Renewable and Sustainable Energy Reviews 76 (2017) 6271

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

The relationship amongst energy consumption (renewable and non- MARK

renewable), and GDP in Algeria
Fethi Amri
University of Chartage, Faculty of Economic Sciences and Management of Nabeul, Street Hammamet, Mrezgua, Nabeul 8000, Tunisia

A R T I C L E I N F O A BS T RAC T

Keywords: This study aims to explore economic growth-energy consumption nexus in Algeria between 1980 and 2012.
Renewable energy consumption With cointegration tests, we demonstrate the existence of long-run link between real gross domestic product,
Non-renewable energy consumption real capital and the two categories of energy consumption i.e renewable energy per capita and non-renewable
Cointegration energy. The long-run and the short-run autoregressive distributed lag (ARDL) estimates indicate that only the
Causality test
non-renewable energy sort and capital can contribute to enhancing economic growth whereas renewable energy
Algeria
does not show any signicant eect. The outcome of causality tests proves a feedback link among non-renewable
energy consumption and gross domestic product, among capital and gross domestic product, and among non-
renewable energy and capital, in both short-run and long-run terms. Moreover, the results reveal a
unidirectional link going from renewable energy to economic growth, capital, and non-renewable energy
respectively, in the long run. Furthermore, the results illustrate the presence of a unidirectional link going from
non-renewable energy to renewable energy in the short term. Our outcome suggests that policy makers in
Algeria should enhance the renewable energy share together with controlling the non-renewable one.

1. Introduction
The exhaustion of non-renewable category of energy, the increased
energy demand, the need to achieve the sustainable development goals,
and the consideration of health eects are the important reasons which
motivate countries to promote the renewable energy sources. In
accordance with International Renewable Energy Agency [1], 164
nations have adopted no less than one class of renewable energy target.
This fact encourages researchers to integrate the renewable energy
variable on the energy consumption-economic growth nexus [2].
The selection of Algeria for observational examination is activated
by the ensuing fundamental reasons. Firstly, Algeria gured from the
countries which have abundant fossil fuels resources. Algeria is
included in the group of countries which have a massive reservoir of
fossil fuel resources. In 2013, Algeria recorded 12.2% of the world oil
reserves and 4.5% of the world natural gas reserves. Algeria has the
seventeenth greatest oil reserves worldwide and the fourth greatest
ones in Africa. It has also the tenth greatest natural gas stock world-
wide and the second greatest one in Africa [3] (Table 1).
Algeria is likewise gured from the biggest fossil fuel producers
worldwide. In 2013, the Algerian oil production was estimated at 1485
thousand barrels per day. Algeria is considered the eighteenth major oil
producer in the world and the third important producer in Africa.
While the Algerian natural gas production was measured by 81.5
billion cubic meters, it has been ranked number ten in the world and
number one in Africa [3] (Table 1).
Secondly, this abundance of fossil fuels resources has created an
economy which is intensely dependent on the natural resource sector.
Indeed, total natural resource rents, oil rent, and natural gas rents
accounted for 28.16%, 21.61% and 6.10% of GDP in 2013, respectively
[4]. According to International Monetary Fund [5], Oil and gas export
represented 98.1% of Algerian exports in 2013. Moreover, Algerian is
one of the biggest consumers of subsidized energy. The Algerian
government has adopted a policy of energy subsidies by xing admin-
istratively the prices of electricity, fuel and set administratively under
its market value [6]. In 2013, the Algerian oil consumption and gas
consumption were estimated at 390 thousand barrels per day and 33.4
billion cubic meters respectively [3] (Table 1). It is considered the
thirty-sixth biggest consumers of oil in the world and the third one in
Africa. It also belongs to the rst twenty-six biggest nations in terms of
gas consumers worldwide and the second substantial one in Africa.
Between 1980 and 2013, oil and natural gas consumption decreased
from 191 (in thousand barrels per day) and 19.8 (in billion cubic
meters) to 390 (in thousand barrels per day) and 33.4 (in Billion cubic
meters), respectively (Fig. 1). In accordance with International Energy
Agency [7], total energy consumption (in thousand tons of oil
equivalents) soared from 12,722 to 31,894 between 1990 and 2013
(Fig. 1). In 2013, the residential sector represented the biggest share of

E-mail address: fethiamri.fsegn@gmail.com.

http://dx.doi.org/10.1016/j.rser.2017.03.029
Received 6 June 2016; Received in revised form 13 December 2016; Accepted 8 March 2017
1364-0321/ 2017 Elsevier Ltd. All rights reserved.
F. Amri Renewable and Sustainable Energy Reviews 76 (2017) 6271

Table 1
Algerian oil and gas statistics in 2013 [3].

Total Proportion in the world Rank in the world Proportion in Africa (%) Rank in Africa

Oil proved reserves (thousand million barrels) 12.2 0.71 17 9.37 4

Oil production (thousand barrels daily) 1485 1.71 18 17.10 3
Oil consumption (thousand barrels daily) 390 0.42 36 10.39 3
Natural gas proved reserves (trillion cubic meters) 4.5 2.41 10 31.69 2
Natural gas: production (billion cubic meters) 81.5 2.39 9 39.85 1
Natural gas consumption (billion cubic meters) 33.4 0.98 28 27.76 2

Fig. 1. Evolution of energy indicators [3].
energy consumption at 42.61% of total energy consumption, trailed by
the transportation and industrial ones at 36.03% and 21.35% respec-
tively.
Thirdly, the Algerian model of energy consumption is based on the
fossil fuels which represented 99.89% in the entirety consumption of
energy in 2013. The biggest share is accorded to the natural gas with
62.59%, followed by oil with 36.91% and coal with 0.39% respectively
(Fig. 2). Renewable energy represents only 0.1% of Algerian energy
consumption. Non-renewable electricity (in billion Kilowatt-hours)
jumped from 5.915 to 42.868 between 1980 and 2012, whereas
renewable electricity went up from 0.248 to 0.616 during the same
period [7] ( Fig. 3 and Table 2).
Fourthly, the growth of energy consumption has been associated
with a decline in natural gas and oil production. Oil production
increased (in thousand barrels per day) from 1138.86 in 1980 to
1989.93 in 2005. Oil production (in thousand barrels per day) declined
from 1989.93 in 2005 to 1485 in 2013. The Algerian natural gas
production (in billion cubic meters) rose from 14.17 in 1980 to 86 in
1999. Natural gas production (in billion cubic meters) went down from
86 in 1999 to 81.5 in 2013 [3] (Fig. 1). Fifthly, this extensive use of oil
and natural gas is associated with C02 emissions repercussion [6]. As
0.05%
0.39% 0.10%
Natural Gas
36.91% Oil
Coal
Hydro electric
62.59%
Renewables
Fig. 2. Energy consumption by fuel (Million tons oil equivalent) in 2013 [7].
Fig. 3. Evolution of non-renewable electricity consumption and renewable electricity
consumption [7].
Table 2
Electricity net generation by fuel (billion kilowatt-hours) 2012 [7].
Nuclear Renewable Fossil fuels Hydroelectric Total
pumped storage
Generation 0 0.616 53.369 0 53.985
from
Share (%) 0 1.14 98.85 0 100
indicated by the Environmental Performance Index (EPI) [8] in 2014,
Algeria was ranked 92 on the list of 178 nations for their ecological
result. As pointed up by U.S Energy Information Administration [7],
carbon dioxide emissions (million tons CO2) moved from 41.1 in 1980
to 125.5 in 2013.
Sixthly, to satisfy the increasing needs of energy demand of its
population, the decrease in oil and natural gas production, and to
promote a clean environment, Algeria has increased its eorts in
matters of development of other alternative energies such as renewable
energy. Algeria forecasts the production of 20% of its energy from
renewable energy by 2030 [1]. To reach this target, Algeria has adopted
numerous options such as the development of manufacturing ability to
create the solar, wind, hydroelectric, geothermal and biomass energies,
along with a progressive legislative framework for the renewable energy
sector, the reinforcement of structures of renewable energy, the
adoption of dierent nancial and tax benets, and the cooperation
with the European Union in relation to non-renewable and renewable
energy in 2013 [9].
To our knowledge, there is no study which is interesting in the
nexus among renewable energy consumption and output (which is
measured by the gross domestic product) in Algeria. The study goal is
to inspect the link among energy consumption and output in Algeria by
assuming dierent components of energy i.e renewable and non-
renewable between 1980 and 2012.
This document is planned as follows. Section 2 describes material
and methods. Section 3 provides the main results and debates, and
Section 4 integrates conclusions and policy implications.

