You are on page 1of 12

Int. j. econ. manag. soc. sci., Vol(4), No (10), October, 2015. pp.

469-480

TI Journals

International Journal of Economy, Management and Social Sciences


www.tijournals.com

ISSN:
2306-7276

Copyright 2015. All rights reserved for TI Journals.

Empirical Analysis of the Relationship between Economic Growth and


Energy Consumption in Nigeria: A Multivariate Cointegration
Approach
Kayode Emmanuel Olaide *
Memorial University of Newfoundland Canada.
*Corresponding author: keo505@mun.ca

Keywords

Abstract

Economic growth
Energy consumption
Vector error correction
Co-integration
Nigeria

Using a neo-classical aggregate production model where capital, labor, real exchange rate and energy are
treated as separate inputs, this study tests for the existence and direction of causality between economic
growth and energy use in Nigeria at both aggregated total energy and disaggregated levels as crude oil, coal,
natural gas and electricity consumption. Using the autoregressive distributed lag (ARDL) co-integration
technique, the empirical findings indicate that there exists long-run co-integration among output, labor,
capital, real exchange rate and energy use in Nigeria at both aggregated and all the disaggregated levels
except for coal. Then using a VEC specification, the short-run dynamics of the interested variables are
tested. This indicates that there exists Granger causality running only from GDP to electricity consumption. I
thus propose policy suggestions to solve the energy and sustainable development dilemma in Nigeria as:
enhancing and guaranteeing energy supply; enhancing energy efficiency to save energy; diversifying energy
sources by exploiting renewable energy and drawing out appropriate policies and measures.

*Student #: 201398088
Econ 6009 (Graduate seminar) project paper
November 2014

1.

Introduction

Energy plays a vital role in the economic development of a country; it enhances the productivity of factors of production and increases living
standards. It is recognized that there is interdependency between economic development and energy consumption. The key question in energy
economics, however, is whether economic growth (EG) leads to energy consumption (EC) or whether EC leads to EG. The causal relationship
between energy consumption (EC) and economic growth (EG) has been well studied in economic literature. Lately, there has been a renewed
interest in examining this relationship due to the impact that energy consumption has on climate change, and because the higher economic
growth rates pursued by developing countries can only be achieved with the consumption of a larger quantity of commercial energy, which is a
key factor of production, along with capital, labour and raw materials. The aim of this paper is to empirically investigate the causal interactions
between EC and EG in Nigeria as a developing economy.
Energy consumption in Nigeria is mainly based on the use of fossil fuels which is non-renewable. Petroleum, a very important source of energy
and economic commodity in Nigeria, has had so many problematic issues since the 1980s. There is the issue of subsidy, the issue of scarcity, the
issue of sharing of revenues accruing from petroleum, the fuel subsidy issue, which in December 2011 generated social and political problems
that paralyzed economic activities nationwide. Also, the issue of probes in the downstream petroleum sub-sector, and recently, the issue of
privatization and deregulation of the Nigerian Oil Industry. It appears these problematic issues may have arisen due to some unfavourable
characteristics of petroleum policies in Nigeria. Similar problems have been encountered in the electricity sector in Nigeria. In Nigeria, the
continuance energy crisis has led to economic lapses, and drastically undermined the effort to achieve rapid and sustained economic growth and
development. Likewise industries have been collapsing due to the same problem thereby leading to massive unemployment and hardship for the
great majority in the country.
The purpose of this study is to re-examine the existence and direction of causality between output growth and energy use in Nigeria at both
aggregated total energy and disaggregated levels as crude oil, natural gas, electricity and coal consumption. This is to be able to check not only
the direct effect of energy consumption on economic growth, but also, the indirect impact of energy consumption on economic growth through
its complementary effect on capital and labor, and also through the influence of the international trade on the Nigerian energy consumption
(being an energy exporter). This is accomplished by proposing a framework based on the neo-classical one-sector aggregate production
technology where capital, labor, real exchange rate (proxy for international trade), and energy are treated as separate inputs. Therefore, this study
is expected to provide useful insight into the relationship between energy consumption and economic development in Nigeria, as a developing
economy. Also, it will contribute to the energy consumption-economic growth nexus literature, as it provides additional empirical evidence on
the relationship within the context of a developing and emerging economy. This study will be useful to stakeholders in the Nigerian energy
industry and market, and policy makers, as it provides evidence on the relationship between energy use and output growth. This research work is
organized as follow: Section I is the introduction; section II of the paper discusses the literature review, where both theoretical and empirical
studies on previous works are looked into. In section III, the methodology of this study is considered. Section IV discusses empirical results,
followed by a conclusion and policy implications in Section V.

2.

