Energy 263 (2023) 125570

Contents lists available at ScienceDirect

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

Primary energy consumption-growth nexus: The role of natural resources,

quality of government, and fixed capital formation
Anam Azam a, Muhammad Ateeq b, Muhammad Shafique c, *, Muhammad Rafiq d, **,
Jiahai Yuan e
a
Center for Agro-Food Economics and Development (CREDA-UPC-IRTA). Parc Mediterrani de La Tecnologia, Edifici ESAB, 08860, Castelldefels, Barcelona, Spain
b
Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong, China
c
Department of Civil and Environmental Engineering, Brunel University London, Uxbridge, United Kingdom
d
Department of Electrical Engineering, University of Engineering and Technology, Taxila, Pakistan
e
School of Economics and Management, North China Electric Power University, Beijing, China

A R T I C L E I N F O A B S T R A C T

Keywords: This present study is the first attempt to analyze the synergy between primary energy consumption, economic
Industrialization growth, natural resources, quality of government, financial development, and gross capital formation in the
Energy demand panel data of 30 developing countries from 1990 to 2017. For this purpose, the study employed pooled
Panel data
regression, fixed effects, random effects, and the heterogeneous panel Dumitrescu-Hurlin causation procedure.
CO2 emissions
Natural resources
Our findings confirm that primary energy consumption, gross capital formation, financial development, and
quality of government positively impact economic growth. The heterogeneous panel causality test confirms the
feedback effect hypothesis in the connection between primary energy consumption and other factors including
economic growth. These results offer new insights into policy implications for the developing countries being
studied.

the role of efficient and reliable utilization of energy in long-term eco­

1. Introduction nomic development [9].
Several studies [10,11] have assessed the causal relationship be­
Energy is considered an important indicator of the global economy tween energy consumption and economic growth in numerous countries
because of its direct contribution to the industrial sector and domestic and regions by using different periods and econometric models. For
consumption [1–3]. Energy has made a substantial contribution to both example, consider the panel ECM [12,13] which confirmed the positive
industrialized and developing countries economic growth and and significant association between energy consumption and growth,
socio-economic development during the last three decades. Therefore, which supports the growth hypothesis [14]. determined that energy
energy demand is boosted due to economic and social development consumption has a positive impact on economic growth. But these
worldwide [4]. According to the International Energy Agency (IEA), studies focus on the omitted-variable bias problems, which make the
demand for primary energy will continue to rise (roughly 1.4% per year empirical results biased and inconsistent. However, the combination of
by 2035) in both developed and developing countries [5]. variables in a multivariate function is crucial for thoroughly assessing
The increase in energy consumption for economic processes leads to the effects of energy consumption on economic growth in the case of
an increase in carbon dioxide emissions (CO2) [6]. More than 87% of developing nations.
energy is derived from fossil fuels, accounting for 89% of CO2 emissions However, this study uses several control variables to explain the link
worldwide [7]. However, growing energy demand motivates GDP, on between the primary energy consumption-growth nexus such as natural
the one hand, but on the other hand, it causes CO2 discharge, which has resources, financial development, fixed capital formation, and quality of
an unfavorable impact on sustainable development and the environment government. Natural resources include gas, oil, coal, minerals, and
[8]. The deterioration of environmental quality has contributed to an forest, and are assessed as a percentage of GDP by total resource rent.
enhancement of environmental protection awareness, with a focus on The function of natural resources may not only affect economic growth

* Corresponding author.
** Corresponding author.
E-mail addresses: [email protected] (M. Shafique), [email protected] (M. Rafiq).

https://doi.org/10.1016/j.energy.2022.125570
Received 18 October 2021; Received in revised form 2 September 2022; Accepted 23 September 2022
Available online 28 September 2022
0360-5442/© 2022 Published by Elsevier Ltd.
A. Azam et al. Energy 263 (2023) 125570

countries are open and dynamic nations that face multifaceted chal­
Abbreviations lenges, e.g., energy insecurity, poor economic growth, low standards of
living, lack of investment in infrastructure and technologies, as well as
GDP Gross domestic product low institutional quality [25]. Recently, more than 170 developing na­
CO2 emissions Carbon dioxide emissions tions have been bestowed with natural resources, but low levels of
QOG Quality of government economic growth and fundamentals of innovation have inaugurated
SDG Sustainable development goals ineffective utilization of energy [26,27]. Moreover, these developing
UN United Nations economies together account for approximately 66% of the population,
OLS Ordinary Least Square 57.2% of primary energy consumption, and 61.8% of carbon emissions,
FEM Fixed effect model but only 36.5% of global growth, reflecting their high level of energy
REM Random effect model and carbon intensity as well as extensive economic development. In
NR Natural resources 2015, the UN projected a new accomplishment of 17 SDGs to endorse
IPS I’m, Pesaran, and Shin global prosperity. However, three of these goals-namely, SDG 7: ensure
IEA International Energy Agency access to affordable, authentic, and modern energy; SDG 8: support
PEC Primary energy consumption comprehensive and long-term economic progress, employment, and
FD Financial development decent work for all; and SDG 16: support passive and broad-based civ­
GCF Gross capital formation ilizations for sustainable development, ensuring access to justice for all
and establishing actual, accountable and general institutes at all levels
[28]. These goals significantly influence energy growth along with other
indicators and are the most effective way to support poor countries to
but also influence primary energy consumption as well. For example, achieve sustainable development [29].
natural resources improve the sources of revenue through employment Against this backdrop, the objective of this study is to explore the
opportunities, which in turn stimulates economic growth. On the other nexus between primary energy consumption, economic growth, natural
hand, natural resource wealth is an essential way for energy production resources, quality of government, financial development, and gross
via offering mandatory capital. Regarding the role of natural resources capital formation in the panel of developing countries by addressing all
in economic development, some studies, for instance Refs. [15–17], of the above-mentioned issues. The literature on the primary energy
claim the opposite. While others [18,19] confirm that natural resources consumption-growth nexus with a multivariate framework is scarce and
are an important element of the global economy because these resources shows disagreement.
maintain the quality of the region’s institutions and systems. Conse­ Therefore, the main contribution and novelty of this study are as
quently, the role of natural resources in the energy consumption-growth follows: (1) To the best of our knowledge, this is the first empirical study
nexus is insufficient, therefore, further analyses are necessary to move of its kind on developing countries in the literature, using the most
toward sustainable development [20]. recent data to provide significant insight into the country’s policy­
We include the quality of government because the role of govern­ makers. As is explained above, these economies merit special attention
ment is imperative in the adaptation of energy innovations and eco­ because of their large and rising populations, which puts a strain on the
nomic development of any country. For instance, respect for democratic environment. (2) we examine the role of natural resources and gov­
principles, rules and laws, legislative authority, and control of corrup­ ernment quality as important factors in the literature on energy growth,
tion are all crucial aspects in promoting the beneficial use of primary which has not been done previously. (3) our study mainly adds to the
energy consumption as well as the country’s development. Very few literature which analyses the connection between energy consumption
studies on the role of institutions in economic growth and environmental on the one hand and GDP growth on the other hand. (4) The study
indicators in the energy-growth nexus have been conducted; therefore, employs a panel regression model including ordinary least squares
pioneering work is still being done. To achieve long-term economic (OLS), a fixed effect model (FEM), and a random effect model (REM);
prosperity, we believe that institutional alignment (economic freedom) also use the Hausman specification test, among others [30] for model
is required to promote the use of renewable energy sources throughout selection. (5) In addition, this study employed [31] panel heterogonous
all economic activities [21]. Most environmental factors, such as CO2 causality approach to infer the causality directions in terms of policies.
emissions, CH4 emissions, and forest area, positively impact institutional The enduring study structure is as follows: Section 2 studied earlier
quality and energy usage based on oil and fossil fuel resources [22]. [23] works. Section 3 describes the data and methodological context, Section
explores the role of institutions in the renewable energy-growth nexus of 4 includes rational results and discussion, and Section 5 proposes a
MENA countries from 1986 to 2015. The findings indicate that there is conclusion.
bidirectional causality between renewable energy growth and in­
stitutions, e.g., law and order and economic growth [9]. finds that 2. Literature review
institutional quality, renewable and non-renewable energy, trade
openness, financial development, and capital have a positive impact on Although several kinds of literature have explored energy
economic growth. Overall, different researchers have used different consumption-growth, natural resource-growth, capital accumulation-
proxies for institutions; consequently, the outcomes remain ambiguous growth, and natural resource-energy consumption nexus separately,
and mixed. there is no study to investigate the association together. However, this
In this study, gross fixed capital formation (GCF) is assumed as study aimed to fill this gap in the literature.
another basic factor of economic growth [24]. Suggest that GCF is an The link between energy consumption and economic growth is a
imperative aspect of any nation’s economic growth. The author further contentious subject [32] and it has been rapidly rising since the pivotal
proposes that it also influences energy consumption through the higher work of [33]. The literature covered a wide range of periods and
accumulation and deployment of GCF for more production of goods and methodological approaches. The energy-growth nexus has been syn­
services, which leads to faster income. The intensity of the factors of thesized into four possible hypotheses. First, the growth hypothesis
capital accumulation, such as savings, foreign direct investments, and states that an escalation in energy usage leads to higher economic
interest rates, determines the impact of capital accumulation on eco­ expansion, so energy is a crucial input for productivity [13,34]. Second,
nomic growth. One of the most fundamental problems restricting the conservative hypothesis assumes that there is unidirectional cau­
countries’ long-term growth is their insufficient capital accumulation. sality from economic development to energy utilization; an increase in
Why did we choose developing countries for our study? Developing economic activities leads to higher energy use, and policies for energy

