Table 5 Estimates of bivariate...

Table 5

Estimates of bivariate VAR-BEKK-GARCH for gold assets and Latin American equity markets during Chinese equity market crash

	BRAZ	GOLD	CHIL	GOLD	MEXI	GOLD	PERU	GOLD
Panel A: Mean equation
Constant	0.000 [0.367]	0.000 [0.863]	0.000 [0.415]	0.000 [0.886]	0.000 [0.879]	0.000 [0.724]	0.000** [0.039]	0.000 [0.794]
$r_{t - 1}^{S}$	0.010 [0.781]	0.011 [0.597]	0.201*** [0.000]	0.050 [0.204]	0.072** [0.050]	0.038 [0.228]	0.108*** [0.001]	0.099 [0.153]
$r_{t - 1}^{G}$	−0.020 [0.752]	−0.008 [0.845]	0.005 [0.828]	−0.007 [0.822]	0.057* [0.077]	−0.009 [0.831]	0.035 [0.318]	−0.011 [0.776]
Panel B: Variance equation
Constant	0.003 [0.133]	0.000 [0.987]	0.002* [0.053]	0.000 [0.991]	0.003*** [0.000]	0.000 [0.988]	0.001*** [0.000]	0.001*** [0.001]
${(e_{t - 1}^{S})}^{2}$	0.217** [0.019]	0.012 [0.497]	0.493*** [0.002]	0.048 [0.279]	0.335*** [0.000]	0.069 [0.135]	0.144*** [0.000]	0.091** [0.046]
${(e_{t - 1}^{G})}^{2}$	−0.501*** [0.002]	0.066* [0.088]	−0.068** [0.042]	0.114*** [0.000]	−0.069* [0.095]	0.123*** [0.000]	−0.241*** [0.000]	0.100*** [0.001]
$h_{t - 1}^{S}$	0.886*** [0.000]	0.020* [0.070]	0.818*** [0.000]	−0.016 [0.604]	0.788*** [0.000]	−0.035 [0.232]	0.893*** [0.000]	−0.027 [0.224]
$h_{t - 1}^{G}$	−0.017 [0.572]	0.992*** [0.000]	0.004 [0.551]	0.990*** [0.000]	−0.019 [0.493]	0.984*** [0.000]	0.056** [0.042]	0.989*** [0.000]
Panel D: Diagnostic tests
LogL	4895.23		5425.02		5381.54		5309.86
AIC	−12.185		−13.492		−13.454		−13.127
SIC	−11.698		−13.005		−12.967		−12.640
Q[12]	25.172 [0.194]	21.209 [0.384]	22.989 [0.289]	21.160 [0.387]	26.142 [0.161]	21.347 [0.376]	29.104* [0.085]	23.373 [0.270]
Q²[12]	17.248 [0.636]	17.283 [0.634]	30.302 [0.122]	16.875 [0.661]	16.206 [0.703]	14.321 [0.813]	14.107 [0.825]	12.138 [0.911]

	BRAZ	GOLD	CHIL	GOLD	MEXI	GOLD	PERU	GOLD
Panel A: Mean equation
Constant	0.000 [0.367]	0.000 [0.863]	0.000 [0.415]	0.000 [0.886]	0.000 [0.879]	0.000 [0.724]	0.000** [0.039]	0.000 [0.794]
$r_{t - 1}^{S}$	0.010 [0.781]	0.011 [0.597]	0.201*** [0.000]	0.050 [0.204]	0.072** [0.050]	0.038 [0.228]	0.108*** [0.001]	0.099 [0.153]
$r_{t - 1}^{G}$	−0.020 [0.752]	−0.008 [0.845]	0.005 [0.828]	−0.007 [0.822]	0.057* [0.077]	−0.009 [0.831]	0.035 [0.318]	−0.011 [0.776]
Panel B: Variance equation
Constant	0.003 [0.133]	0.000 [0.987]	0.002* [0.053]	0.000 [0.991]	0.003*** [0.000]	0.000 [0.988]	0.001*** [0.000]	0.001*** [0.001]
${(e_{t - 1}^{S})}^{2}$	0.217** [0.019]	0.012 [0.497]	0.493*** [0.002]	0.048 [0.279]	0.335*** [0.000]	0.069 [0.135]	0.144*** [0.000]	0.091** [0.046]
${(e_{t - 1}^{G})}^{2}$	−0.501*** [0.002]	0.066* [0.088]	−0.068** [0.042]	0.114*** [0.000]	−0.069* [0.095]	0.123*** [0.000]	−0.241*** [0.000]	0.100*** [0.001]
$h_{t - 1}^{S}$	0.886*** [0.000]	0.020* [0.070]	0.818*** [0.000]	−0.016 [0.604]	0.788*** [0.000]	−0.035 [0.232]	0.893*** [0.000]	−0.027 [0.224]
$h_{t - 1}^{G}$	−0.017 [0.572]	0.992*** [0.000]	0.004 [0.551]	0.990*** [0.000]	−0.019 [0.493]	0.984*** [0.000]	0.056** [0.042]	0.989*** [0.000]
Panel D: Diagnostic tests
LogL	4895.23		5425.02		5381.54		5309.86
AIC	−12.185		−13.492		−13.454		−13.127
SIC	−11.698		−13.005		−12.967		−12.640
Q[12]	25.172 [0.194]	21.209 [0.384]	22.989 [0.289]	21.160 [0.387]	26.142 [0.161]	21.347 [0.376]	29.104* [0.085]	23.373 [0.270]
Q²[12]	17.248 [0.636]	17.283 [0.634]	30.302 [0.122]	16.875 [0.661]	16.206 [0.703]	14.321 [0.813]	14.107 [0.825]	12.138 [0.911]

Note(s): “The number of lags for VAR is decided using AIC and SIC criteria. The Ljung–Box Q statistics of order 12 for autocorrelation are represented by Q(12) and Q²(12) and are applied to the standardized and squared standardized residuals respectively. BRAZ, Brazil; CHIL, Chile; MEXI, Mexico. For brevity, one constant coefficient of variance equation is not reported. p-values are shown in the parentheses. Statistical significance at 1, 5 and 10% is indicated by ^*, ^**, ^*** respectively”

Source(s): Own elaboration

[ViewLarge]

Sharing Unavailable