Abstract

We examine the lack of consensus in the literature about the impact of insider ownership on firm value. We apply non-parametric and semi-parametric methods to analyze various subsamples based on firm size, age, index listing, and institutional ownership. Sample selection is an important determinant of the discord in the literature. Different subsamples lead to widely different estimated ownership-firm value relationships—from upward sloping to downward sloping to flat to V and inverse-V. We replicate and show this for four influential papers (Agrawal and Knoeber, 1996; Himmelberg et al., 1999; McConnell and Servaes, 1990; Morck et al., 1988). By cataloguing systematic differences across subsamples that mirror corresponding disagreements among existing studies, we provide a unifying perspective on the literature.