Friday, January 22, 2010

The BLS where 1-10-10-10-10 = no change

I'm really having a tough time squaring this data in my head. According to the Bureau of Labor Statistics...

"Forty-three states and the District of Columbia recorded over-the-month unemployment rate increases, four states registered rate decreases, and three states had no rate change."


"The national unemployment rate was unchanged in December at 10.0 percent."

So, 84% of the states surveyed registered HIGHER unemployment, but nationally the unemployment rate was unchanged? Someone suggested that this could be caused by a big jump in employment in one or two large states offsetting declines in smaller states.

Let's check that for accuracy - The 4 states registering declines in unemployment were IA, OK, MI and SD. These states are ranked 8th, 28th, 30th and 46th in state population totals, not exactly bursting at the seams with population. Second, the declines in unemployment were all very small just 0.1-0.2% for the month (with the exception being Oklahoma where unemployment fell 0.5%??)

Well, how did the big states do for the month?

California 12.4 12.4 unchanged
Texas 8 8.3
NY 8.6 9.0
Florida 11.5 11.8
Illinois 10.9 11.1
Pennsylvania 8.5 8.9
Ohio 10.6 10.9

MI 14.7 14.6
GA 10.1 10.3
NC 10.7 11.2
NJ 9.7 10.1

Huh, well that's not good. It seems like the individual large states (except Michigan, but note that Michigan's improving unemployment rate is still the highest in the nation) are all seeing meaningful upticks in unemployment but when the data goes into the BLS supergianthypermega computer the national unemployment rate remains unchanged.

I suspect that this may have been caused by a difference in methodology (perhaps they remove workers that drop from the workforce in national data, but not in state data?) but it is this kind of inconsistency that makes it hard to handle BLS data.


No comments: