Instrumental Variables: Intuition and Estimation

This lecture introduces instrumental variables as a technique to address endogeneity issues in regression analysis. It covers the intuition behind instrumental variables, the conditions required for their validity, and the estimation process. The lecture discusses the instrumental variables estimator in the simplest case, the method-of-moments estimation, and the two-stage least squares (2SLS) estimator. It also explores the challenges of weak instruments and the importance of testing the exclusion restriction and relevance condition. An example from Card (1995) on college proximity and schooling is used to illustrate the application of instrumental variables. The lecture concludes with discussions on testing the exogeneity of variables and extensions of the instrumental variables framework.