Modelling the determinants of government expenditure in Nigeria

Figures & data

Figure 1. Nigeria’s total government expenditure (1970–2017).

Trend of Nigeria’s central government expenditure where the horizontal axis shows six years in average. Computed from Statistical Bulletin of the CBN, 2017.

Figure 2. Share of total government expenditure in GDP (1970–2017).

Trend of Nigeria’s central government expenditure share to GDP where the horizontal axis shows two years in average. Computed from Statistical Bulletin of the CBN, 2017.

Table 1. Description of variables and sources of data

Variable	Description and measurement	Source
Government expenditure	Nominal total government expenditure of the central government of Nigeria	CBN
GDP	Nominal GDP	CBN
Oil revenue	Total federally collected revenue from oil	CBN
Public debt	Total debt stock of the central government	CBN
Trade openness	The degree of trade openness is proxy using the ratio of volume of trade (import plus export) to GDP	CBN
Population	Total population	WDI
Inflation	Inflation rate	WDI

Notes: CBN stands for Central bank of Nigeria and WDI stands for World Development Indicators by World Bank. All data are measured in nominal form and in local currency (Naira) and converted into natural logarithms except for inflation and trade openness that are in percentages and ratio, respectively. Source: Drawn by author.

Table 2. Description of variables and sources of data

Variable	Description and measurement	Source
Government expenditure	Real total government expenditure as a share of real GDP	CBN
Economic growth	Real GDP growth rate	CBN
Tax revenue	Total federally collected tax revenue	CBN
Exchange rate	Real effective exchange rate	WDI
Trade openness	The degree of trade openness is proxy using the ratio of volume of trade (import plus export) to GDP	CBN
Population	Growth rate of population	WDI
Oil price	Average yearly crude oil price	OPEC

Notes: CBN stands for Central bank of Nigeria, OPEC stands for Organization of Petroleum Exporting Countries and WDI stands for World Development Indicators by World Bank. Source: Drawn by author.

Table 3. ADF unit root test

Variable	At level		First difference		Order of integration
	Intercept	Trend and intercept	Intercept	Trend and intercept
lnGE	−1.442	−2.976	−7.911***	−8.044***	I(1)
lnGDP	−0.403	−1.943	−6.841***	−6.767***	I(1)
lnPOP	−0.136	−3.221	−7.7155***	−7.6286***	I(1)
lnOILR	−2.422	−1.980	−7.336***	−7.663***	I(1)
lnDEBT	−1.896	−1.215	−6.559***	−5.997***	I(1)
INF	−3.928**	−3.940**	–	–	I(0)
TOP	−0.044	−0.535	−2.994**	−3.404**	I(1)
lnTAXR	−0.550	−2.972	−5.432***	−5.367***	I(1)
lnOILP	−1.166	−1.214	−6.558***	−6.629***	I(1)
EXCR	−2.832*	−2.364*	−4.861***	−4.813***	I(1)
GE/GDP	−2.283	−4.484**	–	–	I(0)
RGDPG	−5.502***	−5.453***	–	–	I(0)
POPG	2.192	2.228	4.654***	6.126***	I(1)

(a) lnGE, lnGDP, lnPOP, lnOILR, lnDEBT, INF, lnTAXR, lnOILP, EXCHR, GE/GDP, RGDPG and TOP stand for log of government expenditure, log of gross domestic product, log of population, log of public debt, inflation, log of tax revenue, log of crude oil price, real exchange rate, real government expenditure as a share of real GDP, real GDP growth rate and trade openness, respectively. (b) ** and ***denote significance at 5% and 1% levels. (c) Optimal lag length is determined by Akaike Information Criterion. (d) I(0) stands for order of integration at order zero while I(1) stands for integration at order one. Source: computed by author using E-view 10.0.

Table 4. PP unit root test

Variable	At level		First difference		Order of integration
	Intercept	Trend and intercept	Intercept	Trend and intercept
lnGE	−1.077	−2.001	−7.886***	−8.036***	I(1)
lnGDP	−0.402	−2.029	−6.841***	−6.767***	I(1)
lnPOP	−0.4007	−1.6427	−6.732***	−7.432***	I(1)
lnOILR	−2.515	−2.172	−7.369***	−7.683***	I(1)
lnDEBT	−1.242	−1.267	−6.558***	−6.645***	I(1)
INF	−3.247**	−3.320**	–	–	I(0)
TOP	−2.396	−2.042	−9.171***	−9.813***	I(1)
lnTAXR	−0.460	−2.915	−6.154***	−6.675***	I(1)
lnOILP	−1.241	−1.267	−6.645***	−6.655	I(1)
EXCR	−2.219	−2.566	−4.333**	−4.263**	I(1)
GE/GDP	−2.051	−4.455**	–	–	I(0)
RGDPG	−5.514***	−5.574***	–	–	I(0)
POPG	2.682	2.687	−5.776***	−5.986***	I(1)

(a) lnGE, lnGDP, lnPOP, lnOILR, lnDEBT, INF, lnTAXR, lnOILP, EXCHR, GE/GDP, RGDPG and TOP stand for log of government expenditure, log of gross domestic product, log of population, log of public debt, inflation, log of tax revenue, log of crude oil price, effective real exchange rate, real government expenditure as a share of real GDP, real GDP growth rate and trade openness, respectively. (b) ** and ***denote significance at 5% and 1% levels. (c) Optimal lag length is determined by Akaike Information Criterion. (d) I(0) stands for order of integration at order zero while I(1) stands for integration at order one. Source: computed by author using E-view 10.0

Table 5. Result of bounds test for co-integration

Bound test for model 1 [1, 1, 0, 1, 0, 1, 0]
Test statistics	Value	Significance	I(0)	I(1)
F-statistic	5.89	10%	2.12	3.23
K	6	5%	3.45	3.61
		1%	3.15	4.43
Bound test for model 2 [1, 1, 1, 1, 1, 1, 0]
F-statistic	5.17	10%	2.17	3.22
K	6	5%	2.55	3.71
		1%	3.424	4.88

Lag length on each variable is selected using AIC criterion. Critical values are generated under the model with unrestricted intercept and no trend.

