Energy 192 (2020) 116628

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Long-run causal relationship between economic growth, transport

energy consumption and environmental quality in Asian countries:
Evidence from heterogeneous panel methods
Samia Nasreen a, *, Mounir Ben Mbarek b, Muhammad Atiq-ur-Rehman c
a
Department of Economics, Government College Women University Faisalabad, Faisalabad, Pakistan
b
Higher Institute of Management, Unversity of Gabes, Gabes, Tunisia
c
Department of Economics, University of the Punjab, Lahore, Pakistan

a r t i c l e i n f o a b s t r a c t

Article history: This paper analyzes data from 18 Asian countries, spanning from 1980 to 2017 to determine panel long-
Received 3 April 2019 run causality between income growth, transport energy consumption and environmental quality. The
Received in revised form empirical methodology of this paper considers structural breaks and cross-sectional dependence issues.
1 October 2019
In this way, our study becomes one of the few energy-GDP studies that address cross-sectional
Accepted 25 November 2019
Available online 26 November 2019
dependence in the unit root, cointegration, causality and elasticity estimation. We attempt to fill the
gap in energy literature by using robust estimates in the case of Asian countries. After these methodo-
logical advances, a bi-directional long-run Granger causality between transport energy consumption,
Keywords:
Transport energy consumption
environment and GDP growth is found. The common correlated effects mean group (CMG) is used to
Environmental quality obtain long-run elasticities. It is revealed that a 1% increase in transport energy consumption and GDP
Economic growth growth deteriorate environmental quality by about 0.57% and 0.46% respectively in Asian countries.
Panel granger causality Furthermore, the magnitude of elasticities varies from country to country. The empirical findings add
Asian countries new dimensions in energy-efficient technologies in the transport sector that positively affect economic
growth without compromising environmental quality.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction (GHG) and other pollutants in the atmosphere. The International

The antagonism between the green economy and energy con-
sumption constitutes a significant global issue. Rapid economic
growth, the process of industrialization, urbanization, population
growth and the growing specialization have accelerated the de-
mand for the transport sector. The expansion in transport demand
has put pressure on the country’s reserves of oil and gas. More than
90% of transport energy consumption is dependent on oil and oil-
related products, all of which are imported [1]. With the rising
demand for transport energy, gas and oil reserves are depleting
rapidly [2]. The increased purchasing power and safe climate drive
will urge people to use more efficient renewable energy sources [3].
Transport sector as the largest consumer of petroleum and other
liquid fuels is a major cause of the increase in Greenhouse gases
* Corresponding author.
E-mail addresses: sami_lcu@yahoo.com (S. Nasreen), mounirbenmbarek.fsegs@
gmail.com (M.B. Mbarek), atiq164@live.com (M. Atiq-ur-Rehman).
Energy Agency [1] estimates show that the transport sector ac-
counts for about 25% global CO2 (carbon dioxide) emission and its
contribution relative to other sectors is projected to increase in the
near future. The increase in number of vehicles on roads is one of
the reasons for emission growth. Currently, the number of vehicles
around the world is estimated to about 1.2 billion, consuming pe-
troleum fuels at about 13.5 billion barrels per annum and emitting
6.1 billion tons CO2 annually in the atmosphere. With growing
energy demand and increasing the number of vehicles on roads, the
global CO2 is projected to increase by about 50% in 2030 to about
80% in 2050 [1].
Asia is the region of a diverse group of countries, with different
levels of economic prosperity and energy resource endowment.
Rapid economic growth in the region increases the demand for
primary energy from 4025 MTOE in 2005e7215 MTOE in 2030. The
transport sector in Asia is rapidly growing and the energy con-
sumption is expected to rise at a rate of 2.9% per annum till 2030.
Among Asian countries, China is the largest consumer of transport
energy (12.3 quadrillions Btu) followed by India (3.3 quadrillions