63
F. Amri
2. Material and methods
2.1. A brief literature review
The studies that looked into the link amongst GDP and energy
consumption in Algeria are limited. For example, Belaid and
Abderrahmani [10] checked the link between economic growth,
petroleum price, and electricity consumption by using cointegration
and causality tests over the 19712010 period. The results of time
series analysis reveal the presence of a feedback linkage between GDP
and energy consumption into the short-term and the long-terms. Also,
there is no causality across energy consumption and petroleum price.
Fuinhas and Marques [11] employed data of Algeria and Egypt
between 1965 and 2010 to check the nexus amongst output and energy
consumption. The outcomes of an autoregressive distributed lag model
demonstrate a bidirectional negative nexus between the two variables
in Algeria. Cher and Kourbali [12] analysed the connection between
the same variables from 1965 to 2008 in the Algerian cross-sectional
study. The outcome provides a lack of cointegration link among them.
In addition, the nal outcomes accept the causal hypothesis going from
output indicator to energy indicator and reject the causality assump-
tion going from energy consumption to GDP. Salman and Atya [13]
focused on time series data of Egypt, Tunisia, and Algeria over the
period 19802010. The outcome predicts an insignicant impact of
GDP on energy consumption in Algeria. Ziramba [14] investigated the
link between economic growth and hydroelectricity consumption in the
cases of Egypt, Algeria, and South Africa from 1980 till 2009. The
causality test results provide a bidirectional linkage amongst economic
growth and hydroelectricity consumption in Algeria. Wolde-Rufael [15]
employed causality test on a sample of nineteen African nations
between 1971 and 2001. They demonstrate unidirectional causality
moving from GDP to energy consumption between 1971 and 2001 in
Algeria. Wolde-Rufael [16] used econometric time series models of
seventeen African economies to test the similar linkage between 1971
and 2001. The outcomes demonstrate the absence of causality among
GDP and electricity consumption in Algeria. Wolde-Rufael [17]
exploited a data of seventeen countries to test the same relationship
between 1971 and 2004. The estimated results of a multivariate model
demonstrate unidirectional causality moving from energy to GDP in
Algeria. Bouoiyour et al. [18] focused on the impact of economic,
foreign direct investment, trade, urbanization, institutions and eco-
nomic growth in energy consumption in Algeria between 1971 and
2012. The outcomes demonstrate the presence of cointegration links
between variables. Besides, the results indicate a small eect of
economic growth.
Except the Algerian case, this theme has attracted a number of
other studies in the case of one or multiple countries. A large number of
these studies have inspected the link among GDP and non-renewable
category of energy (Apergis and Payne [19]; Belloumi [20]; Otzurk
et al. [21]; Belke et al. [22]; Abbas and Choudhury [23]; Ouedraogo
[24]; Bozoklu and Yilanci [25]; Omri [26]; miech and Papie [27];
Alshehry and Belloumi [6]; Iyke [28]; Caraiani et al. [29]; Tang et al.
[30]; Liddle and Lung [31]; Faisal et al. [32]; Chiou-Wei et al. [33];
Adams et al. [34]; Lu et al. [35]; Wolde-Rufael [36]; Sadorsky [37]).
This rst class of studies energy distinguished between four
hypotheses (Bhattacharya et al. [38]; Alper and Oguz [39]). Firstly,
the growth hypothesis proves a unidirectional link going from energy
consumption variable to economic growth one.
Secondly, the neutrality hypothesis indicates the independence
between the two considered variables. Thirdly, the conservation
hypothesis illustrates a unidirectional link going from economic growth
variable to energy consumption one. Fourthly, the feedback hypothesis
proves a bidirectional link between the considered variables.
A set of papers examine the association between renewable category
of energy consumption and GDP (Apergis and Payne [40]; Menegaki
[41]; Pirlogea and Cicea [42]; Yildirim et al. [43]; Ocal and Aslan
64
Renewable and Sustainable Energy Reviews 76 (2017) 6271
[44], Lin and Moubarek [45]; Ohler and Fetters [46]; Koak and
arkgnei [47]; Sebri and Ben Salha [48]; Zeb et al. [49]; Omri et al.
[50]; Sebri [51]; Shahbaz et al. [52]; Inglesi-Lotz [53]; Alper and Oguz
[39]; Bhattacharya et al. [38]; Wesseh and Lin [54]; Destek [55]; Amri
[56]).
A further range of studies investigates the link between the two
categories of energy consumption such as renewable and non-renew-
able, and output (Apergis and Payne [57]; Apergis and Payne [58];
Tugcu et al. [59]; Pao and Fu [60]; Al-mulali et al. [61]; Ben Jebli and
Ben Youssef [62]; Salim et al. [63]; Amri [64]; Bowden and Payne [65];
Dogan [66]; Dogan [2]).
The third category of papers explores the connection among both
types of energy (renewable and non-renewable), and GDP. Some
studies have focused on the case of multiple countries [2]. While a
few numbers of studies have focused on the case of one country. These
latter studies are related with the objective of our study.
Apergis and Payne [57] investigated this nexus for emerging
countries by using panel data method from 1990 to 2007. The results
indicate the presence of feedback linkage among non-renewable energy
and GDP in the short and long-terms. However, there is a unidirec-
tional link moving from renewable energy to economic growth in short
term and feedback link across non-renewable and economic growth.
Apergis and Payne [58] used a data of 80 countries from 1990 to 2007.
The panel cointegration results provide a long-term link between the
three considered variables. The results also provide feedback links
between the energy consumption in both forms and GDP in the long
term and the short term. Moreover, the outcomes demonstrate a
feedback linkage among the two sorts of energy in the short term.
Tugcu et al. [59] used a data for G7 countries between 1980 and 2009.
They have illustrated a variety of results, diering according to the
function productions used. In Brazilian study Pao and Fu [60] scanned
the linkage amongst GDP and dierent indicators of energy between
1980 and 2010. The long-term outcome demonstrates a feedback link
between total renewable kind of energy and gross domestic product, a
unidirectional link moving from non-hydroelectric renewable energy
consumption to GDP, a unidirectional association going from GDP to
total primary energy consumption, and a unidirectional association
going from GDP to non-renewable energy consumption. Furthermore,
the short-term results demonstrate the appearance of a feedback
association among GDP and non-hydroelectric renewable energy
consumption. In a study encompassing 18 American countries, Al-
mulali et al. [61] indicated the appearance of a bidirectional relation-
ship between the two forms of energy consumption on the rst side and
GDP on the second one. Also, the outcomes reveal that renewable
energy consumption ingredient contributes more than non-renewable
energy consumption ingredient. Ben Jebli and Ben Youssef [62]
integrated openness as an additional variable in the nexus among
GDP and the mentioned elements of energy using a data of 69
economies from 1980 till 2010. The short-term results unveil the
absence of a direct link between renewable and non-renewable energy
and GDP and the presence of an indirect link among them throughout
openness. Furthermore, the long-term results indicate a unidirectional
link moving from non-renewable energy consumption to GDP and a
feedback relationship amongst this latter and renewable energy con-
sumption.
Salim et al. [63] demonstrated the appearance of a long-run
relationship among industrial output, economic growth and the two
sources of energy in OECD countries between 1980 and 2011. The
outcome of panel causality tests demonstrates the appearance of a
unidirectional impact going from renewable energy category to eco-
nomic growth. In addition, the outcome demonstrates the presence of
feedback link among economic growth and non-renewable energy in
the short term. Moreover, there is a bidirectional link among both
kinds of energy and industrial output.
Amri [64] focused on the relationship among energy consumption
by sources, economic growth and foreign direct investment by compar-
F. Amri Renewable and Sustainable Energy Reviews 76 (2017) 6271

