Environmental Science and Pollution Research (2019) 26:25218–25234

https://doi.org/10.1007/s11356-019-05737-3

RESEARCH ARTICLE

Do agricultural activities induce carbon emissions? The BRICS

experience
Daniel Balsalobre-Lorente 1 & Oana M. Driha 2 & Festus Victor Bekun 3 & Olawumi Abeni Osundina 4,5

Received: 14 May 2019 / Accepted: 10 June 2019 / Published online: 29 June 2019
# Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
This study investigates the environmental Kuznets curve (EKC) hypothesis for Brazil, Russia, India, China and South Africa
(BRICS) over the period 1990–2014, while considering agricultural activities, energy use, trade openness and mobile use as
driving forces of environmental degradation. The empirical results verify an inverted U-shaped connection between carbon
emissions and economic growth. This study confirms the unfriendly impact of agriculture on the environment. Electricity
consumption and trade openness likewise exhibit similar impacts on carbon emissions. Mobile use however reduces pollution.
A unique revelation from this study is that the interaction between electricity consumption and agricultural activities has an
additional pernicious effect on the environment. The methodologies applied for testing the impact of selected independent
variables on carbon emissions in BRICS are the Dynamic Ordinary Least Squares (DOLS) and the Fully Modify Ordinary
Least Square (FMOLS) for long run regression. Empirical evidence confirms that agriculture exerts a negative impact on the
environment in BRICS countries. This study therefore recommends the adoption of cleaner energy processes and enabling high-
tech and clean foreign investment.

Keywords EKC . Agricultural sector . Electricity consumption . Mobile use

Introduction perpetual increase in human activities that deplete the ozone

layer and degrade the environment and ecosystem at large.
The quest to mitigate global warming and reduce greenhouse Recent report from the Food and Agriculture Organization
gas emissions has been a bane of most—if not all—economies of the United Nation (FAO 2016) asserts that a large share
around the globe. Thus, the call by policymakers and govern- of carbon emissions are from the agriculture sector and
ment officials for a sustainable eco-friendly environment has Brazil, Russia, India, China and South Africa (BRICS) econ-
been on the rise (IPCC 2013). The issue of environmental omies are reputed to emit more than 40% of global total CO2
sustainability is a key priority for government administrators, emissions in 2013 (Liu et al. 2017a). The report also shows
environmentalists and energy economists. However, achiev- that GHG emissions from agriculture accounts for 21% of the
ing the targeted milestone seems near impossible due to the global aggregate and that agriculture is second largest emitter

Responsible editor: Philippe Garrigues

* Daniel Balsalobre-Lorente 1
Department of Political Economy and Public Finance, Economic and
[email protected] Business Statistics and Economic Policy, University of Castilla La
Mancha, Ciudad Real, Spain
Oana M. Driha 2
Department of Applied Economics, International Economy Institute,
[email protected] University of Alicante, Alicante, Spain
3
Department of Economics, Eastern Mediterranean University, North
Festus Victor Bekun Cyprus, Turkey
[email protected] 4
Department of Economics, Eastern Mediterranean University, North
Cyprus, Turkey
Olawumi Abeni Osundina 5
Department of Economics, Babcock University, Ilishan-Remo, Ogun
[email protected]
State, Nigeria
Environ Sci Pollut Res (2019) 26:25218–25234 25219

of GHG emissions. Agricultural GHG emission are primarily limitation of Cyprus being in a political unrecognized region
produced from deforestation, livestock emissions, soil, nutri- and the fact that arable land suffers from drought. This posi-
tion management in regard to the adoption of fossil fuel base tion is also resonated in a similar study conducted by
fertilizer, agricultural machinery and bush burning operation Katircioglu (2004). Also, in a study conducted for China by
as well as burning of biomass fuel. Li and Zhang (2009), it investigated the agricultural induced
It is on this premise that this study primarily focuses on EKC hypothesis between the periods of 1988 to 2006 for a
offering a new perspective to the already existing energy eco- panel of 31 provinces. Several agricultural indicators like fer-
nomics literature by looking into the environmental Kuznets tilizers, livestock, pesticides and poultry were used as measure
curve (EKC) concept which measures the trade-off between for agriculture. The study validated the agriculture-induced
quality of the environment and economic growth (GDP). This environmental degradation and subsequently enjoined
study augments the EKC concept through the inclusion of policymakers to strengthen agricultural environmental
agriculture (%GDP). This study specifically seeks to answer policy. More recently, the study of Chakravarty and Mandal
the following salient questions: (a) Do agricultural operations (2016) for BRICS confirms the detrimental impact of eco-
or activities trigger increased environmental degradation in nomic growth on the environment tested by the fixed effect,
Brazil, Russia, India, China and South Africa (BRICS), a bloc random effect and generalized method of moment (GMM)
of top emerging economies? (b) What sort of causality rela- method for annual time span of 1997–2011.
tionship exists between environmental quality measured Over the last two decades, growth trajectory has been on
through carbon dioxide emission (CO2) and agriculture and the rise, especially in emerging economies like BRICS.
its interaction with energy (electricity) consumption over re- Economist has forecasted that in the near future, BRICS econ-
cent time series data. omies would experience an unprecedented growth path.
This study fills the identified gap in literature by being the However, the economic growth comes with a trade-off. The
first to investigate the theme of EKC framework with these trade-off translates to detrimental effect on the environment
new insights, as opposed to previous studies that focus solely given that these economies like any other would extract her
on the inverted U-shaped pattern of relationship between eco- natural endowment like arable land exploitation like agricul-
nomic growth (GDP) and environmental quality. Plausible tural activities such as fishery, forestry and agronomy among
explanations to the earlier raised pertinent questions will help others in large scale for economic grain. Recently, the BRICS
policymakers and stakeholders simultaneously improve both countries have achieved a considerable rate of GDP growth
the energy mix and agricultural sector in BRICS countries. mainly due to structural changes associated with industriali-
Historically, the process of industrialization is closely asso- zation processes and a sectorial tertiarization (Dong et al.
ciated with heavy consumption of fossil fuels such as natural 2017). The GDP in the BRICS countries has increased from
gas and coal, among others (Gokmenoglu and Taspinarr US$2187 billion (constant in 2010) to US$16,266 billion be-
2018). The growth trajectory while transiting from a predom- tween 1985 and 2016, with an average annual growth rate of
inantly primary sector to an industrial sector by most devel- 6.5% (World Bank 2018). Several studies consider that the
oping and emerging economies such as BRICS exhibits a BRICS have the potential to position themselves economical-
trade-off between output levels and the environment. A major ly and technologically as world leaders (Azevedo et al. 2018;
defining characteristic of underdeveloped countries is their Danish et al. 2019; Danish and Wang 2019a, b). The fact that
heavy reliance on the primary sector of their economies, such these countries present higher levels of economic develop-
as mining, agriculture, forestry and quarrying. Although these ment, this leads them to promote stricter environmental regu-
activities guarantee increased economic output, they also pol- lations and structural changes towards a less polluting indus-
lute the environment. The pivotal role of the primary sector, trial sector and the tertiarization of their economic systems
mainly agriculture, is strengthened by the physiocracy school (Destek and Sarkodie 2019).
of thought which opines that agricultural activity is the pana- According to the existence of structural changes, one of the
cea for long term economic prosperity (Higgs 1897; Sertoglu key identified drivers of such growth path is traced to infor-
et al. 2017). This long-term economic growth however comes mation technology. An important example is the impact of
with severe environmental implications. Empirical validation ICT (information and telecommunication technologies), and
exists that supports the pivotal role of agriculture on economic by extension, the mobile use stretches across all tiers of econ-
growth See (Katircioglu 2004; Katircioglu 2006a, b, 2008). omies (developing, emerging and developed) and produces
For instance, the empirical study of Katircioglu (2006a) vali- positive multiplier effects. This is achievable, given the pene-
dates the significant role of agricultural activity on economic tration of mobile technology in altering the way agricultural
output as suggested by the feedback causality seen from the businesses and general market economies operate, thereby
Granger causality analysis for the case of Cyprus. The study translating into more research and development and conse-
also recommended policymakers to take pragmatic step to quently long-term economic growth. Bastida et al. (2019)
fortify the agricultural sector despite the geographical concluded that ICT technologies contribute positively to
25220 Environ Sci Pollut Res (2019) 26:25218–25234

