基于贝叶斯CAViaR模型的油价风险研究  

Analysis of oil price value at risk using Bayesian CAViaR model

作　　者：陈磊[1] 杜化宇 曾勇[1] 

机构地区：[1]电子科技大学经济与管理学院,成都610054 [2]闽江学院新华都商学院,福州350108

出　　处：《系统工程理论与实践》2013年第11期2757-2765,共9页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(71301019;71202074);教育部人文社会科学基金(08JA790012)

摘　　要：CAViaR模型是常用的VaR估计方法之一,但通常面临参数估计和模型检验的困难.本文发展了贝叶斯CAViaR模型用于分析油价风险,并考察该模型在参数估计、模型选择、VaR预测等方面的作用.采用布伦特原油价格日数据,研究显示贝叶斯CAViaR模型有效控制了估计风险和模型风险,且具有较好的VaR预测绩效,优于传统CAViaR模型.本文同时指出,油价VaR存在自回归特征并受前期正负收益率的不对称影响.不对称斜率CAViaR模型有效刻画了油价VaR的动态变化模式.CAViaR model is usually used to estimate value at risk （VaR）. However, it is difficult to estimate parameters and check model specification for CAViaR model. This paper develops Bayesian CAViaR model, adopts this model to estimate oil price VaR, and analyzes the roles of Bayesian CAViaR model in parameter estimation, model selection and VaR forecast. Using daily data of Brent crude oil price, the results show Bayesian CAViaR model can control estimation risk and model risk effectively, and has the better forecast performance than traditional CAViaR model. This paper also indicates oil price VaR has autoregressive effects and is affected by prior returns. The positive and negative returns have asymmetry effects on VaR. Asymmetric slope CAViaR model is the best model to describe the dynamics of oil price VaR.

关 键 词：贝叶斯CAViaR模型 油价风险 风险值 贝叶斯方法 MCMC 

分 类 号：F830.9[经济管理—金融学]