International Journal of Energy Economics and

Policy
ISSN: 2146-4553

available at http: www.econjournals.com

International Journal of Energy Economics and Policy, 2015, 5(2), 408-414.

Energy Consumption, Energy Prices and Economic Growth:

Causal Relationships Based on Error Correction Model

Augustine C. Osigwe, Damilola Felix Arawomo*

Nigerian Institute of Social and Economic Research, Nigeria. *Email: [email protected]

ABSTRACT
This study examined the Granger causality of energy consumption, oil price and economic growth in Nigeria. Two sub-categories of energy (kerosene
and electricity) were equally considered. The error correction model framework was used to test the granger causality of the variables. The results for
the total energy showed bidirectional causality between energy consumption and economic growth. As regards electricity, bidirectional causality was
found between electricity consumption and economic growth as well as between electricity consumption and electricity price. No causal relationship
exists among kerosene consumption, kerosene price and economic growth. Based on our findings, we recommend that policies that promote energy
consumption and economic growth be introduced. One way of achieving this is through the adoption of appropriate energy pricing framework that
takes cognisance of both the present and the future generation.
Keywords: Energy Consumption, Economic Growth, Kerosene Consumption, Electricity Consumption
JEL Classifications: C22, Q43, Q48

1. INTRODUCTION economic grid. An adequate, secure and affordable energy supply

is thus needed to meet the needs of the business and domestic users,
The Nigerian economy has had a volatile “growth-history.” From including the transport of people and goods. The required energy
1960 to 1970, the gross domestic product (GDP) recorded an that is capable of stimulating growth can only be made available
annual growth of 3.1%. During the oil boom era (1970-78), GDP to economic agents if the right pricing strategy is adopted.
grew positively by 6.2% annually. However, negative growth
rates were recorded in the 1980s. In the period 1988-1997 which Economists have long believed that commodity prices serve
constitutes the period of structural adjustment and economic adequately to guarantee efficient allocation and distribution
liberalisation, the GDP grew at a positive rate of 4.0 (Ekpo and of goods. For instance, Hayek (2009) argued that a free price
Umoh, 2004). GDP annual growth rate in Nigeria averaged system allows for economic coordination via the price signals
6.13% from 2005 to 2014, reaching an all-time high of 8.60% that changing prices send. Price is seen as a label, a signal, a piece
in 2010 and a record low value of 6.3% in 2012. According to of information that is attached to the good and service traded.
Nigeria NBS (2014), the country’s GDP advanced 7.67% in the Choices about methods of production, amounts to be produced
last quarter of 2013. and consume are based on price information.

Energy plays an important role in economic development (Glasure The results of studies that examined the causal relationships
and Lee, 1997). Its output is being used in various forms (cooking, among energy consumption, energy prices and economic growth
lighting, source of motive power for vehicles and other industrial have been mixed as the method of their analysis also varied. Most
equipment and machinery), which affects household and industrial of these studies adopted the Granger causality test (Kraft and
activities in diverse ways. To achieve economic growth it is Kraft, 1978; Akarca and Long, 1980; Yu and Hwang, 1984; Yu
necessary to have access, at affordable prices, to abundant and and Choi, 1985). In view of this, this paper evaluated the causal
different energy types, primarily commercial, which feed into the relationships among energy consumption, energy prices and

408 International Journal of Energy Economics and Policy | Vol 5 • Issue 2 • 2015
 Osigwe and Arawomo: Energy Consumption, Energy Prices and Economic Growth: Causal Relationships Based on Error Correction Model

