处理效应模型的理论拓展及在政策评价中的应用

现有文献在利用处理效应模型评估政策时,模型中的假设条件局限性大多较强,在实际应用中很难验证,且一旦这些假设错误,就会引起参数估计的不一致。本文首先在非参数框架下提出了一种关于处理效应模型的半参数估计方法,其既不对模型中的函数形式做任何假定,也允许误差项的联合分布是广义异方差形式,从而大大减少因模型误设而引起的估计偏误。考虑到处理效应的内生性问题,提出了一个两步估计量。第一步关于选择方程进行非参数估计;第二步在结果方程中,利用工具变量法估计平均处理效应。其次,对估计量的大样本性质进行分析,表明了估计量的一致性和渐近正态性质。再次,通过蒙特卡罗模拟与已有估计方法进行比较,结果表明本文的方法具有较强的稳健性。最后,本文将该方法应用于研究高新技术企业认证政策对企业盈利能力影响,研究发现该政策提升了高新技术企业的盈利能力,并且相比于国有企业,该政策对民营企业促进效应更大。

Theoretical Extension of Treatment Effect Model and Its Application in Policy Evaluation

The assumptions are quite restrictive in the treatment effect model used by current literature on policy evaluation but difficult to verify in practice. Once the assumptions are mistaken, parameter estimation will be inconsistent. Firstly, this paper proposes a semi-parametric estimation method for treatment effect model under a non-parametric framework. In our model, we don’t assume functional forms, and also allow the joint distribution of the error terms to be a general form of heteroscedasticity, which greatly reduces the risk of model misspecification. This paper considers treatment effect under endogeneity, and proposes a two-step approach. The first step uses a non-parametric estimator in the selection equation, and the second step uses an instrumental variables approach to estimate average treatment effect in the outcome equation. Secondly, the proposed estimator is shown to be consistent and asymptotically normally distributed. Furthermore, compared with the existing methods, the Monte Carlo simulation result shows our method is more robust. Finally, we apply our method to empirically estimate the impact of certification policy for high-tech enterprises on their profitability, and find this policy can help promote the profitability, and has a greater impact on private enterprises than state-owned enterprises.