| Abstract: | We investigate the effect of board composition on overall corporate performance while controlling for managerial ownership and other key variables. We recognize that both managerial ownership and board composition may be endogenous to performance, but our work differs from previous in two important respects. First, we measure performance using the market value to book value ratio of common stock equity rather than the more commonly used Tobin's q. Second, recognizing that overall estimates from the IV approach depend greatly on the choice of instruments, we perform sensitivity analysis by using a variety of instruments to proxy for board composition and managerial ownership. Both our OLS and IV estimates indicate a significant curvilinear relation between board composition and performance. However, we find that moderate differences in first-stage regressions, resulting in small changes to first-stage R2s, lead to widely differing overall results. Our results suggest that findings of studies using IV and similar techniques (e.g. two- and three-stage least squares) must be interpreted cautiously. |
