决定人民币均衡汇率因素的统计分析  

Statistical Analysis on RMB Equilibrium Exchange Rate

作　　者：陈淙洁[1] 朱仲义[1] 

机构地区：[1]复旦大学管理学院统计学系,上海200433

出　　处：《数理统计与管理》2009年第1期103-110,共8页Journal of Applied Statistics and Management

摘　　要：本文结合了之前关于均衡汇率的研究以及非参数检验确定了影响人民币汇率的5个因素,并以人民币实际有效汇率为参照基础,采用Engle-Granger两步法得到了人民币均衡汇率的估计模型。结果发现E-G两步法的模型存在较强的共线性。因而在沿用了E-G两步法基本思想的基础上,采用岭回归估计来解决这个问题.从而,对人民币均衡汇率进行估计测算。结果表明,中国在2002年至2005年,人民币实际有效汇率长时期被低估,因而人民币的确面临着一定的升值压力.而2005年的汇率制度改革以及2个百分点的升值使这一情况有了较大的改善。In this paper, 5 factors, which affect the RMB exchange rate, are eventually chosen on the basis of the previous researches and the non-parameter test. Then the model of the RMB equilibrium exchange rate is constructed by means of Engle-Granger two-step method and improved by ridge estimation to solve the problem of Multi-Collinearity. The estimation of RMB equilibrium exchange rate gained from the model shows that RMB exchange rate has been under the appreciation pressure for a long time since real effect exchange rate had been underestimated during the period from 2002 to 2005. What＇s more, the situation becomes better after the system reform and exchange appreciation by 2% in July, 2005.

关 键 词：人民币均衡汇率 非参数检验 线性回归 协整 

分 类 号：O212[理学—概率论与数理统计]