摘要
运用GARCH-CVaR和GARCH-VaR方法,在不同模型、不同分布、不同置信水平的假定下,对人民币汇率风险进行测度,测度结果表明:GARCH模型的种类对CVaR和VaR的计算结果影响不明显,而分布假定和置信水平对CVaR和VaR的计算结果影响显著;同时,还对GARCH-CVaR和GARCH-VaR的风险测度效果进行了对比研究,对比结果表明:分布假定和置信水平会显著影响CVaR对VaR的改进效果.
GARCH-CVaR and GARCH-VaR are applied to measure the risk of RMB/USD exchange rate series under different assumptions of models,distributions and confidence intervals.The results show that the differences from different GARCH models are not quite obvious,but the distributions and the confidence intervals are quite obvious.Meanwhile,comparative analysis on risk measure effects of GARCH-CVaR and GARCH-VaR is also presented.Comparative results show that the distribution assumptions and confidence levels can significantly influence the improved results from VaR to CVaR.
出处
《中南民族大学学报(自然科学版)》
CAS
2011年第2期129-134,共6页
Journal of South-Central University for Nationalities:Natural Science Edition
基金
教育部人文社会科学研究项目(09YJC790209)