2
A. Azam et al. Energy 263 (2023) 125570

reduction may not adversely affect economic growth [35]. Third, the Table 2
feedback hypothesis conjectures that energy usage and economic Summary of literature on energy consumption-growth nexus with the supported
expansion are strongly dependent on each other, and two-way causation hypothesis.
exists between these variables [36]. Fourth, the neutrality hypothesis Author Study area Period Methodology Results
infers that there is no causality between economic expansion and energy (s)
consumption, and any strategy on one indicator will not influence the [54] Turkey 1960–1995 Johansen-Juselius, Growth
other [37]. Table 1 represents the description of the hypothesis that is VECM hypothesis
commonly used in the energy-growth literature. [55] Shanghai 1952–1999 Toda-Yamamoto Growth
causality tests hypothesis
The studies on the relationship between energy consumption and
[56] 6 Central 1989–2004 Pedroni co- Growth
economic growth are summarized in (Table 2). We concluded that the American integration, VECM hypothesis
literature does not provide a clear consensus that mixed results from the Countries
energy use-growth nexus have been observed in past studies. The diverse [34] South America 1980–2005 Pedroni co- Growth
results are related to the different empirical growth models and nation integration, VECM hypothesis
[35] Turkey 1960–2006 Johansen-Juselius, Conservation
samples used, as well as the econometric approach used. VECM hypothesis
The association between natural resources and economic growth is a [37] a) Benin, 1971–2011 Johansen-ML co- a) Growth
contentious subject with different arguments. Some researchers argue b) Congo integration, Hypothesis
that NR increases economic growth, but some recommend that NR re­ Kenya, Granger causality b) Neutrality
Zimbabwe hypothesis
duces growth in the long run. For example [38–40], natural resources
[57] 130 countries 1981–2009 GMM a) Growth
are an important factor that contributes to economic development, po­ Hypothesis
litical stability, social maturity; and energy consumption as well. Inap­ [58] Pakistan 1982–2015 Ordinary least Feedback
propriate use of natural resources and poor technology can demean the squares, Granger hypothesis
country’s resources [41]. [42] scrutinized whether the resource curse causality
[59] BRICS 1990–2013 Pedroni panel co- Conservation
hypothesis is valid in 62 countries from 1990 to 2017 [43]. determined
countries integration, hypothesis
the validity of the resource curse from 1998 to 2013 for China. Findings Granger causality
show that natural resources have a negative influence on economic [60] China a) Granger causality a) Growth
expansion. 1978–1991 Hypothesis
b) b) Growth
For the next-11 economies [44], used the CIPS test and resolved that
1992–2016 Hypothesis
NR rose in economic progress spanning data from 1996 to 2016 [45]. [61] West Africa 1980–2015 Co-integration test, Conservation
used Granger causality to conclude a bidirectional connection between FMOLS, DOLS, hypothesis
studied variables for top resource-rich regions from 1970 to 2013 by Granger causality
employing Granger causality [46]. fail to establish a link between NR [62] China 1978–2016 VECM Granger Feedback
causality tests hypothesis
and economic development for Pakistan from 1971 to 2015.
[63] India 1971–2014 VECM Granger Feedback
In addition, few studies have examined how natural resources affect causality tests hypothesis
energy consumption. For example [47], examined the role of natural [29] Pakistan 1990–2015 Johansen Co- Conservation
resources in carbon emission functions in the case of five European (EU) integration test, hypothesis
Granger causality
regions from 1985 to 2016. The study concluded that natural resources
[64] 16 Latin 1990–2014 AMG, CCEMG, MG Feedback
stimulate energy efficiency by improving the latest energy innovations. America and estimators; hypothesis
Similarly, in another case of EU countries [48], analyzed the association Caribbean Dumitrescu-Hurlin
between natural resources and renewable energy production from 1997 countries panel causality
to 2015. Their findings concluded that natural resource rents have a [65] Ghana 1971–2013 VECM Granger Feedback
causality hypothesis
substantial influence on renewable power production, but oil rents
impede renewable energy production. In addition [49], found growth Notes: VECM: Vector error correction model; GMM: generalized methods of
hypotheses in BRICS countries and [50] determined conservation sup­ moments; FMOLS: fully modified ordinary least squares; DOLS: Dynamic ordi­
positions are effective for the United States of America (USA). Nawaz nary least squares; AMG: Augmented mean group; CCEMG: Common correlated
et al. (2020) [40] investigated the impacts of natural resources on en­ effects mean group; MG: Mean group.
ergy consumption insignificantly in Pakistan using the quantile ARDL
model. technologies, which leads to a rise in energy consumption [51]. In na­
Besides natural resources, financial development is another signifi­ tions where renewable energy-related investments are actively encour­
cant factor that stimulates economic growth. Financial development aged, the role of the financial market is critical. In this background, the
leads to economic expansion through investment in innovative necessity for a well-developed financial system to provide efficient

Table 1
Definition, condition, policies, and graphical depiction of four hypothesis.
Name Growth Hypothesis Conservation Hypothesis Feedback hypothesis Neutrality Hypothesis

Definition A unidirectional relationship from A unidirectional relationship from A bi-directional causal relationship between EC There is no relationship between EC
EC to GDP. An increase in EC GDP to EC. EC increases as a result and Feedback GDP assumes that the two are and GDP
fosters higher GDP of higher GDP interrelated and complement each other
Condition Causality running from EC to GDP unidirectional causality running bidirectional relationship between EC and GDP EC has no significant effect on the
from GDP to EC real output.
Policies Energy-saving policies harm GDP. Restriction on EC or energy EC and GDP are jointly determined and affected neither conservative nor expansive
conservation policies has no at the same time policies concerning EC have any
adverse effect on GDP. effect
Depiction

Notes: E indicates energy use, G refers to economic growth, denotes unidirectional causality, Bidirectional causality, —— represents neutral causality.