Literature review

Various energy crisis and persistently high energy prices, particularly oil prices, have had a significant impact on the economic activity of most
economies. Hence, the causal relationship between energy consumption and economic growth has been a widely studied topic in energy
economics literature; as energy plays a significant role in any economy. The growth hypothesis suggests that energy consumption is an
indispensable component in growth, directly or indirectly as a complement to capital and labour as an input in the production process (Mulegeta
et al. 2010). Since production and consumption activities involve energy as an essential factor inputs, the relationship between energy
consumption and economic growth has been a subject of considerable discussions in the literature (Abdulnasser and Manuuchehr, 2005). The
question as to whether energy consumption has positive, negative or neutral impact on economic activities has motivated the interest of
economists and policy analysts hence the need to find out the direction of causality between energy consumption and economic growth (Eddine,
2009).

Kayode Emmanuel Olaide *

470

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

Empirical studies designed to test the causal relationships between energy consumption and economic growth are generally grouped into three
testable hypotheses. The first hypothesis suggests that energy consumption is a pre-condition for economic growth, given that energy is a direct
input in the production process and also, energy is an indirect input that complements labor and capital inputs (Odhiambo, 2009; Ebohon, 1996).
The second hypothesis assumes that there is a bi-directional or feedback relationship between energy consumption and economic growth. The
implication of the bidirectional relationship is that energy consumption and economic growth are complementary. This implies that an increase
in energy consumption will accelerate economic growth, and in the reverse also, an increase in economic growth will stimulate energy
consumption (Hon, 2009; Omotor, 2008). The third hypothesis which is neutral, assumes that there is no causality between energy consumption
and economic growth and thus policies that are aimed at conserving energy will not retard economic growth (George and Nickoloas, 2011;
Ezatollah et al., 2010). In an attempt to justify the first hypothesis, Odhiambo (2009) applied the autoregressive distributed lag (ARDL) bounds
test approach and Granger non-causality test for Tanzania for the 1971-2006 period. The results of the bounds test revealed a stable long-run
relationship between energy consumption and economic growth. While, the results of Granger non-causality showed the evidence of
unidirectional causality running from energy consumption to economic growth. The results imply that energy conservation policies would have
damaging repercussions on economic growth for Tanzania. Contradicting the first hypothesis, Mehrara (2007) looked at the relationship between
the per capita energy consumption and per capita GDP using the panel data for 11 oil exporting countries for the period 1971-2002. On the basis
of the panel co-integration technique and Granger causality test, the results showed a unidirectional causality from economic growth to energy
consumption for all the countries. His results indicated that energy conservation policies have no damaging effect on economic growth for this
group of countries. Moreover, Esso (2010) examines the long-run causality relationship between energy and economic growth for 7 sub-Sahara
countries over the period 1970-2007. He applied Autoregressive Distributed Lag (ARDL) Bounds testing approach to co-integration. The
findings suggest unidirectional relationship running from energy consumption to real GDP for all countries involved, except CotedIvoire and
Congo. The result of causality indicates bidirectional relationship between energy consumption and real GDP in the case of CotedIvoire and
unidirectional causality from real GDP to energy for Congo. Von (2009), on the basis of panel data from 158 countries for the period 1980 -2004
and employing semi-parametric partially linear panel model, reports that energy consumption leads to increase in economic growth and the
effect of time trend is not significant. This justifies the first hypothesis.
Mawuse and Nezan (2009) on the basis of panel data for 4 West African Economic and Monetary Union (WAEMU) countries for the period
1970-2005 and applying Co-integration test and Vector Error Correction Model (VECM) carried out a study in support of the second hypothesis.
The findings suggest a bidirectional relationship for the panel as a whole, the findings reveal not only feedback between energy consumptiongrowth nexus but also support regional energy policies which must take account some countries specificities. Also, using time series data for the
period 1970-2009 and applying the techniques of Vector Autoregressive (VAR) and Vector Error Correction Model (VECM), Magazzino (2011)
reports the long run bidirectional relationship between energy consumption and economic growth in Italy. Similarly, using time series data from
Malaysia for the period 1971-2008 and applying ARDL bounds testing approach to Co-integration and causality tests within a Vector Error
Correction Model (VECM), Faridul et al. (2011) discovers that energy consumption is affected by economic growth and financial development
in the short run and in the long run. Yu and Jin (1992) used co-integration analysis to test the long-run relationship between energy use and
employment as well as industrial output in the USA. They found that no co-integrating relationship exists between energy use and these two
variables. This justifies the third hypothesis. Also, Stern (1993) examined the causal relationship between energy use and GDP in the USA. He
employed a multivariate vector autoregressive (VAR) analysis and used a weighting index of energy quality, where content of energy use shifts
from lower quality energy such as coal to high quality energy such as electricity, rather than using a measure of total energy use. He also
employed a different test of causality. He found that total energy use does not Granger cause GDP. However, using a weighting measure of
energy, a unidirectional Granger causality is found, running from energy use to GDP. Stern (2000) extends his analysis on the US economy by
introducing co-integration analysis of the relationship between energy and GDP. He found again that total energy use does not seem to have
Granger causality with GDP. However, using quality weighting index of energy, it is found to Granger cause GDP. His co-integration results
thus show that energy cannot be excluded from the co-integration space. In three of the five models estimated he found unidirectional causality
between energy use and GDP where causality runs from energy use to GDP. In the other two models he found a bi-directional causal relationship
between energy use and GDP. Stern results suggest that energy is considered to be a significant input factor that affects GDP in the USA. Also,
using the VEC specification, Ghali and El- Sakka (2004) found out that there is bidirectional Granger causality between output growth and
energy use in Canada. This implies that energy can be considered as a limiting factor to economic growth in Canada. It can be seen that so far,
there is a lack of consensus on the actual relationship between energy consumption and economic or output growth in the energy-growth
literature. This stems mainly from the difference in analytical techniques and the form of the data used.

