中国货币政策规则的比较分析——基于DSGE模型的三规则视角

A Comparative Study of China's Monetary Policy Rules:On the Perspective of Three Rules Based on the DSGE Model

单一的利率规则或数量规则均不足以反映中国货币政策的操作实践。有鉴于此,本文构建了具有并行选择性和包容性特征的货币政策混合规则形式,并基于中型新凯恩斯DSGE模型,通过校准和贝叶斯方法分别估计了利率规则、数量规则和混合规则下三个模型的参数。通过比较不同规则模型在提前一期预测值、预测值标准差和根均方误差上的实证表现,并与现实数据对比,我们发现新的混合规则能更好地拟合中国经济的实际运行。基于等价消费稳态偏离等标准,我们对比了不同规则模型在遭遇外部冲击时的福利表现,发现央行使用具有并行选择特征的混合规则,可以提高调控效率,减少福利损失。概言之,无论在实证(是什么)还是在规范(应该是什么)的意义上,新构建的混合规则均优于利率规则或数量规则。

As the monetary policy operation of the People＇s Bank of China represents the parallel regulations of interest rate and quantity instruments, neither a single interest-rate rule nor a quantity rule can fully reflect the actual implements of monetary policies in China. Moreover, the timing of these two instruments＇ use is not consistent, and thus the central bank＇s specific choice of policy actions is discretionary. We refer to this special monetary policy operation as the mixed operation and the corresponding policy rule as the mixed rule. The literature on the single rule ignores the existence of the mixed rule, which evidently deviates from practice. Hence, it is necessary to establish a reasonable form of the mixed rule and use it to analyze and compare the applicability of various monetary policy rules, which should help us to understand the actions of the central bank, and improve the practices of monetary policy. In this paper, we develop a New Keynesian DSGE model in which monetary policy follows the mixed rule regarding inclusiveness and discretion. Based on this model, we introduce the following shocks in the economy： the technology shock, the consumers＇ preference shock, the investment technology shock, the government expenditure shock, and the monetary policy shock. Using both calibration and Bayesian methods, we separately estimate the parameters in the model under the interest rate rule, quantity rule, and mixed rule and compare the difference in forecasts, forecast standard deviations, and RMSE between data and estimated values under these three rules. The results show that when we compare only the single interest rate and quantity rules, the interest rate rule fits better to the real economy than the quantity rule, and neither are suitable for matching the inflation data. When we compare the three rules simultaneously, we find that a mixed monetary policy better matches the economic performance and inflation data. Moreover, we use the deviation of consumption as the index of welfare fluctuation to measure the regulation efficiency of monetary policy under these three rules in the presence of external shocks. Our analysis indicates that when the technology shock or the investment technology shock happens, the central bank uses the mixed rule to regulate the economy, leading to the fastest recovery and minimum welfare loss, which means that the mixed rule is better fitted than either the sole interest rate rule or quantity rule. When the monetary policy shock happens, the social welfare loss is at its maximum under the quantity rule, followed by the mixed rule, and at its minimum under the interest rate rule. Given the differences in economic applications and model settings, this means that the adoption of rules is more conducive to the improvement of social welfare than the adoption of random and unexpected monetary policy operations. In short, the new mixed rule dominates the interest-rate rule and quantity rule both positively and normatively. Regarding academic research and the operation of monetary policy, our paper suggests that if we study China＇s economy based on the general equilibrium framework, it may be biased to use only the single interest rate or quantity rule to describe the central bank＇s behavior; a mixed rule is better suited to our national conditions. Based on the current economic situation, when the central bank uses monetary policy to intervene in the economy, its selective use of interest rates and quantitative tools for portfolio operations is still the most efficient. Moreover, the central bank should follow the rules of operation to reduce the unexpected policy operation. In other words, our paper takes the first step in the comparative analysis of these three rules. Future research may extend to two aspects. In terms of economic meanings, we may investigate the micro foundation of the central bank＇s monetary policy interventions, and further explore the three monetary policy rules considering issues of action mechanism and timing selection. In relation to empirical studies, we may examine the applicability of monetary policy rules for time-varying parameters at different stages of the economic period and economic cycle, and further test forward-looking monetary policy rules.