摘要
本文考察主方程因变量带有删失结构的第三类Tobit模型的半参数估计问题,基于受限因变量的条件分布提出一种新的半参数两步估计方法,并证明所提出估计量的相合性和渐近正态性.数值模拟的结果显示本文的估计量在一系列设计下均有着良好的表现.通过与文献中的估计量进行比较,结果表明,当主方程因变量的删失比例较大或当样本量较大时,本文的估计量是对现有估计量的一个重要补充.
This paper presents a two-step semiparametric estimation approach to the censored Type-3 Tobit model based on the entire conditional distribution of the limited dependent variables. The estimator is shown to be consistent and asymptotically normal. Simulation designs show that our estimator performs well in a variety of designs. In comparing with existing estimators, we conclude that our estimator is an important complement to the literature when the censoring rate of the main equation is moderately large or/and when the sample size is large.
出处
《中国科学:数学》
CSCD
北大核心
2015年第4期393-409,共17页
Scientia Sinica:Mathematica
基金
国家自然科学基金(批准号:71371199)资助项目