加总偏误的Castle-Ⅳ校正与数值模拟  

Castle-Ⅳ Correction for Aggregation Bias and Numerical Simulation

作　　者：覃琼霞[1] 江涛 Qin Qiongxia;Jiang Tao(School of Economics and Management,Zhejiang Sci-Tech University;School of Economics and Management,China Jiliang University,Hangzhou 310018,China)

机构地区：[1]浙江理工大学经济管理学院,杭州310018 [2]中国计量大学经济与管理学院,杭州310018

出　　处：《统计与决策》2023年第6期33-38,共6页Statistics & Decision

基　　金：浙江省自然科学基金资助项目(LY20G010010);教育部哲学社会科学研究后期资助项目(20JHQ060);国家社会科学基金一般项目(18BJY040);国家自然科学基金面上项目(1874161)。

摘　　要：在加总过程内生性产生的两类单一内生节点和两类多内生节点情形基础上,文章利用Castle过程构建多重Ⅳ,并采用蒙特卡洛方法对多重Ⅳ的加总偏误校正效果进行比较分析。结果发现,单一有效Ⅳ、单一弱Ⅳ和多重有效Ⅳ均有显著的加总偏误校正效果,而单一伪Ⅳ和多重伪Ⅳ依然存在明显的估计偏误。Based on two types of single endogenous nodes and multiple endogenous nodes generated by the endogeneity of the aggregation process,this paper constructs multiple Ⅳs through Castle process,and comparatively analyzes the correction effect on the aggregation bias of multiple Ⅳs by using the Monte Carlo method.The results show that single effective Ⅳ,single weak Ⅳ and multiple effective Ⅳs all have significant aggregation bias correction effects,while single pseudo Ⅳ and multiple pseudo Ⅳs still have obvious estimation bias.

关 键 词：加总偏误 Castle-Ⅳ校正 数值模拟 

分 类 号：F222.1[经济管理—国民经济]