基于多元Laplace分布的外汇期权组合非线性VaR模型

发布时间：2018-03-12 14:26

本文选题：外汇期权　切入点：多元Laplace分布　出处：《系统工程理论与实践》2010年02期 　论文类型：期刊论文

【摘要】：为了克服多元厚尾分布情形下的非线性VaR数值计算的困难,用多元Laplace分布来描述汇率回报分布厚尾性,引入风险函数转换技术和关于多维Laplace多重积分近似计算的结果,来解决多元Laplace分布情形下的反映外汇期权组合价值变化的矩母函数问题;进一步将重要抽样技术发展到多元Laplace分布情形下的外汇期权组合非线性VaR模型中,使得该情形下不再是稀有事件Monte Carlo模拟,从而减少Monte Carlo模拟计算工作量,更精确地估计出组合的损失概率.数值结果表明该算法比常用Monte Carlo模拟法的计算效率更有效,且能很大程度上减少所要估计的损失概率的方差.
[Abstract]:In order to overcome the difficulty of nonlinear VaR numerical calculation in the case of multivariate thick tail distribution, the multivariate Laplace distribution is used to describe the thick tail of the exchange rate return distribution. The risk function conversion technique and the results of the multi-dimensional Laplace multiple integral approximation calculation are introduced. In order to solve the moment generating function problem which reflects the change of portfolio value of foreign exchange options in the case of multivariate Laplace distribution, the importance sampling technique is further developed into the nonlinear VaR model of foreign exchange option portfolio under the condition of multivariate Laplace distribution. In this case, the Monte Carlo simulation is no longer a rare event, thus reducing the computational workload of the Monte Carlo simulation and estimating the loss probability of the combination more accurately. The numerical results show that the algorithm is more efficient than the common Monte Carlo simulation method. And the variance of the probability of loss can be reduced to a great extent.
【作者单位】：浙江大学管理学院;浙江财经学院金融学院;
【基金】：国家自然科学基金(70771099) 中国博士后科学基金(20070421167)
【分类号】：F830;F224

【参考文献】