Wait, wait ... Don’t tell me

A. Sinan Ünür
September 8, 2008

Professor Krugman seems to have a penchant for self-serving starting points when showing graphs. In a recent blog post, he posted a graph of the most comprehensive measure of unemployment called U-6 by the BLS. Interestingly enough, he chose the starting point of his graph to be January 1998 even though the series is available from January 1994. Why not show the full time series?

The answer to that question may be revealed by comparing the graphs. First, here is the graph he showed:

[ Krugman's U6 plot 1/98 - 8/08 ]

Note the compressed vertical axis. This, of course, is due to the fact that the graph was automatically generated by a computer program. However, the compressed axis does exaggerate the magnitude of changes in the series.

Obviously, the rate was low in 1998, spiked after the September 11 terrorist attacks and there is a recent spike. But, let us look at the full span of available data.

[ Full U6 plot 1/94 - 8/08 ]

In addition to showing the full span of the data, the graph above gets rid of the compressed vertical axis.

U6 was higher in January 1994, at a seasonally adjusted rate of 11.8%, than it was in August 2008, at a seasonally adjusted rate of 10.7%. Of course, given the fact that the rate is increasing, this offers very little solace. On the other hand, an objective observer might also note that the spike following the September 11 attacks fell well short of the high reached a year into President Clinton’s term (recall that the Republican’ took control of the Congress following the 1994 elections).

So, why did Professor Krugman exclude the period from January 1994 to December 1997 from his graph?

Recent posts