https://doi.org/10.1016/j.energy.2019.116628
0360-5442/© 2019 Elsevier Ltd. All rights reserved.
2 S. Nasreen et al. / Energy 192 (2020) 116628
Btu). Like India and China, the other economies of the region also
show a substantial increase in transport energy demand from 5.5
quadrillions Btu in 2008 to 8.6 quadrillion Btu in 2017. Energy-
related CO2 emissions in the Asian region will increase from
10065 million tons in 2005e17763 million tons in 2030, with an
annual increment of 2.3%. CO2 emissions from the transport sector
are increasing very rapidly with a growth rate of 2.8% per year. This
growth indicates that the total share of CO2 emission will rise from
12.5% in 2005 to 13.7% in 2030.
Hence, impacts to the environment posed by fuel consumption
by their contribution to global warming through emission of CO2
have been some of the main concerns in recent sustainable trans-
portation policies. For most political decision-makers, the most
pressing issues associated with this sector are traffic fatalities, the
congestion, the greenhouse emissions and petroleum dependence.
These problems are especially acute in developing economies.
Therefore, it is important to provide new insights for a compre-
hensively designed environmental policy frame-work about the
transport energy and environment. Rational and forward-looking
policies can help achieve sustainable growth in the long-run.
A bulk of literature is available on the nexus between energy
consumption, economic growth and environmental quality.
Transport is a major energy-consuming sector which causes nearly
one-fifth of global CO2 emission. There are only a few studies in the
case of Asian countries [4]; [32] that examine a link between
transport energy and environmental quality. According to our
knowledge, there is no comprehensive study in the case of Asian
countries that simultaneously explored the relationship between
transport energy consumption, economic growth and environ-
mental quality using heterogeneous panel methodology. Moreover,
the empirical findings of previous studies are ambiguous and do
not reach any definite conclusion. Therefore, we are strongly
motivated to explore the long-run and causal relationship between
transport energy consumption, economic growth and environ-
mental quality in Asian countries. This attempt enables us to
identify feedback hypothesis/or unidirectional hypothesis/or
neutral hypothesis between environmental quality-transport en-
ergy consumption, environmental quality-economic growth and
transport energy consumption-economic growth. The direction of
the causal relationship between variables also helps policymakers
in making effective policy decisions to develop better transport
systems that efficiently consume energy and pollute fewer emis-
sions in the environment. Furthermore, we consider the specific
impact of transport energy consumption and economic growth on
environmental quality at the country level that helps environ-
mentalists and econometrists to identify whether economic growth
is more harmful to environmental quality or transport energy
consumption adds more to environmental pollution. Therefore, the
contribution of our study in energy-environment literature is
summarized in three key points (i) this is the first comprehensive
study in the case of Asian countries that investigated the long-run
and causal relationship between transport energy consumption,
economic growth and environmental quality. (ii) In the presence of
variables like income growth and energy consumption, the stan-
dard estimation techniques are unable to obtain consistent and
reliable results because the regional and macroeconomic linkages
or local-spillover effects between countries make these variables
cross-sectionally correlated [5]. Asia is an extremely diverse set of
countries that have different attitude, culture and traditions if one
mistakenly assumes that they are homogenous, then the panel
estimates will become inconsistent [6]. Therefore, we use the
methodology suggested by Liddle and Lung [7] to access the long-
run causal nexus between selected variables. This approach allows
a high degree of heterogeneity, unlike the usual causality ap-
proaches that assume homogeneity in long-run causality
parameters. (iii) Westerlund and Edgerton [8] panel cointegration
is applied to examine cointegration between variables. This coin-
tegration methodology efficiently addresses the issue of heteroge-
neity and cross-country correlation.
The organization of this paper is as follows: the literature review
is presented in Section 2, empirical model and data sources in
Section 3, empirical methodology in Section 4, econometric results
and interpretation in Section 5, conclusion and policy implications
in Section 6.
2. Literature review
The existing empirical literature is presented to have an idea of
past empirical findings on the nexus between transport energy,
Carbon dioxide emissions and economic growth. In general, there
exists a consensus that energy consumption from the transport
sector is a principal factor for the increase of CO2 emissions.
However, the empirical findings are inconsistent that might be due
to methodological differences, country coverage and the time
period. The empirical literature on transport energy, income
growth and Carbon dioxide emissions can be categorized into three
major dimensions. A more detailed analysis is presented in the
following section.
2.1. Transport energy-economic growth nexus
Numerous contributions in the literature empirically investigate
the nexus between transport energy and growth. Some published
research includes, for example, Samimi [9]; Liddle [10]; Costantini
and Martini [11]; Liddle [12]; Liddle and Lung [7]; Arvin et al. [13];
Saidi et al. [14], and [21]. Samimi [9] was the first to examine the
short-run and long-run relationship between Australian macro-
economic variables and transport energy. Their findings revealed
that transport energy and Australian economy both are interde-
pendent. Liddle [10]; by using statistical data from the United States
tested the link between transport energy demand, income and
gasoline prices. The study results confirmed the long-run rela-
tionship between the selected variables. He further noted that
transport energy consumption Granger cause price, thus reveal the
absence of any links between income and transport energy in the
short run. Liddle [12] studied the dynamic relationship between
gasoline price, gasoline ingestion, income, and car ownership in
OECD member states using panel co-integration analysis. Empirical
findings revealed that the above-mentioned variables are co-
integrated. The findings confirmed a unidirectional causality
running from income to gasoline consumption.
Liddle and Lung [7] explored the link between per capita GDP
and transport energy consumption for 107 economies. The empir-
ical results show evidence of positive unidirectional relationship
running from the gross domestic product to transport energy
consumption in the long run. Azlina et al. [15] explored some dy-
namic linkages between income, transport energy usage and CO2
emissions in a case study of the Malaysian economy. The empirical
findings show that the Granger causality runs from income to
transport energy consumption. Achour and Belloumi [16] access
the energy consumption link with GDP growth for the Tunisian
economy. The Johansen cointegration results show that GDP
growth affects transport energy consumption but reverse linkage
does not hold. Saidi and Hammami [17] applied a generalised
method of moments (GMM) to confirm that there is feedback
causality between transport energy and GDP growth for 75 coun-
tries. Farhadi [18] empirically evaluated the growth impact of
public infrastructure in a panel of 18 OECD countries. GMM
(Generalised Method of Moments) results showed that growth in
infrastructure positively affects growth in labour productivity and
 S. Nasreen et al. / Energy 192 (2020) 116628
total factor productivity.
Maparu and Mazumder [19] examined the long-run and causal
relationship transport infrastructure and economic development in
India. They applied Vector Auto-Regression and Vector Error
Correction model to find short-run and long-run causality. Results
showed the existence of a long-run relationship between transport
infrastructure and economic development, and the direction of
causality is from economic development to transport infrastruc-
ture. In the case of Egypt, Ibrahiem [20] reported a long-run posi-
tive relationship between transport sector energy usage and GDP
growth by applying the Johansen cointegration approach. Saidi
et al. [14] used GMM to show that there is a feedback causality
between per capita income and transport energy utilization in
MENA countries. Nasreen et al. [21], rejected the neo-classical
assumption about the neutrality of energy for growth in a sample
of 60 countries. The findings further demonstrate that freight
transport and economic growth are interdependent in sample
countries. Rehermann and Pablo-Romero Rehermann and Pablo-
Romero [22] tested the non-linear relationship between GDP per
capita and transport energy consumption for Latin American and
Caribbean countries. The results supported an N-shaped curve,
while the elasticity values of transport energy consumption, con-
cerning GDP per capita, do not show a tendency to decrease in the
long term. Sharif et al. [23], examined the transportation-growth
nexus for the USA. The Quantile-on-Quantile (QQ) approach is
applied for empirical analysis. The empirical results show that the
effect of economic growth on transportation services is positive for
the USA.
2.2. Transport energy- CO2 emissions nexus
Energy consumption and production, energy intensity and the
price of energy all play a vital role in increasing CO2 emission. In this
sense, it acts as an engine of industrial development and economic
growth. Increased industrialization and trade are not only associ-
ated with infrastructural development such as highways, roads,
factories and offices but also with a rise in mobility and thereby an
increase in demand for automobiles, especially in urban areas [24].
These activities, in turn, result in massive consumption of petro-
leum and other liquid fuels. High energy consumption in the
transport sector causes high carbon emissions which have a reverse
effect on the environment [25].
This issue has been treated by some resents studies such as Kolb
and Wacker [26] who noted that transport energy consumption
considered as a principal production of CO2. This view was
confirmed by Ong et al. [27]. Timilsina and Shrestha [28] pointed-
out major sources of CO2 emission in a sample of Asian countries.
They found that energy consumption from the transport sector is
one of the principal factors driving CO2 emission growth in these
countries. They explored further that the total share of transport
energy consumption in the regional carbon dioxide emission for
the most of Asian countries remained constant at about 10% over
the last 25 years. Chandran and Tang [4] found that transport en-
ergy consumption has a significant impact on environmental
quality in the Association of Southeast Asian Nations (ASEAN).
VECM based causality results indicated a two-way causal link be-
