Positive assortative matching in the labor market is the intuitively plausible idea that good workers work at good firms, which may arise because of complementarities in the labour market. Availability of linked-worker firm data allows to estimate the contribution of worker and firm heterogeneity on wage effects (Abowd, Kramarz & Margolis, 1999). But the unobserved component of workers’ wages appears to be either not or negatively correlated with the unobserved component of firms’ average wages. There are three possible explanations/research agendas: (a) highly structural models that attempt to model this prediction; (b) estimating additional match-specific effects; (c) limited mobility bias (LMB). Our paper focuses on LMB, the idea that the correlation is biased downwards because of estimation error and lack of worker movement between firms. We show the importance of limited mobility bias using formulae, simulations and German employer-employee data. Contrary to (almost all of) the existing wage literature, we can recover positive assortative matching with real data.
Presented by:
Martyn Andrews (University of Manchester)
Date & time:
March 4, 2013 4:00 pm - April 3, 2013 4:30 pm
External seminars home