摘要
当今中国和世界正处于百年未有之大变局,维护国家的经济稳定性至关重要.众所周知,严重的通货膨胀是经济不稳定的一个重要因素.因此,建模和预测通货膨胀率成为了亟待解决的问题.在本文中,我们研究了包括中国和美国在内的世界四个主要国家近十年的消费者价格指数(CPI)通胀率,提出含结构性断点波动率与时变参数均值随机波动模型(stochastic volatility in mean model with time-varying parameters and structural breaks in the volatility, SB-TVP-SVM),并给出了相应的贝叶斯估计框架.在以往的大多数研究中,研究者们往往忽略了非平稳特征同时存在于CPI通胀率的条件均值和波动率序列中的可能性.通过引入不可观测的结构性断点, SB-TVP-SVM解决了这一问题,从而得到相比于既有方法更高的序列预测精度.我们模型估计出的结构性断点与过去十年来最大的全球事件高度相关,例如新冠病毒疫情以及俄乌地区冲突.
China and the world today are undergoing great changes that have not been seen in a century, and it is fundamental to maintain the country’s economic stability. As is known, extreme inflation is one of the major sources of economic instability. Modeling and predicting inflation thus become an urgent problem to be solved. In this work, we investigate the consumer price index(CPI) inflation from the recent ten years of four major countries in the world, including China and USA,and propose a novel stochastic volatility in mean model with time-varying parameters and structural breaks in the volatility(SB-TVP-SVM), and a corresponding Bayesian inference framework. In most of the previous studies, researchers usually ignored the potential coexistent non-stationaritiy of both the conditional mean and volatility series for CPI inflation. By introducing unobserved structural break points in the volatility process, SB-TVP-SVM model solved this problem and achieves better prediction accuracy than its competitors. The structural break points estimated from our model are found to be highly related to the biggest global events over the last decade such as the COVID-19 epidemic and the conflict between Russia and Ukraine.
作者
高维清
吴奔
张波
GAO Weiqing;WU Ben;ZHANG Bo(Center for Applied Statistics,Renmin University of China,Beijing 100872,China;School of Statistics,Renmin University of China,Beijing 100872,China)
出处
《计量经济学报》
CSSCI
CSCD
2023年第1期108-127,共20页
China Journal of Econometrics
基金
国家自然科学基金(71873137,72271232,12201628)
中国人民大学科学研究基金(中央高校基本科研业务费专项资金)(21XNLG08)。
作者简介
高维清,中国人民大学统计学院博士研究生,研究方向:金融模型、贝叶斯分析,E-mail:gaowq 143@ruc.edu.cn;吴奔,中国人民大学统计学院讲师,研究方向:高频模型、贝叶斯分析,E-mail:wuben@ruc.edu.cn;通信作者:张波,中国人民大学统计学院教授,研究方向:金融随机分析、数理金融,E-mail:mabzhang@ruc.edu.cn.
