基于CVaR与平均离差波动性限制的投资组合模型

发布时间：2018-04-05 03:16

本文选题：投资组合　切入点：VaR模型　出处：《渤海大学》2014年硕士论文

【摘要】：投资组合问题是当前金融理论的研究热点之一，主要研究在期望收益率确定的前提下寻找资产投资分配方案，或在确定风险的情况下使收益最大化。投资组合的先期研究主要集中在投资组合的收益度量，随后过渡到投资组合的风险度量。首先介绍了投资组合理论，分析了现有的M V、MAD、VaR和CvaR等投资组合模型的性质及优缺点，指出了在投资过程中对多阶段投资组合模型的需求。针对单期组合投资，提出了一种基于CVaR的投资组合模型。对组合资产收益率不作正态分布假设，用CVaR度量投资组合风险函数，用MAD模型作为一个约束条件，实现波动性度量限制。用上凸效用函数作为一个约束条件，表示风险资产交易费用。实验结果表明，该模型满足实际投资要求，符合实际投资规律，与M V模型和原始CVaR模型相比，具有波动性和风险价值最小化的优势。针对多阶段组合投资，提出了一种基于CVaR的投资组合模型。用离散型动态规划把各阶段投资过程整合为一个整体，，对组合资产收益率做正态分布假设，用拉格朗日乘子法将MAD模型写入目标函数，实现波动性度量限制。实验结果表明，该模型较好地完成了多阶段组合投资的资金分配和投资时机，是一种较实用的模型。
[Abstract]:Portfolio problem is one of the hotspots in the current financial theory. It is mainly concerned with finding the asset investment allocation scheme under the premise of determining the expected rate of return or maximizing the income under the condition of determining the risk.The study of portfolio is mainly focused on the measurement of portfolio return and then the risk measurement of portfolio.Firstly, this paper introduces the portfolio theory, analyzes the properties, advantages and disadvantages of the existing portfolio models such as CvaR and MVMADV, and points out the requirement of the multi-stage portfolio model in the process of investment.A portfolio model based on CVaR is proposed for single period portfolio investment.The return rate of portfolio is not assumed to be normal distribution. The CVaR is used to measure portfolio risk function and the MAD model is used as a constraint to realize volatility measurement.The convex utility function is used as a constraint to represent the transaction cost of risky assets.The experimental results show that the model meets the requirements of actual investment and conforms to the law of actual investment. Compared with the original CVaR model and M-V model, the model has the advantages of volatility and minimization of risk value.A portfolio model based on CVaR is proposed for multi-stage portfolio investment.The investment process of each stage is integrated into a whole by discrete dynamic programming. The normal distribution assumption of portfolio return is made, and the MAD model is written into the objective function by Lagrange multiplier method to realize the constraint of volatility measurement.The experimental results show that the model is a more practical model, which completes the allocation and timing of multi-stage portfolio investment.
【学位授予单位】：渤海大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F830.9;O221.3

【参考文献】