reduce electricity usage, improve energy efficiency and thus (Munasinghe 1999; Destek and Sarkodie 2019). Finally, the
help to reducing carbon emissions. Therefore, it is expected third stage is where income per capita reaches a particular
that the use of mobile technology will contribute to the tran- threshold called maximum level and gives rise to a corre-
sition towards a cleaner and more efficient electricity system, sponding increase in clean environment. That is, at the given
with reductions in carbon levels through the adoption of more income level, there is a decline in environmental quality. This
efficient energy actions (EU 2014). Otherwise, these technol- stage is often observed in developed economies around the
ogies are compatible with automation processes in agriculture, globe where there is full consciousness of the importance of
capable of generating an ascending energy efficiency (Alfaris clean and friendly energy sources like renewable energy
et al. 2017). (Shahbaz and Sinha 2018). The technical effects will contrib-
The rest of this study is as follows: ‘Environmental ute to adapt the production to cleaner technologies, which in
Kuznets curve conceptualization and review of related litera- turn will reduce environmental damage (Álvarez et al. 2017;
ture’ presents the EKC conceptual framework and literature Katircioglu and Taşpinar 2017; Danish et al. 2017; Destek and
review, ‘Methodology construction and data source’ details Sarkodie 2019; Dogan and Seker 2016).
the methods applied, ‘Empirical results and discussion’ dis- Figure 1 gives a graphical insight into the conceptualization
cusses the empirical findings, and the concluding remarks and of the ideology of the EKC concept.
policy direction are seen in ‘Conclusion and policy The drive to mitigate global warming, particularly green-
implications’. house gas emissions (CO2), has been a bane for most—if not
all—economies. This is due to the global consciousness of the
implication of the pressure exerted on the ecosystem by hu-
Environmental Kuznets curve man activities. The relationship between output level and in-
conceptualization and review of related equality was discussed in the study of Simon Kuznets (1955),
literature and a phenomenon conceptualized as an inverted U-shaped
pattern between income level and inequality was introduced.
The argument of the inverse association that exists between This concept has been used slightly differently in the energy
per capita income and income inequality was first documented and environmental literature to explain the trade-off between
in the pioneering study conducted by Grossman and Krueger economic output level and environmental quality.
(1991). Environmental economists have since used this phe- The seminal study of Grossman and Krueger (1991) con-
nomenon to conceptualize the nexus between per capita in- ceptualized the inverted U-shape between environmental
come level and environmental quality in recent times. This pollution and per capita income level for the USA.
observed concept helps us to see the implication(s) of eco- Subsequently, several other studies have been documented
nomic growth trajectory on environmental quality. in the energy economics literature on the same theme.
More recent studies have also confirmed the EKC hypoth- Shafik and Bandyopadhyay (1992) conducted a similar study
esis (Sarkodie and Strezov 2018, Solarin et al. 2017, and traced the same inverted U-shaped pattern of relationship
Balsalobre et al. 2018). The growth trajectory consists of three between the quality of the environment and income level.
stages, namely (i) the scale effect stage, (ii) the composition Also, among earlier studies in the 1990s was that of
effect stage and (iii) the techniques stage (Balsalobre and Panayotou (1993) who investigated the nexus between the
Álvarez 2016a, b, Álvarez et al. 2017). quality of the environment and economic output level. The
The first stage is reflective of developing economies that findings from the study reiterated the earlier results of
rely mainly on their primary sector for sustenance. At this Grossman and Krueger (1991) and Shafik and
stage, emphasis is placed on the increase in economic output Bandyopadhyay (1992).1
relative to the quality of the environment. During this stage, The progression of the study on the EKC postulate over the
without changes in the economic structure and technology, years after the ground-breaking study of Simon Kuznets
ascending production also implies an increase in pollution (1955) has attracted more scholars in environmental econom-
levels (Danish et al. 2018). However, at the second stage, ics and related disciplines. More studies on the theme have
the consciousness of the citizenry is awakened to the impor- been conducted in single countries, panels of countries, and
tance of a clean environment; as such, the environment takes more recently, cross countries to either validate or refute the
the upper hand in the trade-off observed between income level existence of the hypothesis. The literature on the theme also
and environmental quality. This is mostly seen in transition observes diverse methodological innovation in estimations.
economies and emerging economies. During this second
stage, structural changes happen, such as a transition from
agriculture to industrial and service sectors. At this stage, the 1
More insight into the EKC hypothesis, interested reader can see Kijima et al.
countries implement more energy-efficient technologies, (2010), Pasten and Figueroa (2012), Hajko et al. (2018) and Shahbaz and
which directly contribute to reduce environmental damage Sinha (2018).
Environ Sci Pollut Res (2019) 26:25218–25234 25221

Fig. 1 Environmental Kuznets

curve conceptualization

For single-country studies, for instance, in Portugal, obvious by the Granger causality estimation where unidirec-
Shahbaz et al. (2010) explore the linkage between income tional causality is seen running from tourism to carbon dioxide
level and CO2 emissions. The study accounts for the role of emissions.
trade openness, urbanization, and energy use. The study For panels of countries, Apergis (2016) conducts a study
affirms the presence of EKC in Portugal. Guangyue and on the EKC hypothesis for 15 countries. The panel regression
Deyong (2011) investigate the theme for the province of affirms the EKC hypothesis for 12 of the sampled 15 coun-
China and reveal a positive correlation between carbon tries. The study further offers diverse implications to the dif-
emission level and economic growth trajectory over the ferent countries. Also, Balsalobre and Álvarez (2016a, b) in-
sampled period and provinces examined. This is also vestigate a panel of 17 countries in the Organization for
resonated in the study of Álvarez et al. (2017) and Bekun Economic Cooperation and Development (OECD) and vali-
et al. (2019). The same outcome is validated in the study of date the inverted U-shaped pattern between greenhouse gas
Balcilar et al. (2010) for G-7 countries. emission and per capita income in the period 1990–2012.
Other single-country studies include Shahbaz et al. (2012). Sinha et al. (2017) explore the trade-off between energy con-
The study confirms the inverse relationship between per capita sumption and environmental pollution, which was disaggre-
income level and quality of the environment in Pakistan. This gated into non-renewable energy and renewable sources,
indicates that Pakistan is still at a point where the focus is on while controlling for openness to the rest of the world and
output level relative to the quality of the environment. Thus, the role of urbanization and find an N-shaped pattern of rela-
Shahbaz (2012) joins the club of studies that confirm the EKC tionship in the sampled region. Katircioglu et al. (2018) stud-
hypothesis. In a similar study, Shahbaz et al. (2013) also ex- ied a bloc of top 10 tourism destinations and the implication of
plore the carbon-income nexus in Romania and corroborate ecological footprint. The study found that tourist’s activities
the EKC hypothesis. The study also validates the pollution induced environmental pollution. Contrary, the study conduct-
haven hypothesis in Romania. Tiwari et al. (2013) examine ed by Liu et al. (2017b) investigated the impact of renewable
the theme in India, using carbon emission from coal as the energy consumption and agriculture on carbon dioxide emis-
measure of environmental degradation. The study accounts sions for ASEAN countries. The study found no empirical
for a possible structural break in the econometric modelling evidence in support of the inverted U-shaped pattern for the
and joins the group of studies that supports the validity of the EKC hypothesis over the sampled period. However, the study
EKC hypothesis in India. The study submits that increased submitted that increased consumption of renewable energy
coal use triggers a higher level of greenhouse gas emissions and agriculture decreases CO2 emissions. Further, empirical
in India. This result is confirmed via Granger causality analy- revelation shows short-run Granger causality relationship ex-
sis. A unidirectional Granger causality is seen running from ist between non-renewable energy consumptions to CO2
coal (energy) consumption to real output level. Katircioglu emissions and to agriculture.
(2014) investigates the income-carbon relationship in In recent times, studies on the EKC hypothesis have aug-
Singapore, incorporating tourism in the conventional EKC mented the conventional EKC framework with other macro-
framework. The study traces a tourism-induced environmental economic variables of interest, such as energy consumption,
quality in Singapore over the considered period. This is made electricity consumption, foreign direct investment, trade
25222 Environ Sci Pollut Res (2019) 26:25218–25234

openness, urbanization and globalization, among others. Development Indicators database. Per capita real income is
Gokmenoglu and Taspinarr (2016) conduct a study in used as measure for economic performance (GDP in current
Turkey, controlling for the role of foreign direct investment USD), agriculture output (%GDP) as proxy for agriculture and
in the conventional EKC. The study affirms the pollution ha- mediation variable for the augmentation of the conventional
ven hypothesis, given that FDI inflows spur an increase in EKC hypothesis, trade openness measured as sum of countries
greenhouse gas emissions in Turkey within the investigated import and export as % GDP, electricity consumption per
period. Katircioglu and Katircioglu (2018) also investigate the capita (kWh per capita). Carbon dioxide emissions were in
EKC hypothesis while accounting for the role of urbanization. metric tonnes per capita and finally the access to mobile sub-
Contrarily, the study does not find any evidence in support of scription as proxy of ICT activities. Table 1 presents a sum-
the existence of EKC in Turkey. Al-Mulali et al. (2015) aug- mary of data source and unit of measurement.
ment the conventional EKC hypothesis in a pollution model The EKC hypothesis in BRICS countries is tested over the
with capital and labour in Vietnam. The study fails to trace the period 1990–2014, confirming the trade-off between econom-
existence of the EKC hypothesis with Pesaran’s ARDL meth- ic growth and environmental quality as earlier investigated
od as estimation approach. Chang (2015) investigates the (inter alia Grossman and Krueger 1991). Equation 1 tests the
theme for G-7 and BRICS countries simultaneously with the non-linear connection between economic growth (GDPpcit)
aid of data envelop analysis. The key findings of the study and per capita carbon emissions (CO2it). Additional explana-
reveal that the G-7 groups are more environmentally con- tory variables are incorporated as the share of agricultural
scious than BRICS countries within the sampled period. sector (AGRICit), electricity consumption (ELECit), trade
Sarkodie and Adams (2018) validated the EKC hypothesis openness (TOit) and mobile subscriptions (MOBit) as proxy
in South Africa, concluding that political institutional quality of ICT activities and digital development. Previous literature
exerts a huge role in the social, governance and economic underlines the key role of ICT and technological readiness in
readiness to mitigate climate change. Sarkodie and Strezov the mitigation of carbon emissions (Byrne et al. 2011; Bastida
(2018) confirmed the EKC in Australia, showing that environ- et al. 2019). To test the joint effect of electric consumption and
mental pollution depends on the share of disaggregate energy agriculture over carbon emissions, the interaction between
sources. Sarkodie and Strezov (2019) validated the EKC hy- these two variables (ELEC*AGRICit) is included in Eq. 1. A
pothesis for China and Indonesia. They also found the pollu- negative effect on environment is expected due to the devel-
tion haven hypothesis for China, India, Indonesia, Iran and oping stage of agriculture and the energy-mix outweighed by
South Africa. fossil fuels in BRICS countries. All these hypotheses are cap-
Literature focusing on the role of agricultural activities on tured in Eq. 1.
the ecosystem (environmental quality) commonly referred to
as the agriculture-induced EKC also exists. These studies ex- Model specification and empirical test process
amine the environmental implications of agricultural activities
such as the use of tractors and bush-burning, among others. The functional form for the conceptualization of the EKC
The literature on agriculture-induced EKC is still growing; as draws from the pioneer study of Grossman and Krueger
such, only a few studies exist. Ullah et al. (2018) show a (1991) and more recently after the studies of Shahbaz et al.
cointegration relationship between CO2 emissions and agri- (2013) and Gokmenoglu and Taspinarr (2018).
culture within the sampled period in Pakistan. The Granger
causality analysis shows diverse directional flows between
CO2 emissions and the various indicators of agricultural ac- CO2it ¼ β1 GDPpcit þ β2 GDPpc2 it þ β3 AGRICit
tivities. Similar studies were conducted in Pakistan
þ β 4 ELECit þ β5 TOit þ β6 ELEC  AGRIC
(Gokmenoglu and Taspinarr 2018), China (Dogan 2018) and
Turkey (Dogan 2016). All aforementioned studies validate the þ β 7 MOBit þ εit ð1Þ
agriculture-induced EKC. Thus, more friendly farming oper-
ations are encouraged in the investigated regions. The β coefficients reflect the relevance of independent var-
iables over carbon emissions as dependent variable (CO2it). εit
is the error term, which is normally distributed with a zero
Methodology construction and data source mean and with constant variance. Sub-index i denotes country
i, and t denotes the time dimension.
This current study examines the income-carbon nexus in This study’s empirical path is in four steps, namely (i) test
Brazil, Russia, India, China and South Africa (BRICS) by for stationarity properties of the series via the Fisher augment-
exploring the role of agricultural activities on environmental ed Dickey-Fuller (ADF) and Phillips-Perron (PP) panel unit
quality within the period 1990–2014 in a balanced panel root tests; (ii) test for cointegration among the variables
framework. All data were retrieved from the World through Fisher and Kao cointegration tests; (iii) long-run
Environ Sci Pollut Res (2019) 26:25218–25234 25223

