股票收益率预测及风险与收益关系研究——基于ARMA-GARCH模型和高频数据

Prediction of Stock Return Rate and Relationship Between Risk and Return——Based on ARMA-GARCH Model and High-frequency Data

以沪深300指数的一分钟为间隔的实时价格为研究样本,利用ARMA模型和基于T分布的GARCH(1,1)模型,对其收益率进行了拟合和预测,同时运用GARCH-M模型,分析风险和收益之间的关系。研究表明,股指波动存在条件异方差性;ARMA模型长期预测效果较好,而GARCH(1,1)-T模型短期预测效果较好;沪深300指数的风险和收益不呈正比,说明我国股市发展不成熟。

By using the high-frequency data of the CSI 300 index, ARMA model and GARCH（1,1）-T mod- el are applied to fit and forecast the rate of return. In addition, GARCH-M model is selected to measure the relationship between risk and return. Experimental results show that the change of return rate has the feature with conditional heteroskedasticity~ ARMA model is suitable for the long term prediction and GARCH（1,1）-T model is suitable for the short term prediction; As risk and return don＇t present positive relationship, our stock market isfft ripe.