中国省份产业经济周期的经验研究——基于贝叶斯潜在多动态因子模型的分析

Provincial Industrial Business Cycles in China

本文利用基于吉本斯抽样估算的贝叶斯潜在多动态因子模型,估算出中国改革开放以来三次产业实际产出波动的全国、地区和省份动态因子,发现全国动态因子和区域动态因子较好得刻画了中国宏观经济波动的基本特征和基本区域经济事实。方差分解的结果表明,全国动态因子是省份第二产业实际产出波动最重要的驱动力,异质性成分是省份第一、第三产业实际产出波动的主要驱动力,区域动态因子影响力弱,省份动态因子对第一、第三产业有一定的影响。不同类型动态因子对各产业影响程度的省份差异在一定程度上和区域以及省份特征相关。

Based on the three industries data in provincial level, this paper employ a Bayesian dynamic latent factor model to estimate the country, region and province dynamic factors, which coincide with the macroeconomic features and im- portant regional facts in China. Based on the results of variance decomposition, a country factor is an important source of volatility for the secondary industry, and the idiosyncratic component is the most important for the volatility for the first and third industries. We find that regional factors only have little influence on economic fluctuations, and provincial factors play a little role in explaining fluctuations in the first and third industries. We also document the relationship between the characteristics of province and the difference in variance decomposition across provinces.