3.

Methodology

3.1 Neo-classical production model


To investigate the relationship between energy consumption and economic growth, this research work makes use of the framework proposed in
Ghali and El-Sakka 2004, and also used in Soytas and Sari, 2007, among others. It is based on the conventional neo-classical one-sector
aggregate production technology where capital, labour, real exchange rate, and energy are treated as separate inputs. That is:
=f(Kt,Lt,ERt,Et)

(1)

Where Y is the real GDP; K is the capital stock; L is the level of employment; ER is real exchange rate; E is total energy consumption in
aggregated level or crude oil consumption, natural gas consumption, electricity consumption and coal consumption at disaggregated level, and
the subscript t denotes the time period. Taking the differential of Eq. (1) we obtain:
=

(2)

Where is the partial derivative of Y with respect to its ith argument. On dividing
Eq. (2) through by and rearranging the resulting expression, we obtain the following growth equation:
= + +

(3)

Where a dot on the top of a variable means that the variable is now in a growth rate form. The constant parameters a, b, c and d are the elasticity
of output with respect to capital, labor, real exchange rate and energy, respectively.
The relationship between output and capital, labor, real exchange rate and energy inputs described by the production function in Eq. (1) suggests
that their long-run movements may be related. Furthermore, if one allows for short-run dynamics in factor-input behavior, the analysis above
would also suggest that past changes in capital, labor, real exchange rate and energy could contain useful information for predicting the future

471

Empirical Analysis of the Relationship between Economic Growth and Energy Consumption in Nigeria: A Multivariate Cointegration Approach
International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

changes of output, ceteris paribus. The long-run and short run dynamics between the variables is examined using the ARDL to test for
multivariate co-integration, and Granger causality within the context of the Vector Error Correction Model.
3.2 Test for co-integration and granger causality
A recently advanced co-integration approach, known as the autoregressive distributed lag (ARDL) [Pesaran et al (2001)], has become popular
among researchers. In Pesaran et al (2001), the co-integration approach, also known as the bounds testing method, is used to test the existence of
a co-integrated relationship among variables. The procedure involves investigating the existence of a long-run relationship in the form of an
unrestricted error correction model for each variable as follows:
= +
+

+
+

,
+

,
+ ,

,
+
+ ,

,
+

,
+
+
+

= +
,

,
+
+ ,

,
+

+
+

,
+
+
+

,
+ ,
+
+
+
+
+

= +
,

= +
,

,
+
+ ,
,
+
+ ,

,
+

+
(4)

= +
,

+
+

,
+
+
+

(5)

(6)

(7)

(8)

Where Y is the natural logarithm of GDP, K is the natural logarithm of Gross capital formation, L is the natural log of the percentage of the
population employed, ER is the natural logarithm of real exchange rate, and E is the natural logarithm of energy consumption. The F-tests are
used to test the existence of long-run relationships. The F-test used for this procedure, however, has a nonstandard distribution. Thus, the Pesaran
et al (2001) approach computes two sets of critical values for a given significance level. One set assumes that all variables are I (0) and the other
set assumes they are all I (1). If the computed F-statistic exceeds the upper critical bounds value, then the null hypothesis (no co-integration) is
rejected. If the F-statistic falls within the bounds set, then the test becomes inconclusive. If the F-statistic falls below the lower critical bound
value, it implies no co-integration. When a long-run relationship exists, the F-test indicates which variable should be normalized. For instance,
the null hypothesis of equation (4) is H 0 : 1 = 2 = 3 = 4 = 5 = 0.
This technique has certain econometric advantages compared with other single co-integration procedures. They are as follows: (i) endogeneity
problems and inability to test hypotheses on the estimated coefficients in the long-run associated with the Engle-Granger method are avoided; ii)
the long and short-run parameters of the model in question are estimated simultaneously; iii) the ARDL approach to testing for the existence of a
long-run relationship between the variables in levels is applicable irrespective of whether the underlying regressors are purely I(0), purely I(1),
or fractionally integrated; iv) It is superior in small sample.
Following Granger (1988), and Engle and Granger (1987), I estimated a VEC model for the Granger causality test for our problem at hand. The
VEC representation is as follow:
= +

