统计研究
統計研究
통계연구
Statistical Research
2014年
3期
99~105
,共null页
平滑转移自回归模型 非线性平稳性检验 伪检验 蒙特卡洛仿真
平滑轉移自迴歸模型 非線性平穩性檢驗 偽檢驗 矇特卡洛倣真
평활전이자회귀모형 비선성평은성검험 위검험 몽특잡락방진
Smooth Transition Autoregressive Model; Nonlinear Stationary Test; Spurious Test; Monte Carlo Simulations
本文提出一种通用非线性单位根检验方法,使用待检序列及其逆序列的Wald统计量的最小值作为检验统计量,将Kapetanios等人提出的受限条件下ESTAR模型非线性单位根检验推广到非0位置参数的情形,也可应用于LSTAR或其他可能的STAR模型、TAR模型或传统的线性AR模型的平稳性检验.并且,推导了检验统计量的极限分布,仿真了检验水平与功效.结果表明,本文提出的检验方法对数据生成过程有广泛的适应性,并且在大多数情况下都能获得较其他方法更佳的检验功效.
本文提齣一種通用非線性單位根檢驗方法,使用待檢序列及其逆序列的Wald統計量的最小值作為檢驗統計量,將Kapetanios等人提齣的受限條件下ESTAR模型非線性單位根檢驗推廣到非0位置參數的情形,也可應用于LSTAR或其他可能的STAR模型、TAR模型或傳統的線性AR模型的平穩性檢驗.併且,推導瞭檢驗統計量的極限分佈,倣真瞭檢驗水平與功效.結果錶明,本文提齣的檢驗方法對數據生成過程有廣汎的適應性,併且在大多數情況下都能穫得較其他方法更佳的檢驗功效.
본문제출일충통용비선성단위근검험방법,사용대검서렬급기역서렬적Wald통계량적최소치작위검험통계량,장Kapetanios등인제출적수한조건하ESTAR모형비선성단위근검험추엄도비0위치삼수적정형,야가응용우LSTAR혹기타가능적STAR모형、TAR모형혹전통적선성AR모형적평은성검험.병차,추도료검험통계량적겁한분포,방진료검험수평여공효.결과표명,본문제출적검험방법대수거생성과정유엄범적괄응성,병차재대다수정황하도능획득교기타방법경가적검험공효.
This paper proposes a general nonlinear unit root test method, using the minimum of Wald statistics of a series and its adverse series as test statistic. The method generalizes the nonlinear stationary test method against constrained ESTAR model proposed by Kapetanios et al. to nonzero position situation, at the same time, it can be applied to test first-, second-order LSTAR model, or other possible smooth transition autoregressive model, or threshold autoregressive model and traditional linear AR model. We derive the limiting nonstandard distribution of the proposed test and simulate the test powers. The results suggest that the new test has wide adaptability to data generating process and is generally superior in terms of power.
