我国GDP周期波动的非对称性研究方法与统计检验

Research into a Statistic Method of Asymmetry in China's GDP Fluctuations and Test

论述了阈值自回归模型(TAR)和冲量阈值自回归(M—TAR)理论方法及其检验统计量的构造,并应用Bootstrap自助法来获得检验统计量的渐近P—值或渐近临界值。运用该方法分析了我国GDP增长率周期波动的"深"和"尖"两方面的非对称性特征,研究发现我国GDP增长率周期波动呈现较强的"深"型波动,而"尖"的特征并不明显。同时通过对价格指数的非对称性研究,发现价格指数周期波动并不存在非对称性,因此以价格名义变量来解释我国GDP增长率周期波动的非对称性是不成立的。

This paper analyses the theory and steucture of Threshold Autoregressive model and Momentum Threshold Autoregressive model,and proposes a method to test this asymmetry in china's GDP fluctuations.We obtain asymptotic P-value or asymptotic critical value of test statistic through using Hansen's Bootstrap method.Then,we use this method to research into the deepness and sharpness features of China's GDP fluctuations,and find a fairly strong evidence of deepness in China's real GDP fluctuations,but we don't find sharpness feature in real GDP.This paper analyses the asymmetry of price series,and finds no asymmetry in price series,and therefore the asymmetry of price can't support the asymmetry of China's GDP fluctuations.