Role of 2008 financial contagion in effecting the mediating role of stock market indices between the exchange rates and oil prices: Application of the unrestricted VAR

Figures & data

Table 1. Descriptive statistics

	Pre-Crisis				Post-Crisis				Overall period
	SI	ER	OP		SI	ER	OP		SI	ER	OP
Mean	19.454	0.376	0.329		319.513	0.491	−0.474		215.154	0.451	0.204
Median	78.640	0.100	2.650		428.305	0.110	0.990		299.610	0.100	1.250
Maximum	1512.670	5.050	26.280		5184.600	10.000	20.320		5184.600	10.000	26.280
Minimum	−3322.090	−1.250	−50.720		−4803.460	−3.800	−30.840		−4803.460	−3.800	−50.720
Std. Dev.	883.762	1.131	13.326		1410.194	1.630	9.445		1215.341	1.435	11.112
Skewness	−1.642	2.758	−1.567		−0.542	2.809	−0.666		−0.608	2.918	−1.204
Kurtosis	6.860	11.497	6.360		6.017	16.255	3.607		7.077	17.516	5.934
Jarque-Bera	65.276	260.864	53.651		41.984	846.300	8.753		125.960	1703.173	100.252
Probability	0.000	0.000	0.000		0.000	0.000	0.013		0.000	0.000	0.000
Sum	1186.700	22.950	20.090		31,312.300	48.100	−46.450		35,930.720	75.400	34.020
Observations	63	63	63		99	99	99		168	168	168

Note: This table explains the descriptive statistics of ER, SI and OP for three different regimes.

Table 2. Coefficient of Variation (CoV)

Variables	Pre-crisis	Post-Crisis	Overall Sample period
Exchange rate	3.005	3.32	3.179
Stock Index	45.42	4.41	5.65
Oil prices	40.46	19.92	54.54

Note: Coefficient of variation is calculated by dividing the standard deviation values of exchange rate, stock indexes, and oil prices during three multiple regimes by their corresponding mean values of every particular regime.

Table 3. Augmented dickey fuller (ADF) unit root test for three regimes

Variables	Pre-crisis		Post Crisis		Overall Period
	ADF unit root		ADF Unit root		ADF Unit root
	at level	I(1)	At level	I(1)	At level	I(1)
Exchange rate	−1.21	−6.76	−1.49	−8.91	−1.55	−8.11
Stock Index	−1.34	−8.17	−1.22	−7.79	−1.56	−9.91
Oil prices	−0.76	−4.76	−1.32	−7.70	−1.61	−11.18

Note: Unit root test purposed by (Dickey & Fuller, 1981). If the p-value is less than 0.05 then we can reject the null hypothesis: Data has a unit root or non-stationary.

Table 4. Philips peron unit root test for three regimes

Variables	Pre-crisis			Post Crisis		Overall Period
	PP unit root			PP Unit root		PP Unit root
	At level	I(1)		At level	I(1)	At level	I(1)
Exchange rate	−1.32	−5.5***		−0.91	−2.87**	−1.120	−3.75**
Stock Index	−1.19	−5.9***		−0.76	−3.95**	−1.55	−3.91**
Oil prices	−0.81	−6.6***		−1.43	−4.10**	−1.20	−5.10**

Note: If p-value is less than 0.05, 0.01 or 0.10 then we can reject the null of non-stationary. Level of significant at 5%, 1% and 10% is denoted by 0.05 or **, 0.01 or *** and 0.10 or * respectively.

Table 5. KPSS unit root test for three regimes

Variables	Pre-crisis		Post-Crisis		Overall Period
	KPSS unit root		KPSS Unit root		KPSS Unit root
	At level	I(1)	At level	I(1)	At level	I(1)
Exchange rate	0.48***	0.31	0.79**	0.11	0.48***	0.31
Stock Index	0.76***	0.21	0.91***	0.21	0.61***	0.39
Oil prices	0.91***	0.19	1.09***	0.31	0.71**	0.41

Note: If test statistics of KPSS is higher than critical values then we cannot accept the null of stationary.

Table 6. Zivot Andrew (ZA) Unit root test for three regimes

Variables	Pre-crisis		Post Crisis		Overall Period
	ZA unit root		ZA Unit root		ZA Unit root
	At level	I(1)	At level	I(1)	At level	I(1)
Exchange rate	−4.76	−9.50**	−3.01	−3.80***	−2.10	−6.70***
Stock Index	−4.90	−11.5**	−3.40	−4.10***	−2.80	−7.70***
Oil prices Structural Break	−5.51	−12.40***	−3.75	−5.10***	−3.44 2014	−8.10***

Note: Unit root estimation in the presence of structure break.: if p-value is less than 0.05, 0.01 or 0.10 then we can reject the null of non-stationary. Level of significant at 5%, 1% and 10% is denoted by 0.05 or **, 0.01 or *** and 0.10 or * respectively.

