The exchange rate risk of Chinese yuan: Using VaR and ES based on extreme value theory
The exchange rate risk of Chinese yuan: Using VaR and ES based on extreme value theory
复制标题
人民币汇率风险：基于极值理论的VaR和ES
DOI：
10.1080/02664760902846114
复制
发表时间：
2010-01-01
影响因子：
1.5
通讯作者：
Zhou, Yanju
中科院分区：
文献类型：
--
作者：
Wang, Zongrun;Wu, Weitao;Zhou, Yanju
本文运用极值理论（EVT）对人民币汇率收益序列的尾部进行估计。我们发现帕累托分布对回归序列尾部数据的拟合程度非常高。实证结果表明，预期不足并不能改善风险价值（VaR）的尾部风险问题。回溯检验的证据表明，基于evt的VaR值低估了美元/人民币、港币/人民币等汇率的风险，这可能是由于人民币对美元和港币的持续升值造成的。但是，与历史模拟和方差协方差法计算的VaR值相比，EVT计算的VaR值可以更准确地衡量JPY/CNY和EUR/CNY汇率的风险。
This paper applies extreme value theory (EVT) to estimate the tails of return series of Chinese yuan (CNY) exchange rates. We find that the degree of fitting Pareto distribution to the data of the tail of return series is extremely high. The empirical results indicate that expected shortfall cannot improve the tail risk problem of value-at-risk (VaR). The evidence of back testing indicates that EVT-based VaR values underestimate the risks of exchange rates such as USD/CNY and HKD/CNY, which may be caused by the continuous appreciation of CNY against USD and HKD. However, compared with VaR values calculated by historical simulation and variance–covariance method, VaR values calculated by EVT can measure the risk more accurately for the exchange rates of JPY/CNY and EUR/CNY.