分整增广GARCH-M模型

Fractional integrated augmented GARCH-M model

ARCH模型是研究异方差和自相关问题的一类重要模型族.长记忆性则是许多时间序列中存在的性质.文章将长记忆性扩展到几乎所有现有的ARCH模型,提出了数十种新模型;并用一个模型概括了迄今几乎所有的ARCH模型,这就是分整增广GARCH＿M模型;从而解决了各种ARCH模型在模型设定检验、长记忆性诊断和参数估计等方面的障碍.最后通过仿真实验和实证研究说明分整增广GARCH＿M模型在实际经济分析中的必要性.

ARCH model proposed by Engle has gained increasing promi nence as the l eading model for describing stochastic volatility and autoregression. Long memor y property exists in many time series. The paper proposes fractional integrated augmented GARCH_M model which can embrace almost all the current ARCH models a n d the dozens of new models. The handicap in model specification test, long memor y test and estimation is overcome. The necessity of FI_Augmented_GARCH_M model i n economic analysis is assessed through a simulation study and empirical study.