应用门限分位点回归模型估计条件VaR  

Evaluation of conditional VaR based on threshold quantile regression model

作　　者：叶五一[1] 缪柏其[1] 

机构地区：[1]中国科学技术大学统计与金融系,安徽合肥230026

出　　处：《系统工程学报》2008年第2期154-160,共7页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(10471135);教育部博士点基金资助项目(20010358022)

摘　　要：在文献中,分位点回归模型是线性的,但是在实际中,这个假设不能很好地满足需要.为此提出了分位点回归的门限模型,用该模型实证分析了单只股票(浦东发展银行)的条件 VaR.选择了一种流动性风险指标作为条件,因此该条件 VaR 也可以看作是流动性调整的 VaR(La-VaR).经过实证分析发现,由门限分位点模型得到的结果能够更好地描述实际市场情况,也能更好地预测市场风险.In most articles, quantile regression model is linear, but in practice, this assumption can not suit the practical demand very well. So in this paper, a new threshold quantile regression model is presented, and an empirical analysis on Pudong Devlopment Bank stock based on this model is given. In this paper, a liquidity risk measure is picked as the condition variable, so the conditional VaR is also a liquidity-adjusted VaR. By the empirical analysis, the result obtained based on threshold quantile regression model can describe the practical market better, and forecast the market risk better.

关 键 词：分位点回归模型 门限分位点回归模型 条件VAR 事后检验 

分 类 号：F830.9[经济管理—金融学]