3
A. Azam et al. Energy 263 (2023) 125570

finance, risk management, and liquidity services has been identified upsurge in energy consumption [6,74]. Financial development boosts
[52]. In addition, the role of fixed capital formation is important in energy consumption because a well-developed financial system may
economic growth. Capital formation in these nations has the potential to offer businesses significantly reduced costs, allowing them to expand
boost economic growth while also assisting in the development of sus­ their production technologies and equipment.
tainable environmental conditions in the Belt and Road countries [53]. Therefore, efficient and independent political indicators play an
Capital accretion increases the state level’s ability to produce more; as a important role in financial and economic development. In this context,
result, it influences economic progress, but the scantiness of capital the quality of government positively affects economic growth and pri­
assertion is a serious issue for sustainable growth [39]. mary energy consumption. However, the improvement of institutions is
Moreover, another factor is the quality of government, which cannot necessary for economic prosperity and reliable energy because an
be ignored in the energy-economic literature. However, there are few effective financial system (investment) and political stability are
studies on the link between economic growth and institutions; for strongly linked with better economic growth as well as promoting the
instance, the role of institutions in economic growth from 1990 to 2014 beneficial use of energy production. Against this backdrop, it might be
by using data from 46 emerging markets and developing economies said that natural resources, financial development, quality of govern­
[64]. The empirical findings demonstrate that institutions have a posi­ ment, and gross capital formation are significant factors for economic
tive but insignificant influence on economic growth, and the causality growth and primary energy consumption. Fig. 1 portrays the inter­
results reveal that the feedback hypothesis is valid in these nations [66]. connected model.
revealed a negative and significant influence of intuitional stability on
economic growth from 1990 to 2016 by employing the ARDL model. In 3.2. Model specification
the case of 14 Central and Eastern European nations [67], explored the
association between the energy-economic development nexus from This study explores the relationship between primary energy con­
1995 to 2017. The results concluded that governance has a constructive sumption and economic growth by including additional factors like
and considerable influence on economic growth and electricity natural resources, quality of government, financial development, and
consumption. gross fixed capital formation. To discuss the role of these explanatory
To summarizes the above-reviewed literature, it is well documented variables on dependent indicators, we construct a model specification by
that there is no serious research examining the association between employing two models: (a) – GDP and (b) - PEC which will meet the core
primary energy consumption-economic growth nexus with several de­ goals of this study. Economic growth is a function of primary energy
terminants such as natural resources, quality of government, gross fixed consumption, natural resources, quality of government, gross fixed
capital formation, and financial development in developing countries. capital, and financial development.
Most of the previous studies use omitted variables to find the relation­
ship. Moreover, we conclude that the studies conducted in the past have GDPit = f (PEC, NR, QOG, GCF, FD) (1)
an inconclusive empirical association between energy use and economic Eq. (1) is re-write by adding the error term as follows:
growth due to the utilization of various tactics, data, and research per­
spectives. Therefore, new approaches FEM and heterogenous causality GDPit = ∝0 + ∝1 PECit + ∝2 NRit + ∝3 QOGit + ∝4 GCFit +∝5 FDit + ε1it (2)
are needed for accurate policies and sustainable development processes. The above equation can be re-write in log-linear form as stated
follows:
3. Data and model
LGDPit = ∝0 + ∝1 LPECit + ∝2 LNRit + ∝3 LQOGit
3.1. Theoretical background + ∝4 LGCFit +∝5 LFDit + ε1it (3)

Before developing the estimation strategy, we need to construct the Here, in Eq. (3) first-panel model is GDP, subscript i = 1, …, N for
theoretical framework of this study, as it will support us in deciding on each nation and t = 1, …, T specifies a period, LGDP is the log of eco­
model indicators. Developing countries are characterized by high eco­ nomic growth, LPEC means the log of total primary energy consumption,
nomic growth, and this faster growth is a result of huge energy con­ LNR is the log of total natural resource rents, LQOG indicates the log of
sumption. However, to understand the importance of energy for the quality of government, LGCF illustrate log of gross capital formation and
development of society and economic growth, the first step is to LFD represent the log of financial development. The parameter ∝0 rep­
empathize with the significance of energy production [68]. Contem­ resents intercept and ε1it indicates an error term. The ∝1 , ∝2 , ∝3 , ∝4 , ∝5
plating theories of production, according to the neo-classical model [69] are coefficients of LGDP regarding pertinent descriptive variables such
proclaims that energy cannot be considered an element of the produc­
tion aspect. The Solow growth theory considers the economy as a closed
loop where productivity is supplied by inputs of labor and capital.
Whereas, internal growth models of public and human capital [70] both
with contributions from neoclassical economists [71], [72] have
demonstrated that energy can be considered a significant factor of
production rather than the assumptions of neoclassical notions.
Furthermore, energy economists, researchers, and policymakers stress
the importance of energy as a production element of capital, labor, and
technology [73]. Energy is a critical resource in the economy because
industrialization, with increasing infrastructure and investment, in­
creases the energy production process. Consequently, the level of eco­
nomic development increases.
Natural resources are an important part of the production process
because they provide the necessary capital that drives economic
expansion and enhances energy output, which is vital for economic
growth and energy consumption in emerging countries. On the other
hand, financial development enhances economic growth through capital
accumulation and technological progress, consequently causing an Fig. 1. Conceptual framework of the study.

4
A. Azam et al. Energy 263 (2023) 125570

as total primary energy consumption, natural resources, quality of Table 4

government, gross capital formation, and financial development. Region-wise distribution of developing countries.
The second-panel model defined primary energy consumption as Continents Number Name of Countries
follows in Eq. (4):
Europe & Central 9 Azerbaijan (AZE), Belarus (BLR), Kazakhstan
PECit = f (GDP, NR, QOG, GCF, FD) (4) Asia (KAZ), Russia (RUS), Ukraine (UKR), Greece
(GRC), Poland (POL), Romania (ROU), Turkey
The above equation is re-written by adding the error term in Eq. (5) (TUR)
as follows: South Asia 3 Bangladesh (BGD), Sri Lanka (LKA), Pakistan
(PAK),
PECit = ∝0 + ∝1 GDP + ∝2 NRit + ∝3 QOGit + ∝4 GCFit +∝5 FDit + ε1it (5) Eat Asia & Pacific 6 China (CHN), Indonesia (IND), India (IND),
Malaysia (MYS), Philippines (PHL), Vietnam
The fowling equation is given in log-linear form as follows: (VNM)
The Middle East & 5 Algeria (DZA), Egypt (EGY), Morocco (MAR),
LPECit = ∝0 + ∝1 LGDPit + ∝2 LNRit + ∝3 LQOGit North Africa Iran (IRN), Iraq (IRQ),
Sub-Saharan Africa
+ ∝4 LGCFit +∝5 LFDit + ε1it (6) 1 South Africa (ZAF)
Latin America & 6 Argentina (ARG), Brazil (BRA), Chile (CHL),
Here in Eq. (6), the second-panel model is LPEC, the subscript i = 1,
Caribbean Colombia (COL), Ecuador (EUC), Peru (PER)
…, N for each state and t = 1, …, T. The parameter ∝0 represents
intercept and ε1it indicates an error term. The ∝1 , ∝2 , ∝3 , ∝4 , ∝5 are co­
efficients of LPEC regarding applicable clarifying variables such as merging several data points (time series and states), its capacity to
economic growth, natural resources, quality of government, gross cap­ control specific assortment, and its ability to raise the level of freedom,
ital formation, and financial development. which gives consistent scrutiny.

3.3. Data and description of variables 4.1. Panel unit root method

In this study, the dataset encompasses a panel of 30 developing The rational analysis is initiated along with the panel unit root
countries from 1990 to 2017. The data was obtained from World technique to avoid the probability of specious deterioration. Some panel
Development Indicators [7], British Petroleum [75], and the University unit root trials were used in previous literature but most of them have
of Gothenburg [76]. The variables and data sources are shown in arithmetic shortcomings concerning the adequacy of size and strength of
Table 3. The selection of these countries and periods was based on data the test [77]. However, three popular unit root [78–80] tests are applied
accessibility. Economic growth and primary energy consumption are [78]. recommended by Augmented Dickey-Fuller (ADF) established
two endogenous variables, and the exogenous variables are natural re­ various root tests [79], (IPS) based on the discrete root, which allows for
sources, quality of government, gross capital formation, and financial diverse autoregressive coefficients and can diverge over cross-sections,
development. and finally [80], (LLC) to identify random tendency constituent in a
The selected countries are distributed across the 6 continents as separately variable that allows for mutual root process across nations,
defined by the World Bank. The considered groups of countries are and its size is common because its greater strength besides slight volume
Europe & Central Asia (ECA), South Asia (SA), East Asia & Pacific (EPA), variation. The null assumption is that indicators have a unit root while
Middle East & North African countries (MENA), and Sub-Saharan Afri­ the alternative hypothesis has no unit root.
can countries (SSA), and Latin America & Caribbean (LAC). A detailed Under the Augmented Dickey-Fuller (ADF) test, all unit root tests
list of these selected countries is given in Table 4. equation is given below:
Table 5 displays a descriptive analysis of selected variables and
outcomes. The standard deviation values imply additional precise and pi
∑
(7)
′
Δyit = ρ yit− 1 + ∅ij Δyit− j + zit φ + εit
full estimation of dispersal. The variable GCF had a high value of the j=1
standard deviation of 6.448080, accompanied by PEC and NR, whilst
QOG had the slightest fickleness. Where, yit is the number of observations for each cross-section in panel i
= 1,‥,N & t = 1,‥, T signifies period, Δ is difference operator, εit in­
4. Estimation approach dicates white noise error term, ρi refers to autoregressive coefficients
and z’it indicates exogenic variables, including continuous and trend.
This study implies panel data to present the general estimation The null hypothesis of the unit root test is (H0: ρi = 0), for all i against the
method because of its colossal benefits, e.g., ensuring extra material by alternative hypothesis (H1: ρ < 0), which supposes at least one indi­
vidual series is (trend) stationary within panel data.
Table 3
Description of Data series.
4.2. Panel co-integration test
Variable Symbol Unit Source

