Inhomogeneous Jump-GARCH models with applications in financial time series analysis

Abstract

This paper discusses the statistical behaviors and applicability of the jump-GARCH model proposed by Chen and Sato (2007), in which jump arrivals are time inhomogeneous and also state dependent. We discuss maximum likelihood estimation and likelihood ratio tests for the model. We investigate the statistical behaviors of the jump-GARCH model through financial time series analysis and showing comparisons of this model with GARCH and traditional jump models. Our results indicate that this model can reveal many important characteristics related with jump dynamics and volatility structures in asset prices.