The impact of computers on manufacturing productivity growth: A multiple-indicators, multiple-causes approach

Abstract

An increase in computer usage could improve product and labor quality. Unfortunately, many quality improvements are not incorporated in price indexes. Thus, a quality bias could distort conventional estimates of the marginal productivity of computers, which are based on the assumption that prices are measured without error. Using detailed industry data, we estimate a multiple-indicators, multiple-causes model that allows us to investigate this relationship, while controlling for measurement errors. Our findings suggest that computers are an important source of quality change and that computers are positively related to productivity growth when adjustments are made for measurement errors.