Localized Shocks, Job Displacement, and Wages: Evidence from Mexico

Preliminary and incomplete: Please do not quote or cite without permission of the authors. 1 Two recent papers highlight the growing focus on wage and employment dynamics. Bergin et al. (forthcoming) describe the importance of understanding the sources and consequences of labor market volatility. Couch and Placzek (forthcoming) represents renewed interest in understanding one particular consequence of this volatility: post-displacement wage dynamics. Bergin et al. (2009) suggest that volatility may be higher in developing countries and focus on Mexico's maquiladora sector as an example. Kaplan et al. (2005), followed by Couch et al. (2009), illustrate the importance of local labor market conditions (across both time and space) in determining post-displacement wages. The differences between the United States and Mexico are consistent with international heterogeneity that emerges from other countries, such as Kuhn (2002), empirical results show much variation, there is little theoretic work seeking to identify the variables that may be driving heterogeneity in volatility and post-displacement experiences. One possible variable that might help explain both the excess volatility and the variation in post-displacement wages found in the literature is a cost to workers of separating from their jobs (a " separation cost "). The analogy for this cost in the U.S. is the driving force behind the " job lock " literature (Gruber and Madrian 1994, 2002 and Bansak and Raphall 2008) that suggests that the cost to workers from leaving their jobs reduces labor market turnover by 20-40%. This kind of cost is distinct from a cost to the employer of adjusting employment (described widely as hiring and firing costs) that have been found to be empirically relevant in the U.S. and other developed countries 2 (Hamermesh 1989 and Hamermesh et al. 1996) but an order of magnitude smaller in Mexico (Robertson and Dutkowsky 2002). Rather than affecting employers, the employee-based costs might result from risk aversion or loss of health insurance. The goal of this paper is to explore the possible consequences of the separation cost in Mexico. In section 1, we develop a very basic labor market model of imperfect competition and a separation cost. This model generates a number of predictions for employment and wage dynamics that we evaluate empirically in the rest of the paper. Specifically, section 2 contains a description of the matched worker-firm data used in section 3 to explore employment dynamics following Davis et al. (1996) and in section 4 to …