Do CEOs of State-Owned Enterprises Matter? Evidence from Brazil, 1973-1993

Abstract

We study if CEOs and their backgrounds matter for the performance of state-owned enterprises (SOEs) exploiting the fact that these companies had different CEOs (with different backgrounds) in different moments in time. We construct a database of Brazilian SOEs between 1973 and 1993 and make three tests. First, we use variance decomposition to study CEO effects. Second, we look at how much of the variation in performance can be explained by CEOs who switch companies. Third, we test if CEO background and ability matter for performance. We find sizable effects of CEOs on performance, especially for military managers and those who attended top universities.