ing developed countries and developing ones between 1990 and 2010. Table 3
The outcome accepts the feedback hypothesis. Summary statistics.
Dogan [66] examined the same relationship in the case of Turkey
Gross Capital Renewable Non Population
from 1990 to 2012. The outcomes of this study indicate the absence of domestic energy renewable
a unidirectional link from renewable energy to GDP and from non- product consumption energy
renewable energy to output, in the short term. Furthermore, the results consumption
illustrate the presence of unidirectional causality from renewable
Maximum 1.24E+11 4.57E+10 0.639 42.868 37,439,427
energy and output, and the existence of bidirectional causality among Minimum 5.08E+10 1.46E+10 0.053 5.915 19,337,723
non-renewable energy and output, in the long term. Mean 7.98E+10 2.19E+10 0.277 18.420 28,767,198
Bowden and Payne [65] employed industrial, residential, and Standard 2.15E+10 8.48E+09 0.159 9.407 5,261,520
commercial data between 1949 and 2006 for the United States. The devia-
tion
outcome proves the existence of neutrality among renewable energy
Coecient 0.269 0.387 0.576 0.510 0.182
and GDP in the case of industrial and commercial sectors. Moreover, of
the feedback hypothesis is validated between non-renewable energy varia-
consumption kind and GDP in the case of commercial and residential tion
sectors. Moreover, the outcome reveals a unidirectional link going from
renewable energy kind to GDP in the case of the residential sector, and
logarithmic form, we establish the following model:
a unidirectional link going non-renewable energy category to GDP in
the industrial sector. ln RGDPCt = 0 + 1 ln RKCt + 2 ln NRENCt + 3 ln RENCt + t (2)
Dogan [2] re-examined the same relationship in the case of Turkey
from 1980 to 2012, by exploiting the Johansen cointegration test, the RGDPC, NRENC, RENC, and RKC are the real gross domestic
ARDL cointegration test, the Hansen-Gregory cointegration test, and product per capita, the per capita non-renewable form of energy, the
VECM Granger causality test in the context of structural break analysis. per capita renewable sort of energy and the real capital stock per capita,
The outcome proves the presence of positive eects of non-renew- respectively.
able energy consumption, capital, and labor on GDP. The outcome
indicates the presence of short-run link going from GDP to renewable
energy, a long-run feedback link between renewable category of energy 3. Results and discussions
and gross domestic product, and long run and short run feedback links
between the non-renewable energy category and output. In our paper, we use the following steps. First, we analyze the
From the previously mentioned literature, there is no study that statistic descriptive of the interested variables. Second, we check the
assessed the nexus among energy consumption by sources and stationarity of the variables. Third, we test the presence of long-term
economic growth by considering a complete dependent country on linkage between variables. Fourth, we use the ARDL approach to check
fossil fuels energy (more than 99% of total energy). To address this the stability of the model and we estimate the long and short-term
knowledge gap, our paper originally tries to examine the nexus among coecients. Finally, we employ the Granger causality method in order
energy consumption by sources and output. Our study breaks down to test the causality among all variables.
long run and short run estimates and the causality links among output,
and the two forms of energy consumption in Algeria.
3.1. Statistic analysis

2.2. Data and methodology
The yearly information from 1980 to 2012 used in our work is: 1)
the gross domestic product (constant 2005 US dollars), 2) the gross
xed capital formation which considered a proxy of capital stock
(constant 2005 US dollars, 3) The population, 4) total renewable
electricity consumption (billions of kilowatt-hours), 5) total non-
renewable electricity Consumption (billion kilowatt-hours). The rst
three variables are gathered from the World Bank Indicators [4]. The
fourth and fth variables are assembled from the U.S. Energy
Information Administration [7].
Our primary aim is to check the relation between gross domestic
product and the two kinds of energy (renewable and non-renewable
The statistic analysis of the four variables is presented in Table 3.
Renewable energy consumption has the highest coecient of variation
(0.652), followed by non-renewable energy consumption (0.510)
because of its high variability. On the other hand, the lowest variability
is recorded by the population variable with a low coecient of variation
of the order of 0.182. As illustrated in Figs. 48, gross domestic
product, population, non-renewable energy consumption have followed
a constant trend in the considered period.
However, for the capital and renewable energy consumption
variables, we have illustrated dierent phases of evolution during the
period from 1980 to 2012.
1.3E+11

one). 1.2E+11
Our model is derived from the Cobb-Douglas production function
where renewable energy, non-renewable energy, labor, and capital are 1.1E+11
considered as additional factors (Bhattacharya et al. [38]; Pao and Fu
1.0E+11
[60]; Shahbaz et al. [52]; Dogan [66]).
9.0E+10
RGDPt = et TFPRK
t

t Lt NRENi RENt

(1)
8.0E+10
where t: 1,,T signies year (19802012). TFP is total factor produc-
tivity. RGDP reects the GDP at constant prices. NREN is the non- 7.0E+10

renewable energy consumption. REN is the per renewable energy 6.0E+10

consumption. L is the labor force. RK is the real capital stock.
denotes the stochastic error term. 5.0E+10
80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12
, , and are the coecients for each variable.
After dividing each side of Eq. (1) by labor and taking them in Fig. 4. Evolution of gross domestic product.

