As it turns out, the three-tenths of a percent decrease in jobless claims announced last week by the Minnesota Department of Employment and Economic Development is the backdrop to an even larger story of (relatively) good economic news for Minnesotans.
While Minnesota has endured record setting increases in unemployment over the past year, the Gopher State has nonetheless experienced one of the lowest increases in unemployment nationwide, particularly since President Barack Obama's signing of the American Recovery and Reinvestment Act of 2009 (ARRA).
A Smart Politics analysis finds Minnesota has experienced the third lowest net increase in unemployment nationwide since the enactment of ARRA in February of this year. While the average state has been burdened with a 15.8 percent (1.1 point) increase in jobless claims during this five month span, Minnesota's unemployment rate is up just 1.3 percent (0.1 points) - from 8.0 to 8.1 percent.
Only two states - Vermont and North Dakota - can boast better numbers. Each of those states have actually seen a decrease in statewide unemployment since February - at a rate of 4.2 percent 2.3 percent respectively.
Critics of President Obama have pointed out that only a fraction of the stimulus package has reached the states, particularly for "shovel ready" projects that could help to alleviate the unemployment crisis that has beset the nation during the past year.
There is no denying the disparate impact the jobs problem has had across the country in recent months.
For example, the statewide average unemployment rate has increased 20.0 percent in the 13 Western states from February through July. That is almost twice the rate of increase as found in the 9 Northern states during this period (11.2 percent). The 12 Midwestern states (15.1 percent) and 16 Southern states (16.2 percent) fall directly in between.
In the Upper Midwest, the job situation in North Dakota (#2, -2.3 percent), Minnesota (#3, 1.3 percent), and South Dakota (#8, 6.5 percent) has fared well since February, compared to the nation as a whole.
Meanwhile, unemployment has been much worse for Minnesota's eastern and southern neighbors. In Wisconsin, unemployment is up 15.4 percent since February, only the 30th best in the country. In Iowa, jobless claims are up 32.7 percent - the third highest rate in the nation, behind only Wyoming (66.7 percent) and West Virginia (50.0 percent).
The worst economic news also tends to be coming out of the red states, not the blue and purple states that helped to elect President Obama last November.
Since February, the average statewide increase in unemployment is notably higher in the 22 McCain states, at 19.7 percent, than the 29 Obama states (and District of Columbia), at 13.0 percent.
In addition to Wyoming and West Virginia, McCain states that have been hit particularly hard are Idaho (31.3 percent increase), Louisiana (29.8 percent), Kansas (25.4 percent), and Arizona (24.3 percent).
Lowest Increase in Unemployment Since February 2009 by State
Rank |
State |
Feb-09 |
Jul-09 |
Change |
Percent |
1 |
Vermont |
7.1 |
6.8 |
-0.3 |
-4.2 |
2 |
North Dakota |
4.3 |
4.2 |
-0.1 |
-2.3 |
3 |
Minnesota |
8.0 |
8.1 |
0.1 |
1.3 |
4 |
North Carolina |
10.7 |
11.0 |
0.3 |
2.8 |
5 |
Virginia |
6.6 |
6.9 |
0.3 |
4.5 |
6 |
Alaska |
7.9 |
8.3 |
0.4 |
5.1 |
7 |
Connecticut |
7.4 |
7.8 |
0.4 |
5.4 |
8 |
South Dakota |
4.6 |
4.9 |
0.3 |
6.5 |
9 |
Mississippi |
9.1 |
9.7 |
0.6 |
6.6 |
10 |
D.C. |
9.9 |
10.6 |
0.7 |
7.1 |
11 |
Maryland |
6.8 |
7.3 |
0.5 |
7.4 |
12 |
Hawaii |
6.5 |
7.0 |
0.5 |
7.7 |
12 |
Maine |
7.8 |
8.4 |
0.6 |
7.7 |
14 |
South Carolina |
10.9 |
11.8 |
0.9 |
8.3 |
14 |
Colorado |
7.2 |
7.8 |
0.6 |
8.3 |
16 |
Washington |
8.3 |
9.1 |
0.8 |
9.6 |
17 |
New York |
7.8 |
8.6 |
0.8 |
10.3 |
18 |
Oregon |
10.7 |
11.9 |
1.2 |
11.2 |
19 |
Florida |
9.6 |
10.7 |
1.1 |
11.5 |
20 |
Montana |
6.0 |
6.7 |
0.7 |
11.7 |
21 |
Georgia |
9.2 |
10.3 |
1.1 |
12.0 |
21 |
Missouri |
8.3 |
9.3 |
1.0 |
12.0 |
23 |
California |
10.6 |
11.9 |
1.3 |
12.3 |
23 |
Delaware |
7.3 |
8.2 |
0.9 |
12.3 |
25 |
Indiana |
9.4 |
10.6 |
1.2 |
12.8 |
26 |
Pennsylvania |
7.5 |
8.5 |
1.0 |
13.3 |
27 |
New Jersey |
8.2 |
9.3 |
1.1 |
13.4 |
28 |
Nebraska |
4.3 |
4.9 |
0.6 |
14.0 |
29 |
Massachusetts |
7.7 |
8.8 |
1.1 |
14.3 |
30 |
Wisconsin |
7.8 |
9.0 |
1.2 |
15.4 |
31 |
Arkansas |
6.4 |
7.4 |
1.0 |
15.6 |
32 |
Utah |
5.1 |
6.0 |
0.9 |
17.6 |
33 |
Ohio |
9.5 |
11.2 |
1.7 |
17.9 |
34 |
Oklahoma |
5.5 |
6.5 |
1.0 |
18.2 |
35 |
Kentucky |
9.3 |
11.0 |
1.7 |
18.3 |
36 |
Tennessee |
9.0 |
10.7 |
1.7 |
18.9 |
37 |
New Hampshire |
5.7 |
6.8 |
1.1 |
19.3 |
38 |
Illinois |
8.6 |
10.4 |
1.8 |
20.9 |
39 |
Rhode Island |
10.5 |
12.7 |
2.2 |
21.0 |
40 |
Alabama |
8.4 |
10.2 |
1.8 |
21.4 |
41 |
Texas |
6.5 |
7.9 |
1.4 |
21.5 |
42 |
Arizona |
7.4 |
9.2 |
1.8 |
24.3 |
43 |
Michigan |
12.0 |
15.0 |
3.0 |
25.0 |
