Macroeconomics of Occupational Mobility: New Facts, Theory, and Quantitative Evaluation

University Of Pennsylvania, Philadelphia PA

Investigators

Abstract

When a worker changes occupations (e.g., sales representative, mechanic, taxi driver) she usually changes the kind of work she performs, with potentially non-trivial loss of specific skills accumulated in the previous occupation. Nevertheless, the fraction of workers changing occupations is remarkably large - close to 20% of workers in the U.S. switch occupations in a given year. Moreover, the size of worker flows vastly exceeds the amount of reallocation needed to accomodate changes in occupational employment shares. Why do workers switch occupations so much? What are the consequences for the aggregate economy? Might there be a role for government intervention? The objective of the proposed research is to provide answers to these questions. The PI uses administrative panel data on 100% of the Danish population to document a new set of facts characterizing the patterns of occupational mobility. He finds that a worker's probability of switching occupation is U-shaped in her position in the wage distribution in her occupation. It is the workers with the highest or lowest wages in their occupations who have the highest probability of leaving the occupation. Workers with higher (lower) relative wage within their occupation tend to switch to occupations with higher (lower) average wages. Higher (lower) paid workers within their occupation tend to leave it when relative productivity of that occupation declines (rises). These previously unknown facts are not implied by existing theories of occupational mobility that mostly treat occupations as horizontally differentiated sets of tasks. The PI suggests that it might be productive to think of occupations as also forming vertical hierarchies. Workers who are unsure of their abilities learn about them by observing their output realizations. Employment opportunities in each occupation are scarce inducing competition among workers for them. Complementarities in the production function between worker's ability and productivity of an occupation induce sorting of workers into occupations according to their expected ability. The PI develops an equilibrium model of occupational choice with these features and show analytically that it is consistent with patterns of mobility described above. The PI plans to embed the features of this new theory into a dynamic general equilibrium model of occupational choice that the PI developed in his earlier work to quantitatively evaluate the contribution of different economic forces - such as learning, fluctuations in occupational productivities or demands, and search frictions in locating jobs in various occupations - to the amount of occupational mobility observed in the data. The intellectual merit of the project lies in (1) discovery of a new set of facts that characterize patterns of occupational mobility, (2) an innovative research strategy of studying occupational mobility as a function of a worker's relative performance in her current occupation, (3) a development of a new theory of occupational mobility that is very tractable, yet consistent with all the features of the data it is designed to account for, (4) a development and estimation of a quantitative model of occupational mobility that would allow better quantitative evaluation of the forces driving occupational mobility. The project will have a broad impact on the economics profession because it will (1) provide a new empirical evidence important for distinguishing between various theories of occupational mobility, (2) provide a new theoretical model of occupational mobility qualitatively consistent with the motivating evidence (3) develop and calibrate a new quantitative model of worker mobility. It will also likely guide our understanding of the reasons why wages grow with occupational experience. Much of the existing literature attributes a sizable amount of this growth to selection: workers with bad occupational matches leave their occupations for better opportunities elsewhere which induces a correlation between occupational match quality and occupational tenure in a cross-section of workers. Given that the PI's findings challenge the importance of selection for wage growth because both bad and good workers leave occupations, human capital is the obvious remainder that can account for the positive relation between wages and occupational tenure. The results of this research will be presented at national and international conferences, and are likely to become a standard entry on the reading lists in advanced courses.