Does anyone know if the actual ground-level data collection is done by the BLS or whether it's some state agencies that theoretically use the same methodology but maybe, in fact, don't? A few of the state lines on the map seem a little suspicious:
- NW Illinois vs. SW Wisconsin & E Iowa, even the manufacturing-heavy parts of the latter two
- E Kentucky vs. S West Virginia
- N Alabama vs. all its neighbors (I mean, Huntsville aircraft/defence will account for some of this, but such a big difference right at the state lines?
- the horizontal line between Mississippi & Louisiana
BLS Local Area Unemployment Statistics -Estimation Methodology The national unemployment rate is based on the Current Population Survey which is conducted for the BLS by the Census Bureau. It is based on 50,000 households surveyed monthly. I think that it is a sliding sample, with new housing units added each month, and old ones removed. My sister used to be a surveyor, and she had to collect data each month. In one case, a house had burned down, and she had to go check each month to confirm that no one lived there. The advantage of the CPS is that it provides a better measure of the population, including those who
are employed; vs. unemployment claims.
At one time the unemployment rate for large states was based on the CPS, but no longer. And the CPS is too small for small areas. In a county of 200,000 it would only amount to about 30 people. To the extent I understand the above, it is based on a state unemployment claims, and also state-based employment models, and then controlled to the CPS.
Some of the areas you mentioned might be legit.
The Tennessee River dips into Alabama between Georgia and Mississippi, so that there is little east-west effect. Chattanooga is offset a little to the east of the AL/GA border and it appears to be red rather than purple. The whole swath trhough Alabama looks questionable but it sort of follows the cities Huntsville-Gadsden-Birmingham-Tuscaloosa-Mobile-Dothan. Mobile is also low. So perhaps there is something odd about Alabama's model for cities.
There is more coal mining in West Virginia than Kentucky. If a mine shuts down, the people are going to move. You can't collect unemployment for years, waiting for a mine to re-open. If a mine does open, people will move back.
Much of the Louisiana area is suburban New Orleans, and could actually be seeing more growth because it didn't get flooded much. There are actually a few counties right on the border that look more like SW Mississippi which has no cities. Further west, you are in Baton Rouge which has had a big increase in population post-Katrina.