文章基本信息

标题：PREDICTIVE CAPACITY OF ARCH FAMILY MODELS
本地全文：下载
作者：Raphael Silveira Amaro ; Paulo Sergio Ceretta ; Kelmara Mendes Vieira 等
期刊名称：Revista de Gestão, Finanças e Contabilidade
印刷版ISSN：2238-5320
出版年度：2016
期号：6893
页码：06-27
语种：
出版社：University of State of Bahia
摘要：In the last decades, a remarkable number of models, variants from the Autoregressive Conditional Heteroscedastic family, have been developed and empirically tested, making extremely complex the process of choosing a particular model. This research aim to compare the predictive capacity, using the Model Confidence Set procedure, than five conditional heteroskedasticity models, considering eight different statistical probability distributions. The financial series which were used refers to the log-return series of the Bovespa index and the Dow Jones Industrial Index in the period between 27 October 2008 and 30 December 2014. The empirical evidences showed that, in general, competing models have a great homogeneity to make predictions, either for a stock market of a developed country or for a stock market of a developing country. An equivalent result can be inferred for the statistical probability distributions that were used.