基于SaRa的面板数据回归模型结构变点估计  

Estimation of Structural Change in Panel Regression Model Based on SaRa

作　　者：徐小平[1] 杨倩男 李拂晓 XU Xiao-ping;YANG Qian-nan;LI Fu-xiao(Department of Applied Mathematics,Xi'an University of Technology,Xi'an 710054,China)

机构地区：[1]西安理工大学理学院,陕西西安710054

出　　处：《数学的实践与认识》2023年第2期183-194,共12页Mathematics in Practice and Theory

基　　金：国家自然科学基金(11801438);陕西省创新能力支撑计划项目(2020PT-023)。

摘　　要：面板数据的变点分析是计量经济学的热门研究课题之一,在金融、医学、质量控制、气象等领域也有着广泛的应用.基于一种快速局部算法SaRa(Screening and Ranking algorithm)研究了面板数据回归模型的结构变点估计问题.首先基于回归系数的估计量建立局部统计量,筛选出可能的变点.其次构造自适应阈值来筛选出最终的变点,并且证明了变点估计量的一致性.Monte Carlo模拟结果显示,当解释变量为外生变量或内生变量,误差项存在序列相关或异方差,提出的方法都能较准确地估计出变点的个数及位置.最后利用该方法分析世界24个低收入和高收入国家自然人口增长率和国际移民存量对人口增长率的影响,说明了方法的有效性。Change point analysis of panel data is one of the hot research topics in econometrics,and it is also widely used in finance,medicine,quality control,meteorology and other fields.In this paper,the estimation of structural change in panel regression model based on SaRa(Screening and Ranking algorithm)is studied.Firstly,local statistics are established based on the estimator of the regression coeficient,and the possible change points are screened out.Secondly,an adaptive threshold is constructed to filter out the final change point,and the consistency of the change point estimators is proved.Monte Carlo simulation results show that when the explanatory variables are exogenous or endogenous variables,there is serial correlation or heteroscedasticity in the error terms,and the method proposed in this paper can accurately estimate the number and location of change points.Finally,the method is used to analyze the impact of natural population growth rate and international migration stock on population growth rate in 24 low-income and high-income countries in the world,which shows the validity of the method.

关 键 词：面板数据回归模型 变点估计 SARA 局部统计量 

分 类 号：O212.1[理学—概率论与数理统计]