tween transport energy ingestion and environmental quality in
Thailand and Malaysia.
Hao et al. [29] demonstrated that freight transport is one of the
major cause of increasing greenhouse gas emissions in China. Ac-
cording to Tongwane [30]; transport energy plays a significant role
in the spread of greenhouse gas emissions in South Africa. Shahbaz
et al. [31], explored the nexus between carbon emissions from the
3
transport sector, transport sector energy demand and transport
infrastructure for Tunisian economy. The results estimated by
applying VECM methodology demonstrate coupling association
between energy consumption and CO2 emission. Mustapa and
Bekhet [32] explored some effective policy options that can help
reduce CO2 emission in Malaysia. The empirical findings estimated
by using linear programming and sensitivity analysis approach
showed that the energy consumption from transport sector pro-
duces 28% of total carbon dioxide emission.
Neves et al. [33] probed the relationship between transport
sector energy consumption by source and CO2 emissions for 15
OECD countries. Autoregressive distributive lag (ARDL) results
explained that transports fossil fuels consumption is an important
factor for boosting Economic growth of OECD countries. Danish and
Baloch [34] investigated the dynamic relationship between road
transport energy consumption and environmental quality in the
case of Pakistan. The statistical results estimated by applying ARDL
are unable to provide any proof of the expected role of transport
energy consumption in environmental pollution. In another study
on Pakistan, Danish et al. [35] examined the relationship between
transport energy consumption, economic growth, and carbon di-
oxide emission (CO2) from the transport sector. The empirical re-
sults estimated by applying autoregressive distributive lag (ARDL)
and vector error correction model (VECM) indicate a strong sig-
nificant impact of transport energy consumption on CO2 emissions
from the transportation sector.
Based on the literature review related to transport energy-
emissions nexus, it is noted that few studies have raised this
issue. In addition, there exists a consensus that energy consump-
tion from the transport sector is a principal factor for the increase of
CO2 emissions.
2.3. Economic growth- CO2 emission nexus
The linkages between economic growth and carbon dioxide
emission have been examined in various empirical studies. The
empirical findings are inconsistent that might be due to method-
ological differences, scope and the time period. Recent studies
include this subject such as Ozturk and Acaravci [36]; Cole et al.,
[37]; Govindaraju and Tang [48]; Kaika and Zervas [69]; Mbarek
et al., [38]; Bella et al. [39]; Shahbaz et al. [31]; Saidi and Hammami
[17]; Liddle [40]; Narayan et al. [41]; [69]; Ajmi et al., [42]; [65];
Heidari et al. [43]; Arvin et al., [13]; Nasreen et al., [44]. The results
found in these studies, however, are contradictory. Some studies
suggest unidirectional causality while others support bidirectional
causal nexus between these two variables.
Empirical findings by Apergis and Payne [45] confirmed a long-
run positive association between GDP growth and environmental
pollution. A unidirectional causality was found from GDP growth to
pollutant emission for 6 selected Central American countries. After
some time, Apergis et al. [46] observed a bidirectional causal link
between economic growth and carbon dioxide emission in 16
industrialized and developing economies. Arouri et al. [47],
empirical results explained that an increase in economic growth
deteriorates environment in MENA countries. The results further
explained that there is a long-term causal link between pollutants
emission and real income in the sample economies. Ozturk and
Acaravci [36] findings confirmed that there exists a long-run causal
nexus between pollutants emission and income growth but no
association was observed in the short run. Govindaraju and Tang
[48] found that there exists a long-run association between carbon
dioxide emission and real GDP in India and China. A unidirectional
causality was revealed from real GDP growth to carbon dioxide
4 S. Nasreen et al. / Energy 192 (2020) 116628
emission for China both in the short-run and long-run. However,
only short-run causality is confirmed in the case of India. Mbarek
et al. [38] and Jebli [49] empirically found that an increase in GDP
growth is a significant determinant of CO2 emission for Tunisia.
Saidi and Hammami [17] found that there is a unidirectional
causal nexus between CO2 emission and GDP growth in a sample of
58 countries. The empirical results confirm that the emission of CO2
Granger causes GDP growth. Using data of 181 countries, Narayan
et al. [41] showed that 68% of countries have a positive and sta-
tistically significant association between GDP growth and carbon
emission whereas in 27% countries, a rise in economic growth
causes a reduction in CO2 emission in the long run. Arvin et al. [13]
found that there is a unidirectional causality between economic
growth and CO2 emission in the short run as well as in the long run
for G20 countries. In this case, causality ran from economic growth
to CO2 emission. Nasreen et al. [44] discovered that any increase in
income Granger cause environmental quality in the South Asian
countries including India, Pakistan, Sri Lanka and Nepal where a
feedback causality between economic growth and environmental
quality was found in case of Bangladesh. Linh and Khanh [50] study
on ASEAN countries supported feedback hypothesis between GDP
growth and contaminant emission. Aye and Edaja [51] used data of
31 developing countries to explore a significant causal relationship
between income growth and environmental pollution. Moreover,
panel threshold cointegration results revealed that income growth
improves environmental quality in low growth regime while
deteriorating environmental quality in high growth regime.
In a recent study, Dogan and Deger [52] analyzed the relation-
ship between carbon emissions and economic growth by using
panel data analysis in E7 countries (Brazil, China, Indonesia, India,
Mexico, Russia and Turkey). The result of the panel analysis sug-
gests that a 1% increase in economic growth leads to an increase of
0.243% in carbon emissions over the long-term in sample countries.
Najid et al. [70] explored the cointegration and causal relation be-
tween real GDP per capita, and CO2 emissions per capita for China.
The autoregressive distributed lag model has been used to confirm
the cointegration among variables. Results reveal short-run income
coefficient is positive while it turns to negative and insignificant in
long run marking that growth and development are supportive in
CO2 emissions reduction in China. Gorus and Mucahit [53] inves-
tigated the causal relationship between economic growth, and CO2
emission through both single- and multi-country Granger causality
analysis in the frequency domain considering eight oil-rich MENA
countries; namely, Algeria, Egypt, Iran, Iraq, Oman, Saudi Arabia,
Tunisia, and the United Arab Emirates. Panel causality test results
show the absence of the causal nexus between economic growth
and CO2 emission. In the case of Indonesia, Sasana and Aminata [54]
analyzed the effect of energy-based economic growth on CO2
emissions. Empirical results estimated by using multiple linear
regression with the ordinary least square method showed that
economic growth positively affected CO2 emissions. In another
study on Indonesia, Nugraha and Osman [55] found similar
findings.
It is clear from the above-mentioned literature that there exists
a piece of limited evidence on the association between transport
energy ingestion, GDP growth and environmental quality in the
Asian countries. This paper extends the empirical literature by
providing evidence on the causal nexus between income growth,
transport energy consumption and environmental quality by
applying robust heterogeneous panel data analysis techniques.
3. Methodology
3.1. Data sources and model development
The present study uses per capita GDP in constant 2010 US $, per
capita, CO2 emission in metric tons, per capita transport energy
consumption1 in kilotonne of oil equivalent (ktoe), real oil prices2 in
US $ per barrel to measure economic growth (EG), environmental
quality (EQ), transport energy consumption (TE) and energy prices
(EP) respectively. The World Development Indicators (WDI) data-
base is a source for data on CO2 emission and economic growth. The
transport energy consumption data is taken from International
Energy Agency while the data on crude oil prices (the simple
average of three spot prices: Dated Brent, West Texas Intermediate,
and the Dubai Fatah deflated by country’s consumer price index) is
extracted from British Petroleum 2018 Statistical Review of World
Energy for 18 Asian countries3 Pakistan, India, Bangladesh,
Indonesia, Iran, Jordan, Japan, Korea, Malaysia, Nepal, Philippines,
Sri Lanka, Thailand, Vietnam, Singapore, Syria, Israel, China, span-
ning from 1980 to 2017. All variables are transformed in the natural
log.
There are three main reasons for the selection of Asian coun-
tries. (i) The empirical contributions on the association between
GDP growth, energy consumption and environmental quality are
relatively scarce about the Asian economies. (ii) Asian countries
especially China and India are among the largest consumers of
energy and largest producers of environmental pollution in the
globe. (iii) The consumption of transport energy is an important
source of environmental degradation while the energy usage for
transport is increasing day by day in Asian countries. In the newly
industrialized and modernized economies, the challenges to sus-
tainable economic growth are vital.
Earlier empirical studies (see e.g. Refs. [4,31,56] explored that
the economic growth, energy consumption, and energy prices are
the important causes of environmental degradation. Following this
argument, the relation between environmental quality and its main
sources is modelled as follows:
EQit ¼ g0 þ f1 EGit þ f2 TEit þ f3 EPit þ mi þ 6it (1)
In Eq. (1) f1 ,f2 and f3 are the long-run parameters for environ-
mental quality with respect to economic growth, transport energy
consumption and energy prices respectively. The expected sign for
f1 and f2 is positive while f3 is expected to be negative. m is
country-specific effect while 6 is stochastic white noise error term.
Descriptive statistics and correlation matrix are displayed in
appendixTable A and B.
3.2. Empirical strategy
3.2.1. Cross-sectional dependence and panel unit root tests
The initial step in empirical methodology is to perform cross-
sectional dependence (CSD) test. This test was developed by
Pesaran [57] follows a standard normal distribution under the null
hypothesis of cross-sectional independence. Pesaran statistic
computes:
1
covers all transport activities except the economic sector to which it
contributes.
2
are used as a proxy for energy prices due to unavailability of data on energy
prices.
3
The selection of 18 Asian countries is based on data availability and the strategy
to use balanced panel.
 S. Nasreen et al. / Energy 192 (2020) 116628 5

