“新汇改”后人民币汇率预测
|
摘要
为了探讨汇率变动的规律,对2015年8月12日至2018年4月18日的人民币兑美元和欧元的汇率日数据进行平稳性检验,并进一步对汇率波动序列数据进行正态性、自相关性检验和BDS检验.再利用BP神经网络对汇率进行精准预测,根据拟合效果图发现,后期人民币兑美元汇率有下降趋势,但是人民币兑欧元汇率短期内呈上升趋势,拟合值与实际值精确度高达99%,在该形势下,我国政府应做好宏观调控,对我国深化汇率制度市场化提出政策建议.
In order to explore the law of exchange rate fluctuations,a stationarity test on the exchange rate data of RMB against the US dollar and Euro between August 18,2015 and April 18,2018 was conducted and then the exchange rate fluctuation sequence data was further subjected to Normality test,Autocorrelation test and BDS test.The BP neural network was adopted to perform accurate prediction of the exchange rate.Based on the fitting effect chart,the exchange rate of RMB against the US dollar underwent a downward trend in the later stage,but the exchange rate of the RMB against the Euro is rising in the short term,with the accuracy of the fitted value and the actual value reaching as high as 99%.Under such a situation,the Chinese government should do a good job in macroeconomic regulation and control,and policy recommendations were put forth for China to deepen the marketization of the exchange rate system.
In order to explore the law of exchange rate fluctuations,a stationarity test on the exchange rate data of RMB against the US dollar and Euro between August 18,2015 and April 18,2018 was conducted and then the exchange rate fluctuation sequence data was further subjected to Normality test,Autocorrelation test and BDS test.The BP neural network was adopted to perform accurate prediction of the exchange rate.Based on the fitting effect chart,the exchange rate of RMB against the US dollar underwent a downward trend in the later stage,but the exchange rate of the RMB against the Euro is rising in the short term,with the accuracy of the fitted value and the actual value reaching as high as 99%.Under such a situation,the Chinese government should do a good job in macroeconomic regulation and control,and policy recommendations were put forth for China to deepen the marketization of the exchange rate system.
引文
[1]中国人民银行.人民币汇率形成机制改革继续朝市场化迈进[EB/OL].[2015-08-11].http://www.pbc.gov.cn/.
[2]肖磊,董树功.对人民币国际化路径的探讨[J].内江师范学院学报,2006,21(4):65-67.
[3]邝坦励.人民币汇率与汇率制度选择[J].内江师范学院学报,2004,19(1):25-28.
[4]何佳.浅析美国金融危机的影响及我国的应对策略[J].内江师范学院学报,2009,24(3):23-25.
[5]ABBATE A,MARCELLINO M.Point,interval and density forecasts of exchange rates with time varying parameter models[J].Journal of the Royal Statistical Society:Series A(Statistics in Society),2017,181(1):28-41.
[6]JOSEPH P.BYRNE,Dimitris Korobilis,Pinho J.Ribeiro.ON THE SOURCES OF UNCERTAINTY IN EX-CHANGE RATE PREDICTABILITY[J].International Economic Review,2018,59(1):67-78.
[7]INGOMAR Krohn,MICHAEL J.MOORE.Dealer activity and macro fundamentals-New evidence from hybrid exchange rate models[J].Journal of International Money and Finance,2018,37(1):98-110.
[8]STAVROS Degiannakis.Multiple days ahead realized volatility forecasting:Single,combined and average forecasts[J].Global Finance Journal,2018,36(1):128-136.
[9]杨帆.新汇改前后人民币汇率中间价的长期记忆性研究[J].改革与开放,2016(10):14-15.
[10]李艳丽,张志翔.人民币远期汇率无偏性的动态检验[J].国际金融研究,2017(11):76-85.
[11]江春,杨宏略,李小林.基于泰勒规则的人民币汇率预测研究:兼论多种汇率决定模型预测比较[J].世界经济研究,2018(04):3-15.
[12]REFENSE A P,Francis G,et al.Stock performance modeling using neural networks[J].1993,56(1):110-124.
[13]JINGTAO YAO,CHEW LIM TAN,HEANLEEPOH.Neural networks for technical analysis:A study on KLCI[J].International Journal of Theoretical&Applied Finance,1999,2(2):221-241.
[14]LU Z,MAY A,LIU K,et al.How to Scale Up Kernel Methods to Be As Good As Deep Neural Nets[J].Eprint Arxiv,2015,39(1):156-179.
[15]单峰.基于神经网络集成的汇率预测研究[D].南京:南京航天航空大学,2004.
[16]徐缘圆.BP神经网络在汇率预测中的应用[J].时代金融,2013,5005(1):147-148.
[17]邓景炜.基于神经网络的外汇汇率预测研究[D].广州:暨南大学,2017.
[2]肖磊,董树功.对人民币国际化路径的探讨[J].内江师范学院学报,2006,21(4):65-67.
[3]邝坦励.人民币汇率与汇率制度选择[J].内江师范学院学报,2004,19(1):25-28.
[4]何佳.浅析美国金融危机的影响及我国的应对策略[J].内江师范学院学报,2009,24(3):23-25.
[5]ABBATE A,MARCELLINO M.Point,interval and density forecasts of exchange rates with time varying parameter models[J].Journal of the Royal Statistical Society:Series A(Statistics in Society),2017,181(1):28-41.
[6]JOSEPH P.BYRNE,Dimitris Korobilis,Pinho J.Ribeiro.ON THE SOURCES OF UNCERTAINTY IN EX-CHANGE RATE PREDICTABILITY[J].International Economic Review,2018,59(1):67-78.
[7]INGOMAR Krohn,MICHAEL J.MOORE.Dealer activity and macro fundamentals-New evidence from hybrid exchange rate models[J].Journal of International Money and Finance,2018,37(1):98-110.
[8]STAVROS Degiannakis.Multiple days ahead realized volatility forecasting:Single,combined and average forecasts[J].Global Finance Journal,2018,36(1):128-136.
[9]杨帆.新汇改前后人民币汇率中间价的长期记忆性研究[J].改革与开放,2016(10):14-15.
[10]李艳丽,张志翔.人民币远期汇率无偏性的动态检验[J].国际金融研究,2017(11):76-85.
[11]江春,杨宏略,李小林.基于泰勒规则的人民币汇率预测研究:兼论多种汇率决定模型预测比较[J].世界经济研究,2018(04):3-15.
[12]REFENSE A P,Francis G,et al.Stock performance modeling using neural networks[J].1993,56(1):110-124.
[13]JINGTAO YAO,CHEW LIM TAN,HEANLEEPOH.Neural networks for technical analysis:A study on KLCI[J].International Journal of Theoretical&Applied Finance,1999,2(2):221-241.
[14]LU Z,MAY A,LIU K,et al.How to Scale Up Kernel Methods to Be As Good As Deep Neural Nets[J].Eprint Arxiv,2015,39(1):156-179.
[15]单峰.基于神经网络集成的汇率预测研究[D].南京:南京航天航空大学,2004.
[16]徐缘圆.BP神经网络在汇率预测中的应用[J].时代金融,2013,5005(1):147-148.
[17]邓景炜.基于神经网络的外汇汇率预测研究[D].广州:暨南大学,2017.