Brad DeLong's Semi-Daily Journal: When Did the Recession of 2001 Begin? Do We Care?: Reminds me of the best abstract in the history of economics:
Paper Title: "Unit Roots in GNP: Do we Know and Should we Care?"
Abstract: No and probably not.
Authors (sue me for forgetting your names) [Larry Christiano and Marty Eichenbaum].
I think we can agree that an economic downturn was inevitable by Jan 20 2001 and indeed by November 2 2000 so it was not George Bush's fault. We can also agree that the concept of "the date a recession began" is a bit metaphysical and debate is silly.
Most importantly, I think that we can all agree that the CEA has no business talking about business cycle timing. The issue is politically important and can't be made mechanical so it is essential that political appointees not be allowed to participate in the debate. This is not because the question is really important and has nothing to do with the integrity of this or that political appointee.
There are many equally useful ways of timing business cycles. There is one way which is much worse than all the others which is to allow politicians to decide dates so as to blame other politicians. Even if Mankiw had good reason to think that the recession began in 2000, out of respect for the valuable principle of non-partisan business cycle timing, he should have kept his mouth shut so long as he was CEA chairman.
I am more enthusiastic about dating business cycles than Robert is. At a very impressionable age--I was nineteen at the time--Marty Feldstein taught me Edmond Malinvaud's (1989) Theory of Unemployment Reconsidered (New York: John Wiley & Sons: 0470268832). Ever since then I have firmly believed that modeling the macroeconomy through a set of linear equations has to be the wrong road: that the economy is a qualitatively different animal when employment is shrinking than when employment is growing and there are still unemployed workers eager to work at prevailing wages, and yet a different animal again when the economy is growing and there is "full employment." I think recession begin and end dates are important things to understand when they are used as markers for the phase transitions the macroeconomy undergoes. (Of course, the NBER's Business Cycle Dating Committee doesn't think about recessions and expansions in these terms--hence someday I want to take over the NBER BCDC and restore it to rationality.)
But Robert is right: recession begin and end dates are stupid things to think about when they are used to assign blame to politicians. Blame Hoover for his inadequate response to the Great Depression. (Blame Roosevelt too, but his response was much less inadequate.) Blame George W. Bush not for the state of the macroeconomy during his tenure, but for his inept policies--and for his attempts to convince the press that everything is fine.