0 1 of long-run causal nexus between selected series. Unlike conven-

X X
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi M1 M ∧ tional causality method, this approach allows a high degree of
CSD ¼ 2T=MðM  1Þ@ lkg A (2) heterogeneity in the long-run relationships. In the non-stationary
k¼1 g¼kþ1
and cointegrated panels, the estimation can be performed in the
form of a dynamic error correction model (ECM):
where lkg is estimated from
X
k1 X
k2
P
P DEQit ¼ hEQj þ bEQik DEQitk þ fEQis DEGits
hkp ngp s¼1 s¼1
p¼1
lkg ¼ lgk ¼ !1=2 !1=2 (3) X
k3 X
k4
P
P P
P þ dEQis DTEits þ qEQis DEPits þ uEQi ECTEQit1 þ εEQit
n2kp n2gp
s¼1 s1
p¼1 p¼1
(4)
The null hypothesis to be tested is
lkg ¼ lgk ¼ corrðhkp ngp Þ ¼ 0ksg for and the alternative hypothesis
X
k1 X
k2
to be tested is lkg ¼ lgk s0 for some ksg. DEGit ¼ hEGj þ bEGik DEGitk þ fEGis DEQits
Bai and Carrion-i-Silvestre [58] introduced a panel unit root test s¼1 s¼1
that handles the multiple structural breaks with cross-sectional X
k3 X
k4
dependence in heterogeneous panels. Structural breaks are þ dEGis DTEits þ qEGis DEPits þ uEGi ECTEGit1 þ εEGit
allowed at level, slope or both with varying time period under s¼1 s¼1
different magnitudes of shift. The computation of structural breaks (5)
is important since the sample covers the period of the Asian
financial crisis and the global financial crisis. Cross-sectional X
k1 X
k2
dependence is captured by common factors as described in Bai DTEit ¼ hTEj þ bTEis DTEits þ fTEis DEGits
and Ng [59]. These common factors may be stationary, non- s¼1 s¼1
stationary or combination of both. After incorporating multiple X
k3 X
k4
structural breaks for each series and common factors, this test pools þ dTEis DEQitk þ qTEis DEPitk þ uTEi ECTTEit1 þ εTEit
modified Sargan-Bhargava (MSB) tests for individuals series. Three s¼1 s¼1
different test statistics: Z, a standardization of individual statistics, P (6)
and Pm combining their p-values are the result of three pooling
statistics. These tests show good performance infinite sample as X
k1 X
k2
confirmed by Monte Carlo simulation [58]. DEPit ¼ hEPj þ bEPik DEPits þ bEPis DEQits
s¼1 s¼1
3.2.2. Panel cointegration test X
k3 X
k4
A possibility of long-run causality is induced due to the presence þ dEPis DEGits þ qEPis DTEits þ uEPi ECTEPit1 þ εEPit
of cointegrated variables. The present study employs Westerlund s¼1 s¼1
and Edgerton [8] panel cointegration test. The test is robust to (7)
different structural breaks and cross-sectional dependence and
allows heterogeneous specification in both the short-run and long- where k1; k2; ……․; k4 refers the lag lengths; εit denotes that the
run portions of the error correction model. We use two tests with random error term which is normally distributed in case of all “t”
the null hypothesis of no cointegration. These tests are derived and “i” with zero mean, constant heterogeneous variance. The ECTs
from Lagrange Multiplier (LM) based unit root tests: the coefficient are the error-correction terms which are obtained from co-
version (4N ) and the t-test version (tN ). There are three different integrating equations. Liddle and Lung [7] proposed that first dif-
cases including no structural break, a shift in level only, a shift in ference terms in Eqs. (4)e(7) may address the cross-sectional
the level and trend. These tests have a tendency to achieve better dependence. Although differenced variables show the short-run
results even in the small samples. causality among the variables, the ECTs value shows the long-run
causal relationship. The short-term causal association is
3.2.3. Long-run estimators computed by F-statistic as well as the significance by changing lags
The long-run elasticities are computed by using common in exogenous variables. However, the long-term causal association
correlated effects mean group (CMG) estimators developed by is computed by the significance of lagged ECT’s t-test. When the
Pesaran [60]. This approach is robust to non-stationary cointe- coefficients concerning group mean test fails to reject causality, but
grated variables and cross-sectional dependences. The failure of ECT is significant, we deduce that the Granger causality test is
CMG estimator in the detection of differences between temporal heterogeneous in the panel.
and general dynamics which are determined by common and
exogenous individual specific time-series factors was pointed out 4. Empirical findings and their interpretation
by [71]. Moreover, the spatial patterns occurring in energy con-
sumption and CO2 nexus cannot be modelled by CMG. The reason is The findings of Pesaran cross-sectional dependence test (CSD)
that the estimator constantly gives efficient slope estimates are reported in Table 1. The cross-sectional dependence hypothesis
without correcting the mechanism of spatial error [71]. To solve is rejected at 1% level of significance for all the variables. Thus, we
these issues augmented mean group (AMG) developed by Eber- conclude that EQit ,EGit ,TEit ,EPit are highly dependent across Asian
hardt and Teal [71] is applied. countries.
The results of Bai and Carrion-i-Silvestre [58] unit root test are
3.2.4. Panel long-run granger causality reported in Table 2. Estimated results of unit root test give strong
The Granger causality approach developed by Holtz et al. [61], evidence in favor of non-stationarity at the level when no break and
and applied by Liddle and Lung [7] is applied to detect the direction break in the trend are introduced in the model. On average one
6 S. Nasreen et al. / Energy 192 (2020) 116628

Table 1 Table 3
Cross-sectional dependence test results. Westerlund and Edgerton cointegration test results.

Variables Statistics P-value abs Model No shift Mean shift Regime shift

EQit 38.81 0.004 0.597 tN 3.432 4.540 5.919

EGit 99.47 0.000 0.817 P-value 0.045 0.001 0.000
TEit 86.97 0.000 0.740 4N 1.647 2.782 2.802
EPit 47.10 0.000 0.421 P-value 0.089 0.023 0.019

Notes: Average absolute values (abs) shows correlation coefficient. Note: The statistics tN and 4N are normally distributed.

Table 2 Table 4
Panel unit root test results. Estimates of breaks.

