Asymmetric impacts of oil price shocks on government expenditures: Evidence from Saudi Arabia

Figures & data

Figure 1. Oil price between 1990Q1 and 2017Q2.

Figure shows crude oil price in US dollars per barrel. The data comes from Datastream hit WDQ76AAZA.

Figure 2. Government expenditure on education as % of total government expenditure 1990Q1–2017Q2.

Figure 3. Government expenditure on health as % of total government expenditure 1990Q1–2017Q2.

Table 1. Descriptive statistics

Var	Obs	Mean	Std. Dev.	Min	Max
Oil price	109	46.57	32.48	11.64	121.11
Educ expend	109	6.96	1.11	4.81	8.18
Health expend	109	4.58	0.2	4.33	4.93

Table 2. Augmented Dickey-Fuller unit root tests

Var.	Educ	D.Educ	Health	D.Health	Oil	D.Oil
Z(t)	−2.362	−5.018	−1.064	−6.545	−2.126	−5.504
p value	0.4002	0.0002	0.9349	0.000	0.5316	0.000

MacKinnon approximate p value for Z(t). The null hypothesis the existence of unit root against the alternative hypothesis of having a stationary series.

Table 3. ARDL estimated coefficients

Coeff.	D.Educ	D.Health	Coeff.	D.Educ	D.Health
L.y	−0.0005***	−0.0005***	L3D.y	2.7168***	0.9848***
	(−0.0001)	(−0.0001)		(−0.4352)	(−0.0281)
lnoil	0.5804***	0.059	L4D.y	−1.1841***
	(−0.2034)	(−0.046)		(−0.3093)
trend	0.0049	0.0029***	L5D.y	0.3051***
	(−0.0049)	(−0.001)		(−0.0935)
LD.y	3.2035***	2.9161***	Cons.	0.0022***	0.0020***
	(−0.0962)	(−0.0273)		(−0.0005)	(−0.0003)
L2D.y	−4.0411***	−2.8999***
	(−0.3169)	(−0.0546)
N	103	105	$F_{pss}$	9.281	16.573
${\chi }_{SC}^2$	1.000	12.000	pvalue${ }_{SC}$	0.243	0.001
${\chi }_{HET}^2$	2.440	1.542	pvalue${ }_{HET}$	0.118	0.214
$F_{ramsey}$	5.266	0.841	pvalue${ }_{ramsey}$	0.002	0.475

Figure 4. Cumulative sum of recursive residuals.

Table 4. Dynamic asymmetric estimation of oil prices on government expenditures on Education and Health

Coeff.	D.Educ	D.Health	Coeff.	D.Educ	D.Health
L.y	−0.048***	−0.057*	L.Doil(+)	0.067	0
	(−0.007)	(−0.029)		(−0.046)	(.)
L.Oil(+)	0.029***	0.036***	L2.Doil(+)	0.140***
	(−0.009)	(−0.007)		(−0.040)
L.Oil(-)	0.027**	0.046***	D.Oil(-)	0.041	−0.008
	(−0.012)	(−0.008)		(−0.032)	(−0.019)
L.Dy	2.896***	1.984***	L.Doil(-)	−0.081**	−0.040**
	(−0.029)	(−0.022)		(−0.034)	(−0.018)
L2.Dy	−2.853***	−1.005***	L2.Doil(-)	0
	(−0.058)	(−0.023)		(.)
L3.Dy	0.957***		Cons.	0.302***	0.003**
	(−0.029)			(−0.045)	−0.001
D.Oil(+)	0	0.058**
	(.)	(−0.026)
N	105	106	$R^2$	0.988	0.990
W${ }_{LR}$	10.021***	15.471***	W${ }_{SR}$	1.591	2.610*
LRoil+	0.611***	0.624**	$F_{pss}$	16.988***	10.695***
LRoil	0.564**	0.801*

Standard errors in parentheses *$p<0.10$, **$p<0.05$, ***$p<0.01$. W${ }_{LR}$, W${ }_{SR}$: Wald test for the null hypothesis of long- and short-run symmetry, respectively. $F_{pss}$ Pesaran et al. (2001) test for cointegration with the null hypothesis of no cointegration.

Table 5. NARDL estimation with only the long-run asymmetries

Coeff.	D.Educ	D.Health	Coeff.	D.Educ	D.Health
L.y	−0.047***	−0.057*	L.DOil(+)	0.071	0.000
	(0.007)	(0.029)		(0.046)	(.)
L.Oil(+)	0.027***	0.036***	L2.DOil(+)	0.128***
	(0.008)	(0.007)		(0.039)
L.Oil(-)	0.024**	0.046***	D.Oil(-)	0.000	0.017
	(0.012)	(0.009)		(.)	(0.013)
L.Dy	2.891***	1.988***	L.DOil(-)	−0.071**	−0.044**
	(0.029)	(0.022)		(0.033)	(0.018)
L2.Dy	−2.844***	−1.009***	L2.DOil(-)	0.000
	(0.058)	(0.024)		(.)
L3.Dy	0.952***
	(0.029)		Cons.	0.295***	0.269**
D.Oil(+)	0.000	0.017		(0.044)	(0.126)
	(.)	(0.013)
N	105	106	$R^2$	0.988	0.990
W_LR	14.040***		W_SR
LRoil+	0.674***	0.637**	Fpss	16.438	10.599
LRoil	0.516***	0.814**

Standard errors in parentheses *$p<0.10$, **$p<0.05$, ***$p<0.01$. W${ }_{LR}$, W${ }_{SR}$: Wald test for the null hypothesis of long- and short-run symmetry, respectively. $F_{pss}$ Pesaran et al. (2001) test for cointegration with the null hypothesis of no cointegration.

Figure 5. Cumulative dynamic multiplier.

Pesaran, M. H., Shin, Y., & Smith, R. J. (2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16(3), 289–326. doi:10.1002/(ISSN)1099-1255

 Web of Science ®Google Scholar