CMStatistics 2021: Start Registration
View Submission - CFE
Title: Modelling non-stationarity robust variance interactions Authors:  Cristina Amado - University of Minho (Portugal) [presenting]
Abstract: A multivariate generalisation of the multiplicative decomposition of the volatility is proposed within the class of conditional correlation GARCH models. The GARCH variance equations are multiplicatively decomposed into a deterministic non-stationary component describing the long-run movements in volatility and a short run dynamic component allowing for spillover effects between assets. The conditional correlations are assumed to be time-invariant in their simplest form or generalised into a flexible dynamic parametrisation. Parameters of the model are estimated equation-by-equation applying the maximisation by parts algorithm in the variance equations in the first step, and the correlation parameters estimated in the second step. An empirical application between four major spot exchange rates against the euro illustrates the usefulness of the model. Our results suggest that neglecting non-stationarity in the form of structural changes in the unconditional variance leads to spurious spillover effects. Furthermore, after modelling the variance equations accordingly, we also find evidence that some transmission mechanism of shocks persists which is supported by the presence of variance interactions robust to non-stationarity.