. 2022 Nov 15;79:103111. doi: 10.1016/j.resourpol.2022.103111

Table A.2.

Estimated parameters of the best-fitting bivariate copulas of the regular vines, whose margins are AR(1)-GJR-GARCH(1,1) models with a normal or Student’s $t$ innovation.

Panel A1: OIL, BTC, and USDT with Normal Margins
Tree	Edge	Copula	$ϑ$	LL	AIC	Kendall’s $τ$

Before COVID-19
$T_{1}$	OIL,USDT	Frank	−0.517	1.453	−0.907	−0.057
	USDT,BTC	180°-rotated Gumbel	1.074	7.183	−12.366	0.069
$T_{2}$	OIL,BTC $\|$ USDT	Frank	0.226	0.254	1.491	0.025
During COVID-19
$T_{1}$	OIL,BTC	Gumbel	1.081	3.436	−4.872	0.075
	BTC,USDT	270°-rotated Gumbel	1.015	11.692	−21.385	−0.015
$T_{2}$	OIL,USDT $\|$ BTC	180°-rotated Clayton	0.032	0.573	0.855	0.016

Panel A2: OIL, BTC, and USDT with Student’s t Margins
Tree	Edge	Copula	$ϑ$	LL	AIC	Kendall’s $τ$

Before COVID-19
$T_{1}$	OIL,USDT	Frank	−0.441	1.446	−0.892	−0.049
	USDT,BTC	180°-rotated Gumbel	1.075	5.988	−9.977	0.070
$T_{2}$	OIL,BTC $\|$ USDT	Clayton	0.037	0.460	1.081	0.018
During COVID-19
$T_{1}$	OIL,BTC	Clayton	0.114	5.043	−8.085	0.054
	BTC,USDT	270°-rotated Gumbel	1.021	4.578	−7.155	−0.021
$T_{2}$	OIL,USDT $\|$ BTC	180°-rotated Clayton	0.037	0.475	1.050	0.018

Panel B1: OIL, BTC, and USD with Normal Margins
Tree	Edge	Copula	$ϑ$	LL	AIC	Kendall’s $τ$

Before COVID-19
$T_{1}$	OIL,USD	Frank	−0.087	0.049	1.901	−0.010
	USD,BTC	Frank	−0.393	0.869	0.261	−0.044
$T_{2}$	OIL,BTC $\|$ USD	Clayton	0.029	0.312	1.377	0.014
During COVID-19
$T_{1}$	OIL,USD	Frank	−0.994	5.105	−8.209	−0.109
	USD,BTC	Frank	−0.956	4.768	−7.535	−0.105
$T_{2}$	OIL,BTC $\|$ USD	180°-rotated Gumbel	1.016	3.301	−4.602	0.016

Panel B2: OIL, BTC, and USD with Student’s t Margins
Tree	Edge	Copula	$ϑ$	LL	AIC	Kendall’s $τ$

Before COVID-19
$T_{1}$	OIL,USD	Frank	−0.110	0.088	1.825	−0.012
	USD,BTC	Frank	−0.431	1.410	−0.820	−0.048
$T_{2}$	OIL,BTC $\|$ USD	Clayton	0.040	0.532	0.936	0.019
During COVID-19
$T_{1}$	OIL,USD	90°-rotated Clayton	0.178	7.125	−12.249	−0.082
	USD,BTC	Frank	−0.938	5.912	−9.824	−0.103
$T_{2}$	OIL,BTC $\|$ USD	Clayton	0.106	3.712	−5.424	0.050

Note: LL and AIC refer to the maximized log-likelihood and the corresponding Akaike information criterion, respectively.