新凯恩斯菲利普斯曲线模型在中国的实证研究:基于VAR的分析

Testing the New Keynesian Philips Curves through Vector Autoregressive Models:Evidence from China

本文采用基于VAR的分析方法,对中国经济的新凯恩斯菲利普斯曲线进行了估计和检验,数据样本为1994年到2005年的季度数据。结果显示,单位产出劳动力边际成本对通货膨胀的影响很小,产出缺口仍然是实际边际成本的有效替代变量;理性预期对通货膨胀的影响是显著的,同时通货膨胀过程也表现出很强的惯性特征,这表明相对于纯前望模型,混合模型能够更好地解释通货膨胀的动态过程。在混合模型中,通货膨胀惯性和预期起主导作用,而产出缺口的作用有限,说明我国控制通货膨胀应当采取循序渐进和稳定预期的策略以保持经济的平稳。此外,对混合模型参数模型的进一步计算还表明在我国约有70%左右厂商定价行为是属于后望形式的,他们的定价规则是以前一期价格和0.72倍的前一期通货膨胀之和作为当期的价格。

Using VAR (Vector Autoregressive) method, we build the new Keynesian Philips Curves for China, with the quarter data from 1994 to 2005 as the sample. The empirical results show that the labor share of income has little influence on inflation and the output gap can work as a meaningful proxy for real marginal cost. The rational expectation is significant in determining inflation; and at the same time there is a strong inertia of inflation in China. So the pure forward looking pricing model is far from a good approximation of the inflation dynamics in China and the hybrid model does instead. In the hybrid model, inflation inertia and expectation have a dominant influence on inflation dynamics, implying that the central bank should take gradual disinflation strategies to keep macroeconomic stabilization. The estimated coefficients also mean that about 70 percent of the firms in China are using backward looking pricing rule, which is based on the sum of price level in the preceding period and the preceding inflation level multiplied by 0.72.