应用GARCH族模型与历史模拟法的商业银行利率风险实证对比研究.docVIP

PAGE PAGE 32 应用GARCH族模型与历史模拟法的商业银行利率风险实证对比研究 摘要：2012年开始，人民银行允许存贷款利率在限定区间内实行浮动机制并逐步放宽浮动区间，这一系列举措进一步推动了我国的利率市场化。因此，在利率市场化背景下，如何准确地对利率风险进行识别、度量和防范成为商业银行经营管理的重要课题。本文在对比利率敏感性缺口、久期缺口和在险价值法的基础上，采用在险价值模型对2007年4月至2016年4月我国上海银行间同业拆借利率的对数收益率进行实证研究，用GARCH族模型和历史模拟法对同业拆借利率的VaR值进行估算，得出以下结论：(1)同业拆借利率存在显著的偏态性、尖峰厚尾性和波动聚集特征；(2)用历史模拟法估计出来的上海银行间同业拆借利率VaR值较小，表明该方法对商业银行利率风险的估计较为保守，高估了风险；(3)对GARCH族模型进行数据拟合得出：GED分布假设更适用于描述同业拆借利率序列的分布状况，且TGARCH模型较为准确地拟合了同业拆借利率序列，并反映出序列存在的“杠杆效应”。 关键词：在险价值模型利率风险 GARCH模型历史模拟法 Research on Measurement of Chinese Commercial Bank Interest Rate Based on GARCH model and Historical Simulation Method Abstract: From 2012, Peoples Bank of China allow the deposit and lending rates changing in the specified range , which promote the marketization of interest rate in China. Therefore, the research of interest rate measurement and prevention has become an important subject in bank management. In this paper, by choosing Shibor O/N data (from 2007.4 to 2016.4) as experimental data and comparing three models including interest-rate sensitive gap model、duration gap model and VaR model, we finally use GARCH model-one of VaR models and historical simulation method to estimate the value at risk of Shibor O/N. And we get the following conclusions : (1) The Shibor data has a characteristic of fluctuation clustering, does not comply with the normal distribution and has an obvious peak and fat tails. (2) VaR estimated by historical simulation method is more conservative.(3) GED distribution is more applicable to reflect risk of interest rate, and TGARCH model is the most suitable model to describe the risk of Shibor comparing with other two models. TGARCH model shows that Shibor data has a characteristic of leverage effect. Keywords: VaR model Interest rate risk GARCH model Historical simulation method 正文目录 一、 引言 （一） 研究背景 上世纪70年代，金融自由化的浪潮席卷了全球，引发了世界范围内金融领域的创新与革命。利率市场化作为金融自由化的核心也在各国逐步推进。迄今为止，世界上大部分发达国家及发展中国家已经完成了利率市场化改革。 我国的利率市场化始于1996年放开全国银行业同业拆借市场，之后又逐步放开国债市场、银行间债券回购与现券交易市场等。2012年