Table 1 Data source and

description Indicator name Code measurement unit Source

Gross domestic product, per capita GDPPC Current 2010 $ USD WDI
Electricity consumption, per capita ELEC kWh per capita terms WDI
Carbon dioxide emissions, per capita CO2 metric tons per capita WDI
Agricultural sector, share GDP AGRIC (%GDP) WDI
Trade openness, share GDP TO (export plus import as % of GDP) WDI
Mobile, per 100 people MOB (Per 100 people) WDI

Source: Authors’ compilation. Note: World Bank (2018): World Development Indicators (WDI) available at:
https://data.worldbank.org/

regression to ascertain the magnitude of cointegration rela- that is built on that draws strength from joint individual test.
tionship; and finally, (iv) detection of the direction of causality The test null hypothesis of no cointegration is tested against an
through the Dumitrescu and Hurlin (2012). alternative of cointegration. The Fisher equation can be esti-
mated as follows:
Stationarity test
r
In empirical analysis, it is a norm to examine the stationarity ^zit ¼ ^ρzit−1 þ ∑ φ j Δ^
^zit−k þ vit ð3Þ
k¼1
traits of variables. This is pertinent in order to circumvent for
spurious analysis and misleading policy direction by exten- With H0 as no cointegration; thus, the ADF test statistics
sion. This is also in the bid to avoid variables integrated of can be given further as
order 2, i.e. ~ I(2). The current study is built on panel frame-
work. That combines both time series and cross-sectional di- pffiffiffiffiffiffiffiffiffiffiffi
mensions. Thus, panel unit root test renders more powerful t ADF þ 6N σ ^v =2^ σ
^2 0v
ADF ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð4Þ
and statistical relevance from the strength of pooling both time σ
^0v 2 3^σ^v 2
series and cross-sectional dimension (Baltagi 2008). Several þ
σ
^v 2 10^
2^ σ
^2 0v
panel unit root tests exist in the econometrics literature. These
test could be divided into two strands namely (a) homogenous here tADF is asymptotically normally distributed with N~(0,1)
panel unit root tests that assumes common intercept for all given by the sequential limit theory.
cross section (see Levin et al. 2002; Im et al. 2003) and (b)
the heterogeneous panel unit root that accommodates for het- Long-run coefficient regression
erogeneity and makes the assumption of different intercept for
the cross section (Fisher Dickey and Fuller 1979 and Fisher To estimate the magnitude of cointegration relationship traced
Phillips and Perron 1988). This study reports the Levin et al. in the course of the study, the Fully Modified Ordinary Least
(2002) panel unit root test specification for the augmented Squares (FMOLS) and the Dynamic Ordinary Least Squares
Dickey-Fuller2 is expressed as follows: (DOLS) estimator for long-run coefficients are employed.
According to Harris and Sollis (2003), FMOLS is a non-
k parametric estimator that is robust to autocorrelation. On the
Δyit ¼ μi þ ρyit−1 þ ∑ α j Δyit−1 þ δi t þ θt þ εit ð2Þ other hand, the DOLS estimator is parametric in nature. DOLS
j¼1
estimators take a lag of first difference term; that is, the lag,
From Eq. 2, ρ = 0 ∀iand the alternative hypothesis of ρ< 0 leads and contemporaneous values of the regressors are aug-
for at least one i. For a null of unit root against an alternative of mented when DOLS is employed.
stationarity. The cointegration parameters are obtained through the pan-
el pool group mean estimators of DOLS and FMOLS as spec-
ified as follows:
Cointegration test N
^ GFM * ¼ N −1 ∑ B* FMi ; where B* FMi is obtained as given
B
This study examines the long-run equilibrium association i¼1

among the variables under review. The Fisher panel in Eq. 1 ∀in the panel.
cointegration test is employed. The test is a residual-based test Similarly, the t ratio is computed as
* −1 N
2
For the want of space more insights into panel unit root test. Interested t −1 BGFM ¼ N 2 ∑ t * ð5Þ
i¼1 B FMI
readers can see bibliography section for references of lead papers.
25224 Environ Sci Pollut Res (2019) 26:25218–25234

In order to estimate cointegration, the equation is specified

as follows:

lnCO2it ¼ β0i þ β1 lnGDPit þ β2 lnAGRICit þ β 3 lnit þ β 4 ELEClnTOit þ β5 lnMOBit

q1 q1 q1 q1 q1
þ ∑ αq ΔlnGDPit þ ∑ λq ΔlnAGRICit þ ∑ φq ΔlnELECit þ ∑ ϕq ΔlnTOit þ ∑ γ q ΔlnMOBit þ εit ð6Þ
q¼−q1 q1¼−q1 q1¼−q1 q1¼−q1 q1¼−q1

From Eq. 6, the lag and lead are indicated by q and − q ΔAGRICit ¼ θ3 j þ λ3 j Σ it−1 þ ∑ θ33ir ΔAGRIC3it−r
r
respectively.
þ ∑ θ31ir ΔCO2it−r þ ∑ θ32ir GDPPCit−r
r r
Panel granger causality
þ ∑ θ33ir ELECit−r þ ∑ θ34ir TOit−r
r r
It is an established fact in econometric literature that regres-
sion does not depict causality between interest variables. The þ ∑ θ35ir MOBit−r þ ε3it ð10Þ
r
need for causality flow in the empirical study is pertinent for
the policy framework construction. Thus, the heterogeneous ΔELECit ¼ θ4 j þ λ4 j Σ it−1 þ ∑ θ44ir ΔELEC4it−r
r
non-causality panel test advanced by Dumitrescu and Hurlin
(2012) is used to investigate causal interactions between the þ ∑ θ41ir ΔCO2it−r þ ∑ θ42ir GDPPCit−r
r r
variables under consideration. The Dumitrescu and Hurlin
(DH) test is suitable where the cross-sectional dimensions þ ∑ θ43ir AGRICit−r þ ∑ θ44ir TOit−r
r r
are growing while the time dimension is non-stochastic. The
test also thrives where T > N, where T represents time and N þ ∑ θ45ir MOBit−r þ ε4it ð11Þ
r
stands for number of observation The DH test is built on the
VAR framework. The equation that reflects the liner form is ΔTOit ¼ θ5 j þ λ5 j Σ it−1 þ ∑ θ55ir ΔTO5it−r
r
expressed as follows:
p
ðpÞ
p
ðpÞ þ ∑ θ51ir ΔCO2it−r þ ∑ θ52ir GDPPCit−r
Z it ¼ ∑ φi yi;t−m þ ∑ β i xi;t−p þ εi;t ð7Þ r r
p¼1 p¼1
ðpÞ
þ ∑ θ53ir AGRICit−r þ ∑ θ54ir ELECit−r
where optimum lag in Eq.7 is denoted byp. φi is the r r
ðpÞ
autoregressive parameter, and βi
represents regression coef- þ ∑ θ55ir MOBit−r þ ε5it ð12Þ
r
ficients which varies within groups.
The dynamic model specification for this study is rendered ΔMOBit ¼ θ6 j þ λ6 j Σ it−1 þ ∑ θ66ir ΔMOB6it−r
r
as follows:
þ ∑ θ61ir ΔCO2it−r þ ∑ θ62ir GDPPCit−r
ΔCO2it ¼ θ1 j þ λ1 j Σ it−1 þ ∑ θ11ir ΔCO2it−r r r
r
þ ∑ θ63ir AGRICit−r þ ∑ θ64ir ELECit−r
þ ∑ θ12ir ΔGDPPCit−r þ ∑ θ13ir AGRICit−r r r
r r
þ ∑ θ65ir TOit−r þ ε6it ð13Þ
þ ∑ θ14ir ELECit−r þ ∑ θ15ir TOit−r r
r r
From the respective causality equations, r denotes the op-
þ ∑ θ16ir MOBit−r þ ε1it ð8Þ
r timum lag as suggested by Akaike information criterion.

