January Unemployment at 9.7%? Unbelievable!!

Economists were expecting 10 – 10.1%, I was expecting something higher, but look at what BLS tells us for the end of January: 9.7%!! I’m pretty shocked, and still trying to digest what all the adjustments made by BLS this month mean to the data. Here’s the chart:

Doesn’t make much sense to me, given that the unemployment filings were so high this month. But even the number of employed people went up:

I’m going to spend a bit of time this morning poring over the numbers, because they’re pretty momentous, if true.

Real-Time Pondering:

• Hmmm. Table B says that 8000 government jobs were lost, while Table A-8 says that 117,000 were added. (9am EST)
• So we lost 20,000 non-farm jobs, meaning that we added 520,000 farm jobs? In January? Methinks the seasonal adjustments have gone a little haywire. (9:14am EST)
• But Table A-8, which doesn’t appear to be consistent with any of the rest of the data, says that we added 59,000 farm jobs, and 559,000 non-farm jobs (9:21am EST)
• So the seasonal adjustment for January last year was about 1.8 million jobs, while this year it was only 1.5 million jobs. That seems like the right direction for changing the seasonal adjustment based on current economic conditions (i.e., you can’t lay off seasonal workers you never hired) (9:34am EST)
• All right – I’m baffled. I don’t understand how we can lose 20,000 non-farm jobs, add at most 60,000 farm jobs, and have the unemployment rate drop when the labor pool expanded. As far as I can tell, we should have stayed at 10%. Maybe I’m not following the adjustments or something, but for now I give up (9:58am EST)
• All right, now we’re getting somewhere. Apparently I’ve been caught trying to reconcile Household survey data with Establishment (business survey) data, and the two aren’t playing well together. I think the Household survey probably had some real issues this time around. (10:24am EST)

[Don’t forget, you can now get the unemployment chart on t-shirts, mugs, and postcards!]

Previous posts in this series:

• Initial unemployment claims as of end of January
• Comparing Mark Zandi’s unemployment predictions to reality
• The December Numbers
• Initial unemployment claims as of end of December
• The new misery index
• YouTube video: The Unemployment Game Show
• The November numbers
• The October numbers
• Initial uemployment claims through the end of November
• Initial unemployment claims through the end of October
• A “million jobs saved?” A drop in the bucket.
• Comparing the administration’s statements to reality
• The September numbers
• The August numbers
• The “million jobs saved” claim
• Initial unemployment filings
• The July numbers.
• The percentage of private-employed workers is steadily decreasing, meaning that the burden of supporting government workers is increasing. How long can this trend continue?
• The June numbers.
• Mark Zandi (Moody’s Economy.com) kind of agreed with the Obama team’s projection back in January. But his predictions weren’t much better.
• Saying that “the recession is worse than anybody thought” is a tired old tune
• Everybody did not “guess wrong” on the stimulus package
• The corrected chart for May.
• The predicted numbers for May from a few days ago, with some thoughts on why unemployment is worse than expected even without the stimulus package (and a hearty discussion in the comments on proper graphing)
• A look at the stimulus package spending – how late it is, and how little thus far has been devoted to job creation (it’s basically gone to pay off states’ social services debts)
• The April numbers
• The original post on the subject, noting that criticisms of the stimulus package may not have been motivated by racism after all.