economic growth based on the error correction model (ECM). and got to its peak in 2010 when it assumed an all high value of
A further rationale for this study stems from our consideration of 38.92 million L. Before 1992, kerosene sold for <1 Naira/L. After
the two sub-categories of energy consumption and prices: kerosene this point, the price of kerosene has maintained a steady rise.
and electricity. Although Asafu-Adjaye (2000) incorporated price
into his energy consumption and economic growth model, it was 3. LITERATURE REVIEW
left at the aggregate level. Moreover, while he proxied the price of
energy with Consumer Price Index, we adopted the actual prices Kraft and Kraft (1978) found evidence in favour of causality
of energy products in this study. running from gross national product (GNP) to energy consumption
in the United States, using data for the period 1947-1974. Their
The remainder of this paper is structured in the following way. findings were later reinforced by other studies. For instance,
Section 2 presents a brief overview of the economic and energy Akarca and Long (1979) established unidirectional Granger
use profiles of Nigeria. Section 3 centres on literature review causality running from energy consumption to employment
whereas Section 4 briefly describes the theoretical framework (their proxy for economic growth) with no feedback, using US
and Methodology adopted. Section 5 presents and discusses the monthly data for the period 1973-1978. They estimated the long-
empirical results while section six concludes the study. run elasticity of total employment relating to energy consumption
to be −0.1356.
2. ECONOMIC AND ENERGY USE PROFILES
However, these findings of Kraft and Kraft (1978) and Akarca and
Figure  1 presents annual trend of GDP, oil price and energy Long (1979) have been subjected to empirical challenge. A number
consumption in Nigeria. It is evident that energy consumption of studies that emerged afterwards found no causal relationships
has rapidly trended upward. From 36070.719 million kt in 1970, between income (proxied by GNP) (Akarca and Long, 1980; Yu
it stood at 113053.066 million kt in 2010. Percentage growth of and Hwang, 1984; Yu and Choi, 1985; and Erol and Yu, 1987a) and
GDP is riddled with high level of instability, depicting negative energy consumption. Regarding the causal relationship between
values in the years before 1984. Ever since then, it has been foot- energy consumption and employment, Erol and Yu (1987b, 1989)
dragging. Oil price clearly presents a different picture, for almost Yu et al. (1988) and Yu and Jin (1992) found evidence in favour
all the periods, it exhibited a random walk. It touched its peak in of neutrality of energy consumption with respect to employment,
2009 before reverting downward. evidence referred to in the literature as the “neutrality hypothesis.”

Figure  2 captures percentage growth of GDP, electricity Glasure and Lee (1997) tested for causality between energy
consumption and electricity price. Electric power consumption consumption and GDP for South Korea and Singapore using the
stood at 1637 million kWh in 1971 and jived to 4997 million kWh standard Granger test, as well as cointegration and error-correction
in 1980, 8291 million kWh in 1990 and 9109 million kWh in modelling. The cointegration and error-correction modelling
2000. It got to its all high value of 23542 million kWh in 2012. results indicated a bidirectional causality between income
Electricity per kWh from 1970 to 1986 did not reach 50 kobo. It and energy for both countries. Conversely, using the standard
rose to 3 Naira, 24 kobo in 1995 and later came down to 1 Naira, Granger causality tests, they found no causal relationships
54 kobo in the following year. Beyond this period, it began to between energy consumption and GDP for South Korea and a
increase until it got to 8 Naira in 2012. unidirectional Granger causality from energy consumption to
GDP for Singapore.
Figure 3 presents percentage growth of GDP, kerosene consumption
and kerosene price. Kerosene consumption rose consistently from Asafu-Adjaye (2000) estimated the causal relationships
1970 to 1988 (14.75 million L to 37.88 million L). From 1977 between energy consumption and income for India, Indonesia,
to 2012, it recorded its lowest value in 2004 (19.94 million L) the Philippines and Thailand, using cointegration and error-

Figure 1: Percentage growth of gross domestic product, oil price and energy consumption

International Journal of Energy Economics and Policy | Vol 5 • Issue 2 • 2015 409
 Osigwe and Arawomo: Energy Consumption, Energy Prices and Economic Growth: Causal Relationships Based on Error Correction Model

Figure 2: Percentage growth of gross domestic product, electricity price and electricity consumption

