Market Responses to Interindustry Wage Differentials

HRUJH -%RUMDV 9DOHULH $ 5DPH\ .HQQHG\ 6.KRRO RI *RYHUQPHQW 'HSDUWPHQW RI (.RQRPL.V +DUYDUG 8QLYHUVLW\8QLYHUVLW\ RI &DOLIRUQLD 6DQ 'LHJR -).6WUHHW /D -ROOD &$ &DPEULGJH 0$ DQG 1%(5 DQG 1%(5 YUDPH\#X.VGHGX JERUMDV#KDUYDUGHGX Katz, 1992;Blackburn and Neumark, 1992).However, the available evidence on the sorting hypothesis is mixed: some studies find evidence of unobserved ability differences, but most studies do not.The cumulative evidence leads many observers to conclude that the persistence of interindustry wage differentials challenges the implications of competitive labor market theory.This paper offers new evidence showing that the market does adjust in response to interindustry wage differentials.Although the tendency for wages to converge across industries is extremely weak, several other economic variables exhibit strong long-run responses to the initial interindustry wage structure.In particular, industries that paid relatively higher wages in 1959 experienced significantly slower employment growth over the subsequent thirty years.The high-wage industries also experienced slower GDP growth and faster growth of the capital-labor ratio and labor productivity.Our evidence rejects the simplest competitive model of labor markets where flows of workers across industries provide an equilibrating mechanism for industry wages.Instead, workers flow in a direction opposite to that predicted by the competitive model.When considering all of the evidence presented, we conclude that noncompetitive wage theories, such as efficiency wages or rent sharing, are more plausible explanations than unobserved ability differences.In fact, practically any theory that generates a rigid interindustry wage structure is consistent with the dynamic correlations documented in this paper.The "story" that explains much of the empirical evidence is that firms in high-wage industries respond to their immutably high wages by substituting capital for labor and by increasing labor productivity.At the same time, the market responds by switching to goods produced by lower-cost industries.The paper proceeds as follows.Section II describes the data and methods used to estimate the industry wage premia.Section III specifies and estimates a simple competitive model of