= +

= +
= +

= +

,
,

,
,

,
,

(9)

,
,

,
,

(10)
(11)
(12)
(13)

Where p is lag length and is decided according to information criterion and final prediction error. The parameters , are the co-integrating
vectors, derived from the long-run co-integrating relationships (i.e. =1Kt+2 Lt+3ER t+4Et+ where is stationary residuals) regression,
and their coefficients , are the adjustment coefficients. The parameters i, (i=1, 2, 3, 4, 5) are intercepts and the symbol denotes the
difference of the variable following it.
In addition to being consistent with the specifications in Equations (2) and (3), the model in Equations (4) (8) describes the inter-temporal
interaction between output and the factor inputs included in the production function. Once the equilibrium conditions represented by the cointegrating relations are imposed, the VEC model describes how, in each time period, output growth is adjusting towards its long-run
equilibrium state. Since the variables are supposed to be co-integrated, then in the short term, deviation of output from its long-run equilibrium
path will feed back on its future changes in order to force its movement towards the long-run equilibrium state. The co-integrating vectors from
which the error-correction terms are derived are each indicating an independent direction where a stable, meaningful long-run equilibrium state
exists. The coefficients of the error-correction terms, however, represent the proportion by which the long-run disequilibrium in the dependent
variables is corrected in each short-term period.
Using the model in Equations (913), Granger causality tests between the variables can be investigated through the following three channels:
i.
ii.

iii.

The statistical significance of the lagged error-correction terms (ECTs) by applying separate t-tests on the adjustment coefficients. This
shows the existence of a long-run relationship
A joint F-test or a Wald 2-test applied to the coefficients of each explanatory variable in one equation. For example, to test whether
energy use Granger-causes output in Eq. (3), we test the following null hypothesis: : , = , == , =0. This is a measure of
short-run causality;
A joint F-test or a Wald 2-test applied jointly to the terms in (i) and the terms in (ii)

Kayode Emmanuel Olaide *

472

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

4.

Empirical results and discussion

4.1 Data and variable definition


This study makes use of annual time series data on real GDP, capital, labor, real exchange rate, and energy consumption for Nigeria during the
period 1980 to 2013. Four forms of energy, which include crude oil, natural gas, coal and electricity are used. The first three are primary while
the fourth is a secondary form of energy. The gross capital formation is used for capital, and the labor is represented by the percentage of the
population gainfully employed. GDP is used as a proxy variable for economic growth. GDP is measured in millions of Nigerian naira. Crude oil
consumption is measured in thousands barrel per day; natural gas consumption in billion cubic feet; coal consumption in thousands of short tons,
and electricity in billion kilowatt hour. The total energy consumption is measured in kilojoule after appropriate conversion of the various
components. The data for the GDP, capital, real exchange rate and labor are obtained from the Central Bank of Nigeria (CBN) Annual Statistical
Bulletin (2013) and the CBN Quarterly Statistical Bulletin (June, 2014) while the data for the various energy consumptions are obtained from the
United States Energy Information Agency (EIA). The variables notations and definitions are as follows.
GDP: Real GDP
GCF: Capital stock
LAB: Labor
RER: Real exchange rate
COC: Crude oil consumption
NGC: Natural gas consumption
ELC: Electricity consumption
CC: Coal consumption
TEC: Total energy consumption
All variables are transformed into their natural logarithm so that their first differences approximate their growth rates. This also helps to adjust
for the differences in units.
All the time series data show some trend.Figure 1 shows the growth trend of the related variables, suggesting that long-run relationship is likely
to be present in the study since all the series tend to move very closely together over time. All the series show trends in their levels, but this
disappears in the first difference. Graphical analysis also reveals that, the interested series (GDP, COC, NGC, ELC, CC and TEC) have linear
relationship.

10

11

.5

LNGDP
12

D.LNGDP
1

13

1.5

14

Fig 1 Time plots of variables

1980

1990

2000
date

2010

2020

1980

1990

Figure1a. Time plot for LNGDP and DLNGDP

2000
date

2010

2020

Empirical Analysis of the Relationship between Economic Growth and Energy Consumption in Nigeria: A Multivariate Cointegration Approach

2.4

-.4

2.6

-.2

2.8

LN G CF

D .LN G CF
0

.2

3.2

.4

3.4

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

1980

1990

2000
date

2010

2020

1980

1990

2000
date

2010

2020

-.15

-.1

4.4

-.05

LN LAB

D .LN LAB
0

4.5

.05

.1

4.6

Figure1b. Time plot for LNGCF and DLNGCF

4.3

473

1980

1990

2000
date

2010

2020

1980

1990

Figure1c. Time plot for LNLAB and DLNLAB

2000
date

2010

2020

Kayode Emmanuel Olaide *

-1

4.5

- .5

LN RER

D .L N R E R

5 .5

.5

6 .5

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

1980

1990

2000
date

2010

2020

1980

1990

2000
date

2010

2020

2000
date

2010

2020

-.1

5.2

LN C O C
5.4

D .L N C O C

.1

5.6

.2

5.8

Figure1d. Time plot for LNRER and DLNRER

1980

1990

2000
date

2010

2020

1980

1990

Figure1e. Time plot for LNCOC and DLNCOC

474

Empirical Analysis of the Relationship between Economic Growth and Energy Consumption in Nigeria: A Multivariate Cointegration Approach