Figure 3: Percentage growth of gross domestic product, kerosene price and kerosene consumption

correction modelling techniques. The results of his study will necessarily miss important indications of both levels and
indicated that, in the short-run, unidirectional Granger causality changes of standard of living.
runs from energy to income for India and Indonesia, while
bidirectional Granger causality runs from energy to income for Adeniran (2008) tested for causal relationship between energy
Thailand and the Philippines. In the case of Thailand and the consumption and GDP in Nigeria using systematic econometric
Philippines, Asafu-Adjaye found that energy, income and prices techniques. The study found that there is a unidirectional causality
are mutually causal. The study results do not support the view that that runs from GDP to electricity consumption. The study also
energy and income are neutral with respect to each other, with the found that GDP granger causes gas consumption. However, his
exception of Indonesia and India where neutrality was observed analyses revealed no causality between oil consumption and GDP.
in the short-run. In a spate study, Al-Iriani (2006) applied In the aggregate, the study disclosed that energy consumption
panel analyses to investigate income and energy consumption granger causes economic growth in Nigeria; implying that a policy
relationship for the Gulf Cooperation Council countries for the to reduce energy consumption aimed at reducing greenhouse gas
period 1971-2002. His study was a foremost attempt towards emissions is likely to have a detrimental impact on the nations
bringing the major energy (oil) exporting countries into the GDP.
mix of countries covered by this line of study. Al-Iriani found
cointegration and unidirectional causality that flows from GDP Rafiq and Salim (2011) examined the short-  and long-run
to energy consumption. causal relationship between energy consumption and GDP of
six emerging economies of Asia. Based on cointegration and
Joyeux and Ripple (2007) employed state-of-the-art panel vector error correction modelling the empirical results their
cointegration techniques to evaluate the nature of the relationship study revealed that there exists unidirectional short- and long-run
between income measures and energy consumption measures causality running from energy consumption to GDP for China, uni-
for seven East Indian Ocean countries. Their general finding directional short-run causality from output to energy consumption
was that income and household electricity consumption are not for India, whilst bi-directional short-run causality for Thailand.
cointegrated. Given this finding, they conclude that standard They found neutrality between energy consumption and income in
of living measures that rely on income measures and do not the case of Indonesia, Malaysia and Philippines. In their study, both
include household-level energy consumption information the generalized variance decompositions and impulse response

410 International Journal of Energy Economics and Policy | Vol 5 • Issue 2 • 2015
 Osigwe and Arawomo: Energy Consumption, Energy Prices and Economic Growth: Causal Relationships Based on Error Correction Model

