192 IV15a IV estimation in R

This video demonstrates practical applications of instrumental variables (IV) in econometrics using R. We begin by loading and exploring the Card dataset, focusing on variables like wages, education, and distance as a potential instrument. The explanation walks through running Ordinary Least Squares (OLS) regression to establish a baseline and highlights potential upward bias due to unobserved ability. The core of the video involves a detailed walkthrough of the two-stage least squares (2SLS) estimation process. This includes performing the first stage regression, checking for instrument relevance and strength using the F-statistic, and then manually constructing the Wald estimator by calculating the ratio of reduced form and first stage coefficients. We also showcase how to use the `ivreg` function in R for a more direct 2SLS estimation, emphasizing that this is the preferred method in practice due to its handling of standard errors. A comparison between OLS and IV results is presented, noting the difference in coefficients and the significantly larger standard errors typically found with IV estimation, which accounts for potential endogeneity. Visit AxiomTutoring.com and subscribe to @AxiomTutoringCourses.