论文部分内容阅读
本文使用H P滤波、时间趋势平稳、ARMA趋势平稳和状态空间分解等趋势分解方法 ,对我国GDP增长率序列进行了趋势分解 ,并对各种周期成分进行了对比检验。我们发现 ,这些分解方法得到的周期成分具有类似的统计性质 ,但就残差序列的白噪声检验来说 ,双变量状态空间模型的分解效果最为显著 ,因此应该采用状态空间模型进一步分析我国的经济周期性质
In this paper, we use trend decomposition of H P filtering, time trend stabilization, ARMA trend stabilization and state space decomposition to analyze the trend of China’s GDP growth rate series, and compare the various periodic components. We find that the periodic components obtained by these decomposition methods have similar statistical properties, but the bivariate state-space model is the most effective for the white noise test of the residual sequence. Therefore, the state space model should be used to further analyze the economy of our country Periodic nature