考虑收益率自相关特征的存货质押动态质押率设定

发布时间：2018-02-07 09:48

本文关键词： AR()-GARCH( )-GED 长期风险VaR预测动态质押率模型评价　出处：《管理科学》2012年03期 　论文类型：期刊论文

【摘要】：异于收益率弱相关的有效金融市场假说,以现货交易为主的质物市场收益率往往存在显著的自相关。从金融时间序列一般规律出发,分析质物市场收益率序列统计特征,以场外现货交易为主的螺纹钢日数据为例,模型化收益率序列自相关性和异方差性特性,建立尖峰厚尾分布下的AR(1)-GARCH(1,1)-GED模型;提出置于多风险窗口下度量未来质押期内钢材价格风险水平,给出同时考虑收益率自相关性和波动率时变性的长期风险VaR计算解析式,得出与银行风险承受能力相一致的质押率;基于失效率法则建立长期风险的碰撞序列函数,回测多风险窗口下长期VaR值。实证分析和回测显示,与现有其他模型相比,引进系数K值后的模型能显著提高银行风险覆盖率,且能显著降低银行效率损失,为银行提供一种动态质押率风险管理框架,模型确定的多风险窗口质押率与未来螺纹钢最低价值呈显著正相关。
[Abstract]:Different from the efficient financial market hypothesis of weak correlation of return rate, there is a significant autocorrelation in the return rate of qualitative goods market, which is dominated by spot trading. Based on the general rules of financial time series, the statistical characteristics of returns series in qualitative goods market are analyzed. Taking the daily data of rebar for over-the-counter spot transaction as an example, the autocorrelation and heteroscedasticity characteristics of yield series are modeled, and the ARP1GARCHN 1D model is established under the distribution of peak and thick tail. A multi-risk window is proposed to measure the risk level of steel price in the future pledge period, and a long-term risk VaR calculation formula considering both the autocorrelation of yield and volatility variability is given. The paper obtains the pledge rate consistent with the risk bearing ability of the bank, establishes the collision sequence function of the long-term risk based on the failure rate rule, and measures back the long-term VaR value under the multi-risk window. The empirical analysis and the re-test show that, compared with other existing models, The model after the introduction of K value can significantly improve the risk coverage of banks and reduce the loss of bank efficiency. It provides a dynamic risk management framework for banks. The multi-risk window pledge rate determined by the model has a significant positive correlation with the future minimum value of rebar.
【作者单位】：西南交通大学交通运输与物流学院;复旦大学金融研究院;同济大学经济管理学院;华夏银行成都地区信用风险管理部;
【基金】：国家自然科学基金(71003082) 全国博士后基金(20080430602) 教育部博士点基金(200806131007) 中央高校基本科研业务费专项资金科技创新项目(SWJTU11CX081) 四川科技计划软科学项目(2010ZR0028) 复旦大学经济学院985平台项目~~
【分类号】：F224;F830;F416.31

【参考文献】