A comparison of tail dependence estimators

Pandemic-driven financial contagion and investor behavior: Evidence from the COVID-19

“…There is a large amount of and still growing body of literature on the copula function due to its flexibility in describing various patterns of dependence structure such as non-linearly, asymmetry, dynamic, and tail dependence ( Abakah et al, 2021 , Chabi-Yo et al, 2018 , Christoffersen et al, 2012 , Hüttner et al, 2020 , Sahamkhadam et al, 2022 , Supper et al, 2020 , Wang and Dyer, 2012 , Wang, Yuan, Wang, 2021 ). A copula captures the dependence structure of a multivariate distribution and is defined as a multivariate distribution function with standard uniform margins.…”

“…GARCH-type models have commonly been used to estimate the marginal distribution of the copula models ( Jayech, 2016 , Koliai, 2016 , Supper et al, 2020 ). However, they cannot adequately approximate the tail behavior of the marginal distribution, yet the tail behavior is essential in measuring financial contagin.…”

Financial contagion and contagion channels in the forex market: A new approach via the dynamic mixture copula-extreme value theory

“…As summarized in Manner and Reznikova (2012) , the choice of copulas with different time-varying modeling is a matter of taste and computational capability of the software. As a compromise of estimation precision and computation cost, we use Patton’s model ( Patton, 2006 ) as the time-varying modeling for the following reasons: (1) It has been widely used to describe the dynamic dependence in financial markets (e.g., Dias and Embrechts, 2012 ; Hoesli and Reka, 2015 ; Luo et al, 2015 ; Fenech and Vosgha, 2019 ; Ji et al, 2019 ; Supper et al, 2020 ). (2) It is more flexible to fit data compared to dynamic copulas that have some restrictions in the dependence structure.…”

Nonparametric estimator of the tail dependence coefficient: balancing bias and variance

“…Among them, we can cite the plateau-finding algorithm (Frahm et al, 2005;Schmidt and Stadtmüller, 2006), or graphical methods (Caillault and Guégan, 2005). Most of the contributions in the field are devoted to the comparison of various methods of TDC estimation in simulation frameworks (Frahm et al, 2005;Schmidt and Stadtmüller, 2006;Poulin et al, 2007;Supper et al, 2020). Yet, there is no theoretical contribution in which the selection rule of the threshold is related to a simple trade-off between the bias and the variance of the estimator.…”