PanelDataNotes-19

PanelDataNotes-19 - Econometric Analysis of Panel Data...

This preview shows pages 1–9. Sign up to view the full content.

Econometric Analysis of Panel Data William Greene Department of Economics Stern School of Business

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

View Full Document
Econometric Analysis of Panel Data 19. Limited Dependent Variables And Models for Count Data
Censoring and Corner Solution Models Censoring model: T(y*) = 0 if y* <  0 and y*  otherwise.  Corner solution: y = 0 if some exogenous  condition is met; y = g(x)+e if the condition is  not met.  We then model P(y=0) and E[y|x,y>0].   (See text, pp. 518-519)     Application: Fair, R., “A Theory of Extramarital     Affairs,” JPE, 1978. Same structural form

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

View Full Document
The Tobit Model Y* = x’ β + ε , y = Max(0,y*),  ε  ~ N[0, σ 2 ] Other censoring limits:         y=Max(L,y*) => y-L=Max(0,y*-L).  Change constant term and LHS  variable Upper censoring:        y=Min(U,y*) => U-y = Max(0,U-Y*). Change constant and LHS variable  and swap signs of estimated coefficients. Censoring limit is person specific: Same as above: Constant term  becomes the variable with a coefficient fixed at 1. Censoring at both extremes: y=Max(L,Min(y*,U)). A minor extension of  the model.  Easy to accommodate. (Already done in major software.) Interval censoring over the entire range.  See yesterday’s notes. (Tobin:  “Estimation of Relationships for Limited Dependent  Variables,” Econometrica, 1958. Tobin’s probit?)
Conditional Mean Functions 2 y* ,   ~N[0, ] y Max(0, y* ) E[y* | ] x E[y| ]=Prob[y=0| ]×0+Prob[y>0| ]E[y|y>0, ]          =Prob[y* >0| ]  E[y* |y* >0, ] ( )          =           ( )          =   ( ) = + ε ε σ = = β φ σ Φ × σ ÷ σ Φ σ Φ σ x β x x x x x x x x x / β β x + β x / β x / x   β β ( ) ( ) " Inverse Mills ratio" ( ) ( ) E[y| ,y>0]= ( ) σφ σ φ σ = Φ σ φ σ σ Φ σ x / β x / β x / β x / β x x + β x / β

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

View Full Document
Conditional Means XB -1.20 -.40 .40 1.20 2.00 -2.00 -1.20 -.40 .40 1.20 2.00 -2.00 EYSTAR EY Variable
Predictions and Residuals What variable do we want to predict? y*?  Probably not – not relevant y?  Randomly drawn observation from the population y | y>0?  Maybe. Depends on the desired function What is the residual? y – prediction?  Probably not.  What do you do with  the zeros? Anything - x β ? Probably not. x β  is not the mean. Generalized residuals – coming below.

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

View Full Document
OLS is Inconsistent - Attenuation E[y | ] =   ( ) ( ) Nonlinear function of x.  What is estimated by OLS regression of y on x? Slopes of the linear projection are approximately
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 01/05/2012 for the course B 55.9912 taught by Professor Willamgreene during the Fall '11 term at NYU.

Page1 / 43

PanelDataNotes-19 - Econometric Analysis of Panel Data...

This preview shows document pages 1 - 9. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online