65
F. Amri Renewable and Sustainable Energy Reviews 76 (2017) 6271

40,000,000
3.2. Stationary tests

36,000,000
To verify the stationary of variables, we employ dierent tests such
as Dickey and Fuller [67], Augmented Dickey and Fuller [67] and
32,000,000
Phillips and Perron [68]. The unit root test is performed in models with
intercept and with intercept and trend.
28,000,000 The outcomes outlined in Table 4 and demonstrate that all variables
at level are I(1). On the contrary, the variables at rst dierence are
24,000,000 stationary.
The above-mentioned traditional unit root tests are not appropriate
20,000,000 in the case of the time series data with a structural break. To avoid the
presence of biased results, some authors (Ben Jebli and Ben Youssef
16,000,000 [69]; Dogan [66]; Iyke [28]) used the Zivot and Andrews [70] unit root
80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12
test which is capable of integrating one break point in the data. Zivot
Fig. 5. Evolution of population. and Andrews [70] tested the non-stationary hypothesis (existence of
unit root) against stationary hypothesis (non-existence of unit root)
45
with the structural break, in specications with intercept, trend and
40
both intercept and trend.
The results of the latter test (see Table 5) conrmed those of the
35 traditional tests. Indeed, the non-stationary hypothesis is accepted for
all variables at the level and the stationary one is approved for all
30
variables at rst dierence.
25
In general, the two categories of unit root tests recommend that real
GDP per capita, real capital, non-renewable energy consumption per
20 capita and renewable energy consumption per capita are I(1) in level
and I(0) in rst dierence. In addition, there is no variable that I(2).
15
These outcomes license to look into the long-run relationship among
10
explored variables.

5 3.3. Cointegration tests

80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12

Fig. 6. Evolution of non-renewable energy consumption. We use the autoregressive distributed lag (ARDL) method to test
the cointegration assumption between the considered variables. This
5.0E+10
method that belongs to Pesaran et al. [71] is better than conventional
4.5E+10 cointegration methodologies. It has several advantages. Firstly, it is
appropriate in the time series data which are integrated in order (1)
4.0E+10 and/or in order (0). Thirdly, this method removes the endogeneity
3.5E+10
problems related to the variables. Fourthly, the short-run and long-run
relationships are evaluated at the same time.
3.0E+10
k k
ln RGDPCt = 0 + j1=1 1j ln RGDPCt j + j 2=1 1j ln RKCt j
2.5E+10
k k
+ j3=1 2j ln RNENCt j + j =1
4
3j ln RENCt j
2.0E+10
+ 1 ln RGDPCt 1 + 2 ln RKCt 1 + 3 ln NRENCt 1
1.5E+10
+ 4 ln RENCt 1 + t (3)
1.0E+10
80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12 : is the rst dierence; : is the error terms, 0 is the intercept term;
Fig. 7. Evolution of capital. 1 4 : are the coecients of the lagged level of each variable; k1 k4 :
are the lagged length of each variable chosen by the Schwarz data
.7 criteria (SIC); the long-run equilibrium relationship among variables is
inspected by the joint signicance of coecients of the lagged level
.6 variables.
Pesaran et al. [51] tested the no cointegration hypothesis between
.5 variables (H0: 1 = 2 = 3 = 4 = 5 = 0 ), against the cointegration one
between variables (H1: 1 2 3 4 5 0 ). The calculated F-
.4 statistic is compared with the lower critical bound value and the upper
critical bound value. This approach marked three possible results: (1)
.3 cointegration link if the calculated F-statistic surpass the upper critical
bound value, (2) no cointegration link if the calculated F-statistic is
.2 smaller than the lower critical bound value, and (3) inconclusive link if
the calculated F-statistic is between the two critical values.
.1 The outcomes of ARDL cointegration test (see Table 6) indicate the
appearance of a cointegration link between real capital per capita,
.0 renewable energy per capita, non-renewable energy per capita, and real
80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12
output per capita. In fact, the calculated F-statistic (6.027) is higher
Fig. 8. Evolution of renewable energy consumption. than the superior critical bound value at 1% (5.61).

66
F. Amri Renewable and Sustainable Energy Reviews 76 (2017) 6271

Table 4
Linear unit root tests.

Variables Augmented DickeyFuller test statistic PhillipsPerron test statistic Dickey Fuller-GLS

Test statistic

Intercept Intercept and trend Intercept Intercept and trend Intercept Intercept and trend

ln RGDPC 0.5602(1) 0.9253 (7) 0.3576(4) 0.9698(3) 0.6859(1) 1.3060(1)

ln RKC 2.1910(4) 0.3742(3) 0.8538(4) 0.1819(1) 1.5406(4) 2.1765(4)
ln RENC 2.4540(0) 3.0258(0) 2.3541(0) 2.9010(1) 1.5635 (0) 2.1924 (0)
ln NRENC 0.0102(0) 1.6315(0) 0.0059 (2) 1.9589(2) 1.5201 (0) 1.6973 (0)
ln RGDPC 2.9733**(0) 3.8535**(0) 3.0348**(3) 3.8864**(3) 2.9986*** (0) 3.2572** (0)
ln RKC 3.8285***(3) 4.4115***(2) 2.9339** (1) 4.0693***(8) 3.0508***(0) 4.1430***(0)
ln RENC 8.7478***(0) 8.7985***(0) 10.2947***(11) 17.5857***(20) 8.3827***(0) 8.8897***(0)
ln NRENC 5.7887***(0) 5.7438***(0) 5.7882***(2) 5.7432***(2) 4.8559***(0) 5.6718***(0)

***
1% symbolizes the signicance at 1% level.
**
Symbolizes the signicance at 5%.

Table 5 Table 8
Nonlinear Zivot-Andrews unit root test. Johansen tests for cointegration results.

Variables Zivot-Andrews unit root test Maximum rank Trace statistic 5% Critical value

Intercept Break year Intercept and trend Break year r=0 51.8333 47.21
r=1 26.5043** 29.68
ln RGDPC 2.810(1) 1990 4.596(1) 1993 r=2 10.0472 15.41
ln RKC 2.201(0) 1987 3.006(0) 1991 r=3 0.1667 3.76
ln RENC 3.774(1) 2003 4.806 (1) 2003
ln NRENC 3.362(0) 1993 3.316(0) 1993 **
Symbolizes the signicance at 5%.
ln RGDPC 4.979**(0) 1995 5.188**(0) 1995
ln RKC 5.509***(0) 1986 5.435**(0) 1986
consolidates the nal outcome of the ARDL cointegration test. The
ln RENC 9.423***(0) 2003 9.322***(0) 2003
ln NRENC 6.571***(0) 1989 6.698***(0) 1996 non-cointegration hypothesis is rejected under the presence of struc-
tural break. There is a long-run link between the variables.
***
1% symbolizes the signicance at 1% level. We use the AIC criteria to x the optimal lag length. The outcomes
**
Symbolizes the signicance at 5%. indicate the presence of two lags in the model. Then, we use the trace
statistics (see Table 8) to determine the number of the cointegrating
Table 6 equations. The null hypothesis of non-cointegration is rejected.
Bound test cointegration results.

