[The Evolution of Unemployment in the United States: 1968-1985]: Comment

Murphy and Topel set as the objective of their article "to confront various theories of unemployment" with the data. An admirable goal, but as my reading of the article suggests that the "sectoral shift" hypothesis is the only theory directly confronted, I find myself, as a believer in the theory, in a somewhat defensive position. The picture they paint does certainly not support the view that the process of labor reallocation to exogenous events accounted for a significant part of the higher unemployment experienced in the U.S. economy in the seventies and eighties. To the contrary, if I read their article correctly, it is their opinion that sectoral shocks account for none of the upward trend or cyclical pattern of unemployment over the last fifteen years. My comments then, will be aimed at resurrecting the sectoral shift hypothesis from the ashes of this article. I will not argue with the facts presented by Murphy and Topel, at least not much, but rather with their interpretation. To begin, let me list two sets of facts that Murphy and Topel find most damaging to the sectoral shift hypothesis. 1. Labor mobility patterns are not consistent with the sectoral shift hypothesis. First, interindustry mobility at the two-digit level declines over the period while unemployment rises. Second, while the incidence of unemployment is much higher for industry movers, the contribution to unemployment of industry stayers is much higher than that of movers. 2. Unemployment over the period is characterized by neutrality, that is, it is not heavily concentrated in particular industries or demographic groups of the economy. In what follows I will argue that (1) is totally consistent with and is in fact essential to the sectoral shift hypothesis, and that (2) is only partially true and can be explained by a variety of factors that are consistent with the theory.