Card (1995) used wage and education data for a sample of men in 1976 to est imate the ret urn to education. He used a dummy variable for whet her someone grew up near a four-year college (nearct) as an instrumental variable for education. The data in CARD.RAW' contains a sample of 3,010 men. We are going to use the following variables: lwage (log wages), erp (experience), black (an indicator for race), south (an indi- cator for residency in the Sout h), south66 (an indicator for residency in the South in 1966), smsa (an indicator for residency in a metropolitan area), smsa66 (an indicator for residency in a metropolitan area in 1966) and reg661-reg669 (regional dummies). Consider the following regression equation for the wage equation: In(wage), -Beduc, + 8" X, + im1,.N (1) where In(wage), is log wages, educ is schooling and x, - south66,, smsa,, smsa66,, regional dummies)". Notice that a constant is not included because it would be collinear with the regional dummies. (exp, exp, black,, Nouth,, 1. Estimate the regression equation (1) using OLS and report the coefficients 3 and 6 (no need to report the regional dummies), st andard errors and R. Do you think that 3 may be biased in this model? Explain. 2 Let's assume that educ is endogenous and nearet is exogenous to e. Is nearet a good instrument for educ? Show with a regression that neared is a good predictor for educ. Explain. 3. Estimate (1) using IV. Report the coefficients (no need to report the regional dummies) and standard errors. Do your results change with respect to question (1.)? Explain.