A0418
Title: Semi-continuous time series with volatility clustering
Authors: Sarka Hudecova - Charles University, Prague (Czech Republic) [presenting]
Abstract: Time series that contain a non-negligible portion of possibly dependent zeros, whereas the remaining observations are positive, are considered. They are treated as GARCH processes consisting of non-negative values. Such models find application in various fields and are, to some extent, related to multiplicative error models (MEM). The aim lies in the estimation of the omnibus model parameters while taking into account the semi-continuous distribution. The hurdle distribution, together with dependent zeros, causes the classical GARCH estimation techniques to fail, so two different estimation approaches are proposed. The resulting two quasi-likelihood estimators are shown to be strongly consistent and asymptotically normal. The empirical properties are illustrated in a Monte Carlo simulation study, demonstrating the computational efficiency of the methods employed. The developed techniques are presented through an actuarial problem concerning sparse insurance claims.
