Previous work on board size effects in closely held corporations has established a negative correlation between board size and firm performance. We argue that this work has been incomplete in analysing the causal relationship due to lack of ownership information and weak identification strategies in simultanous equation analysis. In the present paper we reexamine the causal relationship between board size and firm performance using a dataset of more than 5,000 small and medium sized closely held corporations with complete ownership information and detailed accounting data. We test the potential endogeneity of board size by using a new instrument given by the number of children of the founders of the firms. Our analysis shows that board size can be taken as exogenous in the performance equation. Furthermore, based on a flexible model specification we find that there is no empirical evidence of adverse board size effects in the typical range of three to six board members. Finally, we find a significantly negative board size effect in the minority of closely held firms which have comparatively large boards of seven or more members.
