Factors Affecting Unemployment Rate Aren't Going Away

by: Brad DeLong

Macro Advisors and other mainstream forecasters certainly expect the unemployment rate to decline: a typical glide path has “headwinds” keeping the unemployment rate at 9.2% until the end of this year, and thereafter has it declining from 9.2% to 8.6% in half a year and to 8.0% by the end of 2012.

I really do not see where this forecast is coming from.

It is certainly the case that over an average year in the entire 1948-2010 period the unemployment rate converges 23% of the way back to its sample average level of 6.1%. That regression coefficient starting with 9.2% late this year would get us to 8.7% by mid-2012 and to 8.1% by end 2012.

But since 1990 mean reversion in the American unemployment rate has ebbed away. Since 1990 we have closed on average not 1/4 but rather 1/14 of the gap between the current unemployment rate and its long-run sample average over the course of a year.

The underlying logic of the forecast and the model appears to be that the strong mean-reversion in unemployment we saw in the business cycles from 1948-1990 is still there. However, since 1990 after all three recessions--the recession of 1990-1991, the recession of 2001, and the recession of 2007-2009--the return of unemployment to normal has been held back by special, temporary factors. The forecast is thus that these special, temporary factors are about to end--or, at least, will end six months from now.

And I don't see why the factors that are keeping unemployment from falling are special, temporary, and about to end rather than the new normal, persistent, and likely to endure.