CMStatistics 2017: Start Registration
View Submission - CFE
Title: Volatility transmission in multiply overlapping trading zones Authors:  Andreas Masuhr - University of Munster (Germany) [presenting]
Abstract: Previous volatility spillover models use artificially non overlapping trading zones to identify sources of volatility transmission between these zones. The problem of non overlapping zones is overcome using a copula GARCH approach that allows for multiple overlaps between zones incorporating vine copulas to flexibly model the dependence structure and to meet stylized facts of return data. To handle the relatively large parameter space, the model is estimated by Bayesian methods using a differential evolution MCMC approach. Results are presented for an exchange rate data set containing three trading zones with a single overlap, as well as for simulated data for a multiple overlap setting. Finally, the hypothesis that volatility transmits from preceding trading days as well as from preceding trading zones is supported.