Economic GDP 2010 US$ WDI If variables are incorporated or stable in the same order, the next
growth panel co-integration model is utilized to ascertain the correlation be­
Primary Energy PEC MTOE BP-statistical review tween variables. The co-integration concept was first developed by
consumption
Natural NR % of GDP WDI
Ref. [81] and further proposed by Refs. [82,83]. The main idea of the
resources co-integration model is that if two or more series are separately merged
Quality of QOG − 2.5 (weak) to 2.5 World the Quality of the of order n, then there exists relatively one straight mixture to be inte­
Government (strong) governance Government Institute grated at a lower order, such that ñ < n. The series that are co-integrated
performance University of Gothenburg
demonstrate a strong steady-state long-run relationship, having wide­
Sweden.
Gross fixed GCF 2010 US$ WDI spread tendencies and co-movements. The co-integration theory devel­
capital oped the existence of linear combinations of combined series, which
formation compensate for ordinary speculative trends and provides a climb to
Financial FD % WDI stability associations amid co-integrated series, suggesting these vari­
development
ables exhibit co-movement with each other in the long run.

5
A. Azam et al. Energy 263 (2023) 125570

Table 5
Descriptive statistics.
Variable GDP PEC NR QOG GCF FD

Mean 8.263152 3.927989 1.265930 − 0.737288 22.67969 3.350561

Median 8.404661 3.793229 1.418736 − 0.750776 24.36361 3.328633
Std.Dev. 0.889817 1.447360 1.490894 0.321454 6.448080 0.920882

Two-panel co-integration tests such as Kao [84] and Pedroni [84,85]

∑
P ∑
P
were used in this study. The panel Pedroni co-integration test permits Yit = ∂i + γ(k) β(k) (9)
i iyi,t− k + i ixi,t− k εi,t
diverging intercepts and slopes, whereas the [84] method follows indi­ K=1 K=1
vidual cross-section intercepts and similar coefficients on first-step re­
In Eq. (9) Yi,t and Xi,t are variables for individual i in period t. ∂i is an
gressors. According to the Kao co-integration analysis assumes the
individual fixed effect, γi and βi epitomize autoregressive and
(k) (k)
co-integrating vector is the same overall panels and it has better
power than the [84,85] method when the number of observations regression considerations, which can differ into groups. ε2it Signify an
small-T (high-T) is included in a panel. error term and lag order P that is believed to be the same for all in­
dividuals and the panel must be balanced. Both methods have similar
4.3. Panel regression model form null hypotheses of no causal relationship and alternative hypotheses of
at least one cross-section unit having a causal relationship.
In econometric analysis, panel data regression is utilized in this study
because it assumes the possibility of contemporaneous correlations, 5. Results and discussion
permitting better interpretation from linear models and accounting for
the deviations from spherical errors. Panel data regression can be Table 6 shows the outcomes of the panel unit root method. The re­
compartmentalized into three types: pooled regression model (OLS), sults show that at levels all the series are non-stationary, thus, the null
fixed effect model (FEM), and random-effects model (REM). In the hypothesis can not be rejected. However, all the indicators are signifi­
pooled regression model, all intercepts and coefficients are assumed to cant at the first difference, which means the null hypothesis is rejected at
be the same, and it does not consider time series facets. This regression 1% and 5% levels of significance.
model may mutilate the accurate picture of the association between Y As the series is stationary at the first difference, we can use the panel
and X’s over the cross-sections. FEM allows for some heterogeneity and co-integration technique to find the long-term connection between two
individuality among the cross-sectional units, and it controls for the or more indicators.
impacts of time-invariant variables with time-invariant effects. In this Table 7 details the outcomes of both co-integration methods for the
model, individual-specific effects are correlated with the explanatory GDP model and the PEC model. Regarding the GDP model, results show
variables, and this specific model is estimated. In the REM, individual- that five within-dimension statistics and two between-dimension sta­
specific effects are unrelated to impartial variables, and this follows a tistics are statistically significant out of seven statistics, and the Kao
random distribution with parameters that must be estimated. residual co-integration technique is also significant at a 5% level. Yet,
outcomes demonstrate the rejection of the null hypothesis, so the LPEC,
Yit = ∝0 + ∝i yit− 1 βi Xit + φi + ωt + εit (8) LNR, LQOG, LGCF, and LFD are co-integrated in conjunction with LGDP.
The results show that the LPEC model shows that six statistics are sig­
Where Yit signifies the target variable, Xit is a vector of regression (K,1)
nificant and the Kao residual co-integration test is also positive at a 1%
dimensional and εit is the disturbance term, φi and ωt represents the
significance level; hence, the findings demonstrate rejection of the null
country’s specific effects. The FEM and REM are considered more suit­
hypothesis of no co-integration. This exemplifies epitomizes that the
able for drawing robust and consistent estimations. However, we
variables LGDP, LNR, LQOG, LGCF, and FD are co-integrated with LPEC.
applied the F-test, Hausman test, and likelihood ratio (LR) test to choose
the legitimacy of panel data. The null premise of the F-test is that pooled
OLS is best in favor of the fixed effects alternative, whilst the null pos­ Table 6
Panel unit root test results.
tulates of the Hausman test are that the random effect model is suitable
against the fixed effect alternative. In addition, the null assumption of Method Variables Level First difference
the LR test favoring the pooled OLS estimator is appropriate, while the ADF LGDP 34.634 269.378a
alternative hypothesis supports the fixed effect model. LPEC 66.200 148.413a
LNR 43.202 499.361a
LQOG 62.284 349.852a
4.4. Panel causality test
LGCF 69.720 424.663a
LFD 51.015 369.698a
Panel causality [31] has been employed for a better understanding of
IPS LGDP 5.295 − 12.719a
the associations among variables for panel data. We employ this most LPEC − 0.151 − 20.866a
recent technique due to its flexibility in heterogeneous panels. This LNR − 0.856 − 21.610a
approach has some benefits compared to other panel causality tests: for LQOG − 0.455 − 18.034a
example, i) it is right for panel data; useful and consistent outcomes are LGCF 1.493 − 21.118a
LFD 1.014 17.241b
obtained from panel data [4]. ii) It is useful to address the cross-sectional
−

dependence and heterogeneity. iii) This model regulates the heteroge­ LLC LGDP − 0.302 − 0.302a
LPEC 0.760 21.427a
neity of the regression model as well as the heterogeneity of causal re­ −
LNR 0.182 − 24.304a
lationships. It has two different test statistics 1) Wbar-statistics and 2) LQOG − 1.197 − 18.025a
Z-bar-statistics. The first W-bar-statistics takes normal test statistics; LGCF 7.923 − 23.692a
however, Z-bar defines standard normal distribution. The linear model LFD 0.787 − 11.817a
description established by DH to test connectedness is as follows: Note: a, b implies importance level at 1% and 5%, respectively. GDP, PEC, NR,
QOG, GCF, and FD represent economic growth, Primary energy consumption,
natural resources, quality of Government, gross capital formation, and financial
development.