43 |
Nevada |
10.0 |
12.5 |
2.5 |
25.0 |
45 |
Kansas |
5.9 |
7.4 |
1.5 |
25.4 |
46 |
New Mexico |
5.4 |
7.0 |
1.6 |
29.6 |
47 |
Louisiana |
5.7 |
7.4 |
1.7 |
29.8 |
48 |
Idaho |
6.7 |
8.8 |
2.1 |
31.3 |
49 |
Iowa |
4.9 |
6.5 |
1.6 |
32.7 |
50 |
West Virginia |
6.0 |
9.0 |
3.0 |
50.0 |
51 |
Wyoming |
3.9 |
6.5 |
2.6 |
66.7 |
State Average |
7.6 |
8.8 |
1.1 |
15.8 |
|
National Average |
8.5 |
9.4 |
0.9 |
10.6 |
Follow Smart Politics on Twitter.
While I really enjoy your analyses and think you do a great job in general, I am concerned with your use of a second derivative indicator of unemployment. While Wyoming may have seen a major percentage increase in the percentage of unemployment, I would take 6.5% over MN's 8.1%. A better indicator of economic distress is the point uptick in unemployment (the change column). Your chosen highlight statistic does tell a story, but not a terribly effective one. It is meaningless without the underlying unemployment rate.
Does this stat provide more predictive/interpretive power than a political correlation with the change rate? Is it more predictive/interpretive than a correlation with the overall rate?
Using the percent of change factor does create a wider disparity in outcomes, but I'm not sure that it is an appropriate measure.
The story that it generates is that blue state unemployment is accelerating more slowly than in red states, but five months is a pretty short timeline and the effect of industry concentration (including seasonality) would probably provide a lot of signal noise.
No statistic is perfect, and I love to play with numbers too, but this analysis seems like a derivative too far. Try it one step back, and see what you get.
> The story that it generates is that blue state unemployment is
> accelerating more slowly than in red states, but five months is
> a pretty short timeline and the effect of industry concentration
> (including seasonality) would probably provide a lot of signal
> noise.
The numbers provided in the tables do reflect the seasonally adjusted unemployment data.
The focus of this piece was on the percent change rather than the absolute number to normalize the statistics across the states. For example, the 0.9-point increase in unemployment experienced in Utah (17.6 percent) during these past five months is significantly more pronounced than the 0.9-point increase that took place in South Carolina (8.3 percent), which already had a very high rate in February.
If one rank-ordered the states based on the absolute increase in unemployment (the 'change' column), Minnesota would still rank as the 3rd lowest in the nation.
Good point on the seasonally-adjusted data, but it still doesn't account for industry concentration. And maybe that is fairly included...such as agriculturally-based economies benefit less from the stimulus than an industrialized or high-tech economy (a traditional fault line for red/blue).
But my real critique is that percent change really isn't that important in terms of unemployment. And dividing out the rate of unemployment doesn't normalize states, it doesn't have anything to do with it.
Having a 1.0% unemployment rate tick up to 1.5% is a 50% increase. But in the end, it only means that half a percent of the population became unemployed. (and yes, that isn't necessarily an accurate description)
In another state, the rate may have gone from 6% to 7.5%, only a 25% increase. But you had 1.5% of the population thrown onto the bench. Politically, economically, and personally, that would be more important than the first example.
Unless you take seriously that misery loves company, and that having more fellow people on unemployment makes it more acceptable to be unemployed or that it increases the public's acceptance of unemployment, I would say say the percent change calculation is not that important.
A theory could be put forward that high unemployment begets volatility in the rate of unemployment, in which case your normalization would be appropriate. But you didn't assert such a premise.
The stat that you're pursuing is what impact is the stimulus having on unemployment. My concern is that by using the additional percentage factor, you have lost your connection to the political and economic consequences.
Your method simply inflates the impact on any low-unemployment state, and decreases the impact on any high-unemployment state without justifying the magnification.
You have two sets of important data: the overall rate of UI, and the change in that rate. One tells the story of how bad the economy is, and the other tells the story of what direction it is moving. I think controlling for the magnitude of unemployment in a state is expressly inappropriate without further justification.
Again, I want to tip my cap that I really enjoy your work, I just don't think this one is calculated appropriately.