Model ln RGDPCt = f (ln RKCt , ln NRENCt , ln RENCt )

3.4. Short and long-run ARDL estimation
Bound test F-statistic 6.027***
Signicance 1% In Table 9 we show the short and long-run ARDL estimation. This
LowerI(0) bound 4.29 table recapitulates some diagnostic tests for residuals such as normal-
Upper I(1) bound 5.61 ity, serial correlation, and heteroskedasticity. In the long term, the per
*** capita capital and the per capita non-renewable energy consumption
1% symbolizes the signicance at 1% level.

Table 9
In addition, we use the Gregory and Hansen [72] test and the
Long and short-run estimates. Selected model: ARDL (1, 1, 0, 1).
Johansen [73] test for cointegration to arm the decision of the ARDL
cointegration test. The results of Gregory and Hansen [72] cointegra- Long-run analysis
tion test (see Table 7) conrmed the results of the ARDL cointegration Variables Coefficient. Standard error T- statistic p-values
test in the case of four regressions (break in the constant term, break in
Constant 9.530*** 1.551 6.143 0.000
the constant and the trend, break in the constant and the slope and ln RKC 0.244*** 0.070 3.487 0.003
break in the constant, the slope and the trend). The non-cointegration ln RENC 0.015 0.035 0.429 0.674
assumption is rejected under the appearance of one structural break. ln NRENC 0.246*** 0.058 4.217 0.000
There is a long-run link between the entire variables. Short-run analysis
ln RKC 0.233*** 0.043 5.338 0.000
The outcome of Johansen cointegration [73] test (see Table 8) ln RENC 0.007 0.004 1.457 0.166
ln NRENC 0.142* 0.075 1.900 0.078
ECT(1) 0.319* 0.151 2.117 0.052
Table 7 R2 0.998
GregoryHansen cointegration test results. Adjusted R2 0.994
F-statistic 185.363*** 0.000
Model Break date ADF-statistic DW-statistic 2.250
Diagnostic tests
Break in the constant term. 1998 5.88*** Tests
Break in the constant and the trend. 2002 6.12*** Serial correlation. 1.096 0.365
Break in the constant and the slope 2004 6.47** Normality 1.498 0.472
Break in the constant, the slope and the trend 2001 6.54** Heteroscedasticity 0.708 0.736

*** ***
1% symbolizes the signicance at 1% level. 1% symbolizes the signicance at 1% level.
** *
Symbolizes the signicance at 5%. Symbolizes the signicance at 10%.

67
F. Amri Renewable and Sustainable Energy Reviews 76 (2017) 6271

are positively attached to the GDP. The outcomes indicate that the rise 1.6

in per capita capital by 1% will enhance the level of per capita GDP by
0.244%. This shows that capital brings the output improvement. This 1.2
outcome is analogous to that of Ocal and Aslan [44], Dogan [66], and
Dogan [2] in Turkey and Shahbaz et al. [52] in Pakistan. In addition,
the increment in per capita non-renewable energy consumption by 1% 0.8
will increase the per capita GDP by 0.246%. This indicates that non-
renewable component of energy consumption is the important con-
0.4
tributor to output in Algeria. This evidence is comparable to that of
Dogan [66] and Dogan [2]. However, in the long run, the renewable
category of energy consumption is unrelated to the GDP. The results 0.0
conrm that Algeria has not reached the renewable energy threshold
that allows it to make a positive contribution on output. This
participation is equal to nearly 0.10% in aggregate Algerian energy -0.4
97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
consumption in 2013. This outcome is like that of Dogan [66] and
Dogan [2]. This result is contrary to that of Ocal and Aslan [44], CUSUM of Squares 5% Significance
Shahbaz et al. [52], and Bhattacharya et al. [38].
In the short run, the capital and the non-renewable energy 12
consumption are connected positively to the GDP. 1% increments in
non-renewable energy consumption advance per capita GDP by 0.142. 8
1% increments in capital advance per capita GDP by 0.233. This
outcome is analogous to that of Dogan [66] and Dogan [2] and Shahbaz 4
et al. [52]. Whereas the short run parameter of renewable energy
variable is insignicant. This result is contrary to that of Shahbaz et al.
0
[52] in Pakistani study. This outcome is analogues to the conclusion of
Dogan [66] and Dogan [2].
The equilibrium correction mechanism coecient (ECM) is nega- -4

tive (between 1 and 0) and statistically signicant. This shows that the
deviation of variables from the short to the long-run equilibrium is -8
balanced by 31.9% per year.
The residual diagnostic tests reject the hypothesis of no serial -12
correlation and no autoregressive heteroscedasticity and accept the 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12
hypothesis of normality of residuals.
CUSUM 5% Significance
The presence of a cointegration link among the considered variables
does not imply that the evaluated parameters are steady. The plots of Fig. 9. CUSUM and CUSUM of squares plots of recursive residual.
cumulative sum of recursive residuals (CUSUM) and CUSUMQ (cu-
mulative sum of the square of recursive residuals) statistics (see Fig. 9) k k
ln RENCt = 0 + i =1 1i ln RENCt i + i =1 2i ln RGDPCt i
fall into the limits of 5% signicance. This result indicates the k
constancy of coecient in time. + i =1 3i ln RKCt i
k
+i =1 4i ln NRENCt i + 4ECTt 1 + 4t (7)
t

3.5. Granger causality tests ECTt1 denotes the lagged error correction terms; 1t -4t denotes the
error terms. , , and denote the coecient parameters; is the
Subsequent to setting up the appearance of long-run association rst dierence operator; k is the lagged length of each variable chosen
among the variables utilizing three methods such as ARDL, Gregory by the Schwarz data criteria (SIC);
Hansen and the Johansen cointegration tests, we are interested now in Then, we determine the causality among variables in the short and
the sense of causality [74] between the considered variables. To reach the long run. The long- run causality is investigated by the statistical
this goal, we perform the VECM (vector error correction model) signicance of the ECTt1 term. More precisely, we tested the null
approach established to the following equations. hypothesis H0: j (j = 1, 2, 3, 4) = 0 . The short-run causality is per-
k k formed by the signicance of the Wald test of the lags of each
ln RGDPCt = 0 + i =1 1i ln RGDPCt i + i =1 2i ln RKCt i
additional Variable in Eqs. (4)(7). In Eq. (4), the short run causality
k
+ i =1 3i ln NRENCt i of capital on economic growth is uncovered by testing the null
k hypothesis H0: 2i = 0 against the alternative hypothesis H1: 2i 0 .
+ i =1 4i ln RENCt i + 1ECTt 1 + 1t (4) In Eq. (5) the short run causality of non-renewable energy consump-
tion on capital is discovered by testing the null hypothesis H0: 4i = 0
k k
ln RKCt = 0 + i =1 1i ln RKCt i + i =1 2i ln RGDPCt i against the alternative hypothesis H1: 4i 0 .
k After proving a long-term link between the inspected variables, the
+ i =1 3i ln NRENCt i
execution of the Granger causality test (see Table 10) remains
k
+i =1 4i ln RENCt i + 2ECTt 1 + 2t (5) necessary to help policy makers to develop a suitable energy strategy.
t
First, we check long-run causality through the statistical signi-
k k cance of the ECT parameters. The ECT parameter comprises between
ln NRENCt = 0 + i =1 1i ln NRENCt i + i =1 2i ln RGDPCt i
1 and 0 and is signicant in the case of Eqs. (4) and (7). In the long
k
+ i =1 3i ln RKCt i term, the outcomes demonstrate the appearance of a feedback causality
k
+i =1 4i ln RENCt i + 3ECTt 1 + 3t link amongst income and non-renewable kind of energy, a feedback
t
causality link between capital and non-renewable kind of energy
(6) consumption, and a bidirectional causality link between gross domestic

