Chapter 27
LIMITED DEPENDENT VARIABLES
PHOEBUS J. DHRYMES
Columbia University
Contents
0. Introduction
1. Logit and probit
1 .l Generalities
1.2. Why a general linear model (GLM) formulation is inappropriate
1.3. A utility maximization motivation
1.4. Maximum likelihood estimation
1.5. Goodness of fit
2. Truncated dependent variables
2.1. Generalities
2.2. Why simple OLS procedures fail
2.3. Estimation of parameters by ML methods
2.4. An initial consistent estimator
2.5. Limiting properties and distribution of the ML estimator
2.6. Goodness of tit
3. Sample selectivity
3.1. Generalities
3.2. Inconsistency of least squares procedures
3.3. The LF and ML estimation
3.4. An initial consistent estimator
3.5. Limiting distribution of the ML estimator
3.6. A test for selectivity bias
References
|