Source: Computed by author using E-view 10.0.

Table 6. Estimated long-run coefficient using ARDL approach

Model 1 [1, 1, 0, 1, 0, 1, 0] Dependent variable: LOGGE			Model 2 [1, 1, 1, 1, 1, 1, 0] Dependent variable: GE/GDP
Regressor	Coefficient	P-value	Regressor	Coefficient	P-value
Constant	−17.92**	0.0051	Constant	−0.486	0.2450
LOGGDP	0.453***	0.0008	RGDPG	0.054***	0.0008
INF	0.002*	0.1052	REXCH	−0.003	0.3377
LOGOILR	0.213**	0.0352	POPG	0.396**	0.0105
LOGPOP	3.166**	0.0311	LOGTAXR	0.502***	0.0001
TOP	0.006***	0.0007	LOGOILP	0.405**	0.0010
LOGDEBT	−0.169	0.2195	TOP	−0.006	0.1584

(a) lnGE, lnGDP, lnPOP, lnOILR, lnDEBT, INF, lnTAXR, lnOILP, EXCHR, GE/GDP, RGDPG and TOP stand for log of government expenditure, log of gross domestic product, log of population, log of public debt, inflation, log of tax revenue, log of crude oil price, real exchange rate, real government expenditure as a share of real GDP, real GDP growth rate and trade openness, respectively. (b) *, ** and ***denote significance at 10%, 5% and 1% levels. (c) Optimal lag length is determined by Akaike Information Criterion. Computed by author using E-view 10.0.

Table 7. Estimated short-run error correction model using ARDL approach

Model 1 [1, 1, 0, 1, 0, 1, 0] Dependent variable: LOGGE			Model 2 [1, 1, 1, 1, 1, 1, 0] Dependent variable: GE/GDP
Regressor	Coefficient	P-value	Regressor	Coefficient	P-value
Constant	−17.92***	0.0053	Constant	−0.344	0.2829
lnGE (−1)	−0.765***	0.0000	TOP	−0.004	0.1844
lnGDP	−0.093	0.7015	RGDPG	0.002	0.2820
lnGDP (−1)	0.346**	0.0150	REXCH	−0.005**	0.0473
INF	0.002**	0.0050	POPG	1.087**	0.0037
lnOILR	0.283**	0.0017	lnTAXR	0.217*	0.0948
lnPOP	2.422**	0.0588	lnOILP	−0.002	0.9933
TOP	0.002*	0.0231	GE/GDP (−1)	−0.709***	0.0004
TOP (−1)	0.005**	0.0076	RGDPG (−1)	0.004**	0.0046
lnDEBT	−0.129	0.2310	REXCH (−1)	−0.002	0.3173
LOGOILR (−1)	0.163*	0.0940	POPG (−1)	0.281**	0.0312
ECM (−1)	−0.765***	0.0000	lnTAXR (−1)	0.356**	0.0137
			lnOILP (−1)	0.287**	0.0062
			ECM (−1)	0.709***	0.0000

Notes: lnGE, lnGDP, lnPOP, lnOILR, lnDEBT, INF, lnTAXR, lnOILP, EXCHR, GE/GDP, RGDPG and TOP stand for log of government expenditure, log of gross domestic product, log of population, log of public debt, inflation, log of tax revenue, log of crude oil price, real exchange rate, real government expenditure as a share of real GDP, real GDP growth rate and trade openness, respectively. (b) *, ** and ***denote significance level at 10%, 5% and 1% levels. (c) Optimal lag length is determined by Akaike Information Criterion. Source: computed by author using E-view 10.0.

Table 8. Diagnostic tests result

Model 1			Model 2
Test	Statistic	P-value	777Test	Statistic	P-value
$X_{Auto}^2$ (2)	0.582	0.5642	$X_{Auto}^2$ (2)	0.5789	0.4402
$X_{Hetro}^2$ (2)	0.766	0.6595	$X_{Hetro}^2$ (2)	0.9411	0.5779
$X_{Norm}^2$ (2)	1.099	0.5773	$X_{Norm}^2$ (2)	0.2341	0.6754
$X_{F.form}^2$ (2)	2.761	0.1070	$X_{F.form}^2$ (2)	0.2889	0.5945

Notes: $X_{Auto}^2$, $X_{Hetro}^2$, $X_{Norm}^2$ and $X_{F.form}^2$ are statistics for test of serial correlation, heteroscedasticity, normal errors and functional form specification, respectively.

Source: computed by author using E-view 10.0.

Figure 3. Model 1 CUSSUM and CUSSUM square stability tests.

Figure 4. Model 2 CUSSUM and CUSSUM square stability tests.

Source: Computed by author using E-views 10.0.

CBN. (2017, December). Central bank of Nigeria statistical bulletin. Abuja.

 Google Scholar