3.5

-1

-.5

4.5

LN N GC

D .LN N GC
0

.5

5.5

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

1980

1990

2000
date

2010

2020

1980

1990

2000
date

2010

2020

2000
date

2010

2020

LNELC
2.5

D .LN ELC
.2

.4

.6

3.5

Figure1f. Time plot for LNNGC and DLNNGC

-.2

1.5

475

1980

1990

2000
date

2010

2020

1980

1990

Figure1g. Time plot for LNELC and DLNELC

Kayode Emmanuel Olaide *

-2

-1

LN CC
3

D .L N C C

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

1980

1990

2000
date

2010

2020

1980

1990

2000
date

2010

2020

2000
date

2010

2020

-1

31 .5

32

-.5

3 2 .5

LN T E C

D .L N T E C

33

3 3.5

.5

34

Figure1h. Time plot for LNCC and DLNCC

1980

1990

2000
date

2010

2020

1980

1990

Figure1i. Time plot for LNTEC and DLNTEC

476

477

Empirical Analysis of the Relationship between Economic Growth and Energy Consumption in Nigeria: A Multivariate Cointegration Approach
International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

4.2 Test results for unit roots


When working with time series data, the first question to ask is whether or not the series is stationary. A stochastic process is said to be
stationary if its mean and variance are constant over time, and if the covariance exists between the two time periods and not the actual time at
which the covariance is computed. Since, the VEC specification in Equations (9)(13) requires that some or all the variables are integrated of
order one, I herein investigate the stationarity status of the variables using both the augmented DickeyFuller (ADF) and the Dickey-Fuller
Generalized Least Square (DF-GLS) tests for unit roots. The null hypothesis tested is that the variable under investigation has a unit root against
the alternative that it does not (that is, it is stationary). In the ADF, lag-length is chosen using the Akaike Information Criteria (AIC) after testing
for first and higher order serial correlation in the residuals while in the DF-GLS, the optimal lag is determined using Ng-Perron seq t, Schwarz
Criteria (SC) and AIC. Table 1 reports the results of testing for unit roots in the level variables as well as in their first difference. The table
shows the estimated t- statistics. In the first half of the table the null hypothesis that each variable has a unit root cannot be rejected by both tests,
except for gross capital formation. However, after applying the first difference, both tests reject the null hypothesis, as can be seen in the second
half of the table. Since the data appear to be stationary in first differences, no further tests are performed. I, therefore, maintain the null
hypothesis that each variable is integrated of order one.
Table 1. Results of unit roots test
VARIABLE
ADF
DFGLS
LNGDP
2.230
-2.066
LNGCF
-3.650
-3.029
LNLAB
-1.384
-1.323
LNRER
-2.256
-2.527
LNCOC
-2.013
-2.013
LNNGC
-2.170
-1.201
LNELC
0.310
-1.106
LNCC
-1.809
-1.985
LNTEC
-2.009
-1.413
DLNGDP
-4.233
-3.217
DLNGCF
-5.343
-5.527
DLNLAB
-3.810
-3.201
DLNRER
-3.644
-3.751
DLNCOC
-5.624
-3.674
DLNNGC
-3.989
-3.371
DLNELC
-5.347
-3.672
DLNCC
-8.457
-7.965
DLNTEC
-4.417
-3.492
* The critical values of t-statistics for the ADF are -2.98 and -2.622
(and that of DFGLS, -3.19 and-2.89) at 5% and 10% level of
significance respectively.

4.3 Test results for co-integration


An F deletion test was applied to equations (4) and (8) for each form of energy consumptions and for the total energy consumption in order to
test the existence of a long-run relationship. The results of bounds testing show that there is a long-run relationship between the variables when
Economic growth (GDP) is the dependent variable because its F-statistic exceeds the upper bound critical value at a 5% level of significance,
except for natural gas consumption and total energy consumption. The null hypothesis of no co-integration however, cannot be rejected when
each of the various forms of energy consumption, is used as the dependent variable, because F-statistics is below the lower bound critical value
at a 5% level of significance. Thus, the bounds
test result confirms that long-run unidirectional causality runs from each of crude oil consumption, electricity consumption, and coal
consumption to economic growth. But, there is a no causality between energy growth and each of natural gas consumption and total energy
consumption.
The results of bounds testing are presented in Table 2.

