【作者】 徐斌；

【导师】 刘达禹；

【作者基本信息】 吉林大学 ， 数量经济学， 2022， 硕士

【摘要】 菲利普斯曲线一直是宏观经济领域内的经典议题,它系统性刻画了通胀与产出的同向变动关系,这为宏观经济形势分析、经济下行风险预警与经济复苏政策设计提供了重要的理论基础。然而,自2008年全球金融危机以来,接连触底的经济增速与严重的结构通胀共存,一系列事实向菲利普斯曲线的有效性发起挑战,甚至不少研究认定,菲利普斯曲线发生了结构改变,曲线呈现“扁平化”特征。但有趣的是,在2020年新冠疫情后的经济复苏过程中,世界多国再次出现“高增速与高物价”共存的局面,这似乎预示着菲利普斯曲线再度“复苏”。那么,菲利普斯曲线的特征究竟如何,是我们亟待寻求与讨论的重要议题。基于研究背景,本文选取了1995年第一季度至2021年第四季度的实际GDP数据,创立了一套涵盖经济事实拟合能力检验、平稳性检验、耦合性检验、预测能力检验与测度稳定性检验的产出缺口计量评价体系,利用最适合中国实际的产出缺口对菲利普斯曲线模型进行了分型估计,据此识别通胀变量遴选与模型设定形式对菲利普斯曲线模型估计的影响,同时对菲利普斯曲线的静态特征予以检验。随后,为全面、直观、动态地观测菲利普斯曲线模型中“通胀-产出”的反应系数变化趋势,本文将静态线性模型扩展至时变参数的菲利普斯曲线模型,进一步发掘出菲利普斯曲线发挥功能的条件与适用性。研究结果表明:第一,计量偏误极易导致菲利普斯曲线特征检验的结果存伪,这是由于各类产出缺口的测度方法以及菲利普斯曲线的模型设定形式均是基于当时的经济环境衍生出的结果,因此对于当前的中国而言,需要进行综合的计量评价,这样才能遴选出最优的菲利普斯曲线模型设定方式;第二,Kalman缺口最适合中国的菲利普斯曲线特征检验,相较之下,BK缺口与CF缺口的能力属性较弱,HP缺口的预测能力与测度稳定性不足,小波缺口无法准确拟合出中国的典型经济事件;第三,考虑了投资品价格的GDP平减指数更适合度量通货膨胀水平,而从模型设定视角,利用传统的菲利普斯曲线模型估计出的“通胀-产出”系数最大,而混合菲利普斯曲线模型更适合于通胀预测;第四,从时变参数菲利普斯曲线模型的动态计算结果中发现,菲利普斯曲线往往活跃在经济波动较为剧烈的时期,在经济平稳期的功能比较模糊,由此看来,菲利普斯曲线在经济风险捕捉与经济复苏期的政策设计上能够发挥出最佳效果,这我们认识菲利普斯曲线提供了新的经验证据与事实参考。更多还原

【Abstract】 The Phillips curve has always been a classic topic in the field of macroeconomics.It systematically depicts the relationship between inflation and output in the same direction,which provides an important theoretical basis for macroeconomic situation analysis,economic downside risk warning,and economic recovery policy design.However,since the global financial crisis in 2008,successive bottoming out economic growth and severe structural inflation have frequently coexisted.A series of facts have challenged the validity of the Phillips curve.Many studies have even concluded that the Phillips curve has undergone structural changes.change,the curve exhibits a“flattening” characteristic.But what is interesting is that in the process of economic recovery after the new crown epidemic in 2020,the coexistence of “high growth rate and high prices” appeared again in many countries in the world,which seemed to indicate that the Phillips curve “recovered” again.Then,what are the characteristics of the Phillips curve is an important issue that we need to seek and discuss urgently.Based on the research background,this paper selects the real GDP data from the first quarter of 1995 to the fourth quarter of 2021,and creates a set of economic facts fitting ability test,stationarity test,coupling test,prediction ability test and measurement stability.The output gap measurement and evaluation system tested,using the output gap that is most suitable for China’s actual production gap to estimate the Phillips curve model,according to which to identify the impact of the selection of inflation variables and model setting on the estimation of the Phillips curve model,and at the same time The static characteristics of the Phillips curve are examined.Subsequently,in order to comprehensively,intuitively and dynamically observe the change trend of the “inflation-output” response coefficient in the Phillips curve model,this paper extends the static linear model to the Phillips curve model with time-varying parameters,and further explores the role of the Phillips curve.Condition and applicability of features.The research results show that: first,measurement bias can easily lead to false results of the Phillips curve characteristic test,because the measurement methods of various output gaps and the model setting form of the Phillips curve are based on the economic environment at that time.Therefore,for the current China,it is necessary to carry out a comprehensive measurement evaluation,so as to select the optimal Phillips curve model setting method;second,the Kalman gap is most suitable for China’s Phillips curve characteristic test,In contrast,the capability attributes of the BK gap and the CF gap are weak,the prediction ability and measurement stability of the HP gap are insufficient,and the wavelet gap cannot accurately fit the typical economic events in China;thirdly,the GDP considering the price of investment goods The deflator is more suitable for measuring the level of inflation,and from the perspective of model setting,the "inflation-output" coefficient estimated by the traditional Phillips curve model is the steepest,while the mixed Phillips curve model is more suitable for inflation forecasting;Fourth,from the dynamic calculation results of the time-varying parameter Phillips curve model,it is found that the Phillips curve is often active in periods of relatively violent economic fluctuations,and its function in the stable period of the economy is relatively vague.The policy design in the period of capturing and economic recovery can play the best role,and our understanding of the Phillips curve provides new empirical evidence and factual reference.更多还原