The routinization hypothesis and its empirical applications

Abstract

This doctoral dissertation investigates the routinization hypothesis and its empirical applications. The topic is addressed by means of three different chapters, each of which focuses on different aspects of the subject matter. In the first chapter, I review the most important findings of the recent research activity on the relationship between technological progress and the labor market – to a large extent represented by the literature on the routinization hypothesis. By means of this survey, I show that albeit technical progress is rather unlikely to be detrimental for overall employment growth, the labor demand for different occupational tasks substantially changes as a consequence of technology - which is the main empirical evidence recovered by the literature on the routinization hypothesis. In the second chapter, I map U.S. occupational data into European employment data to assess the effects of exposure to automation on the decline of routine occupations in Europe. I document that, similarly to what found for the United States, higher regional specialization in routine employment is associated to more pronounced employment polarization patterns. Moreover, I find that the effect of exposure to routinization is predominantly associated to within-industry contractions in routine employment. In the third and last chapter, finally, I use German administrative data to assess whether urban agglomeration processes may account for a larger contraction in routine employment in cities. By taking into account the effects of automation and by addressing endogeneity concerns related to measures of employment density, I show that the effect of agglomeration on the contraction of routine tasks is stable, sizeable and highly significant.