Controlling for endogeneity with instrumental variables in strategic management research


Strategic Organization

août 2008, vol. 6, n°3, pp.285-327

Départements : Stratégie et Politique d’Entreprise

Mots clés : Causal relations, Endogeneity, Instrumental variables

This article offers a framework to understand how endogeneity arises and how to control for it with instrumental variables to estimate causal relations with observational data. It builds on the state-of-the-art research in applied and theoretical econometrics to highlight the importance of endogeneity and review the methods that can be used to address it with instrumental variables,The article also discusses when the Heckman two-step procedure can be used, as well as the tests, methods and assumptions that researchers should check when using instrumental variables. To ease implementation of the instrumental variables techniques, the author offers the STATA commands of the exposed tests and methods. Further, an empirical example is provided along with the utilized STATA codes. In the end, this article serves as a 'toolkit' allowing scholars not only to understand whether endogeneity is present in their empirical setting, but also to assess the empirical validity of their work when using instrumental variables