ΔGDPPCit ¼ θ2 j þ λ2 j Σ it−1 þ ∑ θ22ir ΔGDPPC2it−r

r

þ ∑ θ21ir ΔCO2it−r þ ∑ θ22ir AGRICit−r Empirical results and discussion

r r

þ ∑ θ23ir ELECit−r þ ∑ θ24ir TOit−r This section presents and discusses the empirical results.
r r Table 2 presents the summary statistics and correlation matrix
þ ∑ θ25ir MOBit−r þ ε2it ð9Þ for selected variables for BRICS countries between 1990 and
r 2014 (World Bank 2018).
Environ Sci Pollut Res (2019) 26:25218–25234 25225

Table 2 Main statistics and

correlation matrix CO2 GDPPC AGRIC ELEC TO MOB

Mean 5.3559 3954.608 9.9435 2797.568 41.6910 40.0391

Median 2.8443 3059.587 6.1095 2218.884 44.8765 15.8630
Maximum 13.9799 16,007.090 27.6474 6617.136 110.5771 166.0376
Minimum 0.7115 298.2177 2.0977 273.0466 15.1617 0.0000
Std. dev. 4.1039 3662.650 7.6819 2002.311 16.6169 49.0401
Skewness 0.3996 1.4043 0.9089 0.3264 0.4521 1.0540
Kurtosis 1.5794 4.5153 2.4524 1.7215 3.7924 2.8314
Jarque-Bera 13.6164 52.1998 18.4742 10.5614 7.4085 22.9206
Probability 0.0011 0.0000 0.0000 0.0050 0.0246 0.0000
Sum 658.7785 486,416.7 1223.062 344,100.8 5128.001 4924.816
Sum sq. dev. 2054.738 1.64E+09 7199.507 4.89E+08 33,687.14 293,402.2
Correlation matrix
CO2 1.000000
GDPPC 0.429961 1.000000
AGRIC − 0.63312 − 0.64138 1.000000
ELEC 0.964780 0.587085 − 0.76996 1.000000
TO 0.733362 0.078192 − 0.356063 0.620938 1.000000
MOB 0.382221 0.846248 − 0.448132 0.477529 0.283595 1.000000

Source: Authors’ computation

Table 2 reports the basic statistics and correlation matrix variables, while the variables are stationary in first-difference
analysis. All variables exhibit positive averages and great dis- form. Thus, all variables are integrated of order 1 (I(1)).
persion from their means. Also, all variables considered are Next, after confirming that all variables are I(1), the panel
positively skewed. The correlation matrix shows a positive cointegration tests of Kao (1999) and Johansen (1988) are
relationship between per capita GDP and CO2 emissions. A used to scrutinize the equilibrium relationship among vari-
similar trend is seen between trade openness, mobile use, ables. Table 4 contains the results of the proposed
electricity per capita and CO2 emissions. An inverse associa- cointegration test confirming a long-run relationship between
tion is reported between agriculture and CO2 emissions. variables and avoiding biased or spurious regression results.
Inferences drawn from correlations are however not sufficient. The individual (country-by-country basis) cointegration tests
Thus, additional econometric tests are further conducted. are reported in the Appendix Table 7. The results are in har-
First, to explore the unit root properties, the augmented mony with the results from the Kao and Johansen- Fisher
Dickey-Fuller (ADF) and the Phillips-Perron (PP) unit root cointegration tests.
tests are applied and the results are depicted in Table 3. The Table 5 presents pairwise Dumitrescu and Hurlin (2012)
presence of unit roots are confirmed in the level form of the causality test results. This methodology offers efficient results
for both unbalanced panels and cross-sectional dependence.
The pairwise causal relationship among two variables appears
as unidirectional causality (X → Y or Y → X), bidirectional
Table 3 Panel unit root tests (ADF and PP)
causality (X ↔ Y) or no causality (X ≠ Y) among variables.
Variables At level At first difference Finally, Table 6 shows the econometric results of Fully
Modify Ordinary Least Square (FMOLS) and Dynamic
ADF PP ADF PP
Ordinary Least Squares (DOLS) methodologies, which are
CO2 1.0000 1.0000* 0.7579* 0.0028* suitable for long-run estimation and endogeneity problem.
GDPPC 0.9999 0.9999 0.2923* 0.0119* The results confirm an inverted U-shaped EKC connection
AGRIC 0.0862* 0.0022 0.3033* 0.0000* (Fig. 2) between economic growth and environmental degra-
ELEC 1.0000 1.0000 0.6605* 0.5282* dation for BRICS countries during the period 1990–2014. The
TO 0.9623 0.9697 0.0001* 0.0001* adverse effect of agricultural activities over environments is
MOB 0.0000 0.9792* 0.1191* 0.0999* highlighted. Therefore, it can be stated that agriculture is still
in a developing stage in BRICS countries, where scale effect
Authors’ computation dominates the composition and technical effects. Moreover,
*1%, **5%, ***10% significant rejection levels
25226 Environ Sci Pollut Res (2019) 26:25218–25234

Table 4 Kao and Johansen Fisher

panel cointegration tests (1) Kao Cointegration Test
t-Stat Prob.
ADF −3.305* (0.0005)
Residual variance 0.068
HAC variance 0.074
(2) Johansen-Fisher Panel Cointegration Test
Unrestricted Cointegration Rank Test (Trace and Maximum Eigenvalue)
Hypothesized Fisher Stat. Fisher Stat.
No. of CE(s) (from trace test) Prob. (from max-eigen test) Prob.
r=0 204.80* (0.0000) 125.50* (0.0000)
r≤1 126.70* (0.0000) 61.65* (0.0000)
r≤2 90.49* (0.0000) 54.31* (0.0000)
r≤3 49.83* (0.0000) 29.97* (0.0009)
r≤4 29.82* (0.0009) 31.24* (0.0005)
r≤5 7.57* (0.6707) 7.57* (0.6707)

Probabilities are computed using asymptotic chi-square distribution

*
1% statistical significance level

electricity consumption and trade openness also present a di- On the other hand, unidirectional causality is confirmed be-
rect impact over carbon emissions, while mobile use, as proxy tween GDPpc → AGRIC, AGRIC → CO2, MOB → ELEC and
of ICT, mitigates emissions. The main contribution of this MOB → AGRIC. The unidirectional causality running from
study is the evaluation of the joint impact of agriculture and agricultural activities to carbon emissions (AGRIC → CO2)
electricity on carbon emissions under the EKC approach. strengthens the positive connection between agricultural activi-
Figure 2 provides the results of Dumitrescu-Hurlin causality ties and environmental damage, and how in early stages of eco-
tests. nomic development, agricultural dependence reinforces the scale
Bidirectional causality is predominant: CO2 ↔ GDPpc, effect over composition and technical effect (Balsalobre and
CO2 ↔ TO, CO 2 ↔ ELEC, CO2 ↔ MOB, GDPpc ↔ Álvarez 2016a, b). Finally, there is no causality between TO ≠
ELEC, GDPpc ↔ TO, GDPpc ↔ MOB, TO ↔ MOB, TO ELEC.
↔ AGRIC and ELEC ↔ AGRIC. Several studies support The FMOLS and DOLS estimations (Table 6) provide signif-
bidirectional causality between carbon emissions and eco- icant coefficients, confirming the EKC hypothesis for BRICS
nomic growth (Rehdanz and Maddison 2008; Halicioglu (Fig. 3). The findings also reveal that agricultural sector, electric-
2009) and electricity use and economic growth (Yoo 2006; ity use and trade openness damage the environment in BRICS
Chang 2010; Aslan 2014; Lu 2017), confirming a feedback countries. By contrast, mobile use contributes to reduction car-
nexus electricity use and economic growth. These results are bon emissions in BRICS countries. Consequently, agriculture,
in line with Sebri and Ben-Salha (2014) and Apergis and electricity consumption (dominated by fossil sources) and trade
Ozturk (2015) evidence for BRICS. openness have negative effects on the environment in BRICS.
The bidirectional relationship between economic growth As shown in Fig. 4, an inverted U-shaped EKC relationship
and trade openness (GDPpc ↔ TO) is confirmed. This sup- between economic growth and carbon emissions in the BRICS
ports the findings of Klasra (2011), Shahbaz (2012) and Akin countries is validated. This is similar to the results of Tamazian
(2014). Furthermore, based on the results, it can be suggested et al. (2009). The EKC’s inverted U-shaped behaviour corre-
that trade openness infers indirectly over carbon emissions, sponds with coefficients β1 > 0 and β2 < 0, which assume that
when in a developing stage of economic growth scale effect during the first stage of economic growth, carbon emissions in-
is predominant (Jayanthakumaran et al. 2012). In line with crease up to the turning point (X1 = 11,192.05USD), after which
Aydin (2016), the effects of ascending economic growth over the emissions are mitigated by economic growth (see Fig. 4).
carbon emissions will depend mainly on the strength of scale Figure 4 illustrates that India, China and South Africa are still
effects.3 in a developing stage, while Brazil and Russia already passed to a
negative relationship between economic growth and carbon
emissions. Thus, India, China and South Africa need to reinforce
3
The scale effect (Fig. 1) reflects the effect of increased production or con- their economic systems and technical advances to enhance a
sumption resulting from trade opening or freer trade on carbon emissions
(Aydin 2016). Therefore, trade openness increases economic activity and, sustainable development stage. In order to do so, these countries
hence, energy consumption. should adopt new policy strategies focused on attracting cleaner
Environ Sci Pollut Res (2019) 26:25218–25234 25227

Table 5 Pairwise Dumitrescu-

Hurlin panel causality tests Null hypothesis Causality W-stat. Zbar-stat. Prob.

