Thursday, December 3, 2009

A History of Unemployment

Today I decided to create box plots of the United States Unemployment data dating back to 1948 to the most current month of October 2009. Here are the results... (made using R)
If you are new to box plots this website (NetMBA) does an excellent job of explaining it. The abridge version is this: the bolded line is the median, the rest of the lines which create the "box" represent 1st to 3rd quartiles. Any dot/circles outside the box represent outliers. The graph should be read as in January the median unemployment from 1948-2009 was 6.3%.

There is a nice little wave pattern that begins in January, at it's peak, then starts it's decline till May then buckles back up in June then declines as we move towards December and of course repeats. These data do suggest we have months where unemployment is highest (January) and lowest (October). It also suggests that some months have more variability (February having 1.62 standard deviation), while other have wider ranges (June).

I think this says a lot about our economy. During Quarter 4, (Holiday season) when most profits are made, unemployment is low because these business need to meet the higher demand. Consequently, they no longer need these workers after the holiday season, which may account for the step median decline in the unemployment rate from December to January.

What does this mean in our current economic climate? WATCH OUT! January numbers may be VERY interesting...

Other unemployment musings: