中国金融安全混频分层测度及其预警

Mixed-frequency Hierarchical Measurement and Its Early Warning of China’s Financial Security

鉴于同频金融安全测度缺乏实时性和分层动态性,文章选择由年、季、月三种频率构成的22个混频样本数据,使用新构建的混频分层动态因子模型(MF-HDFM)进行估计,测度中国混频金融安全指数体系(MFFSI),并基于MS-AR模型对其进行马尔科夫金融安全状态识别及预警。结果表明:中国MFFSI对中国金融安全状况进行了合理有效的测度;中国金融安全周期波动兼具总层趋同特征和子层部门分异特征;中国金融安全水平变化的原因呈现多样化特征;MS-AR模型较好地识别了中国金融安全状态,并能较好地进行预警。

Given that the measurement on financial security with the same frequency lacks the properties of real time and hierarchical dynamism,this paper selects 22 mixed-frequency sample data composed of annual,quarterly and monthly frequencies and uses the newly constructed mixed-frequency hierarchical dynamic factor model(MF-HDFM)to estimate and measure China’s mixed-frequency financial security index system(MFFSI),and then carries out Markov financial security state recognition and early warning based on MS-AR model.The results show that the measurement of China’s financial security by MFFSI is reasonable and effective,that the cyclical fluctuation of China’s financial security has both the characteristics of overall layer convergence and those of the sub-layers differentiation,that the reasons for the change of China’s financial security level are diversified,and that the MS-AR model can better identify China’s financial security status and give a better early warning.