GDPPC ≠ > CO2 GDPPC ↔ CO2 7.0127* 4.1391 (0.0000)

CO2 ≠ > GDPPC 5.4261* 2.7582 (0.0058)
AGRIC ≠ > CO2 AGRIC → CO2 4.8040** 2.2167 (0.0266)
CO2 ≠ > AGRIC 1.9964 − 0.2269 (0.8205)
ELEC ≠ > CO2 ELEC ↔ CO2 8.8105* 5.7038 (0.0000)
CO2 ≠ > ELEC 5.4727* 2.7987 (0.0051)
TO ≠ > CO2 TO ↔ CO2 7.4753* 4.5417 (0.0000)
CO2 ≠ > TO 5.1879** 2.5508 (0.0107)
MOB ≠ > CO2 MOB ↔ CO2 7.1641* 4.2708 (0.0000)
CO2 ≠ > MOB 7.4722* 4.5390 (0.0000)
AGRIC ≠ > GDPPC GDPPC → AGRIC 2.5944 0.3019 (0.7627)
GDPPC ≠ > AGRIC 5.0076** 2.4169 (0.0157)
ELEC ≠ > GDPPC ELEC ↔ GDPPC 6.3550* 3.5978 (0.0003)
GDPPC ≠ > ELEC 5.2685* 2.6456 (0.0082)
TO ≠ > GDPPC TO ↔ GDPPC 8.4222* 5.4097 (0.0000)
GDPPC ≠ > TO 13.1414* 9.5458 (0.0000)
MOB ≠ > GDPPC MOB ↔ GDPPC 11.5656* 8.1647 (0.0000)
GDPPC ≠ > MOB 7.4509* 4.5584 (0.0000)
ELEC ≠ > AGRIC ELEC ↔ AGRIC 10.2089* 6.9756 (0.0000)
AGRIC ≠ > ELEC 4.9932** 2.4043 (0.0162)
TO ≠ > AGRIC TO ↔ AGRIC 4.7805** 2.2179 (0.0266)
AGRIC ≠ > TO 17.9988* 13.803 (0.0000)
MOB ≠ > AGRIC MOB → AGRIC 11.9874* 8.5344 (0.0000)
AGRIC ≠ > MOB 2.8774 0.5499 (0.5824)
TO ≠ > ELEC TO ≠ ELEC 2.6511 0.3515 (0.7252)
ELEC ≠ > TO 1.3345 − 0.8023 (0.4224)
MOB ≠ > ELEC MOB → ELEC 5.7775** 3.0917 (0.0020)
ELEC ≠ > MOB 3.5691 1.1562 (0.2476)
MOB ≠ > TO MOB ↔ TO 5.0424** 2.4474 (0.0144)
TO ≠ > MOB 5.8738** 3.1761 (0.0015)

Source: Authors computation. Also, the symbols here ≠, → represent no Granger causality, one-way causality and
bi-directional causality respectively. Also, ≠ > means ‘does not Granger cause’
*1%, **5%, ***10% significant rejection levels

and high-tech foreign businesses. Developing countries often emissions into the atmosphere (Cole et al. 1997; Paustian et al.
attract dirty foreign investment due to the lax environmental 1998; Ghorbani et al. 2011; Thamo et al. 2013) and thus to global
regulations and abundance of natural resources (Balsalobre and warming (Lal 2004; Nakagawa et al. 2007; Ozkan et al. 2007).
Shahbaz 2016; Balsalobre et al. 2018). The results also confirm a Havemann (2014) concluded that agriculture is the major source
positive connection between agriculture and carbon emissions of greenhouse gas emissions and recommended the adoption of
(β3 > 0). This is in consonance with conclusions from previous global measures to prevent further environmental disruption.
literature that agricultural activities in BRICS countries lead to The coefficient (β4 > 0) reveals a positive connection be-
increment in carbon emissions (Cole et al. 1997; Paustian et al. tween electricity consumption and carbon emissions in
1998). The damaging effect of agriculture on the environment BRICS,5 enforcing the results of Pao and Tsai (2010). The
through its negative impact on water and land resources is not direct connection between energy use and carbon emissions
new in economic literature. Many studies pointed out the relevant
contribution of traditional agriculture4 to the greenhouse gas 5
The BRICS countries accounted for 33% of world energy use and nearly
40% of global emissions in 2014: the Brazil (1.4%), the Russia (5.2%), India
(6.2%), the China (25.9%) and South Africa (1.2%). Countries on CO2 emis-
4
Agriculture is one of the greatest contributors to global warming, generating sions per capita reveal that China has the lowest carbon intensity (1.88 kg/
nearly one-third of global anthropogenic greenhouse gas emissions (FAO USD) and Russia has the highest CO2 emissions per capita (11.56 tCO2/capita)
2003). (World Bank 2018).
25228 Environ Sci Pollut Res (2019) 26:25218–25234

Table 6 FMOLS and DOLS The positive connection between trade openness and
Dependent variable: CO2 carbon emissions in BRICS countries (β5 > 0) confirms
that trade openness contributes to accelerated carbon
Variable FMOLS DOLS emissions.6 The role of BRICS in the global trade has
increased significantly over the last two decades. Trade
GDPPC 0.000338* 0.000329*
liberalization in emerging economies encourages special-
[3.2320] [4.101756]
ization in ‘dirty industries’, this consequently aggravates
(0.0016) (0.0001)
pollution (Dean 1992a, b). The empirical results rein-
GDPPC^2 − 1.51E−08* − 1.30E−08*
force this theory, showing a direct connection between
[− 3.2200] [− 3.6451] trade openness and carbon emissions in BRICS. In or-
(0.0017) (0.0004) der to moderate the impact of trade openness on the
AGRIC 0.045367* 0.035924** environment, higher environmental control is required
[2.0815] [2.3138] as well as stronger regulation linked to foreign invest-
(0.0398) (0.0225) ment in dirty industries. Moreover, the interaction be-
ELEC 0.001170* 0.001072* tween agriculture and electricity consumption (β6 > 0)
[6.5530] [7.8640] indicates additional pernicious consequences over envi-
(0.0000) (0.0000) ronment (Soni et al. 2013).
TO 0.023142* 0.021370 Finally, mobile use, as proxy of ICT, contributes to
[4.1971] [5.1153] the reduction of carbon emissions in BRICS countries
(0.0001) (0.0000) (β7 < 0). Some reports stipulate that the use of mobile is
ELEC*AGRIC 3.64E−05** 4.42E−05* currently enabling a total reduction of 180 million tons
[2.2927] [3.6717] of carbon emissions a year across the USA and Europe
(0.0238) (0.0004) (GMR 2015).7 In addition, some researchers reveal that
MOB − 0.005101* − 0.006151* ICT activities diminish carbon emissions (Levendis and
[− 1.9258] [− 3.0305] Lee 2013; Qureshi 2013a), higher living standards
(0.0568) (0.0030) (Chavula 2013), externalities in welfare (Carmody
R2 0.994284 0.994617 2013; Qureshi 2013b) or sustainable development
Adjusted R2 0.993696 0.994083 (Byrne et al. 2011). While a lot of attention was paid
S.E. of regression 0.326673 0.315668 to human development and socioeconomic benefits as-
Long-run variance 0.152490 0.090295 sociated with ICT, mobile usage and its impact on en-
Mean dependent var 5.430046 5.355923 vironmental sustainability have been less explored
S.D. dependent var 4.114540 4.103915 (Krishnadas and Radhakrishna 2014).
Sum squared resid 11.41852 11.06076

Source: Authors’ computation. While the symbols [ ] represent T-statis-

tics, numbers in ( ) are probability values Conclusion and policy implications
*1%, **5%, ***10% statistical significant levels
The environmental damage generated by agriculture is linked
to deforestation, livestock emissions, nutrient management,
has been widely explored in economic literature (Ang
soil and fossil fuel use, fertilizer usage, burning biomass and
2007, 2009; Halicioglu 2009). This positive connection
agricultural machinery operation that deplete the environment
between electricity use and carbon emissions suggests
(Liu et al. 2017a).
the overweight of fossil fuel in the energy mix of
It has also been observed as asserted by several stud-
BRICS countries. Under an additional need of energy,
ies (Challinor et al. 2014, Rosenzweig et al. 2014) of
as a consequence of an expanding economic cycle, the
the connection that exists between agriculture and
share of non-renewable sources in the energy mix will
climate change. FAO (2016) states that 21% of total
determine the carbon emissions (Balsalobre and Álvarez
global emissions are generated by agriculture; this
2016a, b; Balsalobre et al. 2018). In other words, for an
makes the agricultural sector the second highest
additional economic growth, the use of fossil fuels will
6
accelerate carbon emissions (Paul and Bhattacharya While in 1990, BRICS accounted for 3% of global trade and, in 2014,
managed to reach about 15% of the global exports and 14% of global imports
2004; Oh and Lee 2004). Therefore, BRICS must pro- of commodities (World Bank 2018).
mote cleaner and more efficient energy processes to 7
The SMARTer-2030 forecasts that ICT might facilitate a 20% reduction of
mitigate the pollution levels due to the high share of global carbon emissions by 2030, i.e. holding pollution to recent levels
non-renewable sources. (Accenture Strategy 2015).
Environ Sci Pollut Res (2019) 26:25218–25234 25229

