VI. Anadolu International Conference on Economics, Eskişehir, Turkey, 13 - 15 May 2022, pp.179
The utilization of multivariate copulas to model the dependence is mainly limited to elliptical and
Archimedean copulas. By nature, both classes are quite restrictive concerning the symmetry and tail
dependencies. Moreover, they are not flexible and not easy to estimate with high-dimensional data.
The vine copulas overcome the restrictive features of the bivariate copulas.. In this context, we apply
Vine-GARCH models in a moving window to examine the time-varying co-dependencies of 12 bank
stocks listed on the Istanbul Stock Exchange for the 2011-2021 period. One of the attractions of the
proposed approach is its potential to construct a rich set of distributions for modeling dependency.
The study aims to show the practical application of Vine copulas on the dependence risk dynamics of
the Turkish banking sector. We estimate Vine copula based value-at-risk (VaR) and expected shortfall
(ES) risk measures for an equally weighted portfolio of considered bank stocks. The obtained results
reveal that the Vine-GARCH outperforms the GARCH models for forecasting risk measures in the
Turkish banking system.