68
F. Amri
Table 10
Results of Granger causality tests.
Dependent variable Short-run
ln RGDPC ln RKC
ln RGDPC 7.048 (0.029)**
ln RKC 14.470(0.000)***
ln RENC 0.520(0.470) 2.847(0.240)
ln NRENC 8.405 (0.015)** 9.701 (0.007)***
Values in parentheses in indicate the p-values and values in brackets are t-statistics.
***
1% symbolizes the signicance at 1% level.
**
Symbolizes the signicance at 5%.
*
Symbolizes the signicance at 10%.
product and capital. In addition, there is a one-way causality link from
renewable kind of energy to GDP, a unidirectional causality link from
renewable energy to capital, and a unidirectional causality link from
renewable energy to non-renewable energy.
Second, we examine short-run causality via the signicance of the
Wald test of the lags of Variables. The short-run analysis proves the
presence of feedback causality among capital and output, feedback
causality among non-renewable variety of energy and GDP, and a
feedback link between capital and non-renewable kind of energy.
Moreover, the outcome reveals a unidirectional link from non-renew-
able energy to renewable energy, the absence of causality among
renewable energy and GDP, and the absence of a linkage among
renewable energy and income.
In both short and long terms, the outcomes show the existence of
the feedback link between non-renewable energy consumption and
output and recommend that economic growth in Algeria is reliant on
non-renewable sort of energy and those energy preservation strategies
could antagonistically aect economic development. This nding
arms the feedback hypothesis, which suggests that an increase of
economic growth prompts an increase of utilization of non-renewable
energy and the other way around. This result is coherent with the result
of Salim et al. [63] and the result of Dogan [2]. This result does not
conform with the result of Iyke [28] on a Nigerian study who
demonstrated the appearance of unidirectional causality going from
non-renewable energy consumption to output. Also, this result rejects
the armations of Pao and Fu [60] who demonstrate the presence of
one-way causality going from GDP to non-renewable energy consump-
tion.
The results demonstrate the presence of the feedback link between
economic growth and capital and suggest that GDP per capita and the
level of capital per capita are interrelated in the short run. The outcome
reveals that any uctuation in economic growth inuences capital, and
any uctuation in capital inuences economic growth. This indicates
that output cannot be accomplished in the absence of additional
capital. These results are in agreement with the ndings of Salim
et al. [63] in a study on the OECD countries, and do not conform to the
Apergis and Payne [40] results.
There is long-run feedback causality among GDP and capital. This
means that an increase of economic growth needs more capital. In the
same time, more capital helps more economic growth. Our result
conforms to the result of Dogan [66]. The result does not conform to
the outcome of Al-mulali et al. [61].
There is a one-way long-run link going from renewable component
of energy to GDP. This means that renewable energy aects the output
in Algeria. This outcome is most certainly not coherent with the
armations of Apergis and Payne [40] who demonstrate the appear-
ance of feedback causality in the short and the long terms. This result is
contrary with the armation of Ben Jebli and Ben Youssef [69] who
indicate the absence of a long-run unidirectional link from renewable
energy to output. This result is in agreement with the conclusions of
Dogan [66] in a Turkish study.
In the short run, there is a lack of causality among renewable energy
69
Renewable and Sustainable Energy Reviews 76 (2017) 6271
Long-run
ln RENC ln NRENC ECT (1)
2.257 (0.078) 5.089 (0.078)* 0.470 [1.949]*
1.533(0.464) 8.349 (0.015)** 0.324 [3.705]***
7.966(0.018)** 4.609 [3.409]***
0.363(0.833) 0.597 [2.346]**
consumption and income. This result is conrms to the result of Dogan
[66]. This result is contrary to the armation of Apergis and Payne
[58].
There is a unidirectional short-run link going from non-renewable
component to renewable one. This means that non-renewable energy
demand incites the exploitation of renewable energy in the short term.
Our result is coherent with the nding of Ben Jebli and Ben Youssef
[69] in the case of Tunisia. The outcome is opposed to the consequence
of Dogan [66] in the case of Turkey.
The Algerian development is strongly related to the abundance of
the fossil fuels resources such as oil and natural gas. However, in recent
years, the Algerian energy sector has been vulnerable to the decrease in
oil and gas production and the increase in oil and gas consumption. In
order to avoid the dependence on primary energy and to vary the
energy provenance, Algeria has initiated a plan of renewable energy
and energy development.
There is a one-way long-run causality going from renewable
variable to non-renewable one, generally characteristic of substitut-
ability among them. This illustrates that the part of renewable category
in total energy consumption will increase in the long term. This is
prone to occur given that the consumption of renewable electricity
which expanded signicantly going from 0.248 billion kilowatt-hours
in 1980 to 0.616 billion kilowatt-hours in 2012.
This outcome is coherent with the nding of Ben Jebli and Ben
Youssef [69] in the case of Tunisia. This result does not conform with
the result of Bento and Moutinho [75] in an Italian study.
There is a two-way short-run link amongst capital and non-renew-
able category of energy consumption. This means that capital and non-
renewable energy are interconnected. Our conclusion does not con-
forms to the result of Dogan [66] who indicates the absence of granger
causality relationship between capital and non-renewable energy. This
nding is contrary to Ben Jebli and Ben Youssef [62] in 69 nations'
study who illustrates the absence of granger causality link between
capital and non-renewable energy in the short run.
There is no short-run causality among renewable energy consump-
tion and capital. Our conclusion does not conform with the result of
Dogan [66] who indicates the presence of unidirectional granger
causality from renewable energy to capital. This nding conforms to
Ben Jebli and Ben Youssef [62] in the case of 69 nations' study who
illustrates the absence of granger causality link between capital and
renewable energy in the short run.
There is one-way long-run causality going from renewable energy
consumption to capital. This means that renewable energy aects
capital in Algeria. Our result is in contradiction with the conclusion of
Salim et al. [63] that proves the presence of one-way long-run link
starting from capital to renewable energy consumption. This result
conrms the armation of Dogan [66].
4. Conclusion and policy implications
The target of this paper is to break down the long and short-run
