一元GARCH模型估计的渐近理论:平稳与非平稳情形

摘要

Models of (Generalized) Autoregressive Conditional Heteroskedasticity(ARCH/GARCH) form the most popular way of parameterizing volatility.This paper contains a survey of estimation methods of univariate GARCH models with a special attention given to the asymptotic results of the quasi-maximum likelihood method and the least absolute deviation method.The estimation for non-stationary GARCH model is also discussed.%ARCH/GARCH模型是刻画波动率最常用的模型.本文综述一元GARCH模型的估计方法,主要讨论准最大似然估计和最小绝对偏差估计方法的渐近性质.此外,本文还讨论了非平稳GARCH模型的估计问题.