CAJ | 학술논문

本文从理论模拟和实证分析两个角度研究了以平稳过程为原假设的KPSS检验在存在结构突变的情形下该检验的有限样本性质,包括实际检验水平和检验功效的分析。结果表明当忽略数据生成过程中的结构突变时,KPSS检验的实际检验水平会发生严重扭曲,从而出现“Perron现象”。本研究针对的是有限样本情形,这不但补充和完善了相关理论,同时对实证分析工作具有很强的借鉴和指导意义。
본문종이론모의화실증분석량개각도연구료이평은과정위원가설적KPSS검험재존재결구돌변적정형하해검험적유한양본성질,포괄실제검험수평화검험공효적분석。결과표명당홀략수거생성과정중적결구돌변시,KPSS검험적실제검험수평회발생엄중뉴곡,종이출현“Perron현상”。본연구침대적시유한양본정형,저불단보충화완선료상관이론,동시대실증분석공작구유흔강적차감화지도의의。
This paper analyzes the characters of KPSS unit root tests in finite sample where exists structural breaks, including size and power. By Monte Carlo simulations, we draw a conclusion that real test size will be distorted and result in Perron Phenomenon if structural break in Data Generation Process （GDP） doesn＇t be considered. Because the purpose of this study is the limited sample of cases, so the conclusions not only supplement and perfect the relevant theory, but also have a strong reference and guidance to empirical analysis.