6
A. Azam et al. Energy 263 (2023) 125570

Table 7 Table 9
Panel co-integration test Results. Pooled regression, fixed effect, and random effect model.
GDP model I PEC model-II Variable Pooled regression Fixed effect Random effect

Pedroni residual Co-integration test Model-I: LGDP

Alternative hypothesis: Common AR coefficients (within-dimension) LPEC 0.001 (0.943) 0.184 (0.000)a 0.180 (0.000)a
LNR 0.038 (0.065)c − 0.002 (0.889) − 0.001 (0.935)
Panel v-statistic − 0.316 − 5.861
LQOG 0.667 (0.000)a 0.089 (0.021)b 0.091 (0.018)b
Panel rho-statistic 3.342 1.945
LGCF 0.016 (0.000)a 0.027 (0.000)a 0.027 (0.000)a
Panel PP-statistic − 1.787b − 22.326a
LFD 0.053 (0.155) 0.229 (0.000)a 0.227 (0.000)a
Panel ADF-statistic − 4.943a − 5.518a
Panel v-statistic (Weighted Statistic) 3.072a − 0.954 Model-II: LPEC
Panel rho-statistic (Weighted Statistic) 3.047 2.535
LGDP 0.003 (0.943) 0.732 (0.000)a 0.693 (0.000)a
Panel PP-statistic (Weighted Statistic) − 1.885b − 9.023a
LNR 0.199 (0.000)a 0.133 (0.000)a 0.133 (0.000)a
Panel ADF-statistic (Weighted Statistic) − 3.448a − 9.023a
LQOG 1.046 (0.000)b 0.279 (0.000)a 0.291 (0.013)b
Alternative hypothesis: Individual AR coefficients (Between dimensions) LGCF 0.002 (0.753) − 0.002 (0.718) − 0.000 (0.894)
LFD 0.316 (0.000)a − 0.187 (0.000)a − 0.173 (0.000)a
Group rho-statistic 4.849 4.699
Group PP-statistic − 2.247b − 9.396a Note: a, b denotes significance levels at 1% and 5%, respectively. GDP, PEC, NR,
Group ADF-statistic − 4.129a − 6.476a QOG, GCF, and FD represent economic growth, Primary energy consumption,
Kao residual Co-integration test natural resources, quality of Government, gross capital formation, and financial
Augmented Dickey-Fuller − 2.233b 5.971a development.
a, b, c
Note: Newey-West automatic bandwidth selection & Bartlett Kernel. in­
dicates 1%, 5%, 10% significance level respectively. Moreover, a 1% increase in QOG is associated with a 0.089% upsurge
in GDP. Our results are in accord with studies [67] in which governance
Since the variables are co-integrated, the fixed effect and random and corruption lead to an acceleration in growth, but the results are
effect models are employed. Before these panel techniques, we employ contradictory to Refs. [65,67] in which institutions and institutional
the LR test, F-test, and Hausmann test specifications to select the best stability do not affect economic growth [91]. The estimated coefficients
fitting effective model for both models, as seen in Table 8. of gross fixed capital formation and financial development have a
The LR test distinguished between pooled ordinary least squares significantly positive effect on economic growth. Our findings are in line
(OLS) and the fixed-effect model for the GDP model. This test indicates with the studies [44,92].
that the fixed effect model is apt because the p-value is less than the 5% The PEC model in Eq. (4) represents that economic growth is opti­
significance level. Then, the F-test was employed to juxtapose pooled mistic and significantly affects primary energy consumption. A 1% in­
OLS and fixed effect, which reports that FEM is suitable at a 1% signif­ crease in economic growth will increase primary energy consumption by
icance level. Next, the Hausmann model is conducted to ascertain 0.732%. Moreover, the coefficient of natural resources is positive and
whether the panel data model is a fixed effect or a random effect. So, the significant; a 1% increase in NR increases PEC by 0.133%. The long-run
results conclude that the random effect model is accurate. Likewise, in association between the quality of government and primary energy use
the PEC model, small p-values within the LR technique reject the null is positively significant, signifying that a 1% rise in the quality of gov­
hypothesis and suggest the fixed effect model is the best choice. The ernment upsurges the utilization of primary energy in developing
small value of the F-test suggests that FEM is superior to the OLS model, countries. Overall, the results suggest that when an economy develops,
and the small Hausmann p-value suggests that suitable measurement is demand for more energy-consuming goods also boosts energy con­
the fixed-effects model. This shows that the fixed-effect model is the best sumption in developing economies. Gross capital formation has no sig­
choice for our analysis. nificant effect on primary energy consumption in developing countries.
The empirical evaluations of panel data for both models are dis­ Moreover, financial development has negative but significant effects on
played in Table 9. The GDP model, from the long-run empirical primary energy consumption, similar to the study [51].
regression of Eq. (1) discloses that the influence of primary energy use The results of panel causality techniques are represented in Table 10.
on economic growth is substantial. This shows that a 1% rise in primary In the case of the economic growth model, we find bidirectional cau­
energy consumption will lead to an increased GDP of 0.184% which sality between primary energy consumption, natural resources, financial
means that primary energy consumption adds to economic growth. development, gross capital formation, quality of government, and eco­
These experiential consequences are consistent with previous studies nomic growth. Regarding the PEC model, unidirectional causality is
such as [87–91]. Further, we note that natural resources have an adverse found from primary energy consumption to natural resources and pri­
and insignificant impact on economic expansion. This may happen mary energy consumption to financial development. In addition, we find
because of weak intuitions, rent-seeking behavior, and weak investment a feedback hypothesis between primary energy consumption and the
in human capital and infrastructure. These results are coherent with the quality of government as well as bidirectional causality between pri­
previous findings [39–42]. mary energy consumption and gross capital formation. Fig. 2 represents
the causality direction among the variables.
The findings indicate that economic development demands more
Table 8 energy, which means that an increase in the utilization of energy would
Specifications tests. further increase economic growth. As a result, implementing energy
Spec.Test Test P- Tested Selection
conservation policies will have a negative influence on economic growth
statistics value because energy is an indispensable factor in the development of these
nations.
Model-I: Hausman 7.124 0.211 Fixed/ Random
GDP Random Moreover, developing countries must manage the efficient and sus­
LR 337.683 0.000 a
OLS/Fixed Fixed tainable use of natural resources, which will assist in creating new jobs
F-test 16.297 0.000a OLS/Fixed Fixed and market opportunities, lowering operational costs, and boosting
Hausman 12.349 0.030b Fixed/ Fixed market stability. In addition, the capital accumulation that occurs in
Random
technological innovation could improve economic development.
Model-II: LR 198.005 0.000a OLS/Fixed Fixed Financial cooperation is an imperative prerequisite for these selected
PEC F-test 30.222 0.000a OLS/Fixed Fixed
economies to ensure their economic development in the long run. More