functions confirmed the direction of causality. These findings, bidirectional relationship between oil consumption and real GDP
according to them, have important policy implications for the growth appears in India. Further, the results proposed that while
countries concerned. The results, they maintained, suggest that nuclear energy stimulates economic growth in both South Korea
while India may directly initiate energy conservation measures, and India, the rapid increase in China economic growth requires
China and Thailand may opt for a balanced combination of additional usage of nuclear energy. In 2015, he changed the focus
alternative polices. of his study by examining the causal relationship between the
same set of variables for four industrialised countries; the US,
Kaplan et al. (2011) studied the causal relationship between energy Canada, Japan, and France. The results showed that there is one-
consumption and economic growth for Turkey during 1971-2006. way causality from nuclear energy consumption to economic
They employed two multivariate models, namely, demand model growth in Japan. Conversely, he found that increasing real GDP
and production model, based on vector ECM. Then, they tested causes additional nuclear energy consumption in France. For the
Granger causality after finding cointegration among variables for US and Canada, he found evidence that supported the neutrality
the both models. The results of their study indicated that energy hypothesis. Naser (2015) finally recommended that policies in the
consumption and economic growth are cointegrated and there developed countries should endeavour to overcome the constrains
is bidirectional causality running from energy consumption to on nuclear energy consumption to face any un-expected hikes in
economic growth and vice versa. This, according to them, means oil prices, which may adversely affect economic growth in such
that an increase in energy consumption directly affects economic oil importing countries.
growth and that economic growth also stimulates further energy
consumption. Consequently, they concluded that energy is a In an attempt to provide reasons for the disparate and often
limiting factor to economic growth in Turkey and, hence, shocks conflicting empirical findings on the relationship between energy
to energy supply will have a negative impact on economic growth consumption and economic growth, Asafu-Adjaye (2000) blamed it
and vice versa. on the variety of approaches and testing procedures employed in the
analyses. According to him, several studies employed simple log-
Shaari et al. (2012) evaluated the relationship between energy linear models estimated by ordinary least squares without any regard
consumptions and economic growth in Malaysia using data from for the nature of the time series properties of the variables involved.
1980 to 2010. They Johansen co-integration was employed for He recognised that most economic time series are non-stationary
the data analyses. Findings of their study showed that energy in levels form and as such, failure to account for such properties
consumptions are related to economic growth. They used the could result in misleading relationships among the variables. These
Granger causality model to measure the causal effect of energy challenges have been adequately addressed in this study.
consumption and GDP. The results indicated that oil and coal
consumption does not Granger cause economic growth and vice 4. THEORETICAL FRAMEWORK AND
versa. Causality, according to the results of their study, runs from METHODOLOGY
economic growth to electricity consumption. They also discovered
that a unidirectional relationship exists between gas and economic 4.1. Theoretical Framework
growth, with causality running from electricity use to economic In view of the objectives of this paper and following Asafu-Adjaye
growth. They took the stand that a policy to reduce gas utilization (2000) we specify the inter linkage among energy consumption,
will harm economic growth in Malaysia. energy price and economic growth. Aside the aggregate energy
consumption and price analysis, this paper also considers two
Abalaba and Matthew (2013) analysed the trend of energy sub-categories of energy consumption and prices: kerosene and
consumption, real output, financial development, monetary electricity.
policy rate and consumer prices and also examined the long-
run relationship and direction of causality between energy 4.2. Methodology
consumption and economic growth with consideration for The Granger causality test within an ECM framework was used
financial development, monetary policy rate and consumer to estimate the inter-linkage among the variables. The ECM
prices. The result showed that all the variables used in the study specification was used given that the variables were integrated of
are characterized by a positive trend. The study provided weak order one (1) and were as well cointergrated.
evidence in support of long-run relationship between energy
consumption and economic growth. They concluded that energy Starting with aggregate energy consumption, oil price and
consumption only has short-run positive impact on the economy economic growth, we have:
but has not enhanced long-run economic growth in Nigeria during
the period under investigation. ∆grgdp t =α 1 + α 2 ∆enrcon t-1 + α 3 ∆oilp t-1 + α 4 ecm1t-1 + ξ1t  (1)

Naser (2014) empirically examined the relationship between ∆enrcon t =α 1 + α 2 ∆grgdp t-1 + α 3 ∆oilp t-1 + α 4 ecm1t-1 + ξ 2t  (2)
oil consumption, nuclear energy consumption, oil price and
economic growth in four emerging economies (Russia, China, ∆oilp t =α 1 + α 2 ∆enrcon t-1 + α 3 ∆grgdp t-1 + α 4 ecm1t-1 + ξ3t  (3)
South Korea, and India) over the period from 1965 to 2010. The
results suggested that there is a unidirectional causality running Then electricity consumption, electricity price and economic
from real GDP to oil consumption in China and South Korea, while growth:

International Journal of Energy Economics and Policy | Vol 5 • Issue 2 • 2015 411
 Osigwe and Arawomo: Energy Consumption, Energy Prices and Economic Growth: Causal Relationships Based on Error Correction Model

∆grgdp t =β1 + β 2 ∆kercon t-1 + β 3 ∆kerprt-1 + β 4 ecm1t-1 + ψ 1t  (4) Kercon: Litters of kerosene.

∆kercon t =β1 + β 2 ∆grgdp t-1 + β 3 ∆kerprt-1 + β 4 ecm2 t-1 + ψ 2t  (5) kerpr: Price of kerosene per litter.