VARIABLES
LNGDP/LNCOC
LNCOC/LNGDP
LNGDP/LNNGC
LNNGC/LNGDP
LNGDP/LNELC

Table 2. Results of ARDL cointegration estimation.


F-STATISTICS
VARIABLES
19.72
LNELC/LNGDP
2.00
LNGDP/LNCC
3.60
LNCC/LNGDP
0.28
LNGDP/LNTEC
8.58
LNTEC/LNGDP

F-STATISTICS
3.51
4.61
2.46
3.58
0.30

*The critical value ranges of F-statistics are 3.96-4.53 and 3.21-3.74 at 5% and 10% level of significance respectively
[Paresh Kumar Narayan (2005)].

4.4 Test results for Vector Error Correction and Granger causality
The optimal lag length for the Vector Error Correction model was determined using the Akaike Information Criteria (AIC) and the Schwarz
Bayesian Information Criteria (SBIC). It was found to be 4.

LAG
AIC
SBIC

0
-8.20329
-7.92305

Table 3. Lag length determination for VEC


1
2
3
-13.5905
-13.9835
-14.7332
-11.6288
-10.3404
-9.40867

Endogenous: LNGDP LNCOC LNNGC LNELC LNCC LNTEC.


Exogenous: _cons

4
-45.9316*
-38.9256*

Kayode Emmanuel Olaide *

478

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

The regression result for the estimation of the long run relationship of the neo-classical production model is shown in Table 4 below.
(1)
LNGDP
0.0938
(0.34)

LNGCF

Table 4. Estimation of the neoclassical production function relationship


(2)
(3)
(4)
LNGDP
LNGDP
LNGDP
-0.510*
-0.0122
-0.186
(-2.06)
(-0.06)
(-0.63)

(5)
LNGDP
-0.418
(-1.81)

LNLAB

-4.557***
(-4.29)

-4.315***
(-4.32)

1.286
(1.33)

-2.745*
(-2.33)

-3.389***
(-3.66)

LNRER

0.121
(0.88)

0.0570
(0.47)

-0.0528
(-0.57)

-0.207
(-1.68)

0.0738
(0.65)

LNCOC

4.197***
(6.77)
1.071***
(7.40)

LNNGC

1.289***
(9.44)

LNELC

-0.381***
(-5.54)

LNCC

1.187***
(8.18)

LNTEC

_cons

8.884
(1.52)

27.68***
(6.08)

4.298
(0.90)

27.78***
(5.15)

-10.54
(-1.58)

34

34

34

34

34

t statistics in parentheses
*
p < 0.05, ** p < 0.01, *** p < 0.001

Column1 shows the result where crude oil consumption is used as the energy input, column2 shows the result for natural gas consumption,
column3 for electricity consumption, column4 for coal consumption, and column5 shows the result for total energy consumption as the energy
input. The result reveals that all else held constant, a 1% increase in crude oil consumption brings about a 4.20% increase in GDP on the
average; a 1% increase in natural gas consumption leads to a 1.07% increase in GDP on the average; a 1% increase in electricity consumption
brings about a 1.29% increase in GDP on the average; a 1% increase in coal consumption leads to a 0.38% decrease in GDP on the average; and
a 1% increase in total energy consumption leads to a 1.19% increase in GDP on the average. The coefficient of each energy input in each case is
found to be statistically significant at the 0.1% level of significance. The result is not surprising as the graphical representation of the series
reveal a linear trend in their levels. However, the estimated residual in each regression is found to be stationary using the ADF. They were also
tested for autocorrelation, using the Breusch-Godfrey (BP) test, and the null hypothesis of no autocorrelation could not be rejected at a lag length
of one.
The result of short- and long-run Granger causality is determined within the VECM framework. The short-run causal effects are demonstrated
through the F-statistics of the explanatory variables and long run causality is tested with the help of statistical significance and sign of the error
correction term. The results show that there is no causality relationship between energy growth and each of the various forms of energy
consumption, except electricity consumption in the short run at both 5% and 10% level of significance; there is a unidirectional causality running
from energy growth to electricity consumption at 5% level of significance. There is also no short run causal relationship between economic
growth and total energy consumption. Also, the results reveal that there is no causal relationship between the various forms of energy
consumption (except electricity consumption) and economic growth in any direction in the long run; as the coefficients of the error correction
term are not statistically significant in all of them. However, there is a long run causality running from economic growth to electricity
consumption at 5% significance level. Also, there is a long run causality from economic growth to total energy consumption at 10% significance
level. The results are shown in Table5 (a and b).