7
A. Azam et al. Energy 263 (2023) 125570

Table 10 find evidence for the long-run relationship among all variables by
Results of Dumitrescu and Hurlin test. applying heterogeneous co-integration [85,86]. Panel regression anal­
Null Hypothesis Dumitrescu-Hurlin ysis is applied to find the estimation of independent variables on the
dependent variable. Further, the Dumitrescu and Hurlin technique is
Zbar-Stat. P-value Causality Results
employed to ascertain the direction of causality among the variables.
LPEC →LGDP 15.1264 0.0000 Yes Feedback The empirical findings obtained from FEM reveal that primary en­
LGDP →LPEC 9.53558 0.0000 Yes
LNR →LGDP 10.4018 0.0000 Yes Feedback
ergy consumption, quality of government, gross capital formation, and
LGDP →LNR 2.25133 0.0244 Yes financial development have a positive and significant effect on the level
LQOG →LGDP 5.05499 4.E− 07 Yes Feedback of economic growth at 1% and 5% significance levels, implying that a
LGDP →LQOG 4.67591 3.E− 06 Yes 1% and 5% increase in primary energy consumption, gross capital for­
LGCF →LGDP 14.9585 0.0000 Yes Feedback
mation, and financial development and quality of government will in­
LGDP →LGCF 15.1639 0.0000 Yes
LFD →LGDP 9.85248 0.0000 Yes Feedback crease GDP by 0.184%, 0.027%, 0.229%, and 0.089%, respectively.
LGDP →LFD 9.32446 0.0000 Yes While natural resources have an adverse and insignificant impact on
LFD →LQOG 3.33270 0.0009 Yes Feedback economic growth. In the second model, economic growth, natural re­
LQOG →LFD 4.04647 5.E− 05 Yes sources, and the quality of government have a positive association with
LNR →LGCF 7.53227 5.E− 14 Yes Feedback
LGCF →LNR 2.04309 0.0410 Yes
primary energy use at a 1% significant level. It shows that a 1% increase
LQOG →LGCF 5.13459 3.E− 07 Yes Feedback in these variables will increase primary energy consumption by 0.732%,
LGCF →LQOG 8.59290 0.0000 Yes 0.133%, and 0.279%.
LQOG →NR 2.73565 0.0062 Yes Feedback The causality results of the growth model indicate bidirectional
LNR →LQOG 3.55505 0.0004 Yes
causality between primary energy consumption, natural resources,
LGCF →FD 12.9339 0.0000 Yes Feedback
LFD →LGCF 3.91775 9.E− 05 Yes financial development, gross capital formation, quality of government,
LNR →LPEC 3.79024 0.0002 Yes Uni-directional and economic growth. Regarding the PEC model, unidirectional cau­
LPEC →LNR 1.37678 0.1686 No sality is found from primary energy consumption to natural resources
LFD →LPEC 9.23741 0.0000 Yes Uni-directional and primary energy consumption to financial development. In addition,
LPEC →LFD 1.59121 0.1116 No
LGCF →LPEC 8.62490 0.0000 Yes Feedback
bidirectional causality between primary energy consumption and the
LPEC →LGCF 4.22302 2.E− 05 Yes quality of government as well as primary energy consumption and gross
LQOG →LPEC 2.02550 0.0428 Yes Feedback capital formation is found.
LPEC →LQOG 8.86528 0.0000 Yes Based on the findings, there are some policies for these countries,
Note: GDP, PEC, NR, QOG, GCF, and FD represent economic growth, Primary which will assist the policymakers in devising sustainable energy stra­
energy consumption, natural resources, quality of Government, gross capital tegies. According to the findings, it is evident that energy is an influ­
formation, and financial development. ential factor that affects economic growth in developing countries.
Hence, policymakers should focus on making long-term energy policies
that develop sustainable growth as well as efficient energy use and in­
crease the energy supply by unscrewing the indigenous energy sources
to eliminate the long-term energy crisis for developing countries [4,93].
Most notably, it is recommended that the present energy structure
should be upgraded to boost the share of renewable in the current en­
ergy mix to strengthen economic progress in developing economies [94,
95]. Natural resources also have a causality associated with primary
energy use and GDP. Therefore, countries should manage natural
resource use more effectively and efficiently, which will help to relieve
pressure on our natural resources, while also creating new jobs and
market opportunities, lowering operational costs, and boosting market
stability in the future.
Similarly, institutions are the fundamental elements of long-term
growth that cannot be envisioned without them. The quality of gov­
ernment directly affects economic progress, yet it has a spill-over effect
on energy needs via economic progress. However, developing countries
should reinforce institutional reforms through precision in governance,
control of corruption, and legal system improvement. For sustainable
Fig. 2. Casual relationship analysis.
energy and economic progress, improving the quality of government can
cause an increase in energy efficiency and solid energy policies.
investment in equipment production and technological development is Furthermore, to achieve the SDGs, these countries should concentrate on
needed for sustainable energy and growth. In addition, stability in efficient institutions and legal frameworks that enable them to reap the
government is a fundamental argument; without it, economic sustain­ advantages of cleaner energy generation.
ability cannot be accomplished. Therefore, governments should over­ The role of fixed capital formation is essential for GDP and PEC in
come institutional impediments, improve their energy strategies, and developing countries. However, higher accretion and implementation of
implement appropriate policy orientations for sustainable development. gross capital formation for more production of goods and services,
which is primarily due to technical innovation and more investment, can
6. Conclusion and policy implications boost energy production as well as the growth of developing regions.
This study has some limitations: First, this inquiry examined the
This study is the first attempt to analyze the synergy among primary primary energy consumption-growth nexus through panel data. This
energy consumption, economic growth, natural resources, quality of opens the possibility of future research for time series analysis instead of
government, financial development, and gross capital formation in the panel-based methodology. Secondly, other relevant factors or alterna­
panel data of 30 developing countries from 1990 to 2017. We apply the tive econometric techniques would contribute to expanding the litera­
panel unit root test to examine the stationary of the variables. We also ture. On the other hand, the association between energy consumption

