统计研究 ›› 2018, Vol. 35 ›› Issue (10): 103-115.doi: 10.19343/j.cnki.11-1302/c.2018.10.009

• • 上一篇    下一篇

空间滞后分位数回归模型的工具变量估计及参数检验

李坤明 方丽婷   

  • 出版日期:2018-10-25 发布日期:2018-10-22

Instrumental Variables Estimation and Parametric Tests of Spatial Lag Quantile Regression Model

Li Kunming & Fang Liting   

  • Online:2018-10-25 Published:2018-10-22

摘要: 本文提出一种遵循空间数据分布特征的空间分位数回归模型,并着重探讨该模型的估计方法和参数检验问题。本文构建了上述模型的一个工具变量估计法,通过数理证明建立了估计量的大样本理论,并基于估计量的渐近分布构造了模型的参数检验方法。本文还通过数值模拟方法和应用实例考察估计方法和参数检验方法的实际应用效果,数值模拟结果显示,估计方法和参数检验方法在有限样本条件下均可以达到较高的精确度和稳定性。在应用实例中,本文利用所构建的理论方法重新检验我国“资源诅咒”效应的存在性,实证结果体现了理论方法的应用价值。

关键词: 空间计量, 分位数回归, 工具变量估计, 资源诅咒

Abstract: In this paper, we propose a spatial quantile regression model following spatial data distribution characteristics, and focus on the model's estimation and parametric test. We construct an instrumental variables estimation method and build the large sample theory of estimators through mathematical proofs. Based on the asymptotic distribution of estimators, a parametric test for the model is also constructed. The numerical simulation method and practical examples are used to investigate the practical application effects of the estimation method and parametric test method. Numerical simulation results show that both the estimation method and the parametric test method can achieve high accuracy and stability under finite sample conditions. In the application example, we use the theoretical method constructed to re-examine the existence of "resource curse" in China. The empirical results show the application value of the theoretical method.

Key words: Spatial Econometrics, Quantile Regression, Instrumental Variables Estimation, Resource Curse