If you go to the bottom of this page from the January BLS employment report:

http://www.bls.gov/news.release/empsit.nr0.htm

You will see that what happened is that the BLS made their annual population control effect to the numbers. Thus, without this effect (ie what the survey alone produced + seasonal adjustments), the household survey actually showed an employment increase of 589,000 jobs and no change in the labor force (instead of the big reduction). This would have left us with a U-3 rate of 9.1% (instead of the 9.0 reported), but would have made the whole report much easier to understand. I just wish they hadn’t buried this major annual adjustment at the bottom of the report.