8
A. Azam et al. Energy 263 (2023) 125570

and economic growth can be conducted at the level of energy con­ [18] Asif M, Majid A. Institutional quality, natural resources and FDI: empirical
evidence from Pakistan. Eurasian Bus Rev 2018. https://doi.org/10.1007/s40821-
sumption per unit of production. This can be addressed in future
017-0095-3.
research. [19] Hassan ST, Danish, Khan SUD, Xia E, Fatima H. Role of institutions in correcting
environmental pollution: an empirical investigation. Sustain Cities Soc 2020.
https://doi.org/10.1016/j.scs.2019.101901.
CRediT [20] Haseeb M, Kot S, Iqbal Hussain H, Kamarudin F. The natural resources curse-
economic growth hypotheses: quantile–on–Quantile evidence from top Asian
Anam Azam: conceptualization, methodology, software, data cura­ economies. J Clean Prod 2021. https://doi.org/10.1016/j.jclepro.2020.123596.
[21] Bhattacharya M, Awaworyi Churchill S, Paramati SR. The dynamic impact of
tion, writing original draft, visualization, validation, and wri­ renewable energy and institutions on economic output and CO2 emissions across
ting—reviewing and editing. Muhammad Ateeq: conceptualization, regions. Renew Energy 2017. https://doi.org/10.1016/j.renene.2017.03.102.
methodology, writing original draft, Muhammad Rafiq: conceptuali­ [22] Azam M, Liu L, Ahmad N. Impact of institutional quality on environment and
energy consumption: evidence from developing world. Environ Dev Sustain:
zation, methodology, software, writing original draft, and wri­
Multidiscip Approach Theor Practice Sustain Dev 2021;23:1646–67.
ting—reviewing and editing. Muhammad Shafique: conceptualization, [23] Saidi H, El Montasser G, Ajmi AN. The role of institutions in the renewable energy-
methodology, resource, and writing—reviewing and editing. Jiahai growth nexus in the MENA region: a panel cointegration approach. Environ Model
Assess 2020. https://doi.org/10.1007/s10666-019-09672-y.
Yuan: conceptualization and supervision.
[24] Igoni S, Anthony Nwadioha N, Odi E R. Appraisal of energy consumption and
carbon dioxide emissions on fixed capital formation in Nigeria. Int J Adv Res 2021.
https://doi.org/10.21474/ijar01/13816.
Declaration of competing interest [25] Tariq G, Sun H, Haris M, Javaid HM, Kong Y. Energy consumption and economic
growth: evidence from four developing countries. Am J Multidiscip Res 2018;7:
100–7.
The authors declare that they have no known competing financial [26] Zhang X, Zhang H, Yuan J. Economic growth, energy consumption, and carbon
interests or personal relationships that could have appeared to influence emission nexus: fresh evidence from developing countries. Environ Sci Pollut
the work reported in this paper. Control Ser 2019;26:26367–80.
[27] Adhikari D, Chen Y. Energy consumption and economic growth: a panel
cointegration analysis for developing countries. Rev Econ Finance 2012;3:68–80.
Data availability [28] UN General Assembly. UN general assembly, transforming our world : the 2030
agenda for sustainable development. Resolution adopted by the general assembly
Data will be made available on request. on 25 september 2015 2015.
[29] Azam A, Rafiq M, Shafique M, Ateeq M, Yuan J. Causality relationship between
electricity supply and economic growth: evidence from Pakistan. Energies 2020;13:
References 837.
[30] Hausman JA, Taylor WE. Panel data and unobservable individual effects.
Econometrica 1981. https://doi.org/10.2307/1911406.
[1] Saldivia M, Kristjanpoller W, Olson JE. Energy consumption and GDP revisited: a
[31] Dumitrescu EI, Hurlin C. Testing for Granger non-causality in heterogeneous
new panel data approach with wavelet decomposition. Appl Energy 2020;272:
panels. Econ Modell 2012. https://doi.org/10.1016/j.econmod.2012.02.014.
115207.
[32] Azam A, Rafiq M, Shafique M, Zhang H, Ateeq M, Yuan J. Analyzing the
[2] Chen J, Xu C, Shahbaz M, Song M. Interaction determinants and projections of
relationship between economic growth and electricity consumption from
China’s energy consumption: 1997–2030. Appl Energy 2021;283:116345.
renewable and non-renewable sources: fresh evidence from newly industrialized
[3] Cai Y, Sam CY, Chang T. Nexus between clean energy consumption, economic
countries. Sustain Energy Technol Assessments 2021. https://doi.org/10.1016/j.
growth and CO2 emissions. J Clean Prod 2018;182:1001–11.
seta.2021.100991.
[4] Azam A, Rafiq M, Shafique M, Yuan J. An empirical analysis of the non-linear
[33] Kraft J, Kraft A. On the relationship between energy and GNP. J Energy Dev 1978:
effects of natural gas, nuclear energy, renewable energy and ICT-Trade in leading
401–3.
CO2 emitter countries: policy towards CO2 mitigation and economic sustainability.
[34] Apergis N, Payne JE. Energy consumption and growth in South America: evidence
J Environ Manag 2021. https://doi.org/10.1016/j.jenvman.2021.112232.
from a panel error correction model. Energy Econ 2010;32:1421–6.
[5] Nguyen HM, Ngoc BH. Energy consumption - economic growth nexus in vietnam:
[35] Ozturk I, Kaplan M, Kalyoncu H. The causal relationship between energy
an ARDL approach with a structural break. J Asian Finance, Econ Bus 2020.
consumption and GDP in Turkey. Energy Environ 2013;24:727–34.
https://doi.org/10.13106/jafeb.2020.vol7.no1.101.
[36] Belke A, Dobnik F, Dreger C. Energy consumption and economic growth: new
[6] Jian J, Fan X, He P, Xiong H, Shen H. The effects of energy consumption, economic
insights into the cointegration relationship. Energy Econ 2011;33:782–9.
growth and financial development on CO2 emissions in China: a VECM approach.
[37] Dogan E. Energy consumption and economic growth: evidence from low-income
Sustainability 2019. https://doi.org/10.3390/su11184850.
countries in Sub-Saharan Africa. Int J Energy Econ Pol 2014;4(2):154–62.
[7] The world bank annual report 2020. 2020. https://doi.org/10.1596/978-1-4648-
[38] Daniele V. Natural resources and economic growth: a curse or a blessing? Riv Ital
1619-2.
degli Econ 2011;16:500–7.
[8] Guan Y, Kang L, Shao C, Wang P, Ju M. Measuring county-level heterogeneity of
[39] Topcu E, Altinoz B, Aslan A. Global evidence from the link between economic
CO2 emissions attributed to energy consumption: a case study in Ningxia Hui
growth, natural resources, energy consumption, and gross capital formation.
Autonomous Region, China. J Clean Prod 2017;142:3471–81.
Resour Pol 2020. https://doi.org/10.1016/j.resourpol.2020.101622.
[9] Le HP. The energy-growth nexus revisited: the role of financial development,
[40] Nawaz K, Lahiani A, Roubaud D. Do natural resources determine energy
institutions, government expenditure and trade openness. Heliyon 2020. https://
consumption in Pakistan? The importance of quantile asymmetries. Q Rev Econ
doi.org/10.1016/j.heliyon.2020.e04369.
Finance 2020.
[10] Azam A, Rafiq M, Shafique M, Zhang H, Yuan J. Analyzing the effect of natural gas,
[41] Auty RM, Auty RM, Mikesell RF. Sustainable development in mineral economies.
nuclear energy and renewable energy on GDP and carbon emissions: a multi-
Oxford University Press; 1998.
variate panel data analysis. Energy 2020:119592.
[42] Abou-Ali H, Abdelfattah YM. Integrated paradigm for sustainable development: a
[11] Shafique M, Azam A, Rafiq M, Luo X. Investigating the nexus among transport,
panel data study. Econ Modell 2013;30:334–42. https://doi.org/10.1016/j.
economic growth and environmental degradation: evidence from panel ARDL
econmod.2012.09.016.
approach. Transport Pol 2021;109:61–71.
[43] Xu X, Xu X, Chen Q, Che Y. The research on generalized regional “resource curse”
[12] Haseeb M, Kot S, Hussain HI, Mihardjo LWW, Saluga P. Modelling the non-linear
in China’s new normal stage. Resour Pol 2016. https://doi.org/10.1016/j.
energy intensity effect based on a quantile-on-quantile approach: the case of
resourpol.2016.04.002.
textiles manufacturing in asian countries. Energies 2020. https://doi.org/10.3390/
[44] Erdoğan S, Yıldırım DÇ, Gedikli A. Natural resource abundance, financial
en13092229.
development and economic growth: an investigation on Next-11 countries. Resour
[13] Hasanov F, Bulut C, Suleymanov E. Review of energy-growth nexus: a panel
Pol 2020. https://doi.org/10.1016/j.resourpol.2019.101559.
analysis for ten Eurasian oil exporting countries. Renew Sustain Energy Rev 2017.
[45] Ben-Salha O, Dachraoui H, Sebri M. Natural resource rents and economic growth in
https://doi.org/10.1016/j.rser.2017.01.140.
the top resource-abundant countries: a PMG estimation. Resour Pol 2018. https://
[14] Salisu A, Oloko T, Okunoye I, Opeloyeru O, Olabisi N. Energy consumption and
doi.org/10.1016/j.resourpol.2018.07.005.
economic growth in oil importing and oil exporting countries: a Panel ARDL
[46] Hassan ST, Xia E, Latif K, Huang J, Ali N. Another look at the nexus among energy
approach. 2018.
consumption, natural resources, and gross capital formation: evidence from
[15] Ahmed K, Mahalik MK, Shahbaz M. Dynamics between economic growth, labor,
Pakistan. Nat Resour Res 2020. https://doi.org/10.1007/s11053-019-09607-0.
capital and natural resource abundance in Iran: an application of the combined
[47] Balsalobre-Lorente D, Shahbaz M, Roubaud D, Farhani S. How economic growth,
cointegration approach. Resour Pol 2016;49:213–21.
renewable electricity and natural resources contribute to CO2 emissions? Energy
[16] Tiba S. Modeling the nexus between resources abundance and economic growth:
Pol 2018. https://doi.org/10.1016/j.enpol.2017.10.050.
an overview from the PSTR model. Resour Pol 2019. https://doi.org/10.1016/j.
[48] Ahmadov AK, van der Borg C. Do natural resources impede renewable energy
resourpol.2019.101503.
production in the EU? A mixed-methods analysis. Energy Pol 2019. https://doi.
[17] Tiba S, Frikha M. The controversy of the resource curse and the environment in the
org/10.1016/j.enpol.2018.11.044.
SDGs background: the African context. Resour Pol 2019. https://doi.org/10.1016/
j.resourpol.2019.04.010.