Table 7. Optimal Lag length criteria

Lag	Pre-crisis			Post-crisis			Overall-sample duration
	AIC	SC	HQ	AIC	SC	HQ	AIC	HC	HQ
0	35.16	35.26	35.20	28.46	28.54	28.49	28.36	28.4	28.3
1	27.21	27.63	27.37	28.00*	28.31*	28.31*	27.04*	28.01*	27.9*
2	26.77*	27.51*	27.06*	28.11	28.64	28.60	27.25	28.25	28.01
3	29.41	29.46	30.42	28.30	28.90	28.65	27.21	28.45	28.10

Note: * sign designates the figure with the lowest AIC, SC and HQ values.

Table 8. Unrestricted Co-Integration Rank test (Trace)

Hypothesized No of CE:	Pre-Crisis		Post Crisis		Overall Sample period
	Eigenvalue	Trace statistics (CV)	Eigen Value	Trace statistics (CV)	Eigen Value	Trace statistics (CV)
None	0.049	3.71	0.09	17.35	0.041	3.6
At most 1	0.0581	3.78	0.05	7.25	0.01	6.5
At most 2	0.00051	0.04	0.01	1.33	0.0004	0.03

Note: Johansson co-integration test for the identification of long-run co-integrating association between the variables.

Table 9. Unrestricted Vector Auto regression modelling results: Dependent variable: Stock Indexes

Variables	Pre-Crisis		Overall sample period		Post crisis period
	Coef.	P-value	Coef.	P-value	Coef.	P-value
Exchange rate(−1)	−1.13	0.02	−0.76***	0.0036	−0.46	0.54
Exchange rate (−2)	−3.72**	0.041	−1.57***	0.0009	−0.01	0.62
Oil prices(−1)	−0.001**	0.029	−0.05	0.53	−0.01	0.341
Oil prices (−2)	−0.0009	0.91	0.14***	0.0012	0.13**	0.041
Stock indexes(−1)	−0.011	0.51	0.02	0.11	−0.07	0.863
Stock index (−2)	−0.04	0.615	−0.12	0.21	−0.03	0.73
Portmanteau test for auto- correlation	3.86(0.9)		2.91(0.7)		4.10(0.17)
Bruesch Pagan test for Hetero-scedasticity	0.87(0.1)		0.91(.8)		0.18(0.09)
DW test value	2.01		1.90		1.90

Note: This table explains the transfer of shocks from the oil prices towards the stock indexes. Moreover, this table also explains transmission mechanism of shocks from the oil prices towards the exchange rate for dissimilar timeframes such as pre-economic contagion, postcontagion and overall time frame. *, **, *** represent the significant level at 10%, 5% and 1% respectively. All variables are first differenced and logarithmic transformed except oil prices during pre-crisis regime. However, all variables are first differenced and logarithmic transformed during post-crisis and overall period.

Table 10. Unrestricted Vector Auto regression modelling results: Dependent variable: Exchange rate

Variables	Pre-Crisis		Overall sample period		Post-crisis
	Coef.	P-value	Coef.	P-value	Coef.	P-value
Exchange rate(−1)	0.50**	0.039	−0.02**	0.03	−0.0003	0.81
Exchange rate (−2)	−0.25*	0.065	0.53***	0.002	0.617***	0.0019
Oil prices(−1)	−0.0002	0.89	0.01	0.772	0.01	0.682
Oil prices (−2)	0.0001	0.91	0.006	0.163	0.006	0.153
Stock indexes(−1)	−0.006***	0.001	−0.01	0.8	−0.07**	0.0172
Stock index (−2)	−0.042***	0.0003	−0.022**	0.045	−0.003	0.119
Portmanteau test for auto- correlation	1.12(0.725)		2.86(0.72)		1.4(0.83)
Bruesch Pagan test for Hetero-scedasticity	0.13(0.7)		1.11(0.71)		0.29(0.81)
DW test statistics	1.90		2.03		2.12

Note: This table explains the transfer of shocks from Stock indices towards the exchange rate. Moreover, this table also explains direct transmission mechanism of shocks from the oil price related shocks towards the exchange rate for three dissimilar regimes such as pre-economic contagion, post-contagion and overall sampling timeframe. *, **, *** represent the significant level at 10%, 5% and 1% respectively. All variables are first differenced and logarithmic transformed except oil prices during pre-crisis regime. However, all variables are first differenced and logarithmic transformed during post-crisis and overall period.

Figure 1. Impulse response function for pre-crisis period.

Figure 2. Impulse response function after crisis.

Figure 3. Mediating role of stock market indices between exchange rate and oil price retunes. a) Summarized results for Pre-Crisis regime, b) Summarized results for Post-Crisis regime.

Dickey, D. A., & Fuller, W. A. (1981). Likelihood ratio statistics for autoregressive time series with a unit root. Econometrica, 49(4), 1057–1072. https://doi.org/10.2307/1912517

 Web of Science ®Google Scholar