The unemployment rate has ticked up again and there’s a round of articles discussing it. One in particular (FlowingData.com) shows a map of the US color coded with the unemployment rate by county. I’ve reproduced that chart below, if you click on it you will be taken to the original website.
But this is only half the story. The stimulus program has already spent about $140 Billion and Recovery.org has an excellent website that shows how the money has been spent. A sample of one of their graphics is below and clicking on it will take you to their website.
The obvious question arises, how do these two images relate to each other? In other words, is the stimulus spending targeting the states with the highest unemployment?
The unemployment graph from FlowingData.com correctly uses the unemployment rate for each county. Consider North Dakota which appears to have a few counties with zero unemployment. But North Dakota has a relatively small population, so you might be wondering if the benign coloring of North Dakota is due to its low population. The answer is “no”, because the rate is being graphed. Sure parts of North Dakota are sparsely populated, but most of the folks who live there have jobs. Contrast this with Alaska which is also a state with a small population. Many of its counties are colored dark red and I’d imagine the population of these counties is no larger than the population of the average county in North Dakota. If FlowingData graphed the total unemployed instead of the unemployment rate then the resulting image would essentially be a population density graph, providing us with little information about unemployment.
But we’re interested in tracking if the stimulus spending is getting to the right places and to answer this we need to compare to the total unemployment. By way of example, 100% unemployment in Rhode Island probably means fewer total people unemployed than 10% unemployment in California. Even in this extreme case it would make sense for more stimulus to be applied to California than Rhode Island.
Recovery.org (and the government’s own Recovery.gov) site shows the stimulus spending to date and the Bureau of Labor Statistics has comprehensive data on unemployment. If you take the $137 Billion in stimulus spending so far and divide by the 14.8 million people who are unemployed you find that on average $9,200 of stimulus has been spent per person unemployed. Our question then becomes how do the states compare to this average? Which states are near the average? Which are significantly above this average and which are significantly below this average? The graph below answers this question.
Yellow is neutral here, these states are close to the average. Green states received more than $9,200 per unemployed person while red states received less. The news is in the deep red and deep green states. Wisconsin received less than 1/3rd of the national average. It is way behind the curve here and its congressmen and senators should get see about attracting more stimulus dollars. Illinois and Indiana are about 1/2 of the national average and, while not quite as dire as Wisconsin, are still way behind the curve and again their governmental representatives should be much more proactive.
Where should the money for Wisconsin, Illinois and Indiana (have) come from? Taking a look at the greener states, the heaviest hitters are Nebraska and New Mexico. While the stimulus per unemployed is large in Alaska, North Dakota and Wyoming, those states have such low populations that the total dollars “wasted” there is relatively inconsequential. Contrast this with Nebraska. It received almost 5x the national average but by population it is a much larger state than North Dakota. Consequently, if it received only half what it did, it would still have received more than twice the national average (per person) but that would have freed up more than $1B to go to Wisconsin. This would have put Wisconsin in the realm of one of the dark tan states like New York. Still below average but better.
You can explore these details yourself by clicking on the graph. This will take you to an interactive version of the page where you can mouse over the state initials to get more details. These additional details are:
- Unemployment Rate – bar graph, height ranges from 0 to 16%
- Total Unemployed – area graph. Area of the square ranges from 0 to 2.2 million
- Total Stimulus Spending – area graph. Area of the square ranges from 0 to $14.8 Billion
- Stimulus Spending per Unemployed – bar graph. Height ranges from $0 to $48,000. (The line at $9,200 shows the national average)
Furthermore the two area graphs are sized so the squares are the same size if the state is close to the national average of $9,200 per unemployed. For example, consider Texas. It received an average of $8,000 per unemployed in stimulus spending. Close to but just a little below the national average. Consequently the square for the total unemployed is just a little larger than the square for the total stimulus dollars received. Compare this to New Mexico. It has a vastly smaller total unemployed than Texas (the red square is smaller) and the size of the green square is significantly larger than the unemployed square which corresponds to its much larger $41,800 stimulus spending per unemployed.
Finally a brief note about the data. Normally I present all the data, favorable and unfavorable, but here I did what I expect was a minor editorial correction. The national totals were calculated by adding up all the state totals, both from the BLS website for unemployment and for stimulus dollars from Recovery.org, with one exception. I omitted the District of Columbia. But I only did this when calculating the $9,200 average. If you mouse over the District of Columbia the detailed stats are legitimate and hence you can see why I discounted them for the overall average. (If I left DC in would have been about $9,800 per person.)
In fact I do not know why the DC average is so large. I suspect it’s an artifact of accounting because many offices that receive stimulus dollars have a branch in DC and for that reason alone a portion of every stimulus program may be erroneously credited to DC. But again I do not know this for sure. It is certainly an oddity worthy of deeper investigation.
Additionally the Recovery.org site is updated continuously. The data used above is a week old. yet today on their website they show that the total stimulus spending to date is less than the data I’m using above. (But DC is still out of whack.)
A note about the graphs.
The colored map of the US was made using the map creator tool at indexmundi.com
The mouse over application was created using Processing available at Processing.org