学习效应、通胀目标变动与通胀预期形成

Learning Effect, Change of Inflation Target and the Formation of Inflation Expectation

本文在国内首次采用新凯恩斯动态随机一般均衡模型测度我国的季度通胀预期,并运用贝叶斯法估计模型参数。研究发现:时变通胀目标和学习机制能增强新凯恩斯DSGE模型对我国宏观经济的刻画程度,对主要宏观变量提供较好的预测;模型测度的通胀预期较好反映1992年以来实际通胀的变化,具有迟缓性、自我实现性、平稳化和近理性特征,同时,比统计调查法、状态空间模型、SVAR模型和马尔科夫范式转换模型等传统方法的通胀预测能力和测度稳健性更强。本文所揭示的学习机制为中央银行的通胀预期管理提供了理论依据,中央银行应提高货币政策的透明度,构造通胀预期形成的名义瞄,积极向公众宣示坚定的反通胀立场,以实现对通胀预期的有效管理。

This thesis applied the New-Keynesian Dynamic Stochastic General Equilibrium Model to measure the quarterly inflation expectation of China by bayesian estimation for the first time. Main results are as following：the public＇s learning behavior and time-varying inflation target can strengthen the ability of New-Keynesian Dynamic Stochastic General Equilibrium Model to describle the relationship between macroeconomy variables in China and has good in-sample and out- sample fit; Inflation expectation estimated by DSGE model exhibits dramatically real inflation process in China since 1992, has the characteristic of being sluggish and self-fulling adjustment, and demonstrates the characteristic of being moderation and near-rational; Compared to current econometric analysis of inflation expectation, such as Statisitic Survey Method, Two-Variable State Space Model, SVAR model and Markov Regime Switching Model, the estimation of inflation expectation on the basis of DSGE model has better predict power of fututer inflation, and remains more robust estimation. The learning behavior revealed by this thesis lends theoretical support to central bank＇ s management of inflation expectation, central bank should improve transparency of monetary policy, construct an anchor for the formation of inflation expectation, actively convey anti-inflation attitude towards the public to effectively manage inflation expectation.