9
A. Azam et al. Energy 263 (2023) 125570

[49] Onishi A, Al-mulali U, Weng-Wai C, Sheau-Ting L, Mohammed AH, Consumption E, [72] Burbidge J, Harrison A. Testing for the effects of oil-price rises using vector
et al. On the relationship between Co2 emissions and economic growth : the autoregressions. Int Econ Rev 1984:459–84.
Mauritian experience. Energy Pol 2012. [73] Azam A, Rafiq M, Shafique M, Yuan J. Mitigating carbon emissions in China: The
[50] Shahbaz M, Gozgor G, Hammoudeh S. Human capital and export diversification as role of Clean Energy, technological innovation and political-institutional quality.
new determinants of energy demand in the United States. Energy Econ 2019. Frontiers in Environmental Science n.d.:182.
https://doi.org/10.1016/j.eneco.2018.11.016. [74] Atil A, Nawaz K, Lahiani A, Roubaud D. Are natural resources a blessing or a curse
[51] Shahbaz M, Hoang TH Van, Mahalik MK, Roubaud D. Energy consumption, for financial development in Pakistan? The importance of oil prices, economic
financial development and economic growth in India: new evidence from a growth and economic globalization. Resour Pol 2020;67:101683.
nonlinear and asymmetric analysis. Energy Econ 2017. https://doi.org/10.1016/j. [75] British Petroleum Company. BP statistical review of world energy. London BPC. BP
eneco.2017.01.023. Statistical Review of World Energy. 67th edition 1981. June 2018. 2018.
[52] Khan H, Khan S, Zuojun F. Institutional quality and financial development: [76] University of Gothenburg. University of Gothenburg 2018. Quality of government
evidence from developing and emerging economies. Global Bus Rev 2020. https:// institute standard dataset. 2018. Version January 2018), https://knoema.com/Q
doi.org/10.1177/0972150919892366. OGISD2018/quality-of-government-institute-standard-dataset-version-jan
[53] Abbas Q, Nurunnabi M, Alfakhri Y, Khan W, Hussain A, Iqbal W. The role of fixed uary-2018.
capital formation, renewable and non-renewable energy in economic growth and [77] Rahman MM, Velayutham E. Renewable and non-renewable energy consumption-
carbon emission: a case study of Belt and Road Initiative project. Environ Sci Pollut economic growth nexus: new evidence from South Asia. Renew Energy 2020.
Control Ser 2020. https://doi.org/10.1007/s11356-020-10413-y. https://doi.org/10.1016/j.renene.2019.09.007.
[54] Soytas U, Sari R, Ozdemir O. Energy consumption and GDP relation in Turkey: a [78] Maddala GS, Wu S. A comparative study of unit root tests with panel data and a
cointegration and vector error correction analysis. Econ Bus Transit: Facilitating new simple test. Oxford Bulletin of Economics and Statistics; 1999. https://doi.
Compet Change Global Environ. Proc. 2001;1:838–44. org/10.1111/1468-0084.0610s1631.
[55] Wolde-Rufael Y. Disaggregated industrial energy consumption and GDP: the case of [79] Im KS, Pesaran MH, Shin Y. Testing for unit roots in heterogeneous panels.
Shanghai, 1952-1999. Energy Econ 2004. https://doi.org/10.1016/S0140-9883 J Econom 2003;115:53–74.
(03)00032-X. [80] Levin A, Lin CF, Chu CSJ. Unit root tests in panel data: asymptotic and finite-
[56] Apergis N, Payne JE. Energy consumption and economic growth in Central sample properties. J Econom 2002. https://doi.org/10.1016/S0304-4076(01)
America: evidence from a panel cointegration and error correction model. Energy 00098-7.
Econ 2009. https://doi.org/10.1016/j.eneco.2008.09.002. [81] Granger CWJ. Some properties of time series data and their use in econometric
[57] Alaali F, Roberts J, Taylor K. The effect of energy consumption and human capital model specification. J Econom 1981;16:121–30.
on economic growth: an exploration of oil exporting and developed countries. [82] Engle RF, Granger CWJ. Co-integration and error correction: representation,
Sheffield Economic Research Paper Series; 2015. 2015015. estimation, and testing. Econometrica: J Econom Soc 1987:251–76.
[58] Siddique HMA, Majeed MT, Khalid A. The relationship between energy [83] Phillips PCB, Ouliaris S. Asymptotic properties of residual based tests for
consumption and economic growth in Pakistan. Bull Energy Econ 2016;4:329–34. cointegration. Econometrica: J Econom Soc 1990:165–93.
[59] Khobai H, Abel S, Le Roux P. A review of the nexus between energy consumption [84] Kao C. Spurious regression and residual-based tests for cointegration in panel data.
and economic growth in the brics countries. 2017. J Econom 1999. https://doi.org/10.1016/S0304-4076(98)00023-2.
[60] Guo WW. An Analysis of energy consumption and economic growth of Cobb- [85] Pedroni P. Critical values for cointegration tests in heterogeneous panels with
Douglas production function based on ECM. IOP Conf Ser Earth Environ Sci 2018. multiple regressors. Oxf Bull Econ Stat 1999. https://doi.org/10.1111/1468-
https://doi.org/10.1088/1755-1315/113/1/012071. 0084.0610s1653.
[61] Twerefou DK, Iddrisu KS, Twum EA. Energy consumption and economic growth: [86] Pedroni P. Panel cointegration: asymptotic and finite sample properties of pooled
evidence from the west african sub region. West Afr J Appl Ecol 2018;26:217–33. time series tests with an application to the PPP hypothesis. Econom Theor 2004.
[62] Huang YS, Shi MS, Yang YQ. Research on co-integration and causal relationship https://doi.org/10.1017/S0266466604203073.
between China’s energy consumption and economic growth. In: IOP conference [87] Streimikiene D, Kasperowicz R. Review of economic growth and energy
series: earth and environmental science, vol. 342. IOP Publishing; 2019, 12018. consumption: a panel cointegration analysis for EU countries. Renew Sustain
[63] Sultan ZA, Tty Alkhateeb. Energy consumption and economic growth: the evidence Energy Rev 2016. https://doi.org/10.1016/j.rser.2016.01.041.
from India. Int J Energy Econ Pol 2019. https://doi.org/10.32479/ijeep.8030. [88] Muhammad B. Energy consumption, CO2 emissions and economic growth in
[64] Le HP, Bao HHG. Renewable and nonrenewable energy consumption, government developed, emerging and Middle East and North Africa countries. Energy 2019.
expenditure, institution quality, financial development, trade openness, and https://doi.org/10.1016/j.energy.2019.03.126.
sustainable development in Latin America and Caribbean emerging Market and [89] Azam M. Relationship between energy, investment, human capital, environment,
developing economies. Int J Energy Econ Pol 2020;10:242. and economic growth in four BRICS countries. Environ Sci Pollut Control Ser 2019.
[65] Kwakwa PA, Alhassan H, Adu G. Effect of natural resources extraction on energy https://doi.org/10.1007/s11356-019-06533-9.
consumption and carbon dioxide emission in Ghana. Int J Energy Sect Manag 2019; [90] Kurniawan R, Managi S. Economic growth and sustainable development in
14(1):20–39. Indonesia: an assessment. Appl Artif Intell 2018. https://doi.org/10.1080/
[66] Mahjabeen, Shah SZA, Chughtai S, Simonetti B. Renewable energy, institutional 00074918.2018.1450962.
stability, environment and economic growth nexus of D-8 countries. Energy [91] Nawaz S. Growth effects of institutions: a disaggregated analysis. Econ Modell
Strategy Rev 2020. https://doi.org/10.1016/j.esr.2020.100484. 2015. https://doi.org/10.1016/j.econmod.2014.11.017.
[67] Bercu A-M, Paraschiv G, Lupu D. Investigating the energy–economic [92] Rahman ZU, Ahmad M. Modeling the relationship between gross capital formation
growth–governance nexus: evidence from central and Eastern European countries. and CO 2 (a) symmetrically in the case of Pakistan: an empirical analysis through
Sustainability 2019. https://doi.org/10.3390/su11123355. NARDL approach. Environ Sci Pollut Control Ser 2019;26:8111–24.
[68] Azam A, Rafiq M, Shafique M, Ateeq M, Yuan J. Investigating the impact of [93] Shafique M, Azam A, Rafiq M, Luo X. Evaluating the relationship between freight
renewable electricity consumption on sustainable economic development: a panel transport, economic prosperity, urbanization, and CO2 emissions: evidence from
ARDL approach. Int J Green Energy 2021. https://doi.org/10.1080/ Hong Kong, Singapore, and South Korea. Sustainability 2020. https://doi.org/
15435075.2021.1897825. 10.3390/su122410664.
[69] Solow RM. A contribution to the theory of economic growth. Q J Econ 1956;70: [94] Shafique M, Azam A, Rafiq M, Luo X. Life cycle assessment of electric vehicles and
65–94. internal combustion engine vehicles: a case study of Hong Kong. Res Transport
[70] Lucas Jr RE. On the mechanics of economic development. J Monetary Econ 1988; Econ 2021:101112.
22:3–42. [95] Shafique M, Luo X. Environmental life cycle assessment of battery electric vehicles
[71] Hamilton JD. Oil and the macroeconomy since world war II. J Polit Econ 1983;91: from the current and future energy mix perspective. J Environ Manag 2022;303:
228–48. 114050.

10