Using administrative panel data on the entire Danish population we document a new set of facts characterizing occupational mobility. For most occupations, mobility is U-shaped and directional: not only low but also high wage earners within an occupation have a particularly large probability of leaving their occupation, and the low (high) earners tend to switch to new occupations with lower (higher) average wages. Exceptions to this pattern of two-sided selection are occupations with steeply rising (declining) productivity, where mainly the lower (higher) paid workers within this occupation tend to leave. The facts conflict with several existing theories that are used to account for endogeneity in occupational choice, but it is shown analytically that the patterns are explained consistently within a theory of vertical sorting under absolute advantage that includes learning about workers’ abilities.
The U-Shapes of Occupational Mobility
Review of Economic Studies, 2015, Vol 82 (2), 659-692. With Fane Groes and Iourii Manovskii.
Occupational mobility is highest for high and low earners, and the former move “up” and the latter “down” as in models of vertical re-sorting.
Explore More
American Economic Journal - Macroeconomics, 2022, 14(4), 1-97, with Michèle Belot and Paul Muller. In a field experiment, we study how job seekers respond to posted wages by randomly assigning wages randomly to pairs of otherwise similar vacancies in a large number of professions, which generates significantly more but not exclusive interest at higher wages. Go to paper
American Economic Review P&P, 2017, 107(5): 158–162 With J. Greenwood, C. Santos & M. Tertilt. In a quantitative equilibrium model of sexual behavior and HIV/AIDS transmission we study policies that encourage long-term partnerships. Go to paper
Publications
Inferring Risk Perceptions and Preferences using Choice from Insurance Menus: Theory and Evidence
Economic Journal, 2021 131: 713-744. With Ericson, Spinnewijn &, Starc Demand for insurance can be driven by high risk aversion or high risk, and we show how to separate the two using observed market shares. Go to paper
Publications
Quarterly Journal of Economics, 2008/123(2), pp. 621-661. With A. Postlewaite. [technical appendix] In a model of social learning, the better informed (wealthier) consumers get preferential service because their consumption signals high quality to others. Go to paper
Publications
With Michèle Belot and Paul Muller. American Economic Journal: Microeconomics, 2025, 17 (1): 130–170. We propose a simple method to elicit time preferences at the individual level even when income and consumption varies over time. We validate the method and apply it to correlate individuals' impatience with job search behavior and success. Go to paper
Econometrica, 2019 87(4): 1081-1113. With J. Greenwood, C. Santos and M. Tertilt A calibrated equilibrium search model of an HIV/AIDS epidemic is developed to analyze the direct impact and the behavioral adjustment to policies. Go to paper
Publications
Review of Economic Studies, 2011, Vol. 78 (3), 872-906. With Jan Eeckhout. Wage and employment data can identify the strength of sorting in search models, though two-sided fixed effects are always mis-specified. Go to paper
joint with L.Brotherhood, C. Santos and M. Tertilt, accepted at Review of Economic Studies. This paper investigates the importance of the age composition for pandemic policy design in a framework with age heterogeneity, individual choice, and incomplete information, emphasizing the value of testing and testing capacity buildup. Optimal social policy follows an asymmetric approach by locking down the young relatively more than the old, and testing substantially lowers economic costs and laxer lockdown. We use the framework to provide systematic insights into pandemics caused by different viruses. Go to paper