Table 5. Results for VECM and Granger causality


(a) F-Statistics (test of short run granger causality)
Dependent variable
LNGDP

LNGDP
-

LNCOC

1.32
(0.33)
1.13
(0.43)
1.55
(0.23)
1.38
(0.30)
1.15
(0.42)

LNNGC
LNELC
LNCC
LNTEC
*Probability > F in bracket

LNCOC
1.19
(0.39)

LNNGC
0.26
(0.99)

LNELC
5.09
(0.005)

LNCC
2.04
(0.12)

LNTEC
0.29
(0.99)

479

Empirical Analysis of the Relationship between Economic Growth and Energy Consumption in Nigeria: A Multivariate Cointegration Approach
International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

(b) ECM t-1 (t-Statistics (test of long run granger causality)


Dependent variable
LNGDP

LNGDP
-

LNCOC

0.11
(0.91)
0.01
(0.94)
0.15
(0.29)
-0.07
(0.47)
-0.01
(0.93)

LNNGC
LNELC
LNCC
LNTEC

LNCOC
0.08
(0.23)

LNNGC
0.48
(0.15)

LNELC
1.01
(0.000)

LNCC
-0.97
(0.25)

LNTEC
0.49
(0.10)

*Probability > |t| in bracket


I tried to investigate into the interaction of the various disaggregated energy forms with one another by estimating a regression model which
includes all of them together in a single long run relationship using the production model above. This was however dropped due to statistical
insignificant of all of the coefficients of the explanatory variables, and multi-collinearity in the VECM. Table6 below shows the result of the
pair-wise correlation between the variables. It reveals a strong correlation between the real GDP and each of the control variables, and also a
strong correlation between the various forms of energy consumption. The sign on each of the partial correlation coefficient is as expected. For
instance, the negative correlation between coal consumption and other forms of energy consumption is as a result of the fact that the
consumption of coal has been on the decline in the country since the discovery of the other forms of energy; even though, coal used to be one of
the source of export earnings for the country. This also translates into the negative correlation between coal consumption and real GDP.
Table 6. pair-wise correlation between variables (correlation matrix)
LNGDP
LNGCF
LNLAB
LNRER
LNCOC
LNNGC
LNELC
LNCC

5.

LNGDP
1
-0.124
-0.2463
-0.4697
0.7612
0.7626
0.9067
-0.7751

LNGCF

LNLAB

LNRER

LNCOC

LNNGC

LNELC

LNCC

1
0.0803
-0.1262
-0.1186
0.1105
-0.1457
0.0937

1
-0.1354
0.1924
0.1852
-0.3969
0.0314

1
-0.6909
-0.6741
-0.433
0.4608

1
0.8316
0.6615
-0.7089

1
0.6844
-0.7297

1
-0.7353

Conclusion and policy implication

This study attempted to analyze the causal relationship between energy use and economic growth in Nigeria. Based on the neo-classical one
sector aggregate production technology, I developed a VEC model after testing for multivariate co-integration between output, capital, labor, real
exchange rate and energy use. The co-integration test indicates that energy enters significantly the co-integration space. However, the short-run
dynamics of the variables show that there is no causality between the various forms of energy consumption and economic growth, except
electricity consumption. The flow of causality runs only in one direction from economic growth to electricity consumption. This result is not
surprising, as the oil and gas subsector has been the least transparent sector in the Nigerian economy, due to corruption and gross
mismanagement. According to the Nigerian National Bureau of Statistics (NBS), the oil and gas subsector accounts for 95% of the Nigerian
export earnings, 75% of the countrys federal government revenue, but contribute less than 30% on the average to the countrys real GDP.
The energy-growth nexus poses important challenges to Nigerian policy makers, considering the high energy consumption growth rate, high
CO2 emissions level and its growth rate. Economic growth rate is expected to keep as high as 78% in the next 20 years. In light of the close
energy-growth nexus, how can Nigeria realize sustainable development and cut down GHG emissions? Since the emissions mainly result from
consumption of fossil fuels, reducing energy consumption seems to be the direct way of handling the emissions problem. However, due to the
negative impact on economic growth, direct measure to reduce energy consumption is not viable in Nigeria. On the other hand, in Nigeria pure
development itself may not be a solution to environmental and ecological problem. Hence, active policies and measures must be implemented.
First of all, enhancing energy supply security and guaranteeing energy supply is of uttermost importance to Nigeria. Particularly in the short run,
proper supply of electric power, natural gas and oil is vital to the function of economic activity. Concerning electricity, long years of epileptic
power supply has hindered the proper development of the small and medium scale enterprises (SMEs) in the country, and also led to the exit
from the country of some multi-national manufacturing companies that are mainly energy dependent. The electricity supply has mainly been
hydro-based; Nigeria can increase the supply of its electric power by making use of the natural gas associated with the exploration of crude oil to
augment the hydro power, instead of flaring it into the atmosphere, which has been the order of the day. The recent privatization of the electric
power supply sub-sector of the economy could also be a move in the right direction, if properly implemented and monitored. As far as crude oil
supply is concerned, the domestic consumption is mainly in form of its bi-product which has been highly subsidized over the years with no
positive impact. The major problem has been that out of the four refineries in Nigeria, only one is semi-functional due to gross mismanagement,
hence, the country has to export crude oil and import the refined products. The existing refineries should therefore be renovated up to their full
capacity utilization levels. The oil companies, especially the multi-national should be encouraged to build their own refineries at least as a joint
venture, as a result of the capital cost. The subsidies on the oil products should be scrapped and the fund diverted to the development of other
sectors, especially agriculture. The bilateral agreement between Nigeria and Algeria to transport the Nigerian natural gas to Europe through the
pipeline in Algeria should be properly implemented instead of leaving it to lie fallow as an ordinary paper agreement over the years. Taking
advantage of the recent Russian crisis could be a way to capture a share of the market in Europe, which is looking for a way to diversify its
source of natural gas supply. Nigeria should also aim at an effective long-term policy to enhance energy efficiency, and to diversify energy
supply with preference on renewable energy such as wind power, solar (which are in great abundance in the country), and others.