Variables EQit EGit TEit EPit Country Mean shift Regime shift
Constant and Trend (No break) Pakistan 2002 2014
Z 1.064 0.515 0.253 0.738 India 2008 2008
Pm 0.288 0.648 1.120 0.543 Bangladesh 1990 2002
P 21.51 25.96 14.76 20.61 Indonesia 2008 1997
Trend Shifts Iran 1988 1990
Z 0.843 0.338 0.046 0.649 Japan 2014 2014
Pm 0.540 0.772 0.149 0.352 Jordan 2007 2007
P 16.75 21.56 18.44 16.98 Korea Dem. 1990 2001
Z* 2.843* 2.065** 1.970** 3.132* Malaysia 1997 1998
Pm * 3.054* 4.873* 2.643* 4.738* Nepal 2005 2005
P* 40.36* 57.72* 46.09* 60.02* Philippines 1998 2010
Sri Lanka 1992 2004
Note: Critical values for the rejection of null hypothesis of unit root at *1%, **5% and Thailand 2000 2000
***10% are 2.326, 1.645 and 1.282 respectively for both Z and Pm while for P test, chi- Singapore 2008 2008
squared critical values are 40.28, 33.92 and 30.81 respectively. Syria 2015 2015
Israel 1988 2001
China 2005 2014
structural break is found four each of four variables
(EQit EGit TEit EPit )4 and up to two structural breaks for each country
time series. economies. Dobnik [63] reported structural breaks in energy-
Westerlund and Edgerton [8] cointegration approach involves growth nexus for 23 OECD countries.
structural breaks and cross-sectional dependence simultaneously Table 5 shows the country-wise impact of economic growth,
for different panel members at different dates. The results dis- transport energy consumption and energy prices on carbon dioxide
played in Table 3 show that both test statistics tN 4N give evidence emission. The results indicate that the transport energy consump-
in favor of long-run association between per capita real GDP, per tion has a positive and significant coefficient in all the selected
capita transport energy consumption, environmental quality and Asian countries except, Japan and the Philippines where the coef-
energy prices when no break, breaks in the mean and break in the ficient is found to positive but insignificant. The positive coefficient
regime are introduced in the model. of transport energy consumption suggests that any rise in the usage
Furthermore, contemporaneously estimated breaks for each of transport energy increases environment pollution in Asian
country are presented in Table 4. By applying Westerlund and economies. These empirical findings are aligned with findings of
Edgerton [8] cointegration approach, we find one structure break Timilsina and Shrestha [28] for Asia; Chandran and Tang [4] for
for each country when mean shift and regime shift specifications ASEAN economies; Hao et al. [29] for China; Mustapa and Bekhet
are taken into account. These breakpoints associated with a huge [32] for Malaysia and Baloch and Danish et al. [35] for Pakistan. The
global crisis. Most of the countries share common break dates, for positive and Significant role of economic growth in deteriorating
example, break in 2008 is coincide with global financial crisis, 1997 environmental quality is reported in Pakistan, India, Indonesia,
with Asian crisis, resulted in episodes of high inflation and unem- Bangladesh, Japan, Jordan, Kore Dem., Thailand, Vietnam, Syria,
ployment, 1980s breakpoints reflected the disturbances throughout Singapore, Israel and China while the negative and significant
the period including Iranian revolution and Iraq-Iran war, 2002 impact of economic growth on the quality of the environment in
breakpoint show the incidence of September 11 attacks, 2005 Malaysia. The empirical results suggest a negative and insignificant
breakpoint is coincide with continued rise in oil prices until 2008 effect of economic growth on environmental quality in Iran. The
financial crisis emerged, 1990 breakpoint show subsequent in- insignificant results in Iran may be due to a declining share of the
crease in oil price due to Iraq’s invasion of Kuwait, 2014 breakpoint industrial sector and the rising share of services sector in GDP
show geopolitics crisis resulted in decline in oil prices, 2007 during the sample period. The empirical findings are in line with
breakpoint describe food price shocks that negatively impact the the results of different studies including Ozturk and Acaravci [36];
majority of Asian countries and 2009 breakpoint is coincided with Arouri et al., [47]; Govindaraju and Tang., [48]; Bella et al. [39];
European sovereign debt crisis. Our findings support the previous Shahbaz et al. [31]; Saidi and Hammami [17]; Liddle [40]; Arvin
empirical studies on energy-growth nexus that reported explicit et al., [13]. A significant and negative impact of energy prices on
break dates. Narayan and Smyth [62] point-out structural breaks in environmental quality is found for all the selected Asian states. On
energy-growth nexus in a sample of G7 economies. Liddle and Lung empirical grounds, our findings are broadly consistent with Shah-
[7] found structural breaks in per capita transport energy con- baz et al., [31]. The Panel estimates give all positive and statistically
sumption and per capita GDP growth nexus for a panel of 107 significant coefficients in Asia. The coefficients of transport energy
consumption and income growth estimated by applying CMG
indicate that 1% increase in transport energy consumption and
4
The break is most recurrent in 2008 in economic growth, 1997 in transport income growth deteriorating environmental quality by about 0.57%
energy consumption, 2010 for environmental quality and 2014 in oil prices.
 S. Nasreen et al. / Energy 192 (2020) 116628 7

Table 5
Long-run heterogeneous estimates.

(Dependent variable: EQt )

Country Variables CMG AMG

TEt EGt EBt TEt EGt EBt

Pakistan Coefficient 0.776 0.013 0.345 1.903 0.256 0.321

Probability 0.005 0.058 0.065 0.000 0.032 0.098
India Coefficient 0.541 0.444 0.547 1.071 0.954 0.430
Probability 0.001 0.019 0.005 0.004 0.050 0.650
Bangladesh Coefficient 0.706 0.230 0.214 0.084 0.054 0.123
Probability 0.070 0.020 0.087 0.963 0.091 0.439
Indonesia Coefficient 0.463 0.014 0.830 1.536 1.086 0.765
Probability 0.401 0.069 0.076 0.000 0.021 0.054
Iran Coefficient 0.247 0.265 0.650 1.575 1.320 0.992
Probability 0.097 0.397 0.004 0.068 0.091 0.231
Japan Coefficient 0.320 1.053 0.478 1.314 1.001 0.864
Probability 0.201 0.003 0.007 0.010 0.003 0.044
Jordan Coefficient 0.597 0.748 0.659 1.677 0.529 0.213
Probability 0.012 0.000 0.054 0.012 0.067 0.091
Korea Dem. Coefficient 1.097 0.135 0.589 1.413 1.490 0.689
Probability 0.000 0.075 0.010 0.000 0.004 0.020
Malaysia Coefficient 0.516 0.101 0.490 0.059 0.114 0.231
Probability 0.000 0.082 0.098 0.910 0.870 0.438
Nepal Coefficient 0.124 0.450 0.129 0.754 1.094 0.745
Probability 0.029 0.469 0.959 0.000 0.004 0.009
Philippines Coefficient 0.659 0.510 0.553 0.755 0.455 0.057
Probability 0.260 0.012 0.080 0.525 0.097 0.665
Sri Lanka Coefficient 0.562 0.107 0.531 0.849 0.669 0.032
Probability 0.054 0.432 0.000 0.680 0.765 0.229
Thailand Coefficient 0.385 0.328 0.233 1.744 0.654 0.869
Probability 0.026 0.076 0.000 0.172 0.046 0.320
Vietnam Coefficient 0.690 0.920 0.578 0.412 0.878 0.992
Probability 0.005 0.054 0.005 0.022 0.084 0.567
Singapore Coefficient 0.967 0.345 0.443 0.785 0.540 0.969
Probability 0.010 0.003 0.067 0.007 0.060 0.084
Syria Coefficient 0.779 0.508 0.689 1.065 0.753 0.561
Probability 0.008 0.040 0.091 0.010 0.062 0.099
Israel Coefficient 0.801 0.769 0.833 0.944 0.652 0.543
Probability 0.093 0.540 0.349 0.667 0.840 0.999
China Coefficient 0.721 0.376 0.491 1.798 0.785 0.633
Probability 0.002 0.001 0.024 0.016 0.005 0.090
Panel statistics Coefficient 0.576 0.464 0.226 0.630 0.442 0.382
Probability 0.001 0.000 0.010 0.006 0.005 0.009

and 0.46% in Asian countries respectively. Similarly, a one percent Table 6

increase in energy prices improves environmental quality by about Panel causality results.

