Heckman-Tobit模型的半参数估计

Heckman-Tobit模型可以同时处理样本选择问题和删失数据问题,是一个重要的微观计量模型.本文根据结果变量的条件生存函数所满足的性质,提出Heckman-Tobit模型的一种半参数估计方法.这种方法通过积分的形式,有效地利用了结果变量整个条件分布的信息.在一些正则性条件下,本文证明了所提出的半参数估计量的相合性和渐近正态性.其渐近性质的成立不依赖于扰动项的具体分布.数值模拟实验的结果表明,本文的半参数估计量具有优越的有限样本性质,且当扰动项服从非正态分布时优于最大似然估计量.

Semiparametric estimation of a Heckman-Tobit model

The Heckman-Tobit model is an important micro-econometric model because it can address both sample selection and censoring. This paper proposes a semiparametric estimation method for the Heckman-Tobit model based on the properties of conditional survival functions of the outcome variable. The virtue of the proposed method is that it effectively exploits information contained in the entire conditional distribution of the outcome variable via integrals. Under several regular conditions, the consistency and asymptotic normality of the proposed semiparametric estimator are established, which do not rely on the specific distribution of the error term. Monte Carlo simulation results show desirable finite sample performances of the semiparametric estimator. In particular, when the error term follows non-normal distributions, the semiparametric estimator is superior to the maximum likelihood estimator.