Solow vs. Solow: Notes on identification and interpretation in the empirics of growth and development

Most empirical studies of long run growth refer to one of the two seminal contributions by Robert Solow (1956, 1957). His work shows that in order to estimate the relative roles of factor accumulation and technology in development, an a priori identification assumption is needed about the nature of technical change. This specific assumption differs across the two Solow papers. I show that starting with the identification assumption made in Solow (1956), one should expect to find that differences in technology rather than differences in factor accumulation explain most if not all of the observed long-run differences in output per worker. The opposite interpretation appears to prevail in parts of the recent literature on the empirics of growth.