| FCI对我国通胀的预测效果分析 |
| |
| 引用本文： | 封思贤,谢启超,张文正.FCI对我国通胀的预测效果分析[J].中国软科学,2012(7):61-70. |
| |
| 作者姓名： | 封思贤 谢启超 张文正 |
| |
| 作者单位： | 南京师范大学商学院,江苏南京,210046 |
| |
| 基金项目： | 国家社会科学基金项目，教育部人文社会科学基金项目，江苏省“青蓝工程” |
| |
| 摘 要： | 对通胀趋势进行预测是治理通胀的关键。本文首先运用广义脉冲响应函数构建了我国的金融状况指数(FCI),然后从多个角度评估了FCI对我国通胀水平的总体预测效果,随后应用Markov机制转换模型将我国通胀状态的转换概率内生化,并对不同状态下的FCI通胀预测效果进行了比较。结果表明:总体上,金融状况指数是我国通货膨胀率的先行指标,能有效预测未来半年内的通胀趋势;在模拟和预测我国通胀率运行趋势方面,Markov机制转换模型优于线性AR(2)模型;高通胀状态下FCI的通胀预测能力强于低通胀状态;低通胀状态下FCI对短期通胀的预测效果优于中长期。
|
| 关 键 词： | 金融状况指数 通胀预测 广义脉冲 Markov机制转换模型 |
A Study on the Predictive Performance of FCI for China's Inflation |
| |
| FENG Si-xian
,
XIE Qi-chao
,
ZHANG Wen-zheng.A Study on the Predictive Performance of FCI for China's Inflation[J].China Soft Science,2012(7):61-70. |
| |
| Authors: | FENG Si-xian XIE Qi-chao ZHANG Wen-zheng |
| |
| Institution: | (School of Business,Nanjing Normal University,Nanjing 210046,China) |
| |
| Abstract: | Inflation prediction is a key to inflation-fighting.At first,this paper puts out a construction of Chinese Financial Conditions Index(FCI) based on Generalized Impulse-response Functions(GIRF).Secondly,it estimates FCI’s general predictive effect for Chinese inflation in different ways.Then,the predictive effects of FCI for inflation in different inflation conditions are given after an analysis of Chinese inflation situation switching by using Markov regime-switching model.The results show that FCI has a relatively strong predictive ability for the trend of Chinese inflation rate in the next six months.Markov regime-switching model is better than linear AR(2)model when predicting Chinese inflation trend.Predictive ability of FCI for Chinese Inflation in high inflation condition is stronger than in low inflation condition.Under the low inflation condition,predictive effect of FCI for inflation in the short term is better than in the medium term and long-term. |
| |
| Keywords: | financial conditions index inflation prediction generalized impulse Markov regime-switching model |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