0.23% in the panel of selected Asian economies for the period Dependent variables Independent variables
1980e2017. The empirical results are aligned with [72] in the case Short-run causality Long-run
of OECD countries. causality
After finding the long-run impacts of economic growth, trans-
DEQit DEGit DTEit DEPit ECTt1
port energy consumption and energy prices on environmental
Eq (4) DEQit 3.731 4.982 3.879 0.307**
quality, Granger causality tests are applied to check the direction of e
(0.023) (0.000) (0.014) [2.360]
long-run causal relationships between variables. From Table 6, the Eq (5) DEGit 1.253 e 1.672 4.995 0.356*
results of Eq (4) show that lagged error correction term (ECTt1 ) has (0.846) (0.324) (0.000) [3.051]
a negative sign with statistically significant t-value. The significance Eq (6) DTEit 0.965 4.390 e 3.659 0.262*
of this error correction term in CO2 emission equation indicates (0.648) (0.006) (0.036) [2.104]
Eq (7) DEPit 1.659 2.991 0.950 e 0.182**
long-run causal relationship and the causality is running from
(0.237) (0.071) (0.533) [2.531]
economic growth, transport energy consumption and energy prices
Note: Values in ( ) are p-values and values in [ ] are t-ratios. * and ** represent
to environmental quality. In Eq. (5), the negative and significant
significance at 1% and 5% level respectively.
ECTt1 implies that there is evidence of long-run causality running
from CO2 emissions, transport energy consumption and energy
prices to economic growth. The speed of adjustment (ECTt1 ) in Eq transport energy consumption (0.262) is slightly lower than the
(6) is found to be statistically significant and reveal that if there is speed of adjustment of environmental quality (0.307) and eco-
any divergence from long-run equilibrium path, CO2 emissions, nomic growth (0.356). The smaller energy consumption adjust-
income growth and energy prices respond positively to correct this ment coefficient compared to economic growth and carbon
disturbance. In Eq (7), the positive and significant ECTt1 demon- emission imply that there is a possibility that energy policies might
strate that variables (transport energy consumption, income not adversely impact long-run economic growth and environ-
growth, and carbon emissions) do not converge towards their long- mental quality [63]. In perspective of environmental quality and
run equilibrium path. Moreover, the error correction term for energy-growth linkages, these results support the feedback
8 S. Nasreen et al. / Energy 192 (2020) 116628
hypothesis between environmental quality-energy consumption,
environmental quality-economic growth and energy consumption-
economic growth. In the empirical literature, our results on envi-
ronmental quality-energy consumption-growth nexus support the
Halicioglu [64] for Turkey, Dobnik [63] for OECD countries.
The short-run causality test results reflect a unidirectional
causality from GDP growth to environmental quality, transport
energy consumption to environmental quality and energy prices to
environmental quality. Similarly, income growth and energy prices
Granger cause transport energy usage in the short-run. Neverthe-
less, a feedback relationship is observed between income growth
and energy prices in Asian countries. found similar results in the
case of Saudi Arabia.
5. Conclusions and policy implications
This study aims at exploring the nexus between transport en-
ergy consumption, environmental quality and economic growth in
the selected Asian countries for the period 1980e2017. For empir-
ical analysis, we employ panel estimation techniques that produce
a robust result in the heterogeneous panel.
Our empirical results indicate that all the variables are station-
ary at first difference and cross-sectionally dependent. The West-
erlund and Edgerton [8] cointegration approach confirm the
cointegration relation between economic growth, transport energy
consumption, energy prices and environmental quality. The CMG
and AMG estimation analysis reveal that increase of transport en-
ergy usage and unprecedented GDP growth are harmful to the
environmental quality of Asian countries. In contrast, the upsurge
in energy prices proves to be beneficial for the environmental
quality of Asian countries. The empirical analysis suggests a bidi-
rectional causal nexus between environmental quality-economic
growth and transport energy consumption in Asian economies.
The transport energy has become an alarming obstruction in the
way of the sustainable and green economy. A major role of trans-
port energy usage in carbon dioxide emission urges the decision-
makers to reconsider their energy and environmental policy. The
policymakers need to encourage more energy-efficient ways of
transportation, so the harmful effects of transportation energy
consumption can be reduced (e.g., CO2 emissions). Moreover, policy
instruments that promote alternative energy sources such as CNG
may also be implemented in Asian countries.
The global coordination should be there to promote clean and
sustainable transport by encouraging the move of shifting of private
motorization to non-motorized transport. The infrastructure
should be properly developed to promote non-motorized trans-
port. The well-built and safe roads will encourage people to cover
shorter distances by bicycles.
Introduction of biofuels would be another option to replace
gasoline and diesel in the transport sector. Biofuels are the fuels
derived from sources like sugar, vegetable oils, starch and animal
fats. The cost-efficient bioenergy industry, especially in developing
countries like Asia, can help to reduce the emission of GHG and
protecting the environment. The bioenergy industry is progressing
in the European countries. The traditional energy sources are
responsible for environmental degradation, so they should be
Table A
Descriptive Statistics
Variables Max. Min. Mean Median
EQit 8.830 2.683 0.361 0.463
EGit 9.514 2.923 7.028 7.948
TEit 6.158 1.217 4.235 6.047
EPit 7.564 1.081 3.787 1.790
replaced by bioenergy. The U.S. government is considering carbon-
neutral biodiesel to protect the environment. Biofuels can be used
as substitutes for petrol because they cause less emission of
greenhouse gases. However, the use of biofuels can cause an up-
surge in the prices of agricultural products due to higher demand
for fuel production.
Smart driving techniques may be considered for fuel-saving and
environmental protection. Eco-driving has been introduced in
many countries for reducing CO2 emissions. The use of public
transport is an ecological and economical mode of travelling. The
people travelling by their own cars mean more automobiles on the
roads and more CO2 emissions. Public transport implies more
people travelling with fewer vehicles. Therefore, Public transport is
a source of communication causing minimum CO2 emissions. The
use of Public transport not only reduces greenhouse gases emission
but also saves energy. It also provides affordable mobility with
lower congestion on the roads. Carpooling can also reduce carbon
emission but the method is not much popular due to privacy and
safety issues.
At the national level, China and India which are larger consumer