Fig. 2 Pairwise Dumitrescu-

Hurlin panel causality tests
scheme

contributor to pollution. Achieving the goal of reduced The empirical evidence points out a positive connec-
greenhouse gas from agricultural activates will not only tion between electricity consumption and carbon emis-
translate into cleaner ecosystem and more source of sions. This result suggests a high proportion of a fossil
livelihood as more farming operation can be undertaken source in the energy mix. Our study reveals that BRICS
(Schleussner et al. 2016; IPCC 2014). In the long run, countries must intensify investment in cleaner electrical
increased living standard was for all farmers in the val- infrastructures, paying more attention to environmentally
ue chain (Hasegawa et al. 2015). friendly projects through the promotion of renewable
Thus, this research explored the impact of agriculture sources (Emir and Bekun 2018; Akadiri et al. 2019) in
on the carbon emission-income function, using FMOLS order to expand their production capabilities, generating
and DOLS econometric techniques, under the EKC ap- higher electrical efficiency, more energy security and a
proach. The results confirm the inverted U-shaped EKC reduction in emissions (Cowan et al. 2014).
pattern between income level and carbon emissions for In addition, empirical evidence is given regarding the
BRICS from 1990 to 2014. This seems to be in line linkage between agriculture and increases in carbon
with the assumption that BRICS countries are among emissions. The interaction between electricity consump-
the largest global carbon emitters. Therefore, environ- tion, agriculture and emissions reinforces previous liter-
mental degradation control might require mitigating ature which concludes that agriculture is highly support-
measures and a closer cooperation among BRICS in ed by fossil fuel consumption in emerging economies
energy and environmental aims. (Soni et al. 2013; FAO 2016; Liu et al. 2017a). Our

Fig. 3 Conceptual scheme:

FMOLS and DOLS estimation
results
25230 Environ Sci Pollut Res (2019) 26:25218–25234

Fig. 4 Inverted U-shaped EKC in

BRICS

estimation results also suggest the promotion of clean It is worth mentioning that the effect of foreign in-
renewable energies, such as solar or wind energy, in vestment and the role of renewable and non-renewable
order to achieve cleaner agricultural production, with energy uses over agriculture should be considered in
positive effects on the environment (Jebli and Youssef further research in order to facilitate additional recom-
2017). To achieve an improvement in agricultural pro- mendations for emerging countries.
duction, it is also essential that governments promote In summary, it is pertinent for pragmatic steps be
more efficient energy infrastructures and subsidize the taken to reinforce environmental regulations in BRICS
conversion to cleaner and more efficient energy sources countries given the findings from this current study.
in the agricultural sector (Mushtaq et al. 2007; Sebri This is necessary in order for the blocs investigated to
and Abid 2012). prioritize global carbon emission reduction from agricul-
Furthermore, a direct connection between trade open- tural operation and non-renewable energy source con-
ness and carbon emission is confirmed, as a conse- sumptions. Thus, it is on this premise that the following
quence of global trade liberalization, which has fostered action directions are rendered for policymakers and
foreign investment in industries with pernicious environ- stakeholders in the countries investigated. So we consid-
mental effects (Shahbaz et al. 2019). Consequently, it is er the need for more efficient and environmental friend-
suggested that to promote cleaner industries and indus- ly energy usage is primary requisite that triggers the
trialization strategies, focus should be on attracting paradigm shift from non-renewable energy consumption
high-tech foreign investments. to renewable energy consumption. The promotion of
Finally, mobile usage mitigates environmental dam- energy technologies is key to reduction of carbon emis-
age. Hence, the adopted model exhibits an optimistic sion especially CO2 emission in order to achieve the
role of mobile technologies in BRICS. In consequence, global climate goal targets. Finally, the energy security
consistent results are provided suggesting the need of is also necessary in BRICS energy sector, with regard to
promoting cleaner energy sources by adapting existing agricultural sector.
policies. Our results validate that mobile technologies As a direction for other researcher on the current
and their application will help to reduce some limita- theme, future studies on this topical issue can query
tions in traditional sectors such as agriculture, being other regions to explore the validity of the relationship
an effective tool for communication and organizational between energy consumption and economic growth nex-
changes (Swaminathan and Swaminathan 2018). us and how its interplay on the ecosystem and environ-
Therefore, policymakers should design and implement ment at large. The need to account for demographic
regulations focused on boosting renewable sources and indicators like population, rule of law, good governance
more efficient and high-tech agriculture infrastructure and other macroeconomic variables not captured in this
reducing environmental pressure. study is also encouraged.
Environ Sci Pollut Res (2019) 26:25218–25234 25231

Appendix
Table A.1 Individual cross-
section results Trace test Max-Eign test

Cross section Stat. Prob. Stat. Prob

Hypothesis of no cointegration
Brazil 126.7487* (0.0001) 47.0277* (0.0071)
China 276.9929* (0.0000)* 99.6695* (0.0000)
India 225.2893* (0.0000) 84.9782* (0.0000)
Russian Federation 181.4312* (0.0000) 89.8399* (0.0000)
South Africa 129.6232* (0.0000) 57.7480* (0.0002)
Hypothesis of at most 1 cointegration relationship
Brazil 79.7210* (0.0066) 32.7093*** (0.0684)
China 177.3234* (0.0000) 66.1185* (0.0000)
India 140.3111* (0.0000) 52.2365* (0.0001)
Russian Federation 91.5913* (0.0004) 35.5788*** (0.0310)
South Africa 71.8752** (0.0339) 28.5084 (0.1910)
Hypothesis of at most 2 cointegration relationship
Brazil 47.0117** (0.0599) 23.9870 (0.1352)
China 111.2049* (0.0000) 57.6978* (0.0000)
India 88.0745* (0.0000) 44.3965 (0.0002)
Russian Federation 56.0125* (0.0071) 22.6724 (0.1879)
South Africa 43.3668* (0.1239) 19.0791 (0.4085)
Hypothesis of at most 3 cointegration relationship
Brazil 23.0248 (0.2448) 15.7285 (0.2411)
China 53.5071* (0.0000) 27.4091* (0.0057)
India 43.6780* (0.0007) 24.4269** (0.0165)
Russian Federation 33.3401** (0.0187) 20.8562** (0.0546)
South Africa 24.2877 (0.1886) 15.6045 (0.2488)
Hypothesis of at most 4 cointegration relationship
Brazil 7.2962 (0.5434) 7.2703 (0.4577)
China 26.0980 (0.0009) 25.5659* (0.0006)
India 19.2510 (0.0129) 16.7843** (0.0196)
Russian Federation 12.4839 (0.1352) 12.2650*** (0.1011)
South Africa 8.6832 (0.3955) 8.5828 (0.3225)
Hypothesis of at most 5 cointegration relationship
Brazil 0.0260 (0.8719) 0.0260 (0.8719)
China 0.5322 (0.4657) 0.5322 (0.4657)
India 2.4667 (0.1163) 2.4667 (0.1163)
Russian Federation 0.2189 (0.6399) 0.2189 (0.6399)
South Africa 0.1005 (0.7513) 0.1005 (0.7513)

**MacKinnon-Haug-Michelis (1999) p values *1%; **5%.***10% statical significance level

References AlFaris F, Juaidi A, Manzano-Agugliaro F (2017) Intelligent homes’

technologies to optimize the energy performance for the net zero
Accenture Strategy (2015) SMARTer2030: ICT solutions for 21st century energy home. Energy Building 153:262–274
challenges. Global e-Sustainability Initiative, Brussels Al-Mulali U, Saboori B, Ozturk I (2015) Investigating the environmental
Akadiri S, Alkawfi MM, Ugural S, Akadiri AC (2019) Towards achiev- Kuznets curve hypothesis in Vietnam. Energy Policy 76:123–131
ing environmental sustainability target in Italy. The role of energy, Álvarez A, Balsalobre D, Shahbaz M, Cantos JM (2017) Energy innova-
real income and globalization. Sci Total Environ 671:1293–1301 tion and renewable energy consumption in the correction of air
Akin CS (2014) The impact of foreign trade, energy consumption and pollution levels. Energy Policy 105:386–397
income on CO2. Emissions Int J Energy Econ Policy 4(3):465–475
25232 Environ Sci Pollut Res (2019) 26:25218–25234

