基于ARMA-GARCH模型的上证指数收益率分析  

Applying ARMA-GARCH Approaches to Analysis the Return Series of Shanghai Stock Exchange Composite Index

作　　者：彭紫君 Peng Zijun(School of Statistics and Mathematics,Zhongnan University of Economics and Law,Wuhan 430073,China)

机构地区：[1]中南财经政法大学统计与数学学院

出　　处：《中南财经政法大学研究生学报》2017年第4期48-55,共8页Journal of the Postgraduate of Zhongnan University of Economics and Law

摘　　要：本文综合ARMA模型在平稳时间序列数据预测方面的优势和GARCH模型在金融数据易变性方面的优势,基于不同的新息分布分析收益率序列的不同特征,并利用样本内统计量和基于循环迭代预测的回测检验方法比较不同模型的预测效果。实证表明上证指数日对数收益率具有显著的异方差性、风险溢价和杠杆效应,因此考虑异方差性的GARCH类模型比基于高斯分布的ARMA模型预测效果有显著提高,基于有偏学生分布的模型具有最优的预测效果。In this paper,the advantages of the ARMA model in the prediction of stationary time series data and the advantages of GARCH model in financial data volatility are discussed. The different characteristics of the yield sequence are analyzed based on different new distributions,and the prediction results of different models are compared by using the statistical quantity in the sample and the back-test method based on the cyclic iterative prediction. The results show that（ 1） the shanghai index daily logarithmic return has significant heteroscedasticity,risk premium and leverage effects. So the models of GARCH considering the heteroscedasticity has the better fitting effects than the models ARMA based on the Gaussian distribution.（ 2） the model that based on the partial student distribution has the best fitting and prediction effects.

关 键 词：异方差性 风险溢价 杠杆效应 GARCH类模型 

分 类 号：F830.9[经济管理—金融学]