connections and to detect the Granger causality directions among
F. Amri
renewable and non-renewable energy, economic growth and capital in
Algeria from 1980 to 2012. The primary reason is that there is an
absence of study that looks into this relationship for Algeria. This
objective is reached by utilizing three cointegration tests (ARDL,
GregoryHansen and Johansen) together with the VECM Granger
causality.
The exact results of ARDL approach demonstrate that short and
long-run associations take place among capital, output, and non-
renewable energy consumption. But, the ndings reject the possible
liaison among income and renewable sort of energy consumption in the
two long and short terms.
The outcomes of the VECM Granger causality approach indicate the
appearance of feedback causality among non-renewable class of energy
and income, between real capital per capita and income per capita, and
among real capital per capita and non-renewable sort of energy
consumption per capita, in the short term and long terms.
Furthermore, there is a unidirectional link going from renewable sort
of energy to real GDP per capita, a one-way link starting from
renewable energy per capita to real capital per capita and a one-way
link going from renewable energy per capita to non-renewable energy
per capita, in the long term. Moreover, there is a short-run unidirec-
tional link going from non-renewable sort of energy to renewable one,
while there is no short-run causality among renewable energy on one
hand, and capital and GDP on the other hand.
This paper involves several policy implications. First, non-renew-
able energy leads to enhance economic growth. Also, the positive eect
on the output of the adoption of non-renewable energy further
accelerates the forward movement of non-renewable energy consump-
tion use. This indicates that the two variables are interrelated.
Consequently, policy makers should adopt a strategy to make the
manipulation of non-renewable energy ecient.
Second, there is a unidirectional long-run contribution of renewable
category of energy on economic growth. In the light of the conse-
quences of this study, it is prescribed that country in focus ought to
expand its investment in renewable energy to build the part of power
utilization from renewable sources. Moreover, Algeria should employ
renewable energy to respond to the growing energy demand linked to
economic growth.
References
[1] International Renewable Energy Agency http://www.ren21.net; 2015.
[2] Dogan E. Analyzing the linkage between renewable and non-renewable energy
consumption and economic growth by considering structural break in time-series
data. Renew Energy 2016;9:112636.
[3] BP statistical review of world energy http://www.bp.com; 2015.
[4] World development indicators http://data.worldbank.org/country/algeria; 2015.
[5] International Monetary Fund http://www.imf.org); 2015.
[6] Alshehry AS, Belloumi M. Energy consumption, carbon dioxide emissions and
economic growth: the case of Saudi Arabia. Renew Sustain Energy Rev
2015;41:23747.
[7] U.S. Energy Information Administration http://www.eia.gov; 2015.
[8] Environmental performance index http://archive.epi.yale.edu.
[9] Renewable Energy Development Centre www.cder.dz.
[10] Belaid F, Abderrahmani F. Electricity consumption and economic growth in
Algeria: a multivariate causality analysis in the presence of structural change.
Energy Policy 2013;55:28695.
[11] Fuinhas JA, Marques AC. Rentierism, energy and economic growth: the case of
Algeria and Egypt (19652010). Energy Policy 2013;62:116571.
[12] Cher SE, Kourbali B. Energy consumption and economic growth in Algeria:
cointegration and causality analysis. Int J Energy Econ Policy 2012;2:23849.
[13] Salman DM, Atya EM. What is the role of nancial development and energy
consumption on economic growth? New evidence from north African countries. Int
J Financ Bank Stud 2014;4:13749.
[14] Ziramba E. Hydroelectricity and economic growth nexus: time series experience of
three African countries. Eur Sci J 2013;9:8496.
[15] Wolde-Rufael Y. Energy demand and economic growth: the African experience. J
Policy Model 2005;27:891903.
[16] Wolde-Rufael Y. Electricity consumption and economic growth: a time series
experience for 17 African countries. Energy Policy 2006;34:110614.
[17] Wolde-Rufael Y. Energy consumption and economic growth: the experience of
African countries revisited. Energy Econ 2009;31:21724.
[18] Bouoiyour J, Selmi R, Ozturki I. The nexus between electricity consumption and
70
Renewable and Sustainable Energy Reviews 76 (2017) 6271
economic growth: new insights from meta-analysis. Int J Energy Econ Policy
2014;4:62135.
[19] Apergis N, Payne JE. Energy consumption and economic growth in Central
America: evidence from a panel cointegration and error correction model. Energy
Econ 2009;31:2116.
[20] Belloumi M. Energy consumption and GDP in Tunisia: cointegration and causality
analysis. Energy Policy 2009;37:274553.
[21] Ozturk I, Aslan A, Kalyoncu H. Energy consumption and economic growth
relationship: evidence from panel data for low and middle income countries.
Energy Policy 2010;38:44228.
[22] Belke A, Dobnik F, Dreger C. Energy consumption and economic growth: new
insights into the cointegration relationship. Energy Econ 2011;33:7829.
[23] Abbas F, Choudhury N. Electricity consumption-economic growth Nexus: an
aggregated and disaggregated causality analysis in India and Pakistan. J Policy
Model 2013;35:53853.
[24] Ouedraogo NS. Energy consumption and economic growth: evidence from the
economic community of West African States (ECOWAS). Energy Econ
2013;36:63747.
[25] Bozoklu S, Yilanci V. Energy consumption and economic growth for selected OECD
countries: further evidence from the Granger causality test in the frequency
domain. Energy Policy 2013;63:87781.
[26] Omri A. An international literature survey on energy-economic growth nexus:
evidence from country-specic studies. Renew Sustain Energy Rev 2014;38:9519.
[27] miech S, Papie M. Energy consumption and economic growth in the light of
meeting the targets of energy policy in the EU: the bootstrap panel Granger
causality approach. Energy Policy 2014;71:11829.
[28] Iyke BN. Electricity consumption and economic growth in Nigeria: a revisit of the
energy-growth debate. Energy Econ 2015;51:16676.
[29] Caraiani C, Lungu CI, Dasclu C. Energy consumption and GDP causality: a three-
step analysis for emerging European countries. Renew Sustain Energy Rev
2015;44:198210.
[30] Tang CF, Tan BW, Ozturk I. Energy consumption and economic growth in Vietnam.
Renew Sustain Energy Rev 2016;54:150614.
[31] Liddle B, Lung S. Energy consumption and GDP causality: importance of a priori
hypothesis testing, disaggregated data, and heterogeneous panels. Appl Energy
2015;142:4455.
[32] Faisal Tursoy T, Resatoglu NG. Energy consumption, electricity, and GDP causality;
The case of Russia, 19902011. Procedia Econ Finance 2016;39:6539.
[33] Chiou-Wei SZ, Zhen Zhu Z, Chen SH, Hsueh SP. Controlling for relevant variables:
energy consumption and economic growth nexus revisited in an EGARCH-M
(exponential GARCH-in-mean) model. Energy 2016;109:3919.
[34] Adams S, Klobodu EKM, Opoku EES. Energy consumption, political regime and
economic growth in sub-Saharan Africa. Energy Policy 2016;96:3644.
[35] Lu HF, Lin BL, Campbell DE, Sagisaka M, Ren H. Interactions among energy
consumption, economic development and greenhouse gas emissions in Japan after
World War II. Renew Sustain Energy Rev 2016;54:106072.
[36] Wolde-Rufael Y. Electricity consumption and economic growth in transition
countries: a revisit using bootstrap panel Granger causality analysis. Energy Econ
2014;44:32530.
[37] Sadorsky P. Renewable energy consumption and income in emerging economies.
Energy Policy 2009;37:40218.
[38] Bhattacharya M, Paramati SR, Ozturk I, Bhattacharya S. The eect of renewable
energy consumption on economic growth: evidence from top 38 countries. Appl
Energy 2016;162:73341.
[39] Alper A, Oguz O. The role of renewable energy consumption in economic growth:
evidence from asymmetric causality. Renew Sustain Energy Rev 2016;60:9539.
[40] Apergis N, Payne JE. Renewable energy consumption and economic growth:
evidence from a panel of OECD countries. Energy Policy 2010;38:65660.
[41] Menegaki AN. Growth and renewable energy in Europe: a random eect model
with evidence for neutrality hypothesis. Energy Econ 2011;33:25763.
[42] Pirlogea C, Cicea C. Econometric perspective of the energy consumption and
economic growth relation in European Union. Renew Sustain Energy Rev
2012;16:571826.
[43] Yildirim E, Sarac S, Aslan A. Energy consumption and economic growth in the USA:
evidence from renewable energy. Renew Sustain Energy Rev 2012;16:67704.
[44] Ocal O, Aslan A. Renewable energy consumptioneconomic growth nexus in
Turkey. Renew Sustain Energy Rev 2013;28:4949.
[45] Lin B, Moubarak M. Renewable energy consumption economic growth nexus for
China. Renew Sustain Energy Rev 2014;40:1117.
[46] Ohler A, Fetters I. The causal relationship between renewable electricity generation
and GDP growth: a study of energy sources. Energy Econ 2014;43:12539.
[47] Koak E, arkgnei A. The renewable energy and economic growth nexus in black
sea and Balkan countries. Energy Policy 2017;100:517.
[48] Sebri M, Ben-Salha O. On the causal dynamics between economic growth, renew-
able energy consumption, CO2 emissions and trade openness: fresh evidence from
BRICS countries. Renew Sustain Energy Rev 2014;39:1423.
[49] Zeb R, Salar L, Awan U, Zaman K, Shahbaz M. Causal links between renewable
energy, environmental degradation and economic growth in selected SAARC
countries: progress towards green economy. Renew Energy 2014;71:12332.
[50] Omri A, Ben Mabrouk N, Sassi-Tmar A. Modeling the causal linkages between
nuclear energy, renewable energy and economic growth in developed and devel-
oping countries. Renew Sustain Energy Rev 2015;42:101222.
[51] Sebri M. Use renewable to be cleaner: meta-analysis of the renewable energy
consumptioneconomic growth nexus. Renew Sustain Energy Rev
2015;42:65765.
[52] Shahbaz M, Loganathan N, Zeshan M, Zaman K. Does renewable energy con-
F. Amri
sumption add in economic growth? An application of auto-regressive distributed
lag model in Pakistan. Renew Sustain Energy Rev 2015;44:57685.
[53] Inglesi-Lotz R. The impact of renewable energy consumption to economic growth: a
panel data application. Energy Econ 2016;53:5863.
[54] Wesseh PK, Lin B. Can African countries eciently build their economies on
renewable energy?. Renew Sustain Energy Rev 2016;54:16173.
[55] Destek MA. Renewable energy consumption and economic growth in newly
industrialized countries: evidence from asymmetric causality test. Renew Energy
2016;95:47884.
[56] Amri F. Intercourse across economic growth, trade and renewable energy con-
sumption in developing and developed countries. Renew Sustain Energy Rev
2017;69:52734.
[57] Apergis N, Payne JE. Renewable and non-renewable electricity consumption
growth nexus: evidence from emerging market economies. Appl Energy
2011;88:522630.
[58] Apergis N, Payne JE. Renewable and non-renewable energy consumption growth
nexus: evidence from a panel error correction model. Energy Econ 2012;34:7338.
[59] Tugcu CT, Ozturk I, Aslan A. Renewable and non-renewable energy consumption
and economic growth relationship revisited: evidence from G7 countries. Energy
Econ 2012;34:194250.
[60] Pao HT, Fu HC. Renewable energy, non-renewable energy and economic growth in
Brazil. Renew Sustain Energy Rev 2013;25:38192.
[61] Al-mulali U, Fereidouni HG, Lee JYM. Electricity consumption from renewable and
non-renewable sources and economic growth: evidence from Latin American
countries. Renew Sustain Energy Rev 2014;30:2908.
[62] Ben Jebli M, Ben Youssef S. Output, renewable and non-renewable energy
consumption and international trade: evidence from a panel of 69 countries. Renew
Energy 2015;83:799808.
[63] Salim RA, Hassan K, Shaei S. Renewable and non-renewable energy consumption
and economic activities: further evidence from OECD countries. Energy Econ
71
Renewable and Sustainable Energy Reviews 76 (2017) 6271
2014;44:35060.
[64] Amri F. The relationship amongst energy consumption, foreign direct investment
and output in developed and developing countries. Renew Sustain Energy Rev
2016;64:694702.
[65] Bowden N, Payne JE. Sectoral analysis of the causal relationship between renew-
able and non-renewable energy consumption and real output in the U.S. Energy
sources, Part B: economics. Plan Policy 2010;5:4008.
[66] Dogan E. The relationship between economic growth and electricity consumption
from renewable and non-renewable sources: a study of Turkey. Renew Sustain
Energy Rev 2015;52:53446.
[67] Dickey DA, Fuller WA. Distribution of the estimators for autoregressive time series
with a unit root. J Am Stat Assoc 1979;74:42731.
[68] Phillips PCB, Perron P. Testing for a unit root in time series regression. Biometrika
1988;75:33546.
[69] Ben Jebli M, Ben Youssef S. The environmental Kuznets curve, economic growth,
renewable and non-renewable energy, and trade in Tunisia. Renew Sustain Energy
Rev 2015;47:17385.
[70] Zivot E, Andrews D. Further evidence of great crash, the oil price shock and the unit
root hypothesis. J Bus Econ Stat 1992;10:25170.
[71] Pesaran MH, Shin Y, Smith RJ. Bounds testing approaches to the analysis of level
relationships. J Appl Econ 2001;16:289326.
[72] Gregory AW, Hansen BE. Tests for cointegration in models with regime and trend
shifts. Oxf Bull Econ Stat 1996;58(3):55560.
[73] Johansen S. Estimation and hypothesis-testing of cointegration vectors in Gaussian
vector autoregressive models. Econometrica 1991;59(6):155180.
[74] Granger CWJ. Some recent developments in a concept of causality. J Econ
1988;39(12):199211.
[75] Bento JPC, Moutinho V. CO2 emissions, non-renewable and renewable electricity
production, economic growth, and international trade in Italy. Renew Sustain
Energy Rev 2016;55:14255.