Probing environmental sustainability pathways in G7 economies: the role of energy transition, technological innovation, and demographic mobility

Abstract

Global warming remains the most devastating environmental issue embattling the global economies, with significant contributions emanating from CO₂ emissions. The continued rise in the level of greenhouse gas (GHG) emissions serves as a compelling force which constitutes the core of discussion at the recent COP26 prompting nations to commit to the net-zero emission target. The current research presents the first empirical investigation on the roles of technological advancement, demographic mobility, and energy transition in G7 pathways to environmental sustainability captured by CO₂ emissions per capita (PCCO2) from 2000 to 2019. The study considers the additional impacts of structural change and resource abundance. The empirical backings are subjected to pre-estimation tests consisting of cross-sectional dependence, second-generation stationarity, and panel cointegration tests. The model estimation is based on cross-sectional augmented autoregressive distributed lag, dynamic common correlated effects mean group, and augmented mean group for the main analysis and robustness checks. The findings reveal the existence of EKC based on the direct and indirect effects of the components of economic growth. The indicators of demographic mobility differ in the direction of influence on PCCO2. For instance, while rural population growth negatively influences PCCO2 in the short-run alone, urban population growth increases PCCO2 in the short-run and long-run periods. Nonrenewable energy, information computer technology (ICT) imports, and mobile cellular subscriptions serve as positive predictors of PCCO2, while ICT exports and renewable energy moderate the surge in PCCO2. Policy implications that enhance environmental sustainability are suggested following the empirical verifications.

Introduction

In the last 3 decades or more, the world has witnessed an unceasing surge in greenhouse gas (GHG) emissions, with ensuing effects threatening the existence of the current ecosystem, peaceful human coexistence, and endangering the sustainability of future generations. For this reason, among many others, environmental issues have remained the most debated and attracted numerous policies globally. As such, sustaining the environment through policies and initiatives that enhance carbon neutrality is recently gaining weight in ecological-related literature and equally receiving significant attention among policymakers, international organizations, and research pundits alike (Wang et al. 2023; Xue et al., 2022; Shao et al., 2021a). Issues relating to a sustainable environment came into being in 2002 under the flagship of carbon neutrality focused on reducing fossil fuel-induced production processes by 6% annually until 2030 (Shen et al., 2021a). Recently, the popularity of the carbon neutrality idea gained increased recognition among the government in the global economy after realizing that it is an essential prerequisite that guarantees the attainment of the sustainable development agenda (SDA) and the maintenance of global warming below 2 °C (Adebayo et al., 2023; Ulla et al., 2023). Consequently, achieving environmental sustainability could be a mirage without maintaining net-zero GHG emissions.

Some notable international organizations and countries have committed to delivering their mandates to pursue carbon neutrality targets. The recognition of this commitment brought world leaders together in September 2015 who jointly agreed on common and ambitious goals of safeguarding the planet from pollution through a set of seventeen (17) goals, thirty (30) indicators, and one hundred and sixty-nine (169) targets (Oke et al., 2022). Specifically, goal 15 of the SDGs advances the need to ensure the protection, restoration, and promotion of terrestrial ecosystems combined with forest management, reduction of desertification, land degradation, and biodiversity loss (Sun et al., 2021c). Besides, the Intergovernmental Panel on Climate Change (IPCC) categorically highlights the inevitability of phasing out fossil fuels to achieve a carbon-free ecosystem (Qin et al., 2021a). Recently, the 2021 United Nations Climate Change Conference (COP26) lay more emphasis on energy transition involving the reduction of the global concentration and reliance on coal power as an energy source, discouraging deforestation, accelerating the global transition to 100 renewable energy, and achieving considerable reduction in methane emissions (Ibrahim, 2022). Besides, the COP26 emphasizes and commits members to pledge and work towards reducing fossil fuel subsidies believed to have reached approximately $5.9 trillion in 2020 and contribute about 89% to GHG emissions globally.

At the national and regional levels, leading global emitters like the European Union, China, and the USA have made pledges and set targets to achieve carbon neutrality by 2050, 2060, and 2050 respectively (Lahiani et al., 2021). In particular, intergovernmental organizations such as the G7 economies are leading the global advocacy for net-zero emissions and providing practical examples for others to follow. For instance, six of the seven G7 countries like France, Canada, Germany, Japan, Italy, the USA, and the UK have all committed to carbon neutrality targets (Safi et al., 2021). The commitment of the G7 to carbon neutrality is evident in the group’s annual convergence since 1970 on issues bothering on addressing the challenges posed by GHG emissions. Specifically, the recent convergence of the G7 economies on climate matters in 2021 produces a communiqué for other countries to follow in the strides to achieve global net-zero emissions by 2050 (International Energy Agency, IEA 2021). Overall, there is a consensus that this target cannot be coincidentally achieved but by deliberate actions towards the environment. As such, nations need to be cautious of the development path they adopt to avoid forfeiting the strides towards environmental sustainability. For instance, achieving economic growth should be pursued in a way that will promote environmental sustainability by reducing the rate of overdependence on traditional energy sources. Hence, seeking more development should be strategized at a decreasing rate of CO₂ emissions. Achieving all these require policy reforms and keeping to the consensus reached in the various international treaties. That notwithstanding, the various environmental policy reforms and international agreements aimed at addressing the issue of global warming and achieving a carbonless environment failed to provide the anticipated results (Ibrahim et al., 2022a; Onifade et al., 2021). Consequently, policymakers are at the crossroad on the most appropriate and effective measures to enhance the achievement of a sustainable or carbon-free environment amidst the growing global concerns for the sustenance of the present and future ecosystem.

From an empirical viewpoint, numerous efforts have evolved towards unraveling the factors that significantly promote or hinder environmental sustainability targets. Notwithstanding the copious empirical studies and the emanating policies so far, the deteriorating state of the environment is not showing any sign of significant improvements (Ulucak & Baloch 2023). What is more evident in most cases is the rising levels of GHG emissions, for which two essential explanations can be provided for such undesirable circumstances. First, the preponderance of empirical studies is neglectful of the structural changes in the production system and how it affects the environment despite the emerging policies from countries globally towards a steady and consistent transition. Numerous empirical studies have shown that the environmental impacts of the three major sectors differ both in developing and developed economies. For instance, the agriculture sector is characterized as low carbon-intensive; the industrial sector is high carbon-intensive, and the service sector moderates CO₂ emissions (Xu et al., 2022; Hou et al., 2021). This is evident in the countries of the G7 whose service sector contributes the highest to the annual GDP (World Bank, 2021). Among the G7 countries, the agriculture sector’s contribution to GDP is mostly negative in Canada, Germany, Italy, and Japan while other states recorded declining contributions. Similar negative and diminishing trends are apparent in industrial sector contributions to GDP while the service sector maintains an increasing value added to GDP (World Bank, 2021). The gradual shifts from low to high carbon-intensive sectors and finally to carbon-mitigating sectors reflect the G7 countries’ commitment to carbon neutrality.

Furthermore, there is massive mobility from the rural to the urban city. For instance, in the last 2 decades, all the G7 countries have continually recorded a declining population in rural areas which is recently reaching a negative point except for Canada (World Bank, 2021). Alternatively, the urban population is geometrically witnessing a persistent increase in all the G7 economies except for Germany and Japan which recorded negative in the last few years (World Bank, 2021). The dichotomy of the population trend in the rural and urban areas shows the intensity of demographic mobility that is becoming the order of the present era. However, the environmental impacts of this population shift remain a less researched scope of study in the literature. Second, it is also worth mentioning that the advocacy for the energy transition from fossil fuels to renewable energy is increasingly gaining acceptance in both developing and developed economies. The G7 countries in particular have committed to phasing out fossil fuel by targeting 100% renewable electricity generation by 2035 (IEA, 2021). Moreover, the composition of energy consumption across the seven countries has seen more increase in renewables as against declining nonrenewables (World Bank, 2021). Beyond the G7 economies, other emerging and developed economies are increasingly buying into the idea of transitioning from nonrenewable to renewable energy. Despite the evolving paradigm shift, an assessment of this transition is in the early stage, especially as most of today’s productive activities rely more on traditional energy sources despite the call for a total shift. Hence, the decision to fully implement the transitional drives would require strong evidence of how effective renewable energy could help nations record more growth with less harm to the environment. This inquiry is yet to be substantially researched in the literature.

Research objectives and contributions to knowledge

The primary aim of the current research is to investigate the role of energy transition, structural change, and demographic mobility in the drive towards achieving a sustainable environment in G7 economies. By energy transition, this study seeks to examine how the advocacy on switching from heavy reliance on fossil fuels is believed to contribute significantly to the rise in environmental pollutants such as carbon emissions to renewable energy perceived as having the potential to bring strides towards a carbon neutral environment to reality. The concept of structural change equally becomes highly important in the pathway to environmental sustainability on the ground that concentrating on the service sector in the pursuit of a development path is confirmed to contribute insignificantly to carbon emissions. Hence, the path to green growth can be easily achieved through the channel of the service sector as against the agricultural and manufacturing sectors that are carbon intensive. The movement of people from the rural to the urban area can to a large extent influence the structure of environmental pollutants in an economy with the urban area contributing more to the overall stock of emissions.

Consequent upon the foregoing, the contributions of this to the extant literature are threefold. First, while extensive works have been conducted to access the drivers of a sustainable environment in recent times, the focus on structural change and demographic mobility has been grossly neglected. In essence, this study will constitute the first strand of the study assessing how structural transition and changes in population status influence the G7 efforts towards achieving carbon neutral environment. Moreover, the effects of energy transition involving the joint impacts of nonrenewable and renewable energy on environmental sustainability are scarce in the literature despite their importance. It is instructive to note that, the consideration of G7 economies is worth lauding for the following reasons. The G7 countries constitute a group of the world’s most industrialized and advanced economies (Hordofa et al., 2021). Their contributions to global development are evident in the various macroeconomic indicators. For instance, the group contributes 40% of the global GDP, 25% of worldwide energy related CO₂, and 30% of global energy consumption (International Energy Agency, 2021). The group’s population accounts for 29% of the world population, while its contributions to GHG stand at nearly 19%.

Third, by employing second-generation estimators like cross-section-dependent autoregressive distributed lag (CS-ARDL), the study can account for the long-run and short-run effects of the explanatory variables. Moreover, the consideration of robustness checks using other estimators like augmented mean group (AMG) and the correlated common effects mean group (CCEMP) is not negligible.

The remainder of the paper is structured thus. The “Literature review” section reviews the relevant literature. The “Method” section focuses on the method. “The “Empirical results and discussion” section presents and discusses the results. The “Conclusion and policy recommendations” section concludes and recommends applicable policies.

Literature review

The rising interests in the drive towards delivering a carbon-neutral economy have spurred countless research studies from different angles ranging from country-specific to panel studies, developed economies to developing economies, among others. While the debates are still ongoing and policy implications emanating from the various empirical findings, the extant studies are grossly filled with mixed reactions and views suggesting a state of inconclusiveness. Consequently, this study reviews available relevant studies on carbon neutrality and draws apparent lacuna from them.

Energy transition-environment nexus

The functional impacts of renewable energy towards environmental sustainability target constitute the core of research interest for Adebayo and Samour (2023) in BRICS from 1990 to 2018. The study relies on panel nonlinear autoregressive distributed lag to draw the positive and negative effects of the exogenous indicators. Findings reveal that expansion in production activities and nonrenewable energy exacerbates the adverse effects of environmental pollutants. On the contrary, environmental sustainability responds positively to changes in renewable energy. Similarly, Chen et al. (2023) examine the role of energy transition in the sustainability plan of the European Union States from 1998 to 2019. The study considers ecological footprint as the outcome variable while the role of energy patents, nonrenewable energy, and natural resources are considered in the empirical model. Feedbacks from the study reveal the significance of energy transition in driving environmental sustainability through the reduction in ecological footprints. Similarly, in addition, patent work reduces carbon, whereas both natural resources and nonrenewable energy escalate the surge in ecological footprint. Yu et al. (2023) find renewable energy to be efficient in driving environmental sustainability in China and India from 1990 to 2019 based on the ARDL estimator. On the other hand, economic growth and transportation driven by fossil fuels hinder the sustainability of the environment in both economies. Zhang et al. (2023) evaluate the asymmetric impacts of renewable energy, nonrenewable energy, and technological innovation in China from 1990 to 2019. The outcomes of the study show that both technological innovation and renewable energy significantly mediate carbon emissions, whereas nonrenewable energy escalates the emissions. Yuan et al. (2022) explore the novel rolling-window Granger causality test to investigate the environmental impacts of renewable energy. Empirical fallouts expose that renewable energy is fundamental in moderating CO₂ emissions towards attaining a sustainable environment. In a similar perspective, Ibrahim et al. (2022b) evaluate the disaggregated impacts of renewable energy amidst the intervening role of natural resource dependence and environmental technology in five selected African countries from 1990 to 2019. The empirical verification based on second-generation estimators reveals that both renewable energy and environmental technology promote environmental sustainability, while natural resource dependence deters it. Stroink et al. (2022) carry out research on the factors efficient for resolving the pervasive challenge of energy transition in EU economies. The empirical findings from the study reveal the interconnection between regional electricity generation and demand through switchable channels as an effective way of resolving the lingering challenge of the energy transition. Besides, Falcone et al. (2021) evaluate the channels that lead to enhancing bioenergy transition in Italy by focusing on reconverting refineries into biorefineries. Following the feedback from the various stakeholders at the national level and informal actors, pursuing the Porto Marghera reconversion is identified as a practicable and efficient solution, whereas the experience from Gela remains complicated.

Focusing on BRICS economies’ pathways to carbon neutrality, Ji et al. (2021) assess how carbon-neutral investment enhances the drive to achieve a carbon-free environment by employing monthly data of the activities of six thousand, five hundred and nineteen (6519) mutual fund managements from 2011 and 2019. The operations and categories of these mutual funds are grouped into green, brown, and black based on their capitals. The empirical results show that green investment funds perform better than other categories (black and brown) in enhancing a carbon-free environment in the sample countries. The need to assess the effectiveness of environmental taxes and eco-innovation in the pursuance of carbon neutrality is the core of the research motivation in Tao et al. (2021) in a panel of the emerging seven (E7) countries for the period covering 1995 to 2018. The study employs a collection of robust estimating techniques consisting of CS-ARDL, AMG, and CCEMG. Feedbacks uncover that environmental taxes and eco-innovation are critical factors that promote the transition to a carbon-neutral economy in E7 countries. The evaluation of the environmental impacts of renewable energy tops the research priorities interest in Li et al. (2021) for the Chinese economy using annual time series data from 1989 to 2019. The estimated model controls for the effects of export diversification and trade openness. Findings expose that long-run connections exist among the variables. More so, renewable energy and export diversification significantly mitigate CO₂ emissions. Contrawise, trade openness and economic growth hinder the transition to carbon.

Technology-environment nexus

The role of technology has come into the limelight of empirical debates on the pathways towards carbon neutral environment. Within this line of perspective, Ahmad et al. (2023) estimate the role of technological innovation in the attainment of a sustainable environment in China based on a 40-year time series. The feedback based on the ARDL estimator reveals that technological innovation drives a sustainable environment by mitigating the surge in carbon emissions. Conversely, economic growth hinders sustainable development by exacerbating the surge in carbon emissions. Sahoo et al. (2023) examine the effects of green technological innovation on the attainment of a sustainable environment in India with a focus on microdata covering 283 manufacturing sectors. The study considers the intervening effects of green knowledge in the estimated model. Findings reveal green technology enhances green knowledge towards promoting green practices that lead to a sustainable environment. Abbasi et al. (2022) probe the effects of technological innovation in the nexuses of financial development, energy use, economic growth, and territory based emissions in Pakistan from 1990Q1 to 2019Q4. The empirical verification based on the study used dynamic autoregressive-distributed lag (ARDL) simulations uncovers that economic growth and financial development drive emissions. Conversely, technological innovations reduce emissions, thereby promoting a sustainable environment.

Furthermore, Shao et al. (2021b) investigate the effects of green innovation on environmental pollution in the next eleven economies from 1980 to 2018. The verification of the empirical nexuses based on the CS-ARDL model reveals that green technology and renewable energy have long-run mitigating effects on the surge in CO₂ emissions. From the energy consumption angle, Koondhar et al. (2021) explore the impacts of bioenergy consumption and CO₂ emissions on the bioeconomic growth of the agricultural sector in China using annual data from 1971 to 2019. The study employs ARDL and dynamic nonlinear ARDL, and the results show that bioenergy consumption drives the growth of the agricultural bioeconomy while fossil fuel retards it. Probing the role of eco-research and development (ENR&D) makes up the core of research objectives in Shao et al. (2021a) for the USA economy. The model is estimated based on long-run estimators comprising fully modified OLS (FM-OLS) and dynamic OLS (DOLS) from 1990 to 2019. The outcomes show that ENR&D and renewable energy enhance the pathway towards carbon neutrality by significantly moderating the pervasive rise in CO₂ emissions. Focusing on the policy effects of carbon neutrality, Chen and Lin (2021) evaluate how carbon trading policy can effectively enhance China’s 2060 commitment towards carbon neutrality using data envelop analysis (DEA) to construct two measures, namely, energy performance index (EPI) and the aggregate factor CO₂ emissions performance index (AFCPI). Feedbacks from the study's analysis reveal that carbon-trading policy holds the potential to stimulate improvements in both indices.

The advocacies towards a carbon-neutral ecosystem must include public and private stakeholders to ensure that the expected benefits are not compromised and all key performance indicators are met. In line with the preceding view, Chen et al. (2021) assess the role of public-private partnership (PPP) in moderating the rise in CO₂ emissions through energy and energy efficiency in China. The study analyzes the earlier stated nexus based on quarterly data from 1991 to 2019, and the findings reveal the existence of a positively significant association of PPP with CO₂ emissions. Additionally, among the covariates considered in the analysis, technological innovation, renewable energy, and energy productivity enhance the carbon neutrality agenda by facilitating a substantial decrease in CO₂ emissions. Shen et al. (2021b) evaluate the influence of globalization on CO₂ emissions in BRICS from 1980 to 2018 by controlling for the impacts of renewable energy and economic growth. The study relies on FM-OLS and DOLS to drive the long-run effects coupled with AMG as robustness for the main techniques. The results show that economic growth and globalization hinder the carbon neutrality agenda. Contrarily, renewable energy proves to be efficient in promoting carbon neutrality through its abating effects on CO₂ emissions.

Furthermore, Qin et al. (2021a) estimate the extent to which green innovation, environmental policy, and the composite risk index can enhance carbon neutrality goals for G7 countries based on yearly data covering 1990 to 2019. The study validates the applicability of EKC for the countries and equally lends empirical support for the role of green innovation, the composite risk index, RER&D, and environmental policy. Contrawise, GDP appears as a dragging factor on the path to attaining carbon neutrality due to its contributions to rising CO₂ emissions. Hu et al. (2021) explore the disaggregated effects of energy consumption and technological innovation on CO₂ emissions in India from 1990 to 2018. The results indicate that disaggregated technological innovations negatively impact CO₂ emissions. Besides, the carbon-inducing roles of economic growth and energy consumption are empirically confirmed.

The role of eco-innovation in the accomplishment of the carbon neutrality agenda is central to the research objective of Iqbal et al. (2021) for selected thirty-seven Organization of Economic Cooperation and Development (OECD). The empirical model considers the functional role of fiscal decentralization and export diversification based on second-generation tests for the preliminary analysis and AMG for the estimation of the empirical model. The results show that the surge in CO₂ emissions is escalated by GDP, fiscal decentralization, and export diversification. Conversely, eco-innovation and renewable energy consumption significantly reduce CO₂ emissions. Similar facilitating roles of renewable energy and economic complexity on carbon neutrality are evident in the study of Zheng et al. (2021a) for sixteen exporting countries for the 1990–2019 study periods.

The EKC approach to carbon neutrality is another viable mechanism that is gaining attention in the carbon-neutral literature. Extending the frontier of knowledge in this perspective, Khan et al. (2021) estimate the nonlinear asymmetric impacts of economic growth, energy consumption, foreign direct investment, and trade accessibility on CO₂ emissions in Pakistan based on annual data covering 1971 to 2018. Findings from the study reveal that adverse shocks from trade accessibility and FDI cause CO₂ emissions. The dual impacts of urbanization and industrialization on CO₂ emissions constitute the center of research focus in Ahmed et al. (2021), Pakistan. The study equally considers other covariates such as FDI and economic growth. Findings from the study show that economic growth and industrialization are not significant despite their negativity. Moreover, FDI and urbanization both exert positive and significant influences on CO₂ emissions. Yue et al. (2021) investigate the dual impacts of tourism and green innovation in the abatement of environmental degradation in Thailand to enhance the achievement of carbon neutrality. The empirical findings reveal that tourism and green innovation significantly moderate the rise in CO₂ emissions.

Demographic mobility-environment nexus

The advancement in economic civilization has aided the connectivity among global economies, with the migration of people from one country to another getting easier than ever. Among many driving factors of human migration, tourism forms a vital pull factor with benefits and costs for the environment. Consequently, Liu et al. (2023) estimate the extent to which urban population (urbanization) drives environmental quality in selected provinces in China from 1995 to 2020. The intervening roles of economic development and energy consumption are equally evaluated based on the PMG-ARDL estimator. Findings show that urbanization is not substantial enough to promote or hinder environmental quality. On the other hand, economic development and energy consumption both mitigate the drive towards achieving environmental quality due to their contributions to a significant rise in carbon emissions. Bu et al. (2022) evaluate how population migration drives carbon emissions in 30 provincial regions in China for the period covering 2000 to 2019. Feedbacks from the analysis reveal that the mobility of population migration from rural to urban increases energy poverty and consumption which by extension escalates the surge in CO₂ emissions. Zhang et al. (2021) evaluate the environmental effects of demographic mobility (rural and urban population) on environmental sustainability in 41 countries in SSA from 1996 to 2018. The empirical verification based on the system GMM reveals both rural and urban populations hinder environmental sustainability in SSA. The militating effects of natural resource rents and nonrenewable energy are empirically verified. Udemba (2021) examines the EKC GDP and FDI-induced carbon neutrality model for the United Arab Emirates (UAE) by employing annual data spanning 1980 to 2018. The contributing roles of selected variables such as population and energy use are included in the model estimation. Findings from the study reveal an inverted U-shaped nexus between economic growth and the environment (proxy by ecological footprint). Besides, the effects of energy use and population increase the surge in ecological footprint while FDI reduces it. Shan et al. (2021) advance the empirical evidence for the empirical impact of population, renewable energy, and technological innovation on CO₂ emissions within the STIRPAT framework from 1990 to 2018 in Turkey. Based on the ARDL method employed, the study revealed that green technology and renewable energy lessen carbon emissions. In contrast, population, energy consumption, and economic growth exacerbate CO₂ emissions.

Research gap in the literature

The prior empirical studies provide some loopholes in the evolving environmental sustainability debates. First, the extant empirical studies on a sustainable environment are filled with mixed reactions on the direction of effects between certain macroeconomic indicators and carbon emission surges. For instance, while Zhang et al. (2021) submit that demographic mobility especially urban population substantially drives the surge in carbon emissions, Liu et al. (2023) find that urban population is insignificant to influence the emission surge. Second, most studies focused mainly on the environmental effects of GDP without precedence for the structural shift in the various sectors. This study focuses on the structural change-environmental sustainability nexus. For instance, GDP constitutes an aggregation of the contributions from different sectors in the economy such that examining the carbon impacts of each will avail the policymakers of the specific direction in decision-making. Third, notwithstanding the massive empirical studies advancing various factors driving or dragging the accomplishment of environmental sustainability, no study has considered structural shift, population demographic, and energy transition in the G7 economies environmental effects.

Method

This section provides details on the methodological approach adopted by the present study, including theoretical underpinning, data description, and model specification, a priori signs of the relationship, econometric procedures, and preliminary analyses.

Theoretical underpinning

The theoretical basis guiding the functional relationship between economic growth, population growth, natural resource abundance, technological advancement, energy consumption, transport services, and CO₂ emissions is exposited thus. First, the impacts of structural change have been empirically affirmed to drive CO₂ emissions differently (Xu et al., 2022). However, the increasing positive nexus between economic growth and CO₂ emissions (Ibrahim &Ajide, 2021) is not static due to the changing activities in each of the sector. Second, population growth constitutes a fundamental determinant of CO₂ emissions considering the dependent nature of human existence on the environment. Consequently, an increase in population implies a corresponding increase in depletion of natural resources or energy consumption which contributes to increased CO₂ emission, thus suggesting the existence of a positive nexus (Adebayo et al., 2022). Third, natural resource abundance plays a significant role in enhancing or hindering the quality of the environment. The nexus’s theoretical argument is that the desire to earn more income from natural resources in the form of rents could continuously motivate their depletion without concern for the adverse environmental effects. Conversely, the stream of income earned from natural resource rents can be reinvested to promote environmental-related technology, which subsequently enhances a quality environment. Based on these diverging views, some empirical studies lend support for positive impacts of natural resource abundance on environmental pollutions (Omokanmi et al. 2022), while adverse effects are equally supported (Majeed et al., 2021). Fourth, the emerging roles of technological advancement in carbon neutrality advocacy have been echoed empirically in recent times with the majority accentuating the supporting impacts of technology in the roadmap towards carbon-neutral ecology (Ibrahim et al. 2022a; Hu et al., 2021). Fifth, the nexus between energy consumption and CO₂ emissions remains one of the most researched areas in environmental literature, with the majority advancing the existence of direct and indirect effects of energy consumption on CO₂ emissions.

The above-illustrated theoretical nexus is further elaborated in the conceptual framework presented in Fig. 1, with consideration of the heterogeneous effects of the regressors. From the relationship depicted, economic growth rate (EG) is decomposed into agricultural (AG), industry (IND), and service (SERV) sectors. The joint activities in the agriculture and industry sectors involving field burning, coal-fired plants, fertilizer manufacturing for food storage and packaging, cement production, and construction contribute significantly to greenhouse gas (GHG) emissions (Athira et al., 2019). The transitional phase in economic advancement involves the extensive contributions of agriculture and manufacturing-driven economy to GHG emissions. Moreover, a service-driven economy posits significant mitigation of the carbon surge. In addition, the succeeding position is well accentuated by the EKC when it hypothesizes an intensive carbon-driven economy at the early phase of development, which gradually reduces at the advanced stage of economic development. Hence, we can assume that the service sector supports the drive towards a carbonless ecological system. From the empirical front, Ge and Lei (2014) find document inverse nexus of the service sector with CO₂ emissions. Besides, some emissions related to the service sector are sometimes an offshoot of the interplay with other sectors. Alluding to this submission, Hou et al. (2021) posit that carbon flows in China attributed to the service sector are usually associated with other sectors’ service demand. Moreover, Hashmi et al. (2020) provide empirical backing for the insignificant contribution of the service sector to GHG emissions. Based on the situations depicted, both direct and indirect nexuses are anticipated in the economic growth—CO₂ emissions nexus.

Fig. 1

Conceptual framework

Demographic mobility, which involves population growth decomposed into the rural population (RUP) and urban population (URP), contributes to CO₂ emissions by increasing carbon footprint. The channel of population contribution to CO₂ emissions when dissected into rural and urban differs. For instance, the urban population (URP) instigates an increase in industrial activities leading to a rise in air pollution, precipitation, and congestion in metropolitan cities. Empirical evidence has alluded to the carbon-inducing effect of the urban population (Fan et al., 2021; Huo et al., 2021; Wang and Wang, 2021). On the contrary, advancement in urban cities through the interplay of technology employed to accommodate accelerated growth could avail people of environmentally friendly social amenities and services that, by extension, moderate CO₂ emissions. Besides, innovation driven by urbanization enhances a green environment and promotes sustainable development. Some studies have alluded to the stimulating effects of urbanization on innovation (Bashir et al., 2021). With respect to the rural population, the low level of income of the household leaves them with the choice of consuming fossil fuel and employing traditional production methods, which negatively impact the environment through the stimulation of CO₂ emissions. Moreover, empirical study has documented that the aging rural population increases household CO₂ emissions (Fan et al., 2021).

The components of technology advancement (TA) comprising ICT imports of goods (ICTM), ICT exports of goods (ICTE), and mobile cellular subscriptions (MCS) have the possibility of promoting green growth and the environment. However, an exceptional case could be made for ICT imports of goods with anticipated retarding effects on the environment due to the likelihood of CO₂ emissions transfer through the importation of carbon-embedded goods (Ibrahim & Ajide, 2021). More so, care should be taken with respect to mobile cellular subscriptions as the indicator could spur the use of energy intensive gadgets which may adversely affect the environment. The impact of energy consumption differs based on energy source such that a moderating effect on CO₂ emissions is evident from renewable energy (RE). In contrast, nonrenewable energy drives CO₂ emissions escalation. Lastly, the measures of natural resource abundance (NRA) comprising oil rents (OR), coal rents (CR), gas rents (GR), mineral rents (MR), and forest rents (FR) all have the possibilities of enhancing or mitigating the increase in CO₂ emissions. An exceptional case may be made for the gas rents since empirical evidence has alluded to its mitigating effects on CO₂ emissions (Ibrahim and Ajide, 2021).

Data description and model specification

Following the preceding theoretical underpinning, this study adopts six exogenous variables comprising economic growth (EG), which include three indicators, namely, agriculture, forestry, and fishing, value added (annual % growth), industry (including construction), value added (annual % growth), and services, value added (annual % growth). To validate or refute sector-level EKC, the square of the preceding GDP (EGsq) indicators is taken. Population growth (PG) data comprise rural population growth (annual %) and urban population growth (annual %). Natural resource abundance (NRA) entails five measures, including oil rents (% of GDP), coal rents (% of GDP), natural gas rents (% of GDP), forest rents (% of GDP), and mineral rents (% of GDP). Energy consumption (EC) serves as a vector of nonrenewable energy captured by fossil fuel energy consumption (% of total) and renewable electricity net generation (kWh). Technology advancement (TA) denotes a superset of three indicators such as ICT goods imports (% total goods imports), ICT goods exports (% of total goods exports), and mobile cellular subscriptions (per 100 people). The outcome variable representing carbon neutrality (CO2N) is captured by CO2 emissions metric tons per capita (PCCO2). Regarding the source of the variables, renewable energy is sourced from the International Energy Administration, while others are collected from the World Bank databank. The study period covers from 2000 to 2019. The choice of the based year (2000) is motivated by the unavailability of some variables (service sector and technology indicators) in the years preceding the base, while 2019 is considered the end year due to the unavailability of variables (after 2019) on many of the indicators. Consequently, the model can be stated as in Eq. (1).

$${\displaystyle \begin{array}{l}{\mathrm{EVS}}_{it}={\varphi}_1+{\varphi}_2{\mathrm{STC}}_{it}+{\varphi}_3\mathrm{STC}{sq}_{it}+{\varphi}_4{PG}_{it}+{\varphi}_5{\mathrm{NRA}}_{it}\\ {}+{\varphi}_6{\mathrm{ENT}}_{it}+{\varphi}_7{\mathrm{TA}}_{it}+{\varpi}_{it}\end{array}}$$

(1)

where EVS denotes environmental sustainability captured by CO₂ emissions per capita (PCCO2). By implication, any decrease in CO₂ emissions suggests a gradual transition to net-zero emissions, while an increase would denote a movement away from it. STC and STCsq are structural change and square, both representing a vector of transition within three sectors, agriculture sector (AGRIC), industrial sector (INDUSTRY), and service sector (SERVICE) and their squares (AGRICSQ, INDUSTRYSQ, and SERVICES). Demographic mobility is proxied by population growth (P), comprising rural population growth (RURALPG) and urban population growth (URBANPG). NRA implies natural resource abundance measured by five indicators entailing oil rents (OILR), coal rents (COALR), natural gas rents (GASR), forest rents (FOREST), and mineral rents (MINERAL). ENT is decomposed into renewable energy (REC) and nonrenewable energy (NEC). TA technology advancement is decomposed into ICT imports of goods (ICTIM), ICT exports (ICTEX), and mobile cellular subscriptions (MCS). TS denotes transport services representing imports (TSIM) and exports (TSEXP). More so, “i” represents a cross-section which comprises economies of the group of seven such as Canada, France, Germany, Italy, the UK, and the USA, while “t” implies time coverage of the study, which stretches from 2000 to 2019. Consequently, Eq. (1) can be stated in six models with full consideration of the heterogeneous indicators.

Model 1

$${\displaystyle \begin{array}{l}\mathrm{PCCO}{2}_{it}={\varphi}_1+{\varphi}_2{\mathrm{AGRIC}}_{it}+{\varphi}_3{\mathrm{AGRIC}\mathrm{SQ}}_{it}+{\varphi}_4{\mathrm{INDUSTRY}}_{it}+{\varphi}_5{\mathrm{INDUSTRY}\mathrm{SQ}}_{it}\\ {}+{\varphi}_6{\mathrm{SERVICE}}_{it}+{\varphi}_7{\mathrm{SERVICE}\mathrm{SQ}}_{it}+{\varpi}_{it}\end{array}}$$

(2)

where Eq. (2) represents the model estimating the structural change effects with consideration for verifying sector-specific EKC.

Model 2

$$\mathrm{PCCO}{2}_{it}={\varphi}_1+{\varphi}_2{\mathrm{RURALPG}}_{it}+{\varphi}_3{\mathrm{URBANPG}}_{it}+{\varpi}_{it}$$

(3)

Equation (3) denotes the model gauging the impact of demographic mobility on environmental sustainability.

Model 3

$$\mathrm{PCCO}{2}_{it}={\varphi}_1+{\varphi}_2{\mathrm{REC}}_{it}+{\varphi}_3{\mathrm{NEC}}_{it}+{\varpi}_{it}$$

(4)

Equation (4) represents the model estimating the effects of energy transition.

Model 4

$$\mathrm{PCCO}{2}_{it}={\varphi}_1+{\varphi}_2{\mathrm{OILR}}_{it}+{\varphi}_3{\mathrm{COALR}}_{it}+{\varphi}_4{\mathrm{GASR}}_{it}+{\varphi}_5{\mathrm{FORESTR}}_{it}+{\varphi}_6{\mathrm{MINERALR}}_{it}+{\varpi}_{it}$$

(5)

Equation (5) implies the model evaluating the impacts of natural resource abundance on carbon neutrality.

Model 5

$$\mathrm{PCCO}{2}_{it}={\varphi}_1+{\varphi}_2{\mathrm{ICTEX}}_{it}+{\varphi}_3{\mathrm{ICTIM}}_{it}+{\varphi}_4{\mathrm{MCS}}_{it}+{\varpi}_{it}$$

(6)

Equation (6) represents the model assessing the impacts of technology on carbon neutrality.

A priori expectations of the nexuses

The expected signs in the nexuses specified above are exposited. Following Eq. (1), since the agricultural sector is hypothesized to drive a significant increase in CO₂ emissions, a positive sign is therefore expected between the two indicators as thus ${\varphi}_2=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{AGRIC}}>0$. Similar positive effects of industry are anticipated on PCCO2, thus, ${\varphi}_2=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{INDUSTRY}}>0$. Considering the anticipated pro-green economic impacts of the service sector, an inverse nexus is hypothesized between the sector and CO₂ emissions, giving the following ${\varphi}_2=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{SERVICE}}<0$. To verify the validity of EKC, five hypotheses are tested as follows;φ₂ = 0 = φ₃ = 0 hypothesizes no economic growth-pollution nexus. φ₂ ≻ 0, φ₃ = 0, hypothesizes economic growth-pollution-inducing nexus. φ₂ ≺ , φ₃ = 0, hypothesizes economic growth-pollution abating nexus. φ₂ ≺ , φ₃ ≻ 0, hypothesizes U-shaped economic growth-pollution nexus. φ₂ ≻ , φ₃ ≺ 0, hypothesizes inverted U-shaped economic growth-pollution nexus. Precisely, negative signs are expected; thus, if the EKC hypothesis must hold, ${\varphi}_3=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{AGRICSQ}}<0,{\varphi}_5=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{INDUSTRYSQ}}<0,{\varphi}_7=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{SERVICESQ}}<0$. The impact of population growth (PG) on CO₂ emissions is anticipated to be positive such that a percentage rise in PG leads to a corresponding increase in the stocks of CO₂ emissions, leading to ${\varphi}_4=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{RURALPG}}>0$ and ${\varphi}_4=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{URBANPG}}>0$ for both rural population growth and urban population growth, respectively. All the five measures of natural resource abundance (NRA) are expected to positively relate with CO₂ emissions making them driving factors to the surge in GHG emissions. Consequently, the relationships can be depicted thus ${\varphi}_5=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{OILR}}>0$, ${\varphi}_5=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{COALR}}>0$, ${\varphi}_5=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{GASR}}>0$, ${\varphi}_5=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{FORESTR}}>0$, and ${\varphi}_5=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{MINERALR}}>0$ denoting the heterogeneous effects of natural resource abundance captured by oil rents, gas rents, coal rents, gas rents, forest rents, and mineral rents, respectively. Relating to energy consumption, renewable energy consumption (REC) reduces CO₂ emissions, while nonrenewable energy consumption (NEC) increases thus ${\varphi}_6=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{REC}}<0$ and ${\varphi}_6=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{NEC}}>0$. Technology advancement has gained much recognition globally regarding its role in moderating the surge in CO₂ emissions. The decomposed effects of technology advancement emanating from ICT imports, ICT exports, and mobile cellular subscriptions can be stated as ${\varphi}_7=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{ICTEX}}<0$, ${\varphi}_7=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{ICTIM}}<0$, and ${\varphi}_7=\frac{\delta {\mathrm{PCCO}}_2}{\delta \mathrm{MCS}}<0$.

Econometric procedures

Cross-section dependence test

The rapid increase in globalization and economic integration in the last few decades has brought the global economies closer than ever and equally enhanced partial or complete dependence among them. This global nexus makes the issue of cross-sectional dependence highly essential in panel studies because failure to accord it the proper importance could lead to misrepresentation and inconsistency of panel regression estimates (Qin et al., 2021a). One of the standard approaches used in resolving this issue is to test for the presence of cross-sectional dependence (CSD) in the panel model using Pesaran et al.’s test (2004). Hence, we employ Pesaran’s CD tests in this study to examine the cross-sectional dependency in the panel of G7 economies. The CSD test based on the null hypothesis of no cross-sectional dependence is given below:

$${\mathrm{CD}}_{\mathrm{test}}=\sqrt{\frac{2T}{N\left(N-1\right)N}}.\sum \limits_{i=1}^{N-1}\sum \limits_{k=i+1}^N{T}_{ik}$$

(8)

Slope heterogeneity test

The homogeneous features of countries may impede estimators from producing unbiased and consistent results, especially when the model involved is panel-based. Given the preceding submission, this study conducts slope homogeneity for the G7 countries using Pesaran and Yamagata’s (2008) test originally advanced by Swamy (1970)¹ in Eqs. (9) and (10) below.

$${\tilde{\Delta}}_{SH}={(N)}^{\frac{1}{2}}{(2k)}^{-\frac{1}{2}}\left(\frac{1}{N}\tilde{S}-k\right)$$

(9)

$${\tilde{\Delta}}_{\mathrm{ASH}}={(N)}^{\frac{1}{2}}{\left(\frac{2k\left(t-K-1\right)}{T+1}\right)}^{-\frac{1}{2}}\left(\frac{1}{N}\tilde{S}-2k\right)$$

(10)

Given that SH and ASH denote slope coefficient homogeneity (SH) and the adjusted SH in that order. In case the results of the estimated model support the null hypothesis of homogeneity in the dataset, the second-generation test becomes the most appropriate. If the results prove the contrary, the first-generation test remains the best choice.

Panel stationarity test

Upon resolving the twin issues of cross-section dependence and slope heterogeneity in a panel model, testing for stationarity becomes the next important step to be considered to avoid the spuriousness of the data employed. To test for unit root, we use the CIPS test credited to Pesaran (2007) and support the results with IPS (2003) developed by Pesaran. The CIPS test efficiently addresses the issue of cross-section dependence and heterogeneity inherent in a panel dataset. The model explaining the CIPS unit root test is given below:

$$\Delta {Q}_{it}={\sigma}_i+{\sigma}_i{R}_{i,t-1}+{\sigma}_i{\overline{Q}}_{t-1}+\sum \limits_{m=0}^{\rho }{\sigma}_{il}\Delta \overline{Q_{t-1}}+\sum \limits_{m=0}^{\rho }{\sigma}_{il}\Delta {\overline{Q}}_{t-1}$$

(11)

In Eq. (11), $\overline{Q}$ represents the cross-section means, which is expanded thus.

$${Q}_{i,t}={\sigma}^1\overline{eg^{i,t}}+{\sigma}^2\overline{pg^{i,t}}+{\sigma}^3\overline{nra^{i,t}}+{\sigma}^4\overline{ec^{i,t}}+{\sigma}^5\overline{ta^{i,t}}+{\sigma}^6\overline{ts^{i,t}}$$

(12)

The test statistic of CIPS is specified as

$$\overline{\mathrm{CIPS}}=\frac{1}{N}\sum \limits_{i=1}^n\mathrm{CADF}$$

(13)

Such that, CADF represents cross-sectional augmented Dickey-Fuller (CADF). The two-unit root tests proposed in this study are placed on the null hypothesis of a unit root in the dataset. The alternative hypothesis states otherwise.

Panel cointegration

The next procedure after the determination of the stationarity status of the dataset is ascertaining whether the nature of the relationship in the model is a long-run nexus or not. We employ the Westerlund (2007) cointegration test to determine the long-run association among the heterogeneous indicators. Two reasons can be advanced for the adoption of Westerlund’s cointegration test in this study. First, the common factor restriction does not impede the test from providing valid and reliable estimates (Khan et al., 2021). Second, the test is appropriate and fits in the presence of cross-sectional dependence (Kapetanios et al., 2011).

Equations (14)–(17) below illustrate the group and panel statistics for the Westerlund test based on the null hypothesis of no cointegration.

$${G}_t=\frac{1}{N}\sum \limits_{i-1}^N\frac{{\sigma^{{}^{\backprime}}}_i}{SE\left({\sigma}^{{}^{\backprime}}\right)}$$

(14)

$${G}_{\sigma }=\frac{1}{N}\sum \limits_{i-1}^N\frac{T{\sigma^{{}^{\backprime}}}_i}{\sigma^{{}^{\backprime}}(1)}$$

(15)

$${P}_T=\frac{{\sigma^{{}^{\backprime}}}_i}{SE\left({\sigma}^{{}^{\backprime}}\right)}$$

(16)

$${P}_{\sigma }=T{\sigma}^{{}^{\backprime}}$$

(17)

Panel regression estimation

Some macroeconomic shocks could cause distortions in the expected nexus between two interrelated macroeconomic variables, such as the oil price shock of 2014–2016 and the previous global financial crisis of 2008–2009. These shocks could hinder efficient estimates if inherent common factors are not duly accounted for in the regression. To resolve these issues, the CS-ARDL estimator credited to Chudik and Pesaran (2013) is employed due to the estimator’s capability to address the problems of slope heterogeneity and endogeneity and subdue the likely effects of cross-section dependence issues (Sun et al., 2021a). The equation depicting the CS-ARDL model is given below.

$$\varDelta {Y}_{i,t}={\sigma}_i+\sum \limits_{l=1}^{\rho }{\sigma}_{il}\varDelta {Y}_{it}-1+\sum \limits_{l=0}^{\rho }{\sigma}_{il}^{\prime }{X}_{s,i,t-1}+\sum \limits_{l=0}^1{\sigma}_{il}^{\prime }{\overline{\mathrm{CSM}}}_{i,t-1}+{\varpi}_{i,t}$$

(18)

CSM, which denotes cross-section means, is disintegrated through ${\overline{\mathrm{CSM}}}_t=\left(\varDelta {Y}_t,{X}_{s,t}\right)$ variables, i.e., regressors denoted as X. To check the robustness of the CS-ARDL estimator, we employ both CCEMG and AMG.

Overview of the empirical indicators in G7 economies

The Group of Seven (G7) is one of the leading intergovernmental organizations formed by the most industrialized economies of the world, comprising Canada, France, Germany, Italy, the UK, Japan, and the USA. Since its formation in 1973, the group has been saddled with deliberating on economic, financial, social, and political matters across regional and global economies. It also discusses issues relating to providing financial and technical assistance to less developed and emerging economies (International Monetary Fund, 2021). The group’s contributions to curtailing the rising level of global warming can hardly be overemphasized. Among many efforts and resolutions to resolve climate issues, G7 members committed to actualizing net-zero emissions by 2050 (IEA, 2021). As an action plan to achieve the target of a carbon-neutral environment by the target year, the group launched a blueprint to accomplish net-zero electricity generation, which serves as a lead-way for other economies to take serious commitment in the efforts towards decarbonizing the planet to a zero level by 2050.

The trend analysis of the various indicators and how they relate among G7 countries is provided in Figs. 2 and 3. Regarding the structural transition indicators (Fig. 2a–c), it is event that the agriculture sector’s contribution to GDP in the different economies of the G7 is presented in Fig. 2b. As apparent from the figure, the agriculture sector is contributing to a declining rate to the economy of the majority of the G7 countries. Similarly, a negative growth rate is recorded in some countries within the group, traceable to Canada, Germany, Japan, and the UK. In contrast, the USA is the only G7 country with positive and rising growth rates in the agriculture sector. The trend in the industry sector in Fig. 2b reveals that all the G7 countries have declining and negative growth rates. The downward and negative slopes in the agriculture and industry sectors accentuate the group’s commitment to pursuing net-zero emissions by 2050 (International Energy Agency, IEA, 2021). This developmental transition from primary to secondary and tertiary implies the transition towards a low-CO₂ emissions production system. The development state of the G7 countries is evident in the prevailing contribution of the service sector to their economies’ GDP growth. Going by the trend in Fig. 2c, it is no gainsaying that the G7 economies are service-oriented. Among the group of countries, the contribution of the service sector varied across the years under consideration. For instance, the UK contributed the most in 2004, followed by Canada, whereas Canada surpassed the UK in 2009. In 2014, the leading contributors were Canada and the USA respectively. The trends vary in 2019 in favor of the USA in 2019. Figure 3a–e presents the level of economic dependence on the various natural resources. As can be seen in Fig. 3a relating to dependence on oil rents, Canada is the economy that depends most on oil resources, followed by the UK and the USA. Besides, the dependence on coal (Fig. 3b) records the highest values in the USA, Canada, and Germany. Natural gas rents (Fig. 3c) record the most significant contribution to GDP in Canada, the USA, and the UK. Forest rents (Fig. 3d) provide significant contributions in all the countries, with Canada, the USA, France, and Germany topping the list. It should be noted that the contributions of forest rents in the G7 countries are at a decreasing rate. The significance of mineral rents to GDP (Fig. 3e) is only feasible in Canada and the USA.

