中国股市长记忆性与趋势变化研究——基于SEMIFAR-FIGARCH模型  

Study on the Lasting Memory of Chinese Stock Market and Its Tendency of Change Based on SEMIFAR-FIGARCH Model

作　　者：张金凤[1,2] 马薇[1] 

机构地区：[1]天津财经大学理工学院,天津300222 [2]天津工业大学理学院,天津300387

出　　处：《求是学刊》2015年第6期55-61,共7页Seeking Truth

基　　金：全国统计科学研究项目"大数据条件下金融风险测度的方法研究";项目编号:2014LY003

摘　　要：文章对中国股市的长记忆性进行研究,在研究中将SEMIFAR模型与FIGARCH模型相结合,建立了既能反映收益率趋势变化情况又能描述收益率和波动长记忆特征的SEMIFAR-FIGARCH模型,利用该模型对我国沪、深两市的收益率和波动率的长记忆性及趋势变化进行实证分析,并与ARFIMA-FIGARCH、ARFIMA-HYGARCH模型结果比较拟合及预测效果。研究结果表明:我国沪、深两市的收益率和波动率均存在长记忆性;其收益率序列存在显著的趋势变化特征;SEMIFAR-FIGARCH模型的拟合和预测效果优于ARFIMA-FIGARCH、ARFIMA-HYGARCH模型,表明SEMIFAR-FIGARCH模型对我国股市有较好的模型解释能力和预测能力。This article studies on the lasting Memory of Chinese stock market, combining SEMIFAR model with FIGARCH model. It established the SEMIFAR-FIGARCH model which reflects both the change of rate of return and describes this rate and the characteristics of the fluctuating lasting Memory. This model is applied into the empirical research of the rate of return and fluctuating rate in the stock market of Shanghai and Shenzhen and compared with ARFIMA- FIGARCH model and ARFIMA-HYGARCH model, showing the predicted effect. Study shows that there is lasting Memory in the rate of return and fluctuating rate in the stock market of Shanghai and Shenzhen; there is obvious trend of change in the quality of its rate of return order; SEMIFAR-FIGARCH model is better than ARFIMA-FIGARCH model or ARFIMA-HYGARCH model, showing SEMIFAR-FIGARCH model is good at interpreting and predicting our stock market.

关 键 词：SEMIFAR-FIGARCH模型 趋势 长记忆 核估计方法 

分 类 号：F832.51[经济管理—金融学]