Learning and Job Search Dynamics during the Great Recession∗

I document two new facts about job search during the Great Recession: (i) search increased after individuals received (and rejected) job offers; and (ii) search decreased with cumulative failed search. To account for these facts I develop a model of sequential search in which Bayesian job seekers learn about the arrival rate of offers through their experiences looking for work. Endogenously-evolving beliefs interact with the intensive margin of search, giving rise to nonmonotonic search dynamics over the spell of unemployment. I decompose the effect of failures to find work into competing income and substitution effects corresponding to rigorous notions of motivation and discouragement and demonstrate that the model endogenously generates negative duration dependence in unemployment exit rates. I structurally estimate the model and show that the mechanism simultaneously accounts for the empirical profiles of search time, offer arrival probabilities and offer acceptance probabilities over the first two years of unemployment.