To investigate the role of information frictions in the labor market, we collect rich and novel panel data on individuals’ labor market expectations and realizations. We find that labor market expectations (earnings; offers) are, on average, highly predictive of actual outcomes. Despite their predictive power, however, deviations of ex post realizations from ex ante expectations are often sizable. The panel aspect of the data allows us to study how individuals update their labor market expectations in response to these unanticipated shocks. We find a meaningful response: individuals revise their offer expectations up (down) by $0.46 for every $1 underestimation (overestimation) of wage offers in the previous period. We embed the empirical evidence that we document on expectations and learning into a model of search on- and off- the job with learning, and estimate the model using our data on expectations and learning. We show that the model with information frictions and learning fits the data better than one estimated assuming rational expectations. Moreover, the two models imply different transition rates, suggesting that inference would be biased if conducted assuming rational expectation. Finally, we compute the gains from removing information frictions. Because of the high rate of learning by workers, we find that the long-run gains from rational expectations are small. In other words, rational expectations is a fairly good assumption in this dimension. We also compare the welfare gains from relaxing information frictions to those from reducing search frictions, and find that the gains from eliminating information frictions are roughly equivalent to raising offer arrival rates by 2-10 percent.
Presented by:
Basit Zafar, Arizona State University
Date & time:
April 23, 2018 3:00 pm - April 23, 2018 4:30 pm
Venue:
ISER large seminar room, 2N2 4.16
External seminars home