Kayode Emmanuel Olaide *

480

International Journal of Economy, Management and Social Sciences Vol(4), No (10), October, 2015.

References
Abdulnasser, H. and Manuuchehr, I. (2005). Energy Consumption and Economic Growth in Sweden: A Leveraged Bootstrap Approach. Journal of Applied
Econometrics a Quantitative Studies, Vol. 2(4), pp.87-98.
Central Bank of Nigeria (2013). Annual Statistical Bulletin. Abuja, Nigeria.
Central Bank of Nigeria: Quarterly Statistical Bulletin volume 3, No 22, June, 2014.
Ebohon, O. J. (1996). "Energy, Economic Growth and Causality in Developing Countries: A Case Study of Tanzania and Nigeria." Energy Policy, Vol.24, No.5,
pp.447-53.
Eddine, H.C. (2009). Long and Short-run Linkages between Economic Growth, Energy Consumption and CO2 Emission in Tunisia. Economic Research Forum,
WP 485.
Engle, R.F., Granger, C.W.J., 1987. Co-integration and error correction: representation, estimation, and testing. Econometrica 55, 251276.
Esso, J.L. (2010). The Energy Consumption-Growth nexus on Seven Sub-Sahara African Countries. Economic Bulletin, Vol. 30(2), pp. 1191-1209.
Ezatolloh, A., Nazari, M., and Nasrindoost, M. (2010). Energy Consumption and Economic Growth in the Iranian Economy: Testing the Causality Relationship.
Middle-East Journal of Scientific Research, Vol. 5(5), pp. 374-381.
Faridul, I., Mohammad, S., and Mahamudal, A. (201.1). Financial Development and Energy Consumption Nexus in Malaysia: A Multivariate Time Series
Analysis. MPRA No. 28403.
George, E.H. and Nickoloas, G.T. (2011). The Effect of Energy Consumption on Economic Growth on countries Economic Efficiency: A Conditional
Nonparametric Approach: MPRA Paper No. 28692.
Ghali, El-Sakka. (2004). Energy Use and Output Growth in Canada: A Multivariate Cointegration Analysis. Energy Economics 26: 225-238.
Magazzino, C. (2011). Energy Consumption and aggregate Income in Italy.MPRA Paper No. 28494 pp. 11-55.
Manuse, O. and Nezon, K. (2009). Energy Consumption and GDP Growth in WAEMU Countries: A Panel Data Analysis. MPRA Paper No. 15521.
Mehrara, M. (2007). Energy Consumption and Economic Growth: The Case of Oil Exporting Countries. Energy Policy, Vol.35, pp.2939-2945.
Mulugeta, S.K., Nondo, C., Schaeffer, P.V., and Gebremedhin, T.G. (2010). Does level of income matter in the energy Consumption and GDP Nexus: Evidence
from Sub-Sahara African Countries. Research Paper#7, Research Institute, West Virginia University.
Nigerian National Bureau of Statistics, Abuja, Nigeria.
Odhiambo, N.M. (2009). Energy Consumption and Economic Growth Nexus in Tanzania: An ARDL Bounds Testing Approach. Energy Policy, Vol.37, pp.
617-622.
Pesaran, M. H., Y. Shin, and R. J. Smith, 2001, Bounds Testing Approaches to the Analysis of Level Relationships. Journal of Applied Econometrics, Vol. 16:
289-226.
Soytas, U., Sari, R., 2003. Energy consumption and GDP: causality relationship in G-7 countries and emerging markets. Energy Economics 25, 3337.
Stern, D., 1993. Energy use and economic growth in the USA, a multivariate approach. Energy Econ.15, 137150.
Stern, D., 2000. A multivariate co-integration analysis of the role of energy in the US macroeconomy22, 267283.
Von, P.N. (2009). Energy Consumption and Income: A Semi-parametric panel Data Analysis. Energy Economics, Vol.32 (3), pp.557-563.
Yu, E., Jin, J., 1992. Co-integration tests of energy consumption, income and employment. Res. Energy 14, 259266.