基于非参数回归的金融传染检验  

Nonparametric regression based testing for financial contagion

作　　者：王霞 付中昊 洪永淼[3,4] 张冬悦 WANG Xia;FU Zhonghao;HONG Yongmiao;ZHANG Dongyue(Lingnan(University)College,Sun Yat-sen University,Guangzhou 510275,China;School of Economics,Fudan University,Shanghai 200433,China;Department of Economics and Department of Statistics and Data Science,Cornell University,Ithaca 14850,USA;The Wang Yanan Institute for Studies in Economics,Xiamen University,Xiamen 361005,China;School of Economics,Dongbei University of Finance and Economics,Dalian 116025,China)

机构地区：[1]中山大学岭南(大学)学院,广州510275 [2]复旦大学经济学院,上海200433 [3]康奈尔大学经济学系与统计学系,伊萨卡14850 [4]厦门大学王亚南经济研究院,厦门361005 [5]东北财经大学经济学院,大连116025

出　　处：《系统工程理论与实践》2020年第6期1398-1418,共21页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(71873151,71903032,71988101);中山大学青年教师重点培育项目(20191951)。

摘　　要：本文借助特征函数的优良性质,基于非参数回归构造了金融传染的检验统计量.与现有文献相比,该统计量不仅避免了模型设定偏误问题,而且能够同时捕获线性和各种形式的非线性传染效应.在原假设成立时,该统计量渐近服从于标准正态分布.数值模拟结果表明,该统计量具有良好的有限样本性质,能够识别多种形式的非线性金融传染.本文进一步应用该统计量探讨了中国金融市场与东亚、拉丁美洲、新兴市场国家之间的传染效应,捕获了传统基于线性测度方法无法刻画的非线性传染效应,说明我国与这些金融市场之间存在显著的非线性传染效应.Based on the conditional characteristic functions via nonparametric estimation,this paper proposes a new test for financial contagion.Compared with the existing tests,the proposed test can not only avoid the model misspecification problem but also capture both linear and various forms of nonlinear financial contagion.The proposed test has a convenient asymptotic standard normal distribution under the null hypothesis.Monte Carlo studies demonstrate the good performance of our test in finite samples.In an empirical application to testing financial contagion,we find significant nonlinear financial contagion between China's stock market and other stock markets,including East Asia,Latin America,emerging markets,which could not be identified by tests based on linear measures.

关 键 词：金融传染 特征函数 非参数回归 非线性 检验功效 

分 类 号：F224.0[经济管理—国民经济]