Research Article

On the Safe-Haven Ability of Bitcoin, Gold, and Commodities for International Stock Markets: Evidence from Spillover Index Analysis

Table 1

Descriptive statistics of all return series.

BitcoinGoldCommoditiesWorldDevelopedEmerging

Entire phase: July 20, 2010–June 23, 2021
Observations270427042704270427042704
Mean0.00470.0001−0.00020.00030.00030.0001
Standard deviation0.07680.00990.01320.00930.00960.0102
Minimum−0.8488−0.0970−0.1252−0.0999−0.1044−0.0694
Maximum1.47420.04930.07620.08060.08410.0557
Skewness2.5801−0.6755−0.7964−1.1364−1.1081−0.5618
Excess kurtosis68.73319.610811.957019.206219.90867.6602
Jarque–Bera489814.35129.49324.830172.932764.92589.1
Q(5)22.39303.540716.509040.133037.904084.6500
Q(10)29.50705.009020.7670115.4700128.250093.6780
Q(20)45.742014.840033.8330152.7000172.0400101.1500
ADF−53.6237−51.2561−53.9454−16.9024−16.9627−31.9317
P-P−53.6050−51.2512−54.0550−50.0002−51.9171−44.5217

Phase I: July 20, 2010–March 13, 2012
Observations379379379379379379
Mean (%)0.01050.0009160.00070.0004620.0004840.000319
Standard deviation0.12630.0123170.0139330.0119930.012340.012693
Minimum−0.4308−0.05679−0.06744−0.05223−0.05256−0.06524
Maximum0.69310.0373090.046610.0430980.0464330.048096
Skewness0.7216−0.56693−0.61513−0.42635−0.37636−0.56629
Excess kurtosis8.98855.4634825.3458735.573095.5534985.407944
Jarque–Bera599.2170116.1377110.8051116.0354111.9143111.8194
Q(5)3.24230.66463.072113.887013.11525.029
Q(10)8.302610.159013.523018.317017.44827.240
Q(20)21.411017.820019.198027.930027.19833.366
ADF−20.5191−18.9369−19.9059−17.7343−18.5337−15.32656
P-P−20.4961−18.9299−19.9029−17.6561−18.5378−15.25227

Phase II: March 3, 2012–January 13, 2020
Observations196119611961196119611961
Mean (%)0.00374−3.87E − 05−0.000380.0002760.0003073.58E − 05
Standard deviation0.0672840.0090580.0116190.0069210.0070410.008796
Minimum−0.84883−0.09704−0.06586−0.04881−0.05029−0.05128
Maximum1.474180.0458970.0761660.0298640.0293830.043654
Skewness4.461687−0.695590.006042−0.61131−0.62609−0.21999
Excess kurtosis137.81112.179036.0543686.6244746.6691655.055585
Jarque–Bera14914747042.443762.28231195.5231228.135361.0696
Q(5)50.5981.877221.81779.17659.23288.070
Q(10)65.7678.920026.11885.49364.14599.324
Q(20)74.94522.66632.858101.8282.003107.44
ADF−22.07981−44.52072−47.35964−36.87325−38.06258−36.33821
P-P−44.39223−44.52230−47.25237−36.39581−37.67187−36.32432

Phase III: January 14, 2020–June 23, 2021
Observations364364364364364364
Mean (%)0.0039110.0003883.04E − 050.0006070.0006250.000476
Standard deviation0.053450.0113640.0190620.015210.0159760.013915
Minimum−0.49728−0.052272−0.12522−0.09997−0.10441−0.06943
Maximum0.191820.049340.0711750.0805870.0840620.055737
Skewness−2.27269−0.817685−1.76762−1.4396−1.3505−0.9786
Excess kurtosis24.373427.09687114.4600716.0975815.777158.544046
Jarque–Bera7241.834295.1252181.442727.5182586.688524.2676
Q(5)14.5298.89854.869842.70646.60515.504
Q(10)22.17216.5789.4020125.96141.4325.195
Q(20)37.94228.85920.095157.10177.0731.924
ADF−21.98285−17.94773−18.70611−6.224532−5.147020−11.19319
P-P−21.80890−17.97224−18.89272−21.87236−22.44193−19.26622

Note. The Jarque–Bera statistic tests for the null hypothesis of normality in sample return distribution. Q (n) is the Ljung–Box statistics of the return series for up to n-th order serial correlation. ADF and P-P are statistics of augmented Dickey–Fuller and Phillips–Perron unit root tests, respectively, based on the least AIC criterion. The symbols , , and indicate rejection at the 1%, 5%, and 10% significance level. The whole sample phase spans from July 20, 2010, to June 23, 2021.