本文主要探讨在扰动项分布对称的假定下,平均处理效应的半参数估计。本文考虑了一种非常普遍形式的异方差,使得我们在估计平均处理效应时,大大扩展了对异方差的处理范围。本文给出了√N收敛速度的一致估计量及其渐进正态性质。本文遵循参数框架下常见的两步估计方法,这种方法广泛地运用于半参数的研究中。一个简单的Monte Carlo模拟将用来对比说明本文中估计方法的实际意义。
In this paper we consider semiparametric estimation of the average treatment effect under a joint symmetry assumption which allows for general forms of heteroscedstieity. Our approaches overcome various drawbacks associated with existing estimators. Our method provides a root-n consistent estimator in a heteroscedastic framework without parametric specification for the error distribution and strict form of heteroscedasticity. Our estimator follows the common two-step approach in a parametric setting, and several others in semiparametric settiongs, and are shown to be consistent and asymptotically normal. A Monte Carlo simulation study indicates the usefulness of our approaches.