THE JOURNAL OF ENERGY

AND DEVELOPMENT

Slim Mahfoudh and Mohamed Ben Amar,

“The Importance of Electricity Consumption

in Economic Growth:
The Example of African Nations,”
Volume 40, Number 1

Copyright 2015
 THE IMPORTANCE OF ELECTRICITY
CONSUMPTION IN ECONOMIC GROWTH:
THE EXAMPLE OF AFRICAN NATIONS

Slim Mahfoudh and Mohamed Ben Amar*

Introduction

A t present, work on energy consumption is topical, especially for African

countries. Used by economic agents for their production activities and for
their daily needs, electricity is an essential component of any nation’s develop-
ment process. According to statistics from the International Energy Agency (IEA)
published in 2009, there is a strong correlation between electricity consumption
and the wealth of a country. Similarly, it is also recognized that a lack of access to
modern energy services is correlated with a high number of people living on less
than U.S. $2 per day.1 At the micro level, empirical studies establish that access to
electricity seems to be one of the most crucial services for improving the welfare
of poor individuals.2 Thus, for African nations to reach the economic level of
developed ones, policies aimed at meeting the challenge of high and sustained

*Slim Mahfoudh, Assistant Professor in the Higher Institute of Business Administration of Sfax
(Tunisia), holds a master’s degree and Ph.D. in economics from the University of Sfax. His principal
research areas are energy economics, economic policies, and economic growth in emerging
countries. The author is also a member of the Research Unit of Development Economy (URED)
and has presented several papers at various international conferences.
Mohamed Ben Amar, Assistant Professor at the Faculty of Economics and Management of Sfax,
received his Ph.D. in economics from the Faculty of Economics and Management of Sfax, and his
M.A. and B.A. in economics from Sfax University. He is also a member of URED. The author’s
research specialization is economics of African nations and econometrics. His most recent works
have appeared in The Journal of Energy and Development, International Journal of Economics, The
International Journal of Economics and Finance, and Revue Européenne du Droit Social.

The Journal of Energy and Development, Vol. 40, Nos. 1 and 2

Copyright Ó 2015 by the International Research Center for Energy and Economic Development
(ICEED). All rights reserved.
99
100 THE JOURNAL OF ENERGY AND DEVELOPMENT

growth are needed, with particular focus being placed on the electricity sector. In
order to make the electricity sector more dynamic and able to advance economic
growth, many African states are engaged in structural adjustment programs. In this
paper we shall try to identify a relationship between electricity consumption as
a factor of energy consumption and economic growth as a factor of development
in 19 African economies. Moreover, based on the Granger test, we will investigate
the direction of causality between electricity consumption and economic growth.

Literature Review

The challenge with the research on the energy-growth causality for developed
countries, which adopt energy as a proxy of energy usage, is that they often present
conflicting results (e.g., D. Stern, Y. Glasure, K. Ghali and M. El-Sakka, C.-Y. Ho
and K. Siu, and J. Payne).3
At the same time, papers on developing countries that take electricity use as an
indicator of energy consumption produce different results. For example, in the
case of Pakistan, A. Aqeel and M. Butt show one-way causation flowing from
electricity use to Pakistan’s economy.4 For the same country, M. Shahbaz and
H. Lean test the relationship between electricity consumption and economic
growth by introducing capital and labor in the production function and conclude
that a dependence exists between electricity consumption and economic growth
and bidirectional causality between the series.5
In the case of the Turkish economy, G. Altinay and E. Karagol assess the
relationship between electricity consumption and economic growth and find that
electricity consumption causes economic growth, based upon the Granger test.6
Conversely, A. Acaravci and I. Ozturk reported unidirectional causality running
from electricity consumption to economic growth for Turkey.7
This array of results also holds true for our area of research—African
nations—where a number of country-specific studies have been conducted. For
South Africa’s economy, N. Odhiambo tested causality between electricity con-
sumption and economic growth and he found a feedback hypothesis between
them.8 Similarly, C. Jumbe, I. Ouédraogo, and A. Kouakou concluded that
a bidirectional relationship exists between electricity consumption and eco-
nomic growth in Malawi, Burkina Faso, and Côte d’Ivoire, respectively.9 On
the other hand, in another article by N. Odhiambo that focused on Kenya’s
economy, the causality between electricity consumption and economic growth
with an intermediate variable (labor participation) was evaluated and the author
found that economic growth is Granger-caused by electricity consumption.10
Similarly, in a third article, N. Odhiambo reached the same conclusion on the
relationship between electricity consumption and economic growth using a bi-
variate system for Tanzania.11
 AFRICAN NATIONS: ELECTRICITY & GROWTH 101

