Many biases plague the estimation of rent sharing in labour markets. Using a Portuguese matched employer-employee panel, these biases are addressed in this paper in three complementary ways: 1) Controlling directly for the fact that firms that share more rents will, ceteris paribus, have lower net-of-wages profits. 2) Instrumenting profits via interactions between the exchange rate and the share of exports in firms? total sales. 3) Considering firm or firm/worker spell fixed effects and highlighting the role of downward wage rigidity. These approaches clarify conflicting findings in the literature and result, in our preferred specification, in a Lester range of pay dispersion of 56%, also shown to be robust to a number of competitive interpretations.
rent sharing instrumental variables matched employer-employee data fixed effects