Fig. 2

a Agriculture value added (annual % growth). b Industry value added (annual % growth). c Service value added (annual % growth)

Fig. 3

a Oil rents (% of GDP). b Coal rents (% of GDP). c Natural gas rents (% of GDP). d Forest rents (% of GDP). e Mineral rents (% of GDP)

As reported in Fig. 4a, nonrenewable energy consumption (fossil fuel) recorded a diminishing trend in the G7 economies, thus supporting the group’s drive and advocacies for a transition from carbon-induced energy to green and clean energy consumption. Furthermore, it is instructive to note that while Fig. 4a indicates the transition path of the G7 economies from fossil fuels or dirty energy sources, Fig. 4b confirms the destination path towards renewable or green energy sources. The G7 countries face massive emigration from rural areas to urban cities (Fig. 5a, b). All the G7 economies currently record negative rural population growth except Canada, which has maintained an infinitesimally low positive rural population growth. Despite the high records of the population in urban cities, a general decline is evident, suggesting an overall population decrease in these countries.

Fig. 4

a Fossil fuel energy consumption (% of total). b Renewable electricity net generation (kWh)

Fig. 5

a Rural population growth (annual %). b Urban population growth (annual %)

The transition towards decarbonization of the economy to guarantee the sustainability of the present and future generations is yielding positive results with more prospects in the G7 economies (Fig. 6). Among the G7 countries, France stands the better chance of hitting net-zero CO₂ emissions by 2050 since it has the lowest stock of CO₂ emissions per capita. Next to France are Italy, Japan, the UK, and Germany. The USA records the highest stock of CO₂ emissions per capita (PCCO₂), followed by Canada. This outcome is not surprising because the USA is ranked as the second-largest carbon emitter globally behind China, with contributions standing at 11% and 27% for both, respectively. It is not surprising to see Canada next to the USA because it is the most dependent on natural resources among the G7 economies (Fig. 3a–e).

Fig. 6

CO₂ emissions per capita

The recognition of the role of technology in enhancing a sustainable environment has motivated a persistent increase in global trade in ICT goods from both export and import angles. Among many others, computers, consumer electronic products, communication tools, and electronic parts, among others, have seen emerging trans-border exchanges in the last few decades. The highest ICT exports are recorded in Canada, followed by Japan, the USA, the UK, France, Germany, and Italy (Fig. 7a). The highest ICT imports are recorded by Canada, followed by the USA, Japan, the UK, Germany, France, and Italy in order of volumes (Fig. 7b). As evident in Fig. 7c, mobile cellular subscriptions show an increasing trend across all the G7 economies, with Italy recording the highest. The general trends in the components of technology (Fig. 7a–c) show that the G7 economies are open to technology and are equally achieving a lot in that regard.

Fig. 7

a ICT goods exports (% of total goods exports). b ICT goods imports (% of total goods imports). c Mobile cellular subscriptions (per 100 people)

Preliminary analysis

The pilot analysis of the indicators employed in this study is presented in Table 1, with two sections comprising descriptive statistics and normality tests. Starting with descriptive statistics, the mean value of per capita CO₂ emissions (PCCO2) which is 16.17 shows the level of CO₂ emissions that individuals contribute to the aggregate CO₂ emissions. More so, it is evident that among the three sectors, service has the highest mean value (1.70), followed by industry (0.83) and agriculture (0.77). This result implies that structural change in G7 countries is driven by the service sector, which is plausible and intuitive given the level of economic advancement in these countries. A service economy is concentrated in services as the main economic activities with less production and agriculture. Some of the activities in a service economy could include trade, finance, tourism, businesses, health schemes, retailing, marketing, information technology, hotels, restaurants, transports, and others. Available statistics show that the contribution of the service sector to the economies of the various countries of the G7 is far more than in another sector. Examples of the service contribution to GDP in the G7 economies are the USA (80.2%), Canada (70.2%), France (78.8%), Italy (73.9%), Germany (68.6), Japan (68.7%), and the UK (80.2) (Central Intelligence Agency, 2021). Besides, the mean value of oil rents (0.33%) is the highest among the measures of natural resource abundance. Next to oil rents are gas rents (0.18%), coal rents (0.07%), mineral rents (0.05), and forest rents (0.04). The importance of oil rents over other indicators signifies the level of dependence of the G7 economies on oil resources.

Table 1 Preliminary analysis

The decomposition of population growth shows a negative mean value for rural population growth (−0.95) against positive urban population growth (0.80). The degeneration in rural population growth as against progression in urban population agrees well with the Lewis dual sector model, which illustrates the transition of labor from the rural sector of the economy to the urban region. Furthermore, the average value of nonrenewable energy (78.04) is more than renewable energy’s mean (2.59). This implies that the G7 economies are still dependent on nonrenewable energy despite rising levels of renewable energy consumption. The mean value of ICT exports and imports shows that the latter is more than the former.

Assessing the normality status of the dataset, a close evaluation of skewness, kurtosis, and Jarque-Bera’s (JB) tests reveals some exciting details. The skewness values, which stand between −0.01 and 3.50 (asymmetric), and the kurtosis, which are mostly above 3 (heavily tailed), imply the variables are not normally distributed. More so, the Jarque-Bera null hypothesis of normal distribution across the dataset is rejected for almost all the variables. Consequently, we can infer that a certain degree of heterogeneous effects exists among the variables, particularly accentuating that the economic structure among the G7 countries broadly differs. The preceding features characterizing the nature of the dataset and the economies of the G7 further underscore the need to employ an econometric method that accounts for divergence in the structure of the cross-sectional units.

The present study examines the level of cross-section dependence (CD) in the stated model using the Pesaran (2004 and 2015) CD tests. From the results in Table 2, it is evident that the null hypothesis of independence among the cross-section units cannot be accepted given the significant nature of the CD tests. Consequently, we conclude that the cross-section unit is dependent on one another to the extent that a shock in one unit instigates a similar shock in another unit. Interestingly, the probabilities of the indicators are significant at a one percent level suggesting a solid level of interdependence among the economies in the G7 countries. A case in point can be cited as the global financial crisis, which started in the USA and extrapolated to other developed and developing nations. The recent 2014/2015 global fall in oil prices and the novel COVID-19 virus are equally notable instances. A close look at the coefficients of the correlation analysis will also reveal some confirmations of the CD test results.

Table 2 Cross-section dependence, correlation analyses, and slope heterogeneity

In addition to the CD and correlation test outcomes, we further conducted a test to examine the homogeneous nature of the variables to ascertain whether they are homogeneously related or not using the Pesaran and Yamagata (2008) test. The results indicated in Table 2 fail to accept the null hypothesis of homogenous series at a one percent significance level. In contrast, we accept the alternative hypothesis, which explains that the data are heterogeneous. What does the presence of CSD and heterogeneity in the model portend for the data analysis? The theoretical implications of cross-sectional dependence coupled with the slope heterogeneity of the coefficients in a model imply that first-generation unit root tests cannot provide reliable and consistent stationarity results for such a model. Alternatively, the second-generation unit root test stands as the best and most appropriate in such a situation.

The results of the stationarity tests consisting of CIPS (2007) and IPS (2003) are presented in Table 3. Based on the results of CIPS, it is evident that the null hypothesis of unit root cannot be accepted at I(0) for AGRIC, INDUSTRY, SERVICE, OILR, RURALPG, REC, and ICTIM, while other series are not significant. However, at the I(1) difference level, all the variables are significant at a one percent level (1%) except MINERALR, which is significant at a ten percent level (10%). Considering IPS (200), the series has stationarity at the I(0) level for AGRIC, INDUSTRY, SERVICE, COALR, ICTEX, and ICTIM at one percent (1%) significant level. Similarly, OILR and FORESTR are significant at five percent (5%) and ten percent (10%) significant levels, meaning that the null hypothesis of unit root is rejected. In contrast, other series fail to reject the null hypothesis of the unit root at level I(0). The subjection of the series to the difference level of the unit root test reveals that all are significant at one percent (1%) level, leading to the rejection of the null hypothesis of a unit root in the series.