of transport energy need to implement regulatory instruments,
such as stringent fuel economy standards. This policy may reduce
transportation energy consumption and thereby help reduce CO2
emissions. Transportation energy consumption can also be reduced
by improving vehicle occupancy rates through carpooling, which
has not been promoted widely in Asia, or through the use of high
occupancy public transit systems. There is a need to introduce
regulatory instruments, such as vehicle fuel economy standards to
reduce fuel demand and CO2 emissions from the transportation
sector. China and South Korea have taken a step to implement fuel
economy standard and India is expected to join them shortly.
Increasing diesel prices, or removing subsidies from diesel, can be
another option to improve environmental quality.
The energy efficiency has a critical role in both achieving
transport energy security, and meeting environmental protection
and economic objectives. The general conclusion is that an increase
in economic growth and technological change has a significant
influence on energy consumption for the transport sector and
creates environmental pollution. Globalization can play a signifi-
cant role in generating and transferring resource-saving and
cleaner production technology from developed to developing
countries. International policies such as the Montreal Protocol and
the Base Convention transfer best available and environmentally
safe technologies to developing countries. At the national level,
Asian countries need to reform their domestic policies that have
negative environmental impacts. They further need to correct their
existing market failure through economic instruments rather than
depend on economic integration and trade liberalization.
An environment-friendly culture should be promoted and the
public should be educated to use public transport instead of private
vehicles. The subsidized public transport can incentivize people to
use public transport and protect environment.
Appendix
Standard Deviation Skewness Kurtosis Inter-Quartile Range
1.321 2.556 4.637 2.012
1.378 3.255 2.907 1.089
0.524 0.248 3.269 1.653
1.505 2.067 4.702 1.461
 S. Nasreen et al. / Energy 192 (2020) 116628
Table B
Correlation Matrix
Variables EQit EGit TEit EPit
EQit 1.000 0.623 0.747 0.489
(0.030) (0.001) (0.076)
EGit 0.623 1.000 0.599 0.342
(0.030) (0.000) (0.001)
TEit 0.747 0.599 1.000 5.760
(0.001) (0.000) (0.065)
EPit 0.489 0.342 5.760 1.000
(0.076) (0.001) (0.065)
Note: values in parenthesis are P-values.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.energy.2019.116628.
References
[1] International Energy Agency, (IEA). Key world energy statistics. Paris: IEA;
2017.
[2] Walmsley MRW, Walmsley TG, Atkins MJ, Kamp PJJ, Neale JR, Chand A. Carbon
emissions pinch analysis for emissions reductions in the New Zealand trans-
port sector through to 2050. Energy 2015;1000:1e8.
[3] Johansson B. Will restrictions on CO2 emissions require reductions in trans-
port demand? Energy Policy 2009;37:3212e20.
[4] Chandran VGR, Tang CF. The impact of transport energy consumption, foreign
direct investment and income on CO2 emissions in ASEAN-5 economies.
Renew Sustain Energy Rev 2013;44:445e53.
[5] Kapetanios G, Pesaran MH, Yamagata T. Panels with non-stationary multi-
factor error structures. J Econom 2011;160:326e48.
[6] Pesaran M, Smith R. Estimating long-run relationships from dynamic het-
erogeneous panel. J Econom 1995;68:79e113.
[7] Liddle B, Lung S. The long-run causal relationship between transport energy
consumption and GDP: evidence from heterogeneous panel methods robust
to cross-sectional dependence. Econ Lett 2013;121:524e7.
[8] Westerlund J, Edgerton D. A simple test for cointegration in dependent panels
with structural breaks. Oxf Bull Econ Stat 2008;70:665e704.
[9] Samimi R. Road transport energy demand in Australia A cointegration
approach. Energy Econ 1995;17(4):329e39.
[10] Liddle B. Long-run relationship among transport demand, income, and gaso-
line price for the US. Transp Res D Transp Environ 2009;14(2):73e82.
[11] Costantini V, Martini C. The causality between energy consumption and
economic growth: a multi-sectoral analysis using non-stationary cointegrated
panel data. Energy Econ 2010;32(3):591e603.
[12] Liddle B. The systemic, long-run relation among gasoline demand, gasoline
price, income, and vehicle ownership in OECD countries: evidence from panel
cointegration and causality modeling. Transp Res D Transp Environ
2012;17(4):327e31.
[13] Arvin MB, Pradhan RP, Norman NR. Transportation intensity, urbanization,
economic growth, and CO 2 emissions in the G-20 countries. Util Policy
2015;35:50e66.
[14] Saidi S, Shahbaz M, Akhtar P. The long-run relationship between transport
energy consumption and transport infrastructure on economic growth in
MENA countries. Transp Res A Policy Pract 2018;111:78e95.
[15] Azlina AA, Hook LS, Nik Hashim NM. Dynamic linkages among transport en-
ergy consumption, income and CO2 emissions in Malaysia. Energy Policy
2014;73:598e606.
[16] Achour H, Belloumi M. Investigating the causal relationship between trans-
port infrastructure, transport energy consumption and economic growth in
Tunisia. Renew Sustain Energy Rev 2016;56:988e98.
[17] Saidi K, Hammami S. The impact of energy consumption and CO 2 emissions
on economic growth: fresh evidence from dynamic simultaneous-equations
models. Sustainable Cities and Society 2015;14:178e86.
[18] Farhadi M. Transport infrastructure and long-run economic growth in OECD
countries. Transp Res A Policy Pract 2015;74(C):73e90.
[19] Maparu TS, Mazumder TN. Transport infrastructure, economic development
and urbanization in India (1990e2011): is there any causal relationship?
Transp Res A Policy Pract 2017;100(C):319e36.
[20] Ibrahiem DM. Road energy consumption, economic growth, population and
urbanization in Egypt: cointegration and causality analysis. Environ Dev
Sustain 2018;20(3):1053e66.
[21] Nasreen S, Saidi S, Ozturk I. Assessing links between energy consumption,
freight transport, and economic growth: evidence from dynamic simulta-
neous equation models. Environ Sci Pollut Control Ser 2018;25:16825e41.
[22] Rehermann F, Pablo-Romero M. Economic growth and transport energy
9
consumption in the Latin American and Caribbean countries. Energy Policy
2018;122(C):518e27.
[23] Sharif A, Shahbaz M, Hille E. The Transportation-growth nexus in USA: fresh
insights from pre-post global crisis period. Transp Res A Policy Pract
2019;121(C):108e21.
[24] Jassie PH, Poon IC, He C. The impact of energy, transport and trade on air
pollution in China. Eurasian Geogr Econ 2006;47(10):1e17.
[25] Alkhathlan K, Javid M. Energy consumption, carbon emissions and economic
growth in Saudi Arabia: an aggregate and disaggregate analysis. Energy Policy
2012;62:1525e32.
[26] Kolb A, Wacker M. Calculation of energy consumption and pollutant emis-
sions on freight transport routes. Sci Total Environ 1995;169(1):283e8.
[27] Ong HC, Mahlia TMI, Masjuki HH. A review on energy pattern and policy for
transportation sector in Malaysia. Renew Sustain Energy Rev 2012;16(1):
532e42.
[28] Timilsina GR, Shrestha A. Transport sector CO 2 emissions growth in Asia:
underlying factors and policy options. Energy Policy 2009;37(11):4523e39.
[29] Hao H, Geng Y, Li W, Guo B. Energy consumption and GHG emissions from
China’s freight transport sector: scenarios through 2050. Energy Policy
2015;85:94e101.
[30] Tongwane M, Piketh S, Stevens L, Ramotubei T. Greenhouse gas emissions
from road transport in South Africa and Lesotho between 2000 and 2009.
Transp Res D Transp Environ 2015;37:1e13.
[31] Shahbaz M, Khraief N, Jemaa MMB. On the causal nexus of road transport CO 2
emissions and macroeconomic variables in Tunisia: evidence from combined
cointegration tests. Renew Sustain Energy Rev 2015;51:89e100.
[32] Mustapa SI, Bekhet HA. Analysis of CO 2 emissions reduction in the Malaysian
transportation sector: an optimisation approach. Energy Policy 2016;89:
171e83.
[33] Neves SA, Marques AC, Fuinhas JA. Is energy consumption in the transport
sector hampering both economic growth and the reduction of CO2 emissions?
A disaggregated energy consumption analysis. Transp Policy 2017;59(C):
64e70.
[34] Danish, Baloch MA. Dynamic linkages between road transport energy con-
sumption, economic growth, and environmental quality: evidence from
Pakistan. Environ Sci Pollut Control Ser 2018;25(8):7541e52.
[35] Danish, Baloch MA, Suad S. Modeling the impact of transport energy con-
sumption on CO2 emission in Pakistan: evidence from ARDL approach. Envi-
ron Sci Pollut Control Ser 2018;25(10):9461e73.
[36] Ozturk I, Acaravci A. CO 2 emissions, energy consumption and economic
growth in Turkey. Renew Sustain Energy Rev 2010;14(9):3220e5.
[37] Cole MA, Elliott RJ, Zhang J. Growth, foreign direct investment, and the
environment: evidence from Chinese cities. J Reg Sci 2011;51(1):121e38.
[38] Mbarek MB, Ali NB, Feki R. Causality relationship between CO2 emissions, GDP
and energy intensity in Tunisia. Environ Dev Sustain 2014;16(6):1253e62.
[39] Bella G, Massidda C, Mattana P. The relationship among CO 2 emissions,
electricity power consumption and GDP in OECD countries. J Policy Model
2014;36(6):970e85.
[40] Liddle B. What are the carbon emissions elasticities for income and popula-
tion? Bridging STIRPAT and EKC via robust heterogeneous panel estimates.
Glob Environ Chang 2015;31:62e73.
[41] Narayan PK, Saboori B, Soleymani A. Economic growth and carbon emissions.
Econ Modell 2016;53:388e97.
[42] Ajmi AN, Hammoudeh S, Nguyen DK, Sato JR. On the relationships between
CO2 emissions, energy consumption and income: the importance of time
variation. Energy Econ 2015;49:629e38.
[43] lu ST, Saeidpour L. Economic growth, CO 2 emissions, and
Heidari H, Katirciog
energy consumption in the five ASEAN countries. Int J Electr Power Energy
Syst 2015;64:785e91.
[44] Nasreen S, Anwar S, Ozturk I. Financial stability, energy consumption and
environmental quality: evidence from South Asian economies. Renew Sustain
Energy Rev 2017;67:1105e22.
[45] Apergis N, Payne JE. CO2 emissions, energy usage, and output in Central
America. Energy Policy 2009;37:3282e6.
[46] Apergis N, Payne JE, Menyah K, Wolde-Rufael Y. On the causality dynamics
between emissions, nuclear energy, renewable energy, and economic growth.
Ecol Econ 2010;69:2255e60.
[47] Arouri MEH, Youssef AB, M’henniH, Rault C. Energy consumption, economic
growth and CO 2 emissions in Middle East and North African countries. En-
ergy Policy 2012;45:342e9.
[48] Govindaraju VC, Tang CF. The dynamic links between CO2 emissions, eco-
nomic growth and coal consumption in China and India. Appl Energy
2013;104:310e8.
[49] Jabli MB. On the causal link between health indicator, output, combustible
renewable and waste consumption, rail transport and CO2 emissions: the case
of Tunisia. Environ Sci Pollut Control Ser 2016;23(16):16699e715.
[50] Linh DH, Khanh TV. CO2 emissions, energy consumption, economic growth
and agriculture development in ASEAN’s members. Empir Econ Rev
2017;7(1):39e56.
[51] Aye GC, Edoja. Effect of economic growth on CO2 emission in developing
countries: evidence from a dynamic panel threshold model. Cogent Econ
Financ 2017;5:1379239. https://doi.org/10.1080/23322039.2017.1379239.
[52] 
DOGAN 
B, DEGER O. The role of economic growth and energy consumption on
CO2 emissions in E7 countries. Theor Appl Econ 2018;2:231e46.
[53] Gorus MS, Mucahit A. The relationship between energy consumption,
10 S. Nasreen et al. / Energy 192 (2020) 116628
economic growth, and CO2 emission in MENA countries: causality analysis in
the frequency domain. Energy 2019;168(C):815e22.
[54] Sasana H, Aminata J. Energy subsidy, energy consumption, economic growth,
and carbon dioxide emission: Indonesian case studies. Int J Energy Econ Policy
2019;9(2):117e22.
[55] Nugraha AD, Osman NH. CO2 emissions, economic growth, energy con-
sumption, and household expenditure for Indonesia: evidence from cointe-
gration and vector error correction model. Int J Energy Econ Policy 2019;9(1):
291e8.
[56] Hossain SP. Panel estimation for CO2 emissions, energy consumption, eco-
nomic growth, trade openness and urbanization of newly industrialized
countries. Energy Policy 2011;39(11):6991e9.
[57] Pesaran MH. General diagnostic tests for cross section dependence in panels.
University of Cambridge, Faculty of Economics; 2004. Cambridge Working
Papers in Economics No. 0435.
[58] Bai J, Carrion-i-Silvestre J. Structural changes, common stochastic trends, and
unit roots in panel data. Rev Econ Stud 2009;76:471e501.
[59] Bai J, Ng S. A PANIC attack on unit root and cointegration. Econometrica
2004;72:1127e77.
[60] Pesaran M. Estimation and inference in large heterogeneous panels with a
multifactor error structure. Econometrica 2006;74(4):967e1012.
[61] Holtz-Eakin D, Newey W, Rosen HS. Estimating vector auto regressions with
panel data. Econometrica 1988;56(6):1371e95.
[62] Narayan PK, Smyth R. Multivariate Granger causality between electricity
consumption, exports and GDP: evidence from a panel of Middle Eastern
countries. Energy Policy 2009;37(1):229e36.
[63] Dobnik Frauke. Energy consumption and economic growth revisited: struc-
tural breaks and cross-section dependence. Essen: RWI; 2011. Ruhr Economic
Papers, No. 303, ISBN 978-3-86788-348-1.
[64] Halicioglu F. An econometric study of CO2 emissions, energy consumption,
income and foreign trade in Turkey. Energy Policy 2009;37:1156e64.
[65] Al-Mulali U, Weng-Wai C, Sheau-Ting L, Mohammed AH. Investigating the
environmental Kuznets curve (EKC) hypothesis by utilizing the ecological
footprint as an indicator of environmental degradation. Ecol Indicat 2015;48:
315e23.
[69] Kaika D, Zervas E. The Environmental Kuznets Curve (EKC) theorydPart A:
concept, causes and the CO 2 emissions case. Energy Policy 2013;62:
1392e402.
[70] Najid A, Liangsheng D, Xian-Liang T, Wang J. Chinese growth and dilemmas:
modelling energy consumption, CO2 emissions and growth in China. Qual
Quantity: Int J Methodol 2019;53(1):315e38.
[71] Eberhardt M, Teal F. Productivity analysis in global manufacturing pro-
duction.Discussion Paper 515. Department of Economics, University of Ox-
ford; 2010.
[72] Mercan M, Karakaya E. Energy Consumption, Economic Growth and Carbon
Emission: Dynamic Panel Cointegration Analysis for Selected OECD Countries.
Procedia Econ Financ 2015;23:587e92.