Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We access a long panel dataset of US workers to document the extent to which individuals are heterogeneous with respect to their pattern of transitions across employment states. We find that heterogeneity is well approximated by three latent types: αs, βs and γs. Workers of type α leave unemployment quickly and, once they find a job, they are likely to keep it for more than 2 years. Workers of type γ find employment slowly and, once they do find a job, they are likely to leave it within 1 year. We use our empirical findings to calibrate a search-theoretic model in which workers are heterogeneous with respect to the parameters governing their employment transitions. We find that αs move quickly out of unemployment to employment because they have large gains from trade, and they are likely to stay on a job for more than 2 years because their productivity is similar in different jobs. In contrast, γs exit unemployment slowly because their gains from trade are small, and they are likely to leave a job within 1 year because they are much more productive in a small fraction of jobs than in the majority of jobs. We find that a negative shock to aggregate productivity leads to a large and persistent increase in unemployment that is mainly driven by γ-workers. The predictions of the model align well with the unemployment dynamics observed during the Great Recession.