Rent-sharing, Holdup, and Wages: Evidence from Matched Panel Data

When wage contracts are relatively short-lived, rent sharing may reduce the incentives for investment since some of the returns to sunk capital are captured by workers. In this paper we use a matched worker-firm data set from the Veneto region of Italy that combines Social Security earnings records for employees with detailed financial information for employers to measure the degree of rent sharing and test for holdup. We estimate wage models with job match effects, allowing us to control for any permanent differences in productivity across workers, firms, and job matches. We also compare OLS and instrumental variables specifications that use sales of firms in other regions of the country to instrument value-added per worker. We find strong evidence of rent-sharing, with a "Lester range" of variation in wages between profitable and unprofitable firms of around 10%. On the other hand we find little evidence that bargaining lowers the return to investment. Instead, firm-level bargaining in Veneto appears to split the rents after deducting the full cost of capital. Our findings are consistent with a dynamic bargaining model (Crawford, 1988) in which workers pay up front for the returns to sunk capital they will capture in later periods.