Productivity Dynamics and Structural Change in the U.S. Manufacturing Sector

The paper investigates structural change among the four-digit (SIC) industries of the U.S. manufacturing sector during 1958-96 within a distribution dynamics framework. Focus is on the transition density of the Markov process that characterizes the value added shares of the industries. This transition density is estimated nonparametrically as well as by maximum likelihood, in which case the functional form of the density is derived from a search theoretic model. The nonparametric and the maximum likelihood fits show striking similarities. The relation of structural change to a relative measure of total factor productivity change is tested by an application of quantile regression and is found to be significantly positive throughout.