统计与信息论坛
統計與信息論罈
통계여신식론단
STATISTICS & INFORMATION TRIBUNE
2014年
6期
40-47
,共8页
同业拆借利率%风险度量%分形理论%ARFIMA-FIGARCH模型
同業拆藉利率%風險度量%分形理論%ARFIMA-FIGARCH模型
동업탁차리솔%풍험도량%분형이론%ARFIMA-FIGARCH모형
inter-bank lending interest rate%risk measurement%fractal theory%ARFIM A-FIGARCH model
随着中国利率市场化改革的加速,利率市场的风险管理问题引发了广泛的关注,作为筹集短期流动性资金的主要工具,同业拆借利率(Shibor)逐渐成为各金融机构决策参考的基准利率。在传统的ARMA-GARCH模型的基础上,引入Hurst指数捕捉Shibor的分形特征,使用扩展后的ARFIMA-FIGARCH模型对Shibor的隔夜和7日利率收益率的VaR进行度量和回测检验。结果显示:隔夜和7日利率收益率都具有反持续性,即收益率过去是上升趋势,则未来倾向于下降;考虑分形特征的ARFIMA-FIGARCH模型,比原模型对Shibor的度量更准确;在同业拆借市场中,Ged分布是解释多头VaR的理想选择,而正态分布是解释空头VaR的理想选择。
隨著中國利率市場化改革的加速,利率市場的風險管理問題引髮瞭廣汎的關註,作為籌集短期流動性資金的主要工具,同業拆藉利率(Shibor)逐漸成為各金融機構決策參攷的基準利率。在傳統的ARMA-GARCH模型的基礎上,引入Hurst指數捕捉Shibor的分形特徵,使用擴展後的ARFIMA-FIGARCH模型對Shibor的隔夜和7日利率收益率的VaR進行度量和迴測檢驗。結果顯示:隔夜和7日利率收益率都具有反持續性,即收益率過去是上升趨勢,則未來傾嚮于下降;攷慮分形特徵的ARFIMA-FIGARCH模型,比原模型對Shibor的度量更準確;在同業拆藉市場中,Ged分佈是解釋多頭VaR的理想選擇,而正態分佈是解釋空頭VaR的理想選擇。
수착중국리솔시장화개혁적가속,리솔시장적풍험관리문제인발료엄범적관주,작위주집단기류동성자금적주요공구,동업탁차리솔(Shibor)축점성위각금융궤구결책삼고적기준리솔。재전통적ARMA-GARCH모형적기출상,인입Hurst지수포착Shibor적분형특정,사용확전후적ARFIMA-FIGARCH모형대Shibor적격야화7일리솔수익솔적VaR진행도량화회측검험。결과현시:격야화7일리솔수익솔도구유반지속성,즉수익솔과거시상승추세,칙미래경향우하강;고필분형특정적ARFIMA-FIGARCH모형,비원모형대Shibor적도량경준학;재동업탁차시장중,Ged분포시해석다두VaR적이상선택,이정태분포시해석공두VaR적이상선택。
With the acceleration of China's market-oriented interest rate reform and the implementation of Basel III ,the risk management in interest market causes widespread concern .As the main tool to raise short-term liquidity ,Shanghai Interbank Offered Rate (Shibor ) has become the benchmark rate w hen financial institutions make decisions ,and is playing an increasingly important role .In this paper ,based on the traditional ARMA-GARCH model ,Hurst index is introduced to capture the fractal characteristics of Shibor ,the expanded ARFIMA-FIGARCH model is used to measure the risk of overnight and 1 week interest rates in Shanghai interbank market , and back-testing is performed for VaR constructed by different models .The results show that Shibor yields have fractal characteristics and anti-persistent ,which means today's earnings brings on tomorrow's losing ;Compared with the original one ,ARFIMA-FIGARCH model is more accurate measure of Shibor after consideration of fractal characteristics ;in the interbank market ,Ged distribution is an ideal choice for explaining long VaR ,and Normal distribution is an ideal choice for short VaR .