Banging My Head Against the January 2010 Unemployment Data February 6, 2010
Posted by geoff in News.trackback
Those who looked at the unemployment post yesterday witnessed me flailing around trying to reconcile the rise in employment indicated by the Household Survey with the fall in employment indicated by the results of the Establishment Survey (the Household survey polls people at home, and the Establishment Survey polls businesses). I wasn’t alone in worrying about this discrepancy – it was noted throughout the blogosphere.
In trying to figure out what was happening, I decided to see how well the two have agreed over the past year, the question being: is this disagreement typical? I certainly didn’t remember it being so bad in the past. So I decided to make a little graph:
This is the non-farm, seasonally adjusted, employment data as adjusted in Jan ’10 by the BLS. In this graph, I divided both sets of data by the value at the starting point (Jan 2009), so that they can be more easily compared. So the value at the outset is 1 for both curves.
As you can see, while the curves have very different shapes, they both trend the same way – when Household data goes down, Establishment data goes down. Except for this latest data, where the Household data goes up, and the Establishment data goes down.
So January 2010 does in fact represent a larger discrepancy than we’ve seen in the past. I guess we’ll have to wait until next month to see which curve was giving a more realistic picture of the employment situation. But just looking at the two, the Establishment Survey appears to yield less noisy results, so for now, that’s where I’m putting my faith.
Gracias Geoff once again for doing the work Americans and cool people just won’t do.
I’m curious, do numbers exist that show the percentage of the total population that are currently employed? It would be interesting to see a table/graph of total population, from say WWII on, breaking out total employed in private and total employed in public, total on welfare, total on social security, and total simply unemployed and not on the government dole. Then color code which party controlled Congress each year, El Nino/La Nina events, women’s average skirt lengths, and seismic activity.
I’m curious, do numbers exist that show the percentage of the total population that are currently employed? It would be interesting to see a table/graph of total population
Yes, the BLS reports include that number. It’s currently at 58.4% (from memory, anyway).
It would be interesting to see a table/graph of total population, from say WWII on, breaking out total employed in private and total employed in public, total on welfare, total on social security, and total simply unemployed and not on the government dole.
It’s always work, work, work with you.
May want to adjust for adult and senior populations, too, by the way.
And then note the Women’s Lib movement, because that’s probably going to be a shift.
How about comparing it to a GDP chart, too, while you’re at it?
[...] expect, however, some big adjustments in the next few months. The BLS marries together two separate surveys: one of households and one of big business, or [...]
Even though unemployment went down in January, there’s an interesting piece from Julian Alssid at the Workforce Strategy Center about how America’s workforce remains ill-prepared to compete with the world…
http://workforcedev.typepad.com/workforcedev/2010/02/unemployment-american-workforce.html