Mankiw on Business Cycle Dating and the 2001 Recession
N. Gregory Mankiw has an interesting post on the question of when exactly did the 2001 recession start? Mankiw points out that four of the time series used in determining when recession start and end have been revised substantially and that the March 2001 dating of the recession is probably too late.
There is quite a bit of validity to this claim, IMO. For example, look at the employment data. Frist up is the payroll survey.
Clearly the time series reaches a (local) peak right about 2001 (the data actually start declining in March of 2001).
Next look at the unemployment rate,
Again, we see that the rise in the unemployment rate is right around the start of 2001 (January 2001 to be exact). Now one might conclude, “This doesn’t point to a recession starting in 2000.” The problem is that employment is a lagging economic indicator. That is, unemployment might continue to go down or not rise for sometime after the end of the recession. Similarly with the start of a recession, unemployment can go down or not decrease even after a recession has started. If the time of the lag is just two or three months this would push back the start of the recession to prior to Bush came into office and could possibly move into the last quarter Clinton was in office…if you have your panties in a knot over which President is in office when recessions start and end.
Mankiw also points to this paper by James Hamilton (a very well known and highly respected time series econometrician) and Marcelle Chauvet. Hamilton and Chauvet put the start of the last recession at September 2000 (see table 6). However, also note that in Table 8 we have the results of the recursive estimation methodology and this puts the start of the recession at March 2001 which coincides exactly with the NBER start date.
James Hamilton’s co-blogger, Menzie Chinn, also looks at this controversy and comes to the conclusion that the recession started in 2001. However, Chinn also notes that given the data reasonable people can disagree about this. Dating business cycles is more of an art than a science (although those who do it try to be as scientific as possible). Saying that the recession started sometime between the last quarter of 2000 and the first quarter of 2001 is almost surely true. Of course for all the partisan bloggers, talking heads, and radio show pundits this isn’t all that helpful as they can’t really point the blame finger at anyone.