基于MS-VAR模型的经济周期转折点的识别与研判

Identification and Judgment of Business-cycle Turning Points based on MS-VAR Model

应用MS-VAR模型从多个一致指标中提取其共同周期,对经济周期的波动进行研究。并与由一致合成指数作为解释变量的MS-AR模型进行比较分析,选取2000年至2011年的我国宏观经济月度数据作为训练集,对历史转折点进行识别,然后再用2012年到2016年数据作检验集以分析模型预测性能。结果表明:包含工业增加值、固定资产投资完成额、进出口总额和发电量的同比增长率指标的MS-VAR模型所提取的共同周期能够较好地代表我国的经济周期;MS-VAR模型更具稳健性,尤其表现在对收缩期或扩张期内暂时性的反向波动的处理上更为有效。

This paper employs the Markov Switching Vector Autoregressive model to extract the common cycle from multiple consistent indicators,and then uses it to study the fluctuation of economic cycle. The Markov Switching Autoregressive model with the consistent composite index as an explanatory variable is for comparison. This paper selects the monthly data of macro economy from 2000 to 2011 as a train set to identify the historical turning points. As a validation set,data from 2012 to 2016 is used to analyze the forecasting performance. The results show that containing year-on-year growth rate indicators of industrial added value,investment in fixed assets,total export-import volume and generating capacity,the MS-VAR model extracts the common cycle that can better represent the fluctuation of China＇s business cycle. And the MS-VAR model is more robust in identification and forecasting,especially in the handling of the temporary reverse fluctuation during the contraction or expansion period.