摘要
人民币汇率升值预期持续强化,观察其变化的重要指标主要是境外人民币无本金交割远期汇率(CNYNDF)。本文建立CNYNDF的高频日汇率随机游走模型,运用ARCH族模型检验该模型的残差后发现,其时间序列数据具有"尖峰厚尾性"和集群性特征。在信息不对称的条件下,市场对人民币升值和贬值存在不同程度的反应,2002年以来,市场预期人民币升值,市场信息具有杠杆作用。2005年7月21日人民币汇率形成机制改革以来,境内人民币即期汇率和远期汇率开始受CNYNDF报价引导,逐步形成稳定的Granger因果关系。为取得人民币汇率定价的主动权,中国需加快外汇交易的微观制度建设,加强汇率与利率改革的协调性,为规避汇率风险创造市场条件。
The CNYNDF （non-deliverable forward） rate is an important reference index for CNY expectation. This paper builds up a random walk model to test the characteristics of the high frequency daily data on CNYNDF. By using a set of ARCH （Autoregressive Conditional Heteroskedastic） models, we find high peak, fat tail and volatility clustering in the residual. Given the asymmetry information, the market＇ s response to appreciation and depreciation differs from each other. It is widely expected that CNY would be appreciated, and market information has a leverage effecL The spot and forward exchange rate were led by the CNYNDF rate and there were granger cause between them since RMB exchange rate regime reform on July 21st 2005. It is urgent for China to build a more efficient foreign exchange system and to recapture leadership in pricing CNY so as to provide with alternatives to manage the risks of RMB exchange rate.
出处
《国际金融研究》
CSSCI
北大核心
2008年第4期52-59,共8页
Studies of International Finance
