, 21.4% for Japan, and 23.6% for Sweden. Unlike poverty and polarization, where the U.S. usually ranks on top, U.S. levels of part-timing and contingency are about average for the OECD. As for women's share of part-time employment (and part-timing as a share of women's employment), the U.S. is at the low end of the OECD.
>I do not know if those two biases can account for the 4-percentage points
>difference between the US and the Canadian/European rates, but it is likely
>that they are a nontrivial part of that difference.
The OECD figs I was quoting are by standardized definitions, so the only possible explanation (other than actual levels of unemployment) for the difference would be coverage of the survey. Do German enumerators capture the Turkish population? French, the Algerians? Brits, the Jamaicans and Pakistanis?
Looked at another way, U.S. employment levels tell the same story the unemployment rate does: In 1996, 75% of the U.S. adult population was working, compared with 59.1% in the EU, 68.5% in Canada, and 74.6% in Japan. In general, I'd say the major labor market problem in the U.S. is too much work for too little pay, not unmeasured unemployment.