∆kerprt =β1 + β 2 ∆kercon t-1 + β 3 ∆grgdp t-1 + β 4 ecm3t-1 + ξ3t  (6) elecon: Electricity consumption is measured with electric power
consumption (kWh).
For kerosene consumption, kerosene price and economic growth:
elepri: Price of electricity per kWh.
∆grgdp t =δ1 + δ 2 ∆elecon t-1 + δ 3 ∆eleprt-1 + δ 4 ecm1t-1 + ζ 1t  (7)
5. EMPIRICAL ANALYSIS
∆elecon t =δ1 + δ 2 ∆grgdp t-1 + δ 3 ∆eleprt-1 + δ 4 ecm2 t-1 + ζ 2t (8)
Reported in Table  1 is the result of both ADF and PP test for
∆eleprt =δ1 + δ 2 ∆elecon t-1 + δ 3 ∆grgdp t-1 + δ 4 ecm3t-1 + ξ3t  (9)
stationary of the variables. It shows that the null hypothesis of
nonstationarity cannot be rejected at the 5% level for the variables
where, grgdp is economic growth, enrcon is aggregate energy
levels. When the first differences of the variables were taken, the
consumption, oilp is crude oil price, kercon is kerosene
null hypothesis of nonstationarity was rejected for all the variables
consumption, kerpr is kerosene price, elecon is electricity
(Table 1).
consumption, elepri is electricity price and ecm is the error-
correction model.
The two methods of unit root test were adopted in order to establish
consistency in the results generated. Since all the variables were
The ECM opens an additional causality channel which is
integrated of order one, I(1), we proceeded to test for cointegration
overlooked by the standard Granger (1969) and Sims (1972) testing
procedures. In the Granger sense a variable X causes another using the Johansen cointegration test (Table  2). Table  2 shows
variable Y if the current value of Y can better be predicted by using that in the aggregate energy model, the null hypothesis of no
past values of X than by not doing so. The Granger causality testing cointegration relationships was rejected against the alternative
procedure involves testing the significance of the conditional on of two cointegrating relationships at the 1% level. Similar results
the optimum lags. were obtained in the case of kerosene and an electricity model,
as the null hypotheses of no cointegration relationship among
The Granger causality of the dependent variables was tested the variables was again rejected against the one cointegrating
as follows: (1) by a simple t-test of the coefficients of the ecm; relationship.
(2) by a joint Wald F-test of the significance of the sum of the
lags of each of the explanatory variables in turn; and (3) by a The existence of cointegrations among energy consumption,
joint Wald F-test of the interaction of the coefficients of ecm with energy prices and economic growth was suggestive of at least one
corresponding variables in the models. The annual time series for direction Granger causality among the variables. This however
Nigeria cover the period 1970 and 2012. The data were sourced would not indicate the direction of the temporary causality. The
from World Development Indicators 2013 and from the energy ECM results would then be required to indicate the direction of
regulatory agencies. causation as well as in identifying the differences between the
long-run and short-run Granger causality.
Variable are defined thus:
Presented in Table 3 is the joint Wald F-statistics of the lagged
grgdp: Growth rate of GDP (%). explanatory variables of the ECM. The results show that there
were short-run causal effects among the variables. Also provided
enrcon: Commercial energy use, (kt of oil equivalent). is the t-statistics for the coefficients of the ECMs which give an
indication of long-run causal effects. The joint Wald F-statistics
oilp: Price of barrel of crude oil. for the interactive terms, that is, the ECM and the explanatory

Table 1: Results of the unit root tests

Variables Augmented dickey fuller (ADF) Phillips Perron (PP)
Levels First differences Levels First differences
grgdp −2.581 −10.505 −3.181 −11.731 I (1)
Aggregate energy variable
enrcon −2.489 −6.130 −2.341 −6.136 I (1)
oilp −2.278 −11.175 −1.637 −10.379 I (1)
Kerosene variable
kercon −2.523 −5.739 −1.928 −8.648 I (1)
kerpr −0.188 −3.574 0.614 −3.397 I (1)
Electricity variable
elecon −1.975 −6.066 −1.949 −8.719 I (1)
elepr 0.995 −8.502 0.977 −8.707 I (1)

412 International Journal of Energy Economics and Policy | Vol 5 • Issue 2 • 2015
 Osigwe and Arawomo: Energy Consumption, Energy Prices and Economic Growth: Causal Relationships Based on Error Correction Model

Table 2: Results of Johansen’s cointegration test (intercept, no trend)

