Identification of Search Models using Record Statistics

Abstract

Identification of Search Models using Record Statistics ∗ Gadi Barlevy Economic Research Department Federal Reserve Bank of Chicago 230 South LaSalle Chicago, IL 60604 e-mail: gbarlevy@frbchi.org March 8, 2005 Abstract This paper shows how record-value theory, a branch of statistics that deals with the tim- ing and magnitude of extreme values in sequences of random variables, can be used to non- parametrically identify the offer distribution of wages workers face. Using NLSY wage data, I show that the data supports the hypothesis that the wage offer distribution is Pareto but rejects that it is lognormal. In addition, I show that my approach can be used to construct a bound on the return to job-specific human capital. Using the same NLSY data, I find that job-specific human capital plays only a minor role in the wage growth of the workers in my sample. Instead, wage growth among the young workers in my sample appears to be driven primarily by the accumulation of general human capital as well as on-the-job search. Key Words: On-the-Job Search, Non-Parametric Identification, Wage Growth, Specific Human Capital I would like to thank Joe Altonji, Susan Athey, Marco Bassetto, Jeff Campbell, Kim-Sau Chung, Zvi Eckstein, Chuck Manski, Francesca Molinari, Chris Taber, Nick Williams, and especially Eric French and H. N. Nagaraja for helpful discussions. I also wish to acknowledge the help of my research assistant Merritt Lyon. Finally, I have benefitted from comments at seminars in Arizona State, the Federal Reserve Bank of Minneapolis, Tel Aviv University, Michigan State, NYU, Penn State, University of Pennsylvania, University of British Columbia, SUNY Albany, Berkeley, the Society of Economic Dynamics, and the NBER Summer Institute. The views expressed here do not necessarily reflect the position of the Federal Reserve Bank of Chicago or the Federal Reserve System.