基于Copula-分位数回归的供应链金融多期贷款组合优化

发布时间：2018-02-24 22:32

本文关键词： 供应链金融多期贷款组合 Copula-分位数回归　出处：《中国管理科学》2017年06期 　论文类型：期刊论文

【摘要】：为优化供应链金融多期贷款组合方案,考虑到供应链金融中呈现出的非对称与非线性等典型特征,以分位数回归拟合单个资产边缘分布、以Copula函数刻画资产间非线性关联关系,建立Copula-分位数回归方法。使用该方法,对供应链金融多期贷款收益进行预测,进而通过优化传统Sharpe比率、广义Omega比率等进行贷款组合选择,给出贷款组合优化方案。选取供应链金融中最常见的质押物:现货铝和铜作为研究对象,实证研究发现:第一,依据AIC准则,在Copula-分位数回归方法中,各贷款期限下的t-Copula函数拟合效果均为最优,表明铝和铜之间具有显著的厚尾相关性;第二,在各贷款期限下,Copula-分位数回归方法均优于Copula-GARCH方法,具体表现在前者拥有更高的Sharpe比率和广义Omega比率,能够获得更好的多期贷款组合效果。
[Abstract]:In order to optimize the multi-period loan portfolio scheme of supply chain finance, taking into account the typical characteristics such as asymmetry and nonlinearity in supply chain finance, the edge distribution of single asset is fitted by quantile regression, and the nonlinear correlation relationship between assets is described by Copula function. Copula- quartile regression method is established. The method is used to predict the multi-period loan income of supply chain finance, and then to select the loan portfolio by optimizing the traditional Sharpe ratio and the generalized Omega ratio, etc. The most common pledge materials in supply chain finance: spot aluminum and copper are selected as the research objects. The empirical results are as follows: first, according to the AIC criterion, in Copula- quantile regression method, The fitting effect of t-Copula function for each loan term is optimal, which indicates that there is a significant thick tail correlation between aluminum and copper. Secondly, the Copula- quantile regression method is superior to Copula-GARCH method in each loan term. The former has a higher Sharpe ratio and a broad Omega ratio, which can obtain better multi-term loan portfolio effect.
【作者单位】：合肥工业大学管理学院;合肥工业大学过程优化与智能决策教育部重点实验室;
【基金】：国家社会科学基金一般项目(15BJY008) 教育部人文社会科学研究规划基金项目(14YJA790015) 国家自然科学基金项目(71671056,71490725)
【分类号】：F832.4;O212.1

【相似文献】