Financial Mathematics, Volatility and Covariance Modelling

9Models with multiplicative decomposition of conditional variances and correlations

Cristina Amado, Annastiina Silvennoinen, and Timo Teräsvirta

1 Introduction

Many daily or weekly volatility series appear nonstationary. In the generalized autoregressive conditional heteroskedasticity (GARCH) framework this nonstationarity has been explicitly modeled by integrated GARCH models (Engle and Bollerslev, 1986) or using a more general version, the Fractionally Integrated GARCH model (Baillie, Bollerslev, and Mikkelsen, 1996). Another strand of literature, see for example Lamoureux and Lastrapes (1990) or Mikosch and Stărică (2004), builds on the assumption that nonstationarity is due to structural changes in the volatility process. One way of adjusting ...