论文部分内容阅读
结构变化检验受回归残差是否平稳影响,而一个时序列的平稳性检验结果又对结构变化发生与否非常敏感,因而结构变化检验与平稳性检验构成一种循环依赖关系。针对此问题,Perron和Yabu(2009)提出了检验结构变化的拟可行广义最小二乘方法,该方法优越性在于不依赖回归残差是否平稳的特征。本文采用此检验对我国8个产出变量的增长趋势是否发生结构变化进行了实证研究。结果显示,其中7个指标发生了结构变化,在此基础上本文对导致变量增长趋势发生偏移的冲击进行了详细的分析。
The structural change test is affected by the regression residuals. However, the stationary test results of a time series are very sensitive to the structural changes. Therefore, the structural change test and the stationarity test form a cyclic dependence. In response to this problem, Perron and Yabu (2009) proposed a quasi-feasible generalized least squares method to test structural changes. The superiority of this method is that it does not depend on whether the residuals of the regression are stable or not. This paper uses this test to make an empirical study on the structural change of the growth trend of eight output variables in our country. The results show that seven of the indicators have undergone structural changes. Based on this, the paper analyzes in detail the impact of offsetting the growth trend of variables.