非平稳技术冲击、时变通胀目标与中国经济波动——基于动态随机一般均衡的分析

Time-Varying Inflation Target,Non-stationary Technology Shock and China’s Economic Fluctuation

本文构建了一个包含时变通胀目标和非平稳技术冲击的动态随机一般均衡(DSGE)模型,基于中国1992—2010年的季度数据,采用贝叶斯推断方法,估计了模型的结构参数。模拟仿真结果的比较分析发现,构建的DSGE模型能够更好地拟合中国经济,主要经济变量的理论波动相比较而言与实际经济波动更为一致。从模型的预测误差方差分解来看,时变通胀目标和非平稳技术冲击对主要经济变量的波动都具有相应的解释能力。

This paper incorporates time-varying inflation target shock and non-stationary technology shock into a constructed DSGE mode[ to analyze China＇ s economic fluctuation. The DSGE mode[ consists of household, firm and government sectors. In addition, there exist two nominal frictions and three actual frictions in the model. Nominal frictions include sticky product price and households＇ monetary demand constraint. Actual frictions are monopolistic competition production market, consumer＇ s habit, and the cost arising from adjusting investment. The whole theoretical economic system is non-stationary because there are two non-stationary variables in the model, i. e. non-stationary technology and nominal monetary supply. Before solving the rational expectation model and estimating the structural parameters, some transformation procedures are conducted in order to convert the model＇ s variables into stationary ones. Not all parameters in the model are identifiable. We calibrate some parameters that are commonly used in current research literatures. MCMC based Bayesian inference is then adopted to estimate the rest parameters using China＇ s economic quarterly data spanned 1992：1 through 2010： 4. For the purpose of comparing alternative model＇ s specifications, we take the similar steps to estimate two alternative specifications： the one that ignores non-stationary technology shock and the other that specifies inflation target to be constant. Time-varying inflation target is replaced with constant inflation target. The comparison of simulations among different specified models indicates that DSGE model including the two exogenous shocks can imitate China＇ s economy much better. Moreover, the fluctuation of main simulated economic variable is more consistent with actual one. Finally, we use impulse response analysis and forecast error variance decomposition to implement some policy analysis. The impulse response analysis shows that positive inflation target shock could be interpreted as expansionary moneta^ry policy. Meanwhile, the forecast error variance decomposition implies the time-varying inflation target shock and non-stationary technology shock have expository ability to predict main economic variables＇ fluctuation. In conclusion, the research of this paper suggests that when DSGE paradigm is applied to studying such issues as economic fluctuation and optimal monetary policy in China, non-stationary technology shock and time-varying inflation target are two necessary ingredients.