Endogeneity and the family involvement–firm performance relationship: on the daunting search for instrumental variables

One of the predominant themes in the family business literature is the relationship between family involvement in ownership, management and governance and firm performance. Empirical research in this field is typically plagued by tricky endogeneity problems. From a statistical point of view, endogeneity is a major concern as it violates the central Ordinary Least Squares assumption that the regressor is uncorrelated with the error term, giving rise to inconsistent and biased estimates. A common remedy for this problem is the use of instrumental-variables (IVs) techniques. However, finding suitable IVs that satisfy the conditions of both ‘validity’ and ‘relevance’ is a real challenge. This recurrent difficulty invited us to provide family business scholars with a structured discussion of the problem and to make a number of suggestions for practical solutions, which rely on recent developments in applied econometrics to cope with instruments that turn out to be 'imperfect' and/or 'weak'.