Endogeneity and Structural Equation Estimation in Political Science

Abstract This article outlines the current situation with respect to ‘endogeneity and structural equation estimation’. It begins by exploring why econometrics, and particularly the treatment of endogeneity, proved so valuable for political scientists. The most prominent methods used to deal with endogeneity are explained. The article also includes a critique of the critical conditions needed to support the use and interpretation of this method and discusses frequently used diagnostics developed to assess the severity of the problems. The estimation method and diagnostics are shown with a running example with data from a previous study of US congressional politics. In many instances, empirical evidence will continue to be the classic observational study, replete with endogeneity problems. Rather than forgoing analyses of these data and questions because of the reservations about a particular method, it is better to use and to improve those methods in the context of these studies and data.