Search Results
Title: Maximum Likelihood Estimation With Sample Selection: An Application to the Labor Supply of Women
Resulting in 1 citation.
| 1. |
Wang, Boqing |
|
Maximum Likelihood Estimation With Sample Selection: An Application to the Labor Supply of Women Ph.D. Dissertation, Washington State University, 1994 Cohort(s): NLSY79 Publisher: UMI - University Microfilms, Bell and Howell Information and Learning Keyword(s): Labor Economics; Labor Supply; Modeling; Sample Selection; Women; Work Hours/Schedule Sample selection bias has been a focus in econometrics since the 1950s. However, previous methods did not provide efficient and robust estimates for a three-equation system with sample selection. To find efficient and robust estimates, a full information maximum likelihood (FIML) estimate model based on a trivariate logistic distribution is developed in this study. This FIML model is applied to the three-equation model of female labor supply. The data used in this study are from the 1986 National Longitudinal Survey of Youth. The empirical results in this study show that the three-equation FIML model of female labor supply provides an efficient and robust estimate for a woman's hours of work while the Heckman two-stage method does not. The three-equation FIML model was extended to a four-equation FIML model with double selection biases. The FIML models developed in this study have wide application in econometric analysis. |
|
|
Bibliography Citation
Wang, Boqing. Maximum Likelihood Estimation With Sample Selection: An Application to the Labor Supply of Women. Ph.D. Dissertation, Washington State University, 1994. |