Lecture 21 Multiple linear IV model and 2-stage least squares

Lecture 21 Multiple linear IV model and 2-stage least squares

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon
Economics 326 Methods of Empirical Research in Economics Lecture 21: Multiple linear IV model and two-stage least squares (2SLS) Vadim Marmer University of British Columbia May 5, 2010
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Multiple linear IV model I In empirical research, we often have to estimate models that include multiple endogenous and exogenous regressors. I Example: ln Wage i = γ 0 + γ 1 Age i + γ 2 Sex i + β 1 Educ i + β 2 Children i + U i . I Exogenous regressors: age, sex, and a constant. I Endogenous regressors: education and children (family size). 1/11
Background image of page 2
Multiple linear IV model I Consider the following model: y i = γ 0 + γ 1 X 1 , i + . . . + γ k X k , i + β 1 Y 1 , i + . . . + β m Y m , i + U i , where I y i is the dependent variable. I γ 0 is the coe¢ cient on the constant regressor: EU i = 0 . I X 1 , i , . . . , X k , i are the k exogenous regressors: Cov ( X 1 , i , U i ) = . . . = Cov ( X k , i , U i ) = 0 . I Y 1 , i , . . . , Y m , i are the m endogenous regressors: Cov ( Y 1 , i , U i ) 6 = 0 , . . . , Cov ( Y k , i , U i ) 6 = 0 . 2/11
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
I There are k + 1 + m unknown coe¢ cients y i = γ 0 + γ 1 X 1 , i + . . . + γ k X k , i + β 1 Y 1 , i + . . . + β m Y m , i + U i . I The exogeneity conditions EU i = 0 and Cov ( X 1 , i , U i ) = . . . = Cov ( X k , i , U i ) = 0 give us only k + 1 equations: E [ y i γ 0 γ 1 X 1 , i . . . γ k X k , i β 1 Y 1 , i . . . β m Y m , i ] = 0 , E [ X 1 , i ( y i γ 0 γ 1 X 1 , i . . . γ k X k , i β 1 Y 1 , i . . . β m Y m , i )] = 0 , . . . E [ X k , i ( y i γ 0 γ 1 X 1 , i . . . γ k X k , i β 1 Y 1 , i . . . β m Y m , i )] = 0 . I There are more unknowns than equations. Thus, the knowledge of the true covariances between X ±s, Y ±s and y is not su¢ cient to recover the unknown coe¢ cients γ 0 , γ 1 , . . . , γ k , β 1 , . . . , β m . I
Background image of page 4
Image of page 5
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 03/24/2012 for the course ECON 326 taught by Professor Whisler during the Spring '10 term at The University of British Columbia.

Page1 / 12

Lecture 21 Multiple linear IV model and 2-stage least squares

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online