For Ghana’s economy, P. Adom researched the causal relationship between

electricity consumption and economic growth and concluded that electricity
consumption is Granger-caused by economic growth, implying a growth-led en-
ergy hypothesis.12 For the same country, P. Kwakwa added another component
into the investigation of the relationship between electricity consumption and
economic growth—the consumption of fossil fuels.13 The author found that
electricity consumption and fossil fuel consumption Granger cause economic
growth. Moreover, P. Adom et al. provided another examination of the relation-
ship between electricity consumption and economic growth for Ghana by using an
electricity demand function and employing autoregressive distributed lag (ARDL)
bounds testing.14
In the case of Nigeria, S. Solarin and M. Bello studied the electricity–growth
nexus by introducing capital and labor into the production function.15 They con-
firmed the presence of a growth hypothesis, which required the need for explo-
ration of new energy sources to support economic growth.
Finally, in the case of Algeria, F. Bélaid and F. Abderrahmani tested the
causal relationships between electricity consumption, petroleum prices, and
economic growth.16 They found a feedback effect between electricity consumption
and economic growth and a neutral hypothesis was indicated for electricity
consumption and petroleum prices. We now turn our attention to our modeling
methodology.

Modeling Strategy

Our objective in this section is to study the effects of electricity consumption on

per-capita gross domestic product (GDP). The first step is to select an adequate
model; we have followed the recommendation by Mankiw, Romer, and Weil and
have chosen the Solow-Swan model,17 which is given as:

Yt ¼ K at ðAt Lt Þ1a ð1Þ

g t þrt u
with Lt ¼ Lo e and At ¼ A0 e
nt
; where Y is the real output; K is the stock of
physical capital; L is the labor; A is the factor reflecting the level of technology and
the efficiency in the economy; n is the rate of the labor force growth; g is the rate
of technological progress, which is supposed constant; r is the vector representing
electricity consumption and other factors that can affect the level of technology
and efficiency in the economy; u is the vector of coefficients related to these
variables; and the subscript t indicates time.
The technological progress (At) can explain the existing relationship between
electricity consumption and economic growth. We consider the fact that electricity
consumption can encourage a more rapid formation of capital and this is due to
new technologies.
102 THE JOURNAL OF ENERGY AND DEVELOPMENT

Additionally, this relationship between electricity consumption and growth can

be explained by the Schumpeterian process of “creative destruction,” whereby the
negative shocks are catalysts of reinvestment and modernization in equipment.
These shocks create, thanks to technological progress, a higher level of growth.
The evolution of the economy is represented by equation (2):
dK t
K_t ¼ ¼ sk Yt  dK t ð2Þ
dt
where sk is the rate of investment in physical capital. We suppose that K_ t = It – dKt
and It = St, with d being the rate of physical capital depreciation.
We consider that the production by unit of labor and the physical capital stock
by unit of labor are given by:
Yt Kt
yt ¼ ; kt ¼
At Lt At Lt

We can write the evolution of physical capital stock by unit of labor as:


_ d Kt
kt ¼
d t At Lt

The evolution of the physical capital by unit of effective labor is as follows:


_ d Kt
kt ¼
d t At Lt

K_ t ðAt Lt Þ  ðAt Lt Þ9K t

k_t ¼
ðAt Lt Þ2

_ 
_ K_ t At Lt þ Lt A_ t K t
kt ¼ 
At Lt At Lt At Lt

_ 
_ sk Yt  dK t At L_t
kt ¼  þ kt
At Lt At Lt
with:
A_ t
¼ g: the rate of technical progress supposed exogenous, and
At
L_t
¼ n: the rate of demographic growth.
Lt Yt Kt
Therefore, k_t ¼ sk d  ðg þ nÞk t
At Lt At Lt
k_t ¼ sk yt  ðn þ g þ dÞk t ð3Þ
 AFRICAN NATIONS: ELECTRICITY & GROWTH 103

The evolution of the production by unit of effective labor is as follows:

K at ðAt Lt Þ1a
yt ¼
At Lt
K at
yt ¼ ðAt Lt Þ1a
At Lt
K at
yt ¼ ðAt Lt Þ: ðAt Lt Þa
At Lt
K at
yt ¼
ðAt Lt Þa
 a
Kt
yt ¼
ð At Lt Þ
yt ¼ k at ð4Þ

The substitution of equation (4) into equation (3) produces equation (5):

k_t ¼ sk k at  ðn þ g þ dÞk t ð5Þ

At the equilibrium, we have: k_t ¼ 0. This result leads us to the following relation
as provided by equation (6):
sk k a ¼ ðn þ g þ dÞk ð6Þ

The determination of the steady state of the economy is given as follows, while
taking into account the relation in equation (6):
sk k a ¼ ðn þ g þ dÞk

sk k a1 ¼ ðn þ g þ dÞ
sk
k 1a ¼
ðn þ g þ dÞ
 1=
 sk 1a
k ¼ ð7Þ
ð n þ g þ dÞ

The economy converges toward a steady state represented by the relation in

equation (7). According to the relation in equation (4), we have:
104 THE JOURNAL OF ENERGY AND DEVELOPMENT
 
Yi  a
¼ k i
Ai Li
 a
yi ¼ ðAi Þ k i ð8Þ

Equation (8) represents the output by worker to the equilibrium and for every
country.
At the equilibrium, the technological progress is represented by equation (9):

Ai ¼ A0 eri ui ð9Þ

where r represents the variables that correspond to the factors that can influence
the technological progress. In our study, r regroups the variables reflecting the
electricity consumptions.
The substitution of equation (7) and equation (9) into equation (8) produces
equation (10):
 1a
a
 sk
ðyÞ ¼ A0 e ru
ð10Þ
nþgþd

To have a linear relation, we apply the logarithm:

2 !1ab 3
1b b
a
 a 1a 1ab
b
sk sh sk sh
LnðyÞ ¼ Ln4A0 er u 5

ð11Þ
nþgþd nþgþd

For the production per capita through time and by country, while adding the
indications of times and individual countries, we can rewrite equation (10) into the
following equation:
    a   a  
Ln yi;t ¼ Ln A0;i þ ui ri;t þ Ln sk i;t  Ln ni;t þ g þ d ð12Þ
1a 1a
   
where A0;i , sk i;t , ni;t , g, and d are, respectively, the constant of every nation, the
physical capital reserves, the growth rate of the labor force, the growth rate of
technological progress, and the rate of investment depreciation. The rates g and d
are supposed to be constant through time and across countries with their sum equal
to 0.05.18
As we already have seen, the level of electricity consumption is considered
a key factor in explaining the different levels of the economic development among
nations. Electricity consumption allows for the accumulation of physical capital
stock while incorporating new technologies.
 AFRICAN NATIONS: ELECTRICITY & GROWTH 105

Econometrically, the model can be written as equation (13):

gdpi;t ¼ ai þ becit þ gcapit þ dlabit þ eit ð13Þ
with gdpi,t denoting GDP per capita; ai representing individual effects with ai
2 ℝ; eci,t denoting electricity consumption; capi,t representing physical capital
stock; labi,t is a grouping of the growth rate of the labor force, the growth rates of
technological progress, and investment depreciation. Knowing that the growth
rates of technological progress and of depreciation are supposed to be constant
through time and across countries, their sum (g +d) is equal to 0.05. (b, g, d) is
a vector of the coefficients to estimate and eit stands for the error term.
Presentation of the Variables: The variables that will be presented cover the
period 1990–2010 and are collected for a panel of 19 African nations: Algeria,
Botswana, Cameroon, Congo, Côte d’Ivoire, Democratic Republic of Congo, Egypt,
Ethiopia, Gabon, Ghana, Kenya, Sudan, Morocco, Mozambique, Senegal, South
Africa, Tanzania, Tunisia, and Zambia. The variables in our study are the real gross
domestic product per capita as an endogenous variable, the electricity consumption
per capita, the physical capital stock, and the labor force. All the variables are
extracted from the World Bank’s World Development Indicators.19
Dependent Variable: In our empiric analysis, we use the logarithm of gross
domestic product per capita as the dependent variable.
Physical Capital Stock: We calculate the physical capital stock by the loga-
rithm of gross fixed capital formation as a percentage of GDP.
Labor Force: According to the World Bank, the logarithm of the labor force is
estimated as the total active population.
Electricity Consumption: This is measured as the logarithm of electric power
consumption—kilowatt hour (kWh) per capita.

Estimation Methodology

Panel Unit Root Test: A series of unit root tests is the first step in our sta-
tionary analysis for the panel data estimation. We examine our data by performing
several unit root tests in a panel framework based on K. Im et al. (hereafter referred
to as the IPS test).20 These are the more frequently used when the temporal di-
mension is limited. The authors propose tests permitting the detection of the
presence of the unit root in the models using Fisher-augmented Dickey-Fuller
(ADF) statistics.
The results of the IPS unit root test are given in table 1 and show that this
standardized statistic converges slightly toward the standard normal distribution,
which allows a comparison of the critical values of the distribution N (0,1). The
verification of the non-stationary properties for all variables of the panel leads us
to study the existence of a long-term relation between the variables.
106 THE JOURNAL OF ENERGY AND DEVELOPMENT

Table 1
a
TEST RESULTS USING THE UNIT ROOTS IPS METHOD

Variables gdp ec cap lab

Level –1.173 –1.246 –0.278 –0.669

First difference –3.411* –6.022* –2.218** –7.512*

a
gpd = gross domestic product per capita; ec = energy consumption; cap = physical capital; lab =
labor; * = significance at the 1-percent level; ** = significance at the 5-percent level; and
*** = significance at the 10-percent level.
Source: Calculations by authors based upon results of the IPS test.

Cointegration Test: After testing for unit roots, the long-run relationships
are assessed using the panel cointegration technique based upon the work of
P. Pedroni (see table 2).21 The Pedroni approach allows for heterogeneity among
individual members of the panel. The estimated residual indicates the devia-
tion from the long-run relationship. With the null of no cointegration, the pa-
nel cointegration is a test of unit roots in the estimated residuals of the panel.
The Pedroni test utilizes seven different statistics: the panel v-statistic, panel rho-
statistic, panel PP-statistic, and panel ADF-statistic (which are panel cointegration
statistics and are based on the within approach) and the group rho-statistic, group
PP-statistic, and group ADF-statistic (which are group panel cointegration sta-
tistics and are based on the between approach). In the presence of a cointegrating
relationship, the residuals are expected to be stationary. The panel v-test is a one-
sided test and the null of no cointegration is rejected when the test has a large

Table 2
a
COINTEGRATION TEST RESULTS USING THE PEDRONI METHOD

Statistics With Trend

Panel v-Statistic 6.579*

Panel rho-Statistic 4.175
Panel PP-Statistic 0.713
Panel ADF-Statistic –2.573*
Group rho-Statistic 6.354
Group PP-Statistic –0.469
Group ADF-Statistic –3.152*

a
* = significance at the 1-percent level; ** = significance at the 5-percent level; and
*** = significance at the 10-percent level.
Source: Calculations by authors based upon results of the Pedroni test.
 AFRICAN NATIONS: ELECTRICITY & GROWTH 107

positive value. The other statistics reject the null hypothesis of no cointegration
when they have large negative values.
The simulations from the Pedroni methodology demonstrate that for values of
T higher than 100, the seven statistics give comparable results in terms of po-
tentiality. For smaller sample sizes, the panel-ADF and group-ADF tests have
better properties than the other tests. Therefore, from the Pedroni cointegration test
results, we confirm the existence of a cointegration relationship.
To estimate systems of cointegrated variables on the panel data, it is necessary
to apply an efficient estimation methodology. We identify the following techniques:
the fully modified ordinary least squares (FMOLS) method used by Pedroni and the
dynamic ordinary least squares (DOLS) method. C. Kao and M.-H. Chiang sug-
gest a fully modified (FM) and DOLS estimators in a cointegrated regression and
show that their limiting distribution is normal.22 The authors suggest that the
DOLS estimator may be more promising than OLS or FM estimators in estimating
the cointegrated panel regressions. We provide the results of the estimations via
the FMOLS and DOLS methods in table 3. The lags used in the DOLS method are
r = –1 and r = –2.
The results of our estimation confirm that the level of electricity consumption
exercises a positive and statistically significant effect on the level of GDP per
capita in our sample of 19 African countries. The result can explain and justify the
contribution of the evolution of electricity consumption in African countries
during the latter years of their economic growth—an evolution that is due to
a number of economic reforms.23 Thus, the outcomes of these reforms are justified
by the positive effects upon physical capital accumulation and the improved
quality of the labor force on economic growth.
The effect exercised by electricity consumption, confirmed by our econometric
study, can also contribute to the transformation of the continent’s potential into
concrete achievements and development in the coming years. African nations have

Table 3
REGRESSION RESULTS OF THE FULLY MODIFIED ORDINARY LEAST SQUARES
a
(FMOLS) AND THE DYNAMIC ORDINARY LEAST SQUARES (DOLS) METHODS
(The endogenous variable is gross domestic product per capita–gdp)

Variables ec cap lab

FMOLS 0. 36* 0. 066* 0. 257*

DOLS 0. 152* 0. 104* 0. 312*

a
The lags used in the DOLS method are r = –1 and r = –2; ec = energy consumption; cap =
physical capital; lab = labor; * = significance at the 1-percent level; ** = significance at the 5-percent
level; and *** = significance at the 10-percent level.
Source: Calculations by authors based upon results from the FMOLS and DOLS methods.
108 THE JOURNAL OF ENERGY AND DEVELOPMENT

achieved high performance in terms of economic growth, but the most important
issues are the continuity of that growth, on the one hand, and, on the other hand,
the development of the endogenous factors of wealth creation.
Causality Test: Our next step was to conduct Granger causality tests on the
panel data.24 As the variables are I(1) and are cointegrated, an error-correction
model such as the one presented can be used to identify the direction of causality.
The optimal lag structure of one year is chosen using the Schwarz-Bayesian cri-
terion and the 2-year lags of the dependent variables are used as instruments in the
estimation. The significance of the causality results are determined by the Wald F-test.
The results displayed in table 4 show that there is a unidirectional causality
from GDP to electricity consumption at the 10-percent significance level, that is,
an increase in GDP will bring about an increase in electricity use. This result
confirms the conservation hypothesis.25 In this situation, energy-saving policies
will not have a negative impact on economic growth.

Concluding Remarks

Throughout the world, economic growth is viewed as a metric for develop-

ment, which thus necessitates the examination of factors impacting this growth.
Many recent studies demonstrate that one of the variables influencing economic
growth is energy consumption.
In this paper we have tried to show the existence of a relationship between
electricity consumption and economic growth. First, and according to the FMOLS
and DOLS models, we have identified that electricity consumption positively and
significantly affects economic growth. Additionally, based upon our Granger test
results, we were able to identify that it is GDP variation that causes electricity
consumption. From this, we can assert the existence of a unidirectional-causality
relationship between electricity consumption and economic growth.

Table 4
a
CAUSALITY TEST RESULTS

Direction of causality
gdp ec cap lab

gdp - 2.659*** 5.655* 5.178*

ec 0.174 - 1.151 4.966*
cap 7.626* 3.103** - 0.147
lab 1.497 0.779 0.161 -

a
gpd = gross domestic product per capita; ec = energy consumption; cap = physical capital; lab =
labor; * = significance at the 1-percent level; ** = significance at the 5-percent level; and
*** = significance at the 10-percent level.
 AFRICAN NATIONS: ELECTRICITY & GROWTH 109

Thus, it can be concluded that policy makers in African nations that face
economic challenges must provide more electricity to boost economic activity.
Likewise, for African states attempting to further their development, policy makers
need to prioritize the provision of increasing electricity supplies to keep up with
growing demand and thus not dampen economic growth. Finally, according to the
results of this research, we can say that African countries are now forced to provide
more electricity to their people not just because it is a vital and critical component
for improved standards of living but also to achieve their economic objectives.
NOTES
1
International Energy Agency (IEA), World Energy Outlook 2002 (Paris: IEA, 2002).
2
Ibid.
3
D. I. Stern, “A Multivariate Cointegration Analysis of the Role of Energy in the US Macro-
economy,” Energy Economics, vol. 22, no. 2 (2000), pp. 267–83; Y. U. Glasure, “Energy and
National Income in Korea: Further Evidence on the Role of Omitted Variables,” Energy Eco-
nomics, vol. 24, no. 4 (2002), pp. 355–65; K. H. Ghali and M. I. T. El-Sakka, “Energy Use and
Output Growth in Canada: A Multivariate Cointegration Analysis,” Energy Economics, vol. 26, no.
2 (2004), pp. 225–38; C.-Y. Ho and K. W. Siu, “A Dynamic Equilibrium of Electricity Con-
sumption and GDP in Hong Kong: An Empirical Investigation,” Energy Policy, vol. 35, no. 9
(2007), pp. 2507–513; and J. Payne, “On the Dynamics of Energy Consumption and Output in the
US,” Applied Energy, vol. 86, no. 4 (2009), pp. 575–77.
4
A. Aqeel and M. S. Butt, “The Relationship between Energy Consumption and Economic
Growth in Pakistan,” Asia-Pacific Development Journal, vol. 8, no. 2 (2001), pp. 101–10.
5
M. Shahbaz and H. H. Lean, “The Dynamics of Electricity Consumption and Economic
Growth: A Revisit Study of Their Causality in Pakistan,” Energy, vol. 39, no. 1 (2012), pp. 146–53.
6
G. Altinay and E. Karagol, “Electricity Consumption and Economic Growth: Evidence for
Turkey,” Energy Economics, vol. 27, no. 6 (2005), pp. 849–56.
7
A. Acaravci and I. Ozturk, “Electricity Consumption and Economic Growth Nexus: A Mul-
tivariate Analysis for Turkey,” Amfiteatru Economic, vol. 14, no. 31 (2012), pp. 246–57.
8
N. M. Odhiambo, “Electricity Consumption and Economic Growth in South Africa: A Tri-
variate Causality Test,” Energy Economics, vol. 31, no. 5 (2009), pp. 635–40.
9
C. B. L. Jumbe, “Cointegration and Causality between Electricity Consumption and GDP:
Empirical Evidence from Malawi,” Energy Economics, vol. 26, no. 1 (2004), pp. 61–8.;
I. Ouédraogo, “Electricity Consumption and Economic Growth in Burkina Faso: A Cointegration
Analysis,” Energy Economics, vol. 32, no. 3 (2010), pp. 524–31; and A. K. Kouakou, “Economic
Growth and Electricity Consumption in Cote d’Ivoire: Evidence from Time Series Analysis,” Energy
Policy, vol. 39, no. 6 (2011), pp. 3638–644.
10
N. M. Odhiambo, “Electricity Consumption, Labour Force Participation Rate and Economic
Growth in Kenya: An Empirical Investigation,” Problems and Perspectives in Management, vol. 8,
no. 1 (2010), pp. 31–38.
110 THE JOURNAL OF ENERGY AND DEVELOPMENT
11
N. M. Odhiambo, “Energy Consumption and Economic Growth Nexus in Tanzania: An ARDL
Bounds Testing Approach,” Energy Policy vol. 37, no. 2 (2009), pp. 617–22.
12
P. K. Adom, “Electricity Consumption-Economic Growth Nexus: The Ghanaian Case,” In-
ternational Journal of Energy Economics and Policy, vol. 1, no. 1 (2011), pp. 18–31.
13
P. A. Kwakwa, “Disaggregated Energy Consumption and Economic Growth in Ghana,” In-
ternational Journal of Energy Economics and Policy, vol. 2, no. 1 (2012), pp. 34–40.
14
P. K. Adom, W. Bekoe, and S. K. K. Akoena, “Modeling Aggregate Domestic Electricity
Demand in Ghana: An Autoregressive Distributed Lag Bounds Cointegration Approach,” Energy
Policy, vol. 42 (March 2012), pp. 530–37.
15
S. Solarin and M. Bello, “Multivariate Causality Test on Electricity Consumption, Capital, Labour
and Economic Growth for Nigeria,” Journal of Business and Economics, vol. 3, no. 1 (2011), 1–29.
16
F. Bélaı̈d and F. Abderrahmani, “Electricity Consumption and Economic Growth in Algeria:
A Multivariate Causality Analysis in the Presence of Structural Change,” Energy Policy, vol. 55
(April 2013), pp. 286–95.
17
R. M. Solow, “A Contribution to the Theory of Economic Growth,” Quarterly Journal of
Economics, vol. 70, no. 1 (1956), pp. 65–94, and T. W. Swan, “Economic Growth and Capital
Accumulation,” Economic Record, vol. 32 no. 2 (1956), pp. 334–61.
18
N. G. Mankiw, D. Romer, and D. N. Weil, “A Contribution to the Empirics of Economic
Growth,” The Quarterly Journal of Economics, vol. 107, no. 2 (1992), pp. 407–37.
19
World Bank, World Development Indicators (Washington, D.C.: The World Bank, 2014).
20
K. Im, M. Pesaran, and Y. Shin, “Testing for Unit Roots in Heterogeneous Panels,” Journal of
Econometrics, vol. 115, no. 1 (2003), pp. 53–74.
21
P. Pedroni, “Panel Cointegration: Asymptotic and Finite Sample Properties of Pooled Time
Series Tests with an Application to the PPP Hypothesis,” Econometric Theory, vol. 20, no. 3
(2004), pp. 597–625.
22
C. Kao and M.-H. Chiang, “On the Estimation and Inference of a Cointegrated Regression in
Panel Data,” Center for Policy Research Working Papers no. 2, Center for Policy Research,
Maxwell, School, Syracuse University, Syracuse, New York, 1999.
23
International Energy Agency (IEA), op. cit.
24
C. W. J. Granger, “Some Aspects of Causal Relationships,” Journal of Econometrics, vol. 112,
no. 1 (2003), pp. 69–71.
25
C. T. Tugcu, I. Ozturk, and A. Aslan, “Renewable and Non-renewable Energy Consumption
and Economic Growth Relationship Revisited: Evidence from G7 Countries,” Energy Economics,
vol. 34, no. 6 (2012), pp. 1942–950.