Hypothesized Eigenvalue Trace statistic 5% critical value P**
Aggregate energy model
None* 0.477879 38.04101 29.79707 0.0045
At most 1* 0.406701 17.24562 15.49471 0.0270
At most 2 0.016727 0.539783 3.841466 0.4625
Trace test indicates 2 cointegrating eqn (s) at the 0.05 level
Kerosene model
None* 0.475294 31.83208 29.79707 0.0287
At most 1 0.249427 11.83967 15.49471 0.1648
At most 2 0.090633 2.945207 3.841466 0.0861
Trace test indicates 1 cointegrating eqn (s) at the 0.05 level
Electricity model
None* 0.477385 33.80661 29.79707 0.0164
At most 1 0.251849 10.44584 15.49471 0.2480
At most 2 1.19E‑05 0.000429 3.841466 0.9853
Trace test indicates 1 cointegrating eqn (s) at the 0.05 level. *Denotes rejection of the hypothesis at the 0.05 level

Table 3: Causality between energy consumption, energy price and economic growth of using ECM framework
Dependents Short‑run effects Sources of causation
variables Wald F‑statistics t‑ratio Wald F‑statistics
Total energy grgdp enrcon oilp ect1 only grgdp enrcon oilp
grgdp ‑ 8.61** 1.45 −3.21** ‑ 4.19** 3.49
enrcon 7.34** ‑ 0.34 −1.29 3.92** ‑ 0.37
oilp 1.52 1.40 ‑ −1.26 0.76 0.89 ‑
Kerosene grgdp kerocon keropr ect only grgdp kerocon keropr
grgdp ‑ 1.87 1.87 1.90 ‑ 2.41 4.71
kerocon 1.74 ‑ 0.65 1.56 3.71 ‑ 1.79
keropr 0.17 0.41 ‑ −1.89 4.00 1.69 ‑
Electricity grgdp eletcon elecpr ect only grgdp elctcon elecpr
grgdp ‑ 12.1*** 1.20 −2.17** ‑ 12.9** 0.21
elctcon 5.23** ‑ 7.67** −3.92** 4.43** ‑ 1.23
elecpr 0.45 4.51** ‑ −3.78** 1.90 2.38 ‑
ECM: Error correction model, 1: ECT-error correction term in the error-correction model. Significance at the UUU1% level, the 0.05% leveland the 0.10% level. *, **, ***represents 0.10,
0.05 and 0.01

variables is also captured in the models. This indicated that the consumption in the same way that electricity consumption granger
variables bear the burden of short-run adjustment to re-establish causes the price of electricity.
long-run equilibrium.
6. CONCLUSION
Starting with the short-run result of the total energy (Table 3), it
showed that the F-statistics for energy (in the economic growth This study examined the Granger causality of energy consumption,
equation) was significant. Similarly, the F-statistics of economic oil price and economic growth in Nigeria. Two sub-categories
growth (in the energy consumption equation) was equally of energy (kerosene and electricity) were as well considered.
significant in both short and long-run. The result implied that The ECM framework was used to test for the granger causality
there was bidirectional Granger causality between economic among the variables. The results showed for the total energy
growth and energy consumption. Also indicated in the result equations, bidirectional causality between energy consumption
was the fact that oil price had a neutral effect on both energy and economic growth. As regards electricity, bidirectional
consumption and economic growth. For kerosene, the result causality was discovered between electricity consumption and
showed that there was no causal relationship among grgdp, economic growth. Similarly there was bidirectional causality
kerocon and keropr. Moreover, none of the interactive terms were between electricity consumption and electricity price. No causal
statistically significant, implying that the long-run relationship relationship was dictated among kerosene consumption, kerosene
among kerosene consumption, kerosene price and economic price and economic growth. In general, the study does not support
growth were weak. the hypothesis of a neutral relationship between energy and
economic growth, except for kerosene.
The economic growth equation indicated that electricity
consumption granger-causes economic growth in both short-run Based on the findings, we therefore, recommend that policies that
and long-run while the electricity consumption equation showed promote energy consumption and economic growth be introduced.
that economic growth granger-cause electricity consumption. One way of achieving this is through the adoption of appropriate
This indicated bidirectional granger causality. This result further energy pricing framework that takes cognisance of both the present
revealed that price of electricity granger causes electricity and the future generation.

International Journal of Energy Economics and Policy | Vol 5 • Issue 2 • 2015 413
 Osigwe and Arawomo: Energy Consumption, Energy Prices and Economic Growth: Causal Relationships Based on Error Correction Model

REFERENCES models and cross-spectral methods. Econometrica, 37(3), 424-438.

Hayek, F.A. (2009), Denationalisation of money: The argument refined.
Abalaba, B.P., Matthew, A.D. (2013), Energy Consumption and economic Auburn: Ludwig von Mises Institute.
growth nexus: new empirical evidence from Nigeria. International Joyeux, R., Ripple, R.D. (2007), Household energy consumption versus
Journal of Energy Economics and Policy, 3(4), 412-423. income and relative standard of living: a panel approach. Energy
Adeniran, O. (2008), Does energy consumption cause economic growth? Policy, 35, 50-60.
An empirical evidence from Nigeria. United Kingdom: University Kaplan, M., Ozturk, I., Kalyoncu, H. (2011), Energy Consumption and
of Dundee. economic growth in Turkey: cointegration and causality analysis.
Akarca, A.T., Long, T.V. (1979), Energy and employment: a time series Romanian Journal of Economic Forecasting, 2, 31-41.
analysis of the causal relatonship. Resources Energy, 2, 151-162. Kraft, J., Kraft, A. (1978), On the relationship between energy and GNP.
Akarca, A.T., Long, T.V. (1980), On the relationship between energy and Journal of Energy Development, 3, 401-403.
GNP: a re-examination. Journal of Energy Development, 5, 326-331. Naser, H. (2014), Oil market, nuclear energy consumption and economic
Al-Iriani, M.A. (2006), Energy-GDP relationship revisited: an example growth: evidence from emerging economies. International Journal
from GCC countries using panel causality. Energy Policy, 34, of Energy Economics and Policy, 4(2), 288-296.
3342-3350. Naser, H. (2015), Can nuclear energy stimulates economic growth?
Asafu-Adjaye, J. (2000), The relationship between energy consumption, Evidence from highly industrialised countries. International Journal
energy prices and economic growth: time series evidence from Asian of Energy Economics and Policy, 5(1), 164-173.
developing countries. Energy Economics, 22, 615-625. Rafiq, S., Salim, R.A. (2011), The linkage between energy consumption
Ekpo, A.H., Umoh, O.J. (2004), Heath and education sectors in Nigeria's
and income in six emerging economies of Asia: an empirical analysis.
National Development, 1970-2000. International Journal of Social
International Journal of Emerging Markets, 6(1), 50-73.
Science, 3(1), 124-135.
Shaari, M.S., Hussain, N.E., Ismail, M.S. (2012), Relationship between
Erol, U., Yu, E.S.H. (1987a), Time series analysis of the causal
energy consumption and economic growth: empirical evidence for
relationships between US energy and employment. Resources
Energy, 9, 75-89. Malaysia. Business Systems Review, 2(1), 17-28.
Erol, U., Yu, E.S.H. (1987b), On the causal relationship between Yu, E.S.H., Choi, J.Y. (1985), The causal relationship between energy and
energy and income for industraialised countries. Journal of Energy GNP: an international comparison. Journal of Energy Development,
Development, 13, 113-122. 10, 249-272.
Erol, U., Yu, E.S.H. (1989), Spectral analysis of the relationship between Yu, E.S.H., Choi, P.C.Y., Choi, J.Y. (1988), The relationship between
energy and income for industrialised countries. Journal of Energy energy and employment: a re-examination. Energy Systems Policy,
Development, 13, 113-122. 11, 287-295.
Glasure, Y.U., Lee, A.R. (1997), Cointegration, error-correction, and the Yu, E.S.H., Hwang, B.K. (1984), The relationship between energy and
relationship between GDP and energy: the case of South Korea and GNP: further results. Energy Economics, 6, 186-190.
Singapore. Resource and Energy Economics, 20, 17-25. Yu, E.S.H., Jin, J.C. (1992), Cointegration tests of energy consumption,
Granger, C.W.J. (1969), Investigating causal relations by econometric income and employment. Resources Energy, 14, 259-266.

414 International Journal of Energy Economics and Policy | Vol 5 • Issue 2 • 2015