粘性价格、粘性信息与中国菲利普斯曲线

Sticky Prices,Sticky Information and Chinese Phillips Curve

本文在引入预期和平滑机制的泰勒规则基础上建立状态空间模型,通过Kalman滤波估计不同时期对当期通胀率和产出缺口增长率的预期,进而估计中国粘性信息模型(SIPC)和双粘性模型(DSPC)。研究显示,中国企业平均每4到9个季度更新1次信息,每3个季度更新1次价格。通过进一步比较各模型在解释中国通胀动态时的表现,本文发现SIPC对通胀的解释力严重不足;DSPC主要体现出粘性价格成分,其估计结果同粘性价格模型(NKPC)相似且拟合效果很好;而混合NKPC对通胀的解释力则稍强一些。

In this paper, we present a methodology to estimate sticky -information and dual-sticki- ness models. Using this methodology, we estimate past expectations of current inflation and the output gap, and then estimate average durations of information stickiness under different analytical framework. Evidence shows that, on average, firms in China update their information once every 4 - 9 quarters and update their prices once every 3 quarters on average. In addition, we also compare SIPC, DSPC, NKPC and hybrid NK- PC. We find that SIPC cannot provide an accurate representation of Chinese inflation, DSPC fits well and captures inflation dynamics as well as NKPC, and hybrid NKPC has slightly stronger explanatory power than other models.