Ang JB (2007) CO2 emissions, energy consumption, and output in Danish, Wang Z (2019a) Dynamic relationship between tourism, eco-
France. Energy Policy 35(10):4772–4778 nomic growth, and environmental quality. J Sustain Tour 0:1–16
Ang JB (2009) CO2 emissions, research and technology transfer in Danish, Wang Z (2019b) Does biomass energy consumption help to
China. Ecol Econ 68(10):2658–2665 control environmental pollution? Evidence from BRICS countries.
Apergis N (2016) Environmental Kuznets curves: new evidence on both Sci Total Environ 670:1075–1083
panel and country-level CO2 emissions. Energy Econ 54:263–271 Danish, Zhang B, Wang B, Wang Z (2017) Role of renewable energy and
Apergis N, Ozturk I (2015) Testing environmental Kuznets curve hypoth- non-renewable energy consumption on EKC: evidence from
esis in Asian countries. Ecol Indic 52:16–22 Pakistan. J Clean Prod 156:855–864
Aslan A (2014) Causality between electricity consumption and economic Danish, Wang B, Wang Z (2018) Imported technology and CO2 emission
growth in Turkey: an ARDL bounds testing approach. Energy in China: collecting evidence through bound testing and VECM
Sources 9:25–31 approach. Renew Sustain Energy Rev 82:4204–4214
Aydin L (2016) Intra-BRICS trade opening and its implications for car- Danish BMA, Mahmood N, Zhang JW (2019) Effect of natural resources,
bon emissions: a general equilibrium approach. J Econ Dev Stud renewable energy and economic development on CO2 emissions in
4(2):207–218 BRICS countries. Sci Total Environ 678:632–638
Azevedo VG, Sartori S, Campos LMS (2018) CO2emissions: a quanti- Dean JM (1992a) Trade and the environment: a survey of the literature.
tative analysis among the BRICS nations. Renew Sustain Energy In: Bank, T.W (ed) World development report 1992. World devel-
Rev 81:107–115 opment report office. World Bank, Washington, pp 13–17
Balcilar M, Ozdemir ZA, Arslanturk Y (2010) Economic growth and Dean JM (1992b) Trade policy and the environment: developing country
energy consumption causal nexus viewed through a bootstrap concerns. OECD, The Environment Directorate
rolling window. Energy Econ 32(6):1398–1410 Destek MA, Sarkodie SA (2019) Investigation of environmental Kuznets
Balsalobre D, Álvarez A (2016a) Economic growth and energy regulation curve for ecological footprint: the role of energy and financial de-
in the environmental Kuznets curve. Environ Sci Pollut Res 23(16): velopment. Sci Total Environ 650:2483–2489
16478–16494 Dickey DA, Fuller WA (1979) Distribution of the estimators for
Balsalobre D, Shahbaz M (2016) Energy consumption and trade open- autoregressive time series with a unit root. J Am Stat Assoc
ness in the correction of GHG levels in Spain. Bull Energy Econ 74(366a):427–431
4(4):310–322
Dogan N (2016) Agriculture and environmental kuznets curves in the
Balsalobre D, Shahbaz M, Roubaud D, Farhani S (2018) How economic case of Turkey: evidence from the ARDL and bounds test. Agric
growth, renewable electricity and natural resources contribute to Econ – Czech 62:566–574
CO2 emissions? Energy Policy 113:356–367
Dogan N (2018) The impact of agriculture on CO2 emissions in China.
Baltagi B (2008) Econometric analysis of panel data. John Wiley and
Panoeconomicus 66(2):257-271
Sons, United Kingdom
Dogan E, Seker F (2016) The influence of real output, renewable and
Bastida L, Cohen JJ, Kollmann A, Moya A, Reichl J (2019) Exploring the
non-renewable energy, trade and financial development on carbon
role of ICT on household behavioural energy efficiency to mitigate
emissions in the top renewable energy countries. Renew Sustain
global warming. Renew Sust Energ Rev 103:455–462
Energy Rev 60:1074–1085
Bekun FV, Emir F, Sarkodie SA (2019) Another look at the relationship
Dong K, Sun R, Hochman G (2017) Do natural gas and renewable energy
between energy consumption, carbon dioxide emissions, and eco-
consumption lead to less CO2emission? Empirical evidence from a
nomic growth in South Africa. Sci Total Environ 655:759–765
panel of BRICS countries. Energy 141:1466–1478
Byrne E, Nicholson B, Salem F (2011) Information communication tech-
nologies and the millennium development goals. Inf Technol Dev Dumitrescu EI, Hurlin C (2012) Testing for granger non-causality in
17(1):1–3 heterogeneous panels. Econ Model 29(4):1450–1460
Carmody P (2013) A knowledge economy or an information society in Emir F, Bekun FV (2018) Energy intensity, carbon emissions, renewable
Africa? Thintegration and the mobile phone revolution. Inf Technol energy, and economic growth nexus: new insights from Romania.
Dev 19(1):24–39 Energy and Environment 30(3):427-443
Chakravarty D, Mandal SK (2016) Estimating the relationship between EU (2014) European Commission staff working document, cost-benefit
economic growth and environmental quality for the BRICS analyses & state of play of smart metering deployment in the EU-27.
economies-a dynamic panel data approach. J Dev Areas 50(5): Smart Grids Taskforce SWD 189:2014
119–130 FAO (2016) Food and Agriculture Organization: The State of Food and
Challinor AJ, Watson J, Lobell DB, Howden SM, Smith DR, Chhetri N Agriculture. http://www.fao.org/3/a-i6030e.pdf
(2014) A meta-analysis of crop yield under climate change and Ghorbani R, Mondani F, Amirmoradi S, Feizi H, Khorramdel S, Teimouri
adaptation. Nature Climate Change 4, 287–291 M, Sanjani S, Anvarkhah S, Aghel H (2011) A case study of energy
Chang CC (2010) Multivariate causality test of carbon dioxide emissions, use and economical analysis of irrigated and dryland wheat produc-
energy consumption and economic growth in China. Appl Energy tion systems. Appl Energy 88:283–288
87:3533–3537 GMR (2015) Gesi Mobile report: how mobile communications technol-
Chang MC (2015) Room for improvement in low carbon economies of ogy is enabling carbon emissions reduction. Available at https://
G7 and BRICS countries based on the analysis of energy efficiency www.carbontrust.com/media/672238/mobile-carbon-impact-
and environmental Kuznets curves. J Clean Prod 99:140–151 ctc856.pdf. (Accessed 29 Jan 2019)
Chavula HK (2013) Telecommunications development and economic Gokmenoglu KK, Taspinarr N (2016) The relationship between CO2
growth in Africa. Inf Technol Dev 19(1):5–23 emissions, energy consumption, economic growth and FDI: the case
Cole CV, Duxbury J, Freney J, Heinemeyer O, Minami K, Mosier A, of Turkey. J Int Trade Econ Dev 25(5):706–723
Zhao Q (1997) Global estimates of potential mitigation of green- Gokmenoglu KK, Taspinarr N (2018) Testing the agriculture-induced
house gas emissions by agriculture. Nutr Cycl Agroecosyst 49(1–3): EKC hypothesis: the case of Pakistan. Environ Sci Pollut Res 25:
221–228 1–13
Cowan WN, Chang T, Inglesi-Lotz R, Gupta R (2014) The nexus of Grossman GM, Krueger AB (1991) Environmental impacts of a North
electricity consumption, economic growth and CO2 emissions in American free trade agreement. National Bureau of Economic
the BRICS countries. Energy Policy 66:359–368 Research 3914
Environ Sci Pollut Res (2019) 26:25218–25234 25233

