CAJ | 학술논문

在推导ADF检验模式下趋势项和漂移项伪 t检验量极限分布基础上，提出修正的系数检验量。研究表明，它们与DF检验模式下检验量具有相同的极限分布；构造漂移项和趋势项检验的Bootstrap实现方法并证明了有效性。将蒙特卡洛模拟技术与临界值检验方法进行对比，结果表明Bootstrap方法能够明显降低检验的水平扭曲，在检验功效方面也有一定优势。模拟也显示临界值检验的局限性和Bootstrap方法的稳健性。
재추도ADF검험모식하추세항화표이항위 t검험량겁한분포기출상，제출수정적계수검험량。연구표명，타문여DF검험모식하검험량구유상동적겁한분포；구조표이항화추세항검험적Bootstrap실현방법병증명료유효성。장몽특잡락모의기술여림계치검험방법진행대비，결과표명Bootstrap방법능구명현강저검험적수평뉴곡，재검험공효방면야유일정우세。모의야현시림계치검험적국한성화Bootstrap방법적은건성。
The paper first derives the limit distribution of pseudo t test for drift and trend based on ADF test mode ,and brings forward the modified coefficient test .Results show these tests share the same limit distribution with the DF test mode .Then the Bootstrap test and its validity for drift and trend is given and testified . Results from Monte Carlo Simulation indicate that the Bootstrap method can significantly reduce the size distortion and also has certain advantages in test power over critic value method .Simulation also show s the limitation of the critic value method and the robustness of the Bootstrap method .