基于Wild Bootstrap的ST-ECM模型的协整检验问题研究

Research on the Cointegration Test of the ST-ECM Model Based on the Wild Bootstrap Method

研究目标:解决异方差条件下平滑转移误差修正模型(ST-ECM)的非线性协整检验问题,完善ST-ECM模型的建模理论。研究方法:将Wild Bootstrap方法应用到ST-ECM模型的非线性协整检验问题中,构建异方差条件下ST-ECM模型的Wild Bootstrap非线性协整检验方法,推导异方差条件下ST-ECM模型非线性协整检验统计量的极限分布,并验证其有限样本性质。研究发现:本文提出的基于Wild Bootstrap的ST-ECM模型的非线性协整检验方法具有较好的功效和较低的检验水平扭曲。研究创新:将Wild Bootstrap方法应用到ST-ECM模型的非线性协整检验问题中,提出了统计性质优良的异方差条件下ST-ECM模型的非线性协整检验方法。研究价值:为实证研究者研究具有异方差特征的ST-ECM模型的非线性协整检验问题提供了有效的检验方法。

Objectives： Solving the nonlinear cointegration problem in Smooth Transition Error Correction Model （ST-ECM） on the condition of heteroscedasticity residual, in order to improve and perfect the modelling theory of the ST-ECM model. Research Methods： This paper proposes a Wild Bootstrap cointegration test procedure in ST-ECM model, derives the test statistics＇ asymptotic distribution and investigates their finite sample property. Research Findings： The Monte Carlo simulation results show that our proposed statistics have reason- able powers and less size distortion compared to other tests. Research Innovations： The paper proposes a good-properties statistic to solve the nonlinear cointegration problem in ST-ECM model on the condition of heteroscedasticity residual. Research Value： This paper provides an effective method for empirical research study when the STECM model has heteroscedasticity residual.