统计研究 ›› 2008, Vol. 25 ›› Issue (11): 64-71.

• 论文 • 上一篇    下一篇

我国保险投资组合的模拟和金融风险测量研究

陈辉 陈建成   

  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-11-15 发布日期:2008-11-15

Modeling and Financial Risk Measure for the China’s Insurance Investment Portfolio

Chen Hui & Chen Jiancheng   

  • Received:1900-01-01 Revised:1900-01-01 Online:2008-11-15 Published:2008-11-15

摘要: 本文利用Copula函数的概念研究了保险投资组合多元金融数据的统计模拟。根据我国保险投资的特殊性,我们选用沪深300指数、基金指数、企债指数和国债指数四种风险资产来模拟保险投资组合中的股票、基金、企债和国债收益。基于模拟的结果分别利用传统近似方法(Add-VaR、N-VaR和H-VaR)和Copula方法计算了投资组合的总风险;相对于Copula-VaR方法,Add-VaR显著高估了风险,N-VaR显著低估了风险,H-VaR对于Copula-VaR的近似效果比较好,但其也高估了风险,即H-VaR相对于Copula-VaR是一种比较保守的方法。另外,我们分析了投资组合权重变化和Copula函数的选择对投资组合总风险的影响。

关键词: Copula函数, 在险价值, 保险投资组合, 蒙特卡罗模拟

Abstract: This paper concerns the statistical model of the dependence structure of multivariate financial data using the concept of copulas. According to the specific characteristics of China’s insurance investment, we use four stock market indices: HS300 Index, Fund Index, Corporate Bond Index and Treasury Bond Index in China to simulate the daily returns of stock, fund, corporate bond and Treasury bond respectively. Based on the simulation results, we use the traditional approximation methods and Copula method measured the risk of investment portfolio. According the risk measure results, we find that Add-VaR overestimates the total risk and N-VaR underestimates the total risk relative to the Copula-VaR; H-VaR provides a fairly close approximation, but it also overestimates the total risk and is a conservative method. In addition, we explore the impact of investment assets mix change and Copula selection on total risk. 

 

Key words: Copula, VaR, Insurance investment portfolio, Monte Carlo simulation