This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Strategy n (1) Purge ED of the influence of everything but MS, FEM, BLK (and the other variables). Predict ED using all exogenous information in the sample ( X and Z ). n (2) Regress LWAGE on this prediction of ED and everything else. n Standard errors must be adjusted for the predicted ED ˜™™™™™™ ™ 8/61 Part 13: Endogeneity OLS ˜™™™™™™ ™ 9/61 Part 13: Endogeneity The weird results for the coefficient on ED happened because the instruments, MS,FEM,BLK are all dummy variables. There is not enough variation in these variables. ˜˜™™™™™ ™ 10/61 Part 13: Endogeneity Source of Endogeneity n LWAGE = f( ED, EXP,EXPSQ,WKS,OCC, SOUTH,SMSA,UNION ) + n ED = f( MS,FEM,BLK, EXP,EXPSQ,WKS,OCC, SOUTH,SMSA,UNION ) + u ˜˜™™™™™ ™ 11/61 Part 13: Endogeneity Remove the Endogeneity n LWAGE = f( ED, EXP,EXPSQ,WKS,OCC, SOUTH,SMSA,UNION ) + u + n Strategy p Estimate u p Add u to the equation. ED is uncorrelated with when u is in the equation. ˜˜™™™™™ ™ 12/61 Part 13: Endogeneity Auxiliary Regression for ED to Obtain Residuals ˜˜™™™™™ ™ 13/61 Part 13: Endogeneity OLS with Residual (Control Function) Added 2SLS ˜˜™™™™™ ™ 14/61 Part 13: Endogeneity A Warning About Control Function ˜˜™™™™™ ™ 15/61 Part 13: Endogeneity The Problem ˜˜™™™™™ ™ 16/61 1 2 Cov( , ) , K variables Cov( , ) , K variables is OLS regression of y on ( , ) cannot estimate ( , ) consistently. Some other estimator is needed. Additional structure: = + wh = + + = ≠ endogenous y X Y X Y Y X Y Y Z V β δ ε ε ε β δ Π ere Cov( , )= . An estimator based on ( , ) may be able to estimate ( , ) consistently. instrumental variable ( Z X IV) Z ε β δ Part 13: Endogeneity Instrumental Variables p Framework: y = X + , K variables in X . p There exists a set of K variables, Z such that plim( Z’X/n ) but plim( Z’ /n) = The variables in Z are called instrumental variables. p An alternative (to least squares) estimator of is b IV = ( Z’X )1 Z’y p We consider the following: n Why use this estimator? n What are its properties compared to least squares? p We will also examine an important application ˜˜™™™™™ ™ 17/61 Part 13: Endogeneity IV Estimators Consistent b IV = ( Z’X )1 Z’y = ( Z’X /n)1 ( Z’X /n) β + ( Z’X /n)1 Z’ε /n = β + ( Z’X /n)1 Z’ε /n β Asymptotically normal (same approach to proof as for OLS) Inefficient – to be shown. ˜˜™™™™™ ™ 18/61 Part 13: Endogeneity The General Result By construction, the IV estimator is consistent. So, we have an estimator that is consistent when least squares is not....
View
Full Document
 Fall '10
 H.Bierens
 Econometrics, Regression Analysis, Endogeneity

Click to edit the document details