THE Unemployment Rate (part 2)

The national unemployment rate is about as useful a number in a nation as large as the United States as a national traffic report would be. In this series of articles we’re going to peel back the layers on THE unemployment rate in an attempt to get a number that’s a little more useful. The first article in this series laid some ground work by giving an overview of what is meant by the seasonally adjusted unemployment rate. This time we’ll use that quantity and examine the unemployment rate in various states.

The image below comes courtesy of the Bureau of Labor Statistics website and shows graphically the unemployment rate by state across the nation for April 2009 (the most recent graph). The link above contains not only this image, but a table that gives the numerical unemployment rate for each state. The image shows what we expected, namely the unemployment rate varies significantly from state to state. Michigan and Oregon are the highest (seasonally adjusted) at 12.9% and 12% respectively while Nebraska and North Dakota are the lowest at 4.4% and 4.0% respectively)

1244654297009_mapBut this only provides a snapshot at a single point in time. It’s only natural to ask the follow up question, are we pulling out of this slump? One way to get a sense of that is to animate this graph. Fortunately the BLS website provides this map data as far back as 1978. Below is an animation of the unemployment rate by state from April 2008 to April 2009. One thing that stands out is it appears that unemployment is spreading from California to neighboring states. There are many reasons why this may be the case. Las Vegas (for example) is only a 4 hour drive from Los Angeles and it’s easy to imagine that fewer Californians would be visiting Las Vegas as the California unemployment rises. And there are other things that may tie neighboring economies into each other (especially since housing is at the root of this recession), however I do not know for sure if proximity to California had an effect on Oregon’s and Nevada’s unemployment rates. Indeed, it’s hard to detect a coherent spreading pattern in unemployment on the east coast as many of the changes there seem to be simultaneous. One bright spot though is that Indiana’s unemployment rate improved slightly in April 2009.


As useful as this animation may be, the “chunkiness” of the data makes it less useful that it could otherwise be. The darker ends of the spectrum cover larger ranges of unemployment. Most of the color blocks cover one percentage point, but the 2nd darkest covers 3 and the darkest one covers 50 (although since the top end of the actual data is less than 13% in practice it too is only covering 3 percentage points). Anyone who cares about the difference between 7.0% and 9.9% will not be able to distinguish these from the maps the BLS supplies. This affects both fluctuations and total unemployment rate. Any states that are improving (or getting worse) in this range show no change in this animation. Additionally, if every state in the country were at 7.1% unemployment the national map would look the same as if every state were at 9.9% unemployment and one is much much worse than the other.

Audio engineers call this clipping. But when they envision clipping the animation above is not the image that springs to mind. What does spring to mind is closer to the image below, which shows both the actual nationwide unemployment rate over the same time period and a “clipped” version rounded to illustrate the effect of the shading in the graph above. While the graphical chart and the animation has its uses, the clipping hides small fluctuation, and if you’re looking for early signs of a turn-around you won’t find it here until it is well underway.


Because of this we’ll continue to show the data in a more traditional graph. To reduce clutter in the graph, let’s examine the unemployment rates for the 4 states that have been hit the worst and the least in this current recession. How would you define “worst” and “least” in this context? For our purposes I’ll focus on:

  • highest unemployment rate (worst absolute)
  • biggest change (for the negative) in unemployment rate (worst relative)

The first one ought to be self explanatory. Determine which states have the highest unemployment rate now and you got your list. The second one is designed to catch states that, even if their unemployment rate now isn’t horrible, if it doubled or went through a similarly large change because of the recession it could also qualify as “worst” hit. (There are counterparts to these on the “least hit” side.) Of course it’s possible for some states to be in both of these categories. Indeed, if overlap didn’t occur then the graph below would be for 16 states: 4 worst absolute, 4 worst relative, 4 least absolute and 4 least relative. But there is significant overlap, the graph below only covers 11 states.


This is a graph of these 11 states, graphed in an “absolute” mode. Meaning that it shows the actual unemployment rate and not the relative increase in unemployment. Michigan and Oregon have the highest statewide unemployment at 12.9 and 12% respectively. From a relative standpoint it is possible to see that Oregon stands out as its unemployment rate more than doubled since Jan 2009 2008.

This graph shows a more nuanced picture than the nationwide animated map. One of the first things that stands out is that single line in the middle (on the right hand side). The lines at the top clearly belong to the “worst” hit states (both definitions) and the lines at the bottom belong to the “least” hit states, but what is this line in the middle? It belongs to Alaska and is one of the “least” hit states. Alaska is #4 in the least hardest hit relatively speaking but only because its unemployment rate was relatively high back in Jan 08. In an absolute sense Alaska’s unemployment rate falls in the middle of the nation, but in a relative sense, since its unemployment rate started out high in a relative measure it didn’t increase that much. An oddity to be sure, but that’s the point of showing all the data. With data we can debate the merits of including Alaska in a chart such as this.

Next, note that the unemployment rate has increased for every state in the country. From the left edge to the middle (for all of 2007), the lines remain relatively flat. But from the middle of the graph to the right side (all of 2008 and YTD 2009) all the lines move upward. Even those least hit had an increase.

Finally, the unemployment rate appears to be leveling out. Even though the national chart shows no sign of slowing down, the individual states shown here are leveling off or even slightly decreasing in unemployment rate.

In looking at the national number above we see an unemployment rate that shows no sign of slowing down, but digging a little deeper we see signs that the rate is beginning to slow in some of the hardest hit states and in some of the least hardest hit states as well. Indeed, practically every state in the chart above shows signs of slowing unemployment rates. A positive sign. The moral? When we look a little deeper into the numbers a more useful story can sometimes be told.

Next time we’ll examine the unemployment rate by various demographic sectors.

2 comments for “THE Unemployment Rate (part 2)

  1. September 12, 2016 at 4:15 pm

    Do you have any video of that? I’d like to find out more details.

Leave a Reply

Your email address will not be published. Required fields are marked *