Guangyue X, Deyong S (2011) An empirical study on the environmental Klasra MA (2011) Foreign direct investment, trade openness and eco-
Kuznets curve for China’s carbon emissions: based on provincial nomic growth in Pakistan and Turkey: an investigation using bounds
panel data. Chin J Popul Resour Environ 9(3):66–76 test. Qual Quant 45:223–231
Hajko V, Sebri M, Al-Saidi M, Balsalobre D (2018) The energy-growth Krishnadas N, Radhakrishna RP (2014) Green information technology:
nexus: history, development, and new challenges. In: Menegaki AN literature review and research domains. J Manag Syst 24(1):57–79
(ed) The economics and econometrics of the energy-growth nexus, Kuznets S (1955) Economic growth and income inequality. Am Econ
pp 1–46 Rev 45(1):1–28
Halicioglu F (2009) An econometric study of CO2 emissions, energy Lal R (2004) Carbon emission from farm operations. Environ Int 30(7):
consumption, income and foreign trade in Turkey. Energy Policy 981–990
37(3):1156–1164 Levendis J, Lee SH (2013) On the endogeneity of telecommunications
Harris R, Sollis R (2003) Applied time series modelling and forecasting. and economic growth: evidence from Asia. Inf Technol Dev 19(1):
John Wiley and Sons, Chichester 62–85
Hasegawa T, Fujimori S, Shin Y, Tanaka A, Takahashi K, Masui T (2015) Levin A, Lin CF, Chu CSJ (2002) Unit root tests in panel data: asymptotic
Consequence of climate mitigation on the risk of hunger. and finite-sample properties. J Econ 108:1–24
Environmental Science Technologies 49:7245–7253 Li HP, Zhang JB (2009) An empirical test on the “Ekc” relationship
Havemann T (2014) Investing in agriculture; Jumping Kutnet’s curve. between the agricultural non-point source pollution and economic
Clarmondial GmbH, Weinrebenstrasse 20 9708 Mannedorf development in China [J]. Resour Environ Yangtze Basin 6
Switzerland Liu X, Zhang S, Bae J (2017a) The nexus of renewable energy-
Higgs H (1897) The physiocrats: six lectures on the French economists of agriculture-environment in BRICS. Appl Energy 204:489–496
the 18th century. Macmillan and Company, Limited. (Augustus M. Liu X, Zhang S, Junghan B (2017b) The nexus of renewable energy-
Kelly Reprint, 1989), New York agriculture-environment in BRICS. Appl Energy 204:489–496
Im KS, Pesaran MH, Shin Y (2003) Testing for unit roots in heteroge- Lu WC (2017) Electricity consumption and economic growth: evidence
neous panels. J Econ 115:53–74 from 17 Taiwanese industries. Sustainability 9(1):50–55
IPCC (2013). Intergovernmental Panel on Climate Change Fifth assess- Munasinghe M (1999) Is environmental degradation an inevitable conse-
ment report (AR5). Retrieved from: https://www.ipcc.ch/report/ar5/ quence of economic growth: tunneling through the environmental
IPCC (2014) Intergovernmental Panel on Climate Change 2014: mitiga- Kuznets curve. Ecol Econ 29:89–109
tion of climate change Contribution of Working Group III to the Mushtaq K, Abbas F, Ghafour A (2007) Energy use for economic growth:
Fifth Assessment Report of the Intergovernmental Panel on cointegration and causality analysis from the agriculture sector of
Climate Change. Cambridge University Press, Cambridge, NY Pakistan. Pak Dev Rev 46:1065–1073
Nakagawa H, Harada T, Ichinose T, Takeno K, Matsumoto S, Kobayashi
Jayanthakumaran K, Verma R, Liu Y (2012) CO2 emissions, energy
M, Sakai M (2007) Biomethanol production and CO2 emission
consumption, trade and income: a comparative analysis of China
reduction from forage grasses, trees, and crop residues. JARQ
and India. Energy Policy 42:450–460
41(2):173–180
Jebli MB, Youssef SB (2017) The role of renewable energy and agricul-
Oh W, Lee K (2004) Causal relationship between energy consumption
ture in reducing CO2 emissions: evidence for North Africa coun-
and GDP: the case of Korea 1970-1999. Energy Econ 26(1):51–59
tries. Ecol Indic 74:295–301
Ozkan B, Fert C, Karadeniz F (2007) Energy and cost analysis for green-
Johansen S (1988) Statistical analysis of cointegration vectors. J Econ
house and open-field grape production. Energy 32:1500–1504
Dyn Control 12(2–3):231–254
Panayotou T (1993) Empirical tests and policy analysis of environmental
Kao C (1999) Spurious regression and residual-based tests for degradation at different stages of economic development, ILO
cointegration in panel data. J Econ 90(1):1–44 Technology and Employment Program Working Paper 238.
Katircioglu ST (2004) Co-integration and causality between GDP, agri- Pao HT, Tsai CM (2010) CO2 emissions, energy consumption and eco-
culture, industry and services growth in North Cyprus: evidence nomic growth in BRIC countries. Energy Policy 38(12):7850–7860
from time series data. Rev Soc Econ Bus Stud 5(6):173–187 Pasten R, Figueroa E (2012) The environmental Kuznets curve: a survey
Katircioglu ST (2006a) Agriculture in North Cyprus. J Econ Adm Sci of the theoretical literature. Int Rev Environ Resour Econ 6(3):195–
Osmangazi Univ 1(1):61–70 224
Katircioglu ST (2006b) Causality between agriculture and economic Paul S, Bhattacharya RN (2004) Causality between energy consumption
growth in a small nation under political isolation: a case from and economic growth in India: a note on conflicting results. Energy
North Cyprus. Int J Soc Econ 33(4):331–343 Econ 26(6):977–983
Katircioglu ST (2008) The direction of causality between growth, agri- Paustian K, Cole CV, Sauerbeck DR, Sampson N (1998) CO2 mitigation
culture, industry and services sector: time series population data by agriculture: an overview. Clim Chang 40:135–162
from Turkey: 1924 - 2003. Adv Manag 1(3):19–23 Phillips PC, Perron P (1988) Testing for a unit root in time series regres-
Katircioglu S (2014) Testing the tourism-induced EKC hypothesis: the sion. Biometrika 75(2):335–346
case of Singapore. Econ Model 41:383–391 Qureshi S (2013a) What is the role of mobile phones in bringing about
Katircioglu S, Katircioglu S (2018) Testing the role of urban development growth? Inf Technol Dev 19(1):1–4
in the conventional environmental Kuznets curve: evidence from Qureshi S (2013b) Networks of change, shifting power from institutions
Turkey. Appl Econ Lett 25(11):741–746 to people: how are innovations in the use of information and com-
Katircioglu ST, Taşpinar N (2017) Testing the moderating role of finan- munication technology transforming development? Inf Technol Dev
cial development in an environmental Kuznets curve: empirical ev- 19(2):97–99
idence from Turkey. Renew Sust Energ Rev 68:572–586 Rehdanz K, Maddison D (2008) Local environmental quality and life-
Katircioglu S, Gokmenoglu KK, Eren BM (2018) Testing the role of satisfaction in Germany. Ecol Econ 64(4):787–797
tourism development in ecological footprint quality: evidence from Rosenzweig C, Elliott J, Deryng D et al (2014) Assessing agricultural
top 10 tourist destinations. Environ Sci Pollut Res 25(33):33611– risks of climate change in the 21st century in a global gridded crop
33619 model intercomparison PNAS 111(9):3268-3273
Kijima M, Nishide K, Ohyama A (2010) Economic models for the envi- Sarkodie SA, Adams S (2018) Renewable energy, nuclear energy, and
ronmental Kuznets curve: a survey. J Econ Dyn Control 34(7): environmental pollution: accounting for political institutional quali-
1187–1201 ty in South Africa. Sci Total Environ 643:1590–1601
25234 Environ Sci Pollut Res (2019) 26:25218–25234

Sarkodie SA, Strezov V (2018) Assessment of contribution of Australia’s importance of biomass energy consumption. J Clean Prod 217(1):
energy production to CO2. Emissions and environmental degrada- 603–614
tion using statistical dynamic approach. Sci Total Environ 639:888– Sinha A, Shahbaz M, Balsalobre D (2017) Exploring the relationship
899 between energy usage segregation and environmental degradation
Sarkodie SA, Strezov V (2019) Effect of foreign direct investments, eco- in N-11 countries. J Clean Prod 168:1217–1229
nomic development and energy consumption on greenhouse gas Solarin SA, Al-Mulali U, Musah I, Ozturk I (2017) Investigating the
emissions in developing countries. Sci Total Environ 646:862–871 pollution haven hy-pothesis in Ghana: an empirical investigation.
Schleussner CF et al (2016) Differential climate impacts for policy- Energy 124, 706–719
relevant limits to global warming: the case of 1.5 °C and 2 °C. Soni P, Taewichit C, Salokhe VM (2013) Energy consumption and CO2
Earth Syst Dynam 7:327–351 emissions in rainfed agricultural production systems of Northeast
Sebri M, Abid M (2012) Energy use for economic growth: a trivariate Thailand. Agric Syst 116:25–36
analysis from Tunisian agriculture sector. Energy Policy 48:711– Swaminathan M, Swaminathan MS (2018) ICT and agriculture CSIT.
716 Trans ICT 6(3–4):227–229
Sebri M, Ben-Salha O (2014) On the causal dynamics between economic
Tamazian A, Chousa JP, Vadlamannati KC (2009) Does higher economic
growth, renewable energy consumption, CO2 emissions and trade
and financial development lead to environmental degradation: evi-
openness: fresh evidence from BRICS countries. Renew Sust Energ
dence from BRIC countries. Energy Policy 37(1):246–253
Rev 39:14–23
Sertoglu K, Ugural S, Bekun FV (2017) The contribution of agricultural Thamo T, Kingwell RS, Pannell DJ (2013) Measurement of greenhouse
sector on economic growth of Nigeria. Int J Econ Financ Issues 7(1): gas emissions from agriculture: economic implications for policy
547–552 and agricultural producers. Aust J Agric Resour Econ 57(2):234–
Shafik N, Bandyopadhyay S (1992) Economic growth and environmental 252
quality: time-series and cross-country evidence, 904. World Bank Tiwari AK, Shahbaz M, Hye QMA (2013) The environmental Kuznets
Publications, Washington DC curve and the role of coal consumption in India: cointegration and
Shahbaz M (2012) Does trade openness affect long-run growth? causality analysis in an open economy. Renew Sust Energ Rev 18:
Cointegration, causality and forecast error variance decomposition 519–527
tests for Pakistan. Econ Model 29:2325–2339 Ullah A, Khan D, Khan I, Zheng S (2018) Does agricultural ecosystem
Shahbaz M, Sinha A (2018) Environmental Kuznets curve for CO2 emis- cause environmental pollution in Pakistan? Promise and menace.
sion: a literature survey. J Econ Stud 46(1):106–168 Environ Sci Pollut Res 25(14):13938–13955
Shahbaz M, Jalil A, Dube S (2010) Environmental Kuznets curve (EKC): World Bank (2018) World Bank database. https://datacatalog.worldbank.
times series evidence from Portugal. Working paper series org/dataset/world-development-indicators
Shahbaz M, Lean HH, Shabbir MS (2012) Environmental Kuznets curve Yoo SH (2006) The causal relationship between electricity consumption
hypothesis in Pakistan: cointegration and Granger causality. Renew and economic growth in the ASEAN countries. Energy Policy 34:
Sust Energ Rev 16(5):2947–2953 3573–3582
Shahbaz M, Mutascu M, Azim P (2013) Environmental Kuznets curve in
Romania and the role of energy consumption. Renew Sust Energ Publisher’s note Springer Nature remains neutral with regard to jurisdic-
Rev 18:165–173 tional claims in published maps and institutional affiliations.
Shahbaz M, Balsalobre D, Sinha A (2019) Foreign direct investment–
CO2 emissions nexus in Middle East and North African countries: