基于CARR模型与GARCH模型对VaR的比较研究

Comparative study of VaR based on CARR and GARCH model

研究了在一般情形下和极端风险下的风险度量,分别采用基于极差、收益率为变量建模的CARR模型、GARCH模型应用于VaR的计算,结合深证成指的实际数据进行实证分析,分别对比在不同分布下GARCH模型和CARR模型计算出的VaR,最终得出基于在广义伽马分布下CARR模型算出的VaR值,能更加真实地反映深证股市极端情形下风险程度,而基于T分布下的GARCH模型更加真实地反映深证股市一般情形下的风险程度.

The paper studies the risk measurement in general and extreme cases, calculates the VaRs under the CARR Model and GARCH Model which are respectively based on the variables of range and rate on return, conducts the empirical analysis on the actual data of SZSE Component Index, compares the VaRs calculated by GARCH Model and CARR Model in different distri- butions, and finally obtains VaR value calculated by CARR Model in Generalized Gama Distribution, and the value can more truly reflect the risk degree of the Shenzhen component stock market in extreme cases, and the VaR value calculated by GARCH Model in the T-distribution can more truly reflect the risk degree of the Shenzhen component stock market in general cases.