基于GDFM模型的我国通胀动态预测研究

发布时间：2018-04-26 05:23

本文选题：通货膨胀预测 + 货币政策　；参考：《湖南大学》2013年硕士论文

【摘要】：通货膨胀是全球性难题,各国央行均害怕高通货膨胀会给经济带来损失。为有效地控制高通货膨胀的发生,稳定物价水平,许多国家使用新的技术工具控制通货膨胀工具——通货膨胀预测。特别是在那些实行了通货膨胀目标制的国家,这种工具越来越受重视。而在没有实行通货膨胀目标制的国家,通货膨胀预测也被其央行视为货币政策的重要参考。对于央行来说,由于货币政策的效果存在时滞,在通货膨胀发生时制定货币政策无法达到控制通胀的效果。但如果中央银行在制定货币政策时,能定期提前预测准确的通货膨胀率,并在不同时点上掌握通货膨胀的发展趋势,从而能制定超前的货币政策,维持长期物价的稳定,则有利于维持国内经济长期平稳的发展,并提高货币政策的有效性。本文在国外通货膨胀预测研究已有的基础上,探讨一种新的通货膨胀预测方法,即GDFM模型预测方法。GDFM模型是一种统计类预测模型,其利用频域分析方法和滞后算子多项式刻画经济系统中多个因子的动态性,并允许随机扰动项同期非正交,充分考虑了经济系统中各变量间的相互联系,更好地刻画经济系统的真实情况。在本文的实证部分,根据我国2001年1月至2010年12月的居民消费价格指数及其分类、工业生产者出厂价格指数、汇率、出进口、货币供给、消费税等37个与通货膨胀有关的经济变量月度数据,建立GDFM模型对我国通货膨胀进行样本外预测。实证结果表明,GDFM模型很好地预测我国通货膨胀。通过在绝对误差,相对误差以及均方根误差预测评价指标上的比较,得出GDFM模型的预测效果优于ARIMA(4,1,12)预测模型的结论。
[Abstract]:Inflation is a global problem, and central banks fear that high inflation will cause economic losses. In order to effectively control the occurrence of high inflation and stabilize the price level, many countries use new technical tools to control inflation-inflation forecast. Especially in countries where inflation targets are in place, the tool is gaining ground. In countries that do not target inflation, inflation forecasts are seen by central banks as an important reference for monetary policy. For central banks, because of the delay in the effect of monetary policy, when inflation occurs, monetary policy can not achieve the effect of controlling inflation. However, if the central bank can regularly predict the accurate rate of inflation in advance when formulating monetary policy, and grasp the development trend of inflation at different points in time, so that it can formulate an advanced monetary policy and maintain long-term price stability. It will help to maintain a stable domestic economy, and improve the effectiveness of monetary policy. In this paper, a new inflation forecasting method, I. E. GDFM model forecasting method. GDFM model is a kind of statistical forecasting model, is discussed on the basis of the existing research on inflation prediction abroad. The method of frequency domain analysis and the polynomial of delay operator are used to characterize the dynamics of many factors in the economic system, and the random perturbation terms are allowed to be nonorthogonal simultaneously, and the interrelation among the variables in the economic system is fully considered. Better portray the real situation of the economic system. In the empirical part of this paper, according to the consumer price index and its classification from January 2001 to December 2010, industrial producer price index, exchange rate, import, money supply, Based on the monthly data of 37 inflation-related economic variables, the GDFM model is established to forecast the inflation in China. The empirical results show that the GDFM model can well predict inflation in China. By comparing the prediction indexes of absolute error, relative error and root mean square error, it is concluded that the prediction effect of GDFM model is better than that of Arima (41 / 12) model.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F822.5

【参考文献】