中文摘要：

本文基于广义虚拟经济视角,分析了通胀预期理论的演化与发展、学习型通胀预期测度,以及如何基于学习型预期对新凯恩斯菲利普斯曲线进行扩展。首先,广义虚拟经济视角下的学习预期更能反映经济个体的＂有限理性＂＂黏性信息＂约束及学习动态,可以为通胀预期提供更有说服力的微观基础。其次,广义虚拟经济视角下的通胀预期可以从学习机制和信息维度两个方面来测度,基于信息维度的SVR通胀预期相对于基于学习机制的适应性学习预期是一种更＂高级＂的预期学习方式。最后,在广义虚拟经济视角下,学习预期的引入可以使菲利普斯曲线扩展出新的预期增广形式,通过对比不同预期系数值和不同学习预期系数值,还可以进一步分析货币政策实施和通胀预期管理方式,特别地,如果适应性学习预期系数显著大于SVR学习预期,则中央银行应引导经济个体对通胀的学习行为,通过提高货币政策的透明度和信息披露水平来加快预期学习速度,反之,则中央银行还应加强与其他宏观经济政策的政策协调,增强通胀预期形成的信息维度。

英文摘要：

This article is based on the generalized virtual economy perspective to analyze the evolution and development of inflation expectation theory, the measurement of learning expectation, and how to extend New-Keynesian Phillips Curve based on learning expectation. Firstly, based on the generalized virtual economy perspective, learning expectation is able to reflect the bounded rationality, sticky information and learning dynamics much better. Secondly, based on the generalized virtual economy perspective, inflation expectation can be measured in two forms： learning mechanism and information dimension. Compared with inflation expectation with adaptive learning mechanism, SVR inflation expectation with information dimension is a more advanced form. Finally, the introduction of learning expectation can extend Phillips Curve to some new expectations-augmented forms, and the monetary policy and expectation management can be analyzed by comparing coefficients of different expectations and learning expectations. Especially, if the coefficient of adaptive learning expectation is larger significantly than SVR expectation, the central bank should guide the learning behavior of agent and accelerate the speed of learning by improving transparency of monetary policy and level of information disclosure, otherwise, central bank should also augment information dimensions of expectation forming.by improving the coordination and cooperation with other macroeconomic policy.