Table 3 Results of stationarity test for first and second-generation estimators

Beneath Table 3, the cointegration test results based on Westerlund (2007) examine the validity of the null hypothesis, which posits that there is no long-run relationship. It is evident from the outcomes of both group statistics (Gt and Ga) and panel statistics (Pt and Pa) that the null hypothesis cannot be accepted while the alternative hypothesis positing the existence of long-run cointegration is unconditionally accepted at the one percent (1%) significant level.

Empirical results and discussion

This section presents and discusses the outcomes of the estimated models illustrating the functional relationship among the selected variables in the determination of carbon neutrality status in the G7 economies. Two subsections are presented on the results: the main empirical results consisting of the outcomes from the CS-ARDL estimator and the robustness results comprising outcomes from the CCEMG and AMG estimators.

Main empirical outcome: CS-ARDL estimator

The emanating short-run and long-run results in Table 4 reveal that the agricultural sector and industry sector exert positive effects on per capita CO₂ emissions. However, the impacts of the service sector are negatively significant on per capita CO₂ emissions. The effects of the square of the sectors (agriculture, industry, and services) are negatively significant on per capita CO₂ emissions (PCCO2). The positive and negative effects of the agriculture, industry, and service sectors validate the sector-based EKC hypothesis. The impact of the rural population is significantly negative on PCCO2 in the short-run but positive in the long-run without significance. Furthermore, the effects of the urban population are positive and significant in both the short-run and long-run. Renewable energy consumption has negative and significant short-run and long-run impacts on PCCO2, while nonrenewable energy exerts positive impacts in both periods. The lag of PCCO2 exerts adverse and significant effects on the current stock of CO₂ emissions throughout the models.

Table 4 Economic growth-population growth-energy consumption-CO₂ emissions nexuses

Focusing on the discussion of the economic intuitions in the preceding empirical results, it should be noted that the positive effect of agriculture on PCCO2 implies that the various activities in the sector such as bush burning, movement of farm instruments and equipment during tilling and planting, utilization of fertilizers and pesticides, and clearing of farmlands escalate the stock of CO₂ emissions. Based on this outcome, it could be inferred that agriculture-related activities in the rural areas of G7 economies contribute significantly to the surge in CO₂ emissions. This is despite the sector’s minimal contributions to GDP in the G7 economies comprising Italy (2.1%), the USA (0.9%), the UK (0.6%), Canada (1.6%), France (1.7%), Germany (0.7%), and Japan (1.1%) (World Economic Forum, 2020). The result is consistent with Raihan et al. (2023), Koondhar et al. (2021), and Xiong et al. (2020), which find that the agricultural sector positively promotes CO₂ emissions. Despite the reported positive agriculture-CO₂ emissions nexus, Jiang et al. (2021) and Wang et al. (2020) find a negative relationship. Similarly, the positive industry-PCCO2 nexus implies that the former contributes to the rising stock of PCCO2 in the G7 economies. Industrial sectors’ activities contribute to CO₂ emissions from various channels such as the production of cement, emission of smoke into the atmosphere, electricity production and distribution, incineration, and management. Relative to the members’ countries of the G7, it is instructive to note that the contribution of the industrial sector to GDP is still high despite the apparent structural changes. Consequently, it is not strange to see the industrial sector contributes significantly to the surge in CO₂ emissions in G7 countries. As evident in this study, the industrial sector’s contribution to CO₂ emissions is equally established in the findings of Rehman et al., (2023), Raihan (2023), Sikder et al. (2022), and Chontanawat (2020).

The service sector’s mitigating impact on PCCO₂ is quite intuitive, going by the fact that the activities in the service sector involve less energy-intensive activities. It is thus not surprising to see that the majority of the economic activities in the G7 are concentrated in the service sector. The moderating impacts of the service sector on CO₂ emissions as uncovered in this study are inconsonant with prior studies such as Xu et al. (2022), Huo et al. (2021), Zheng et al. (2021b), and Hashmi et al. (2020), which advance that the service sector moderates CO₂ emissions and is equally less carbon-intensive. The agriculture-induced EKC hypothesis, supported in this study, corroborates (Jiang et al. (2021)), which confirms the EKC existence in the agriculture sector. Besides, the joint validation of EKC for the agriculture, industry, and service sectors is consistent with similar empirical results in Wang et al. (2021b).

Relating to the impacts of rural population growth, which is negative in the short-run, it could be inferred that the transition of people from rural to urban areas in the various economies of the G7 in search of greener pasture reduces production activities in the transition period, which could result in a reduction in the stock of CO₂ emissions per person. This reduction would only last for a few periods of migration to urban cities (short-run). As the rural area balances up its workforce and resumes its full operation in production activities, the emitting effects become evident. The carbon-inducing effects of urbanization agree well with our prior sign because industrial activities are implemented in mass in urban cities due to the rising demand from the increasing population. Besides, as proposed by Rostow, the mass production stage of economic growth occurs most often in urban cities. Consequently, it is not unexpected that urban population growth instigates a significant rise in CO₂ emissions. Some previous studies also support the positive urban population-CO₂ emissions nexus (Wei et al., 2023, Sikder et al., 2022; Fan et al., 2021; Huo et al., 2021; Wang and Wang, 2021). In a similar vein, the inducing and mitigating effects of nonrenewable energy and renewable energy on CO₂ emissions as exposed in this study are plausible. For instance, energy consumption from nonrenewable energy remains the core source for most developed countries with the G7 inclusive. In particular, coal consumption constitutes the largest contributor to electricity generation. In contrast, oil consumption still tops the source of energy consumption by the industrial sectors in developed economies like the G7. It is thus not amazing to see the countries trapped in the circle of CO₂ emissions despite policy efforts for decades. As reported in this study, the moderating effects of renewable energy on PCCO2 are in line with the fact that renewable energy is less carbon-intensive and equally promotes a green economy. The findings on both nonrenewable and renewable energy in the present study corroborate the findings of Voumik et al. (2023), Ali et al. (2022), and Yuping et al. (2021), which report that while nonrenewable energy retards environmental quality, renewable energy promotes it.

In Table 5, the outcomes of the impactful roles of natural resource abundance are unveiled for both short-run and long-run periods. Going by the results, it is instructive to note that oil rents, coal rents, forest rents, and mineral rents drive CO₂ emissions per capita (PCCO₂). In contrast, gas rents inversely relate to PCCO₂. In terms of technology, ICT exports negatively relate to PCCO₂ in the short-run and long-run. ICT imports positively and significantly impact PCCO₂ in the short-run, while the effects are not significant in the long-run. Mobile cellular subscription positively and significantly interacts with PCCO₂. Transport services from both import and export angles positively drive PCCO₂ in the short-run and long-run.

Table 5 Natural resource abundance-technology-transport services- CO₂ emissions nexuses

Regarding the plausibility of the preceding results to economic intuition, the driving effects of oil rents, coal rents, forest rents, and mineral rents on PCCO₂ can be justified because earnings from natural resources propel their continuous depletion with less consideration for the environmental impacts. The adverse effects of natural gas rents are understandable because the combustion of natural gas results in less than 50% to 60% CO₂ emissions, and considering the technological advancement status of the G7 economies, producing less than the highlighted percentages is plausible. The moderating impacts of ICT exports on PCCO₂ can be explained because the employment of technological devices in production processes enhances efficiency and reduces carbon emissions during the process. In addition, ICT enhances less energy-intensive equipment in the production process, thereby limiting the surge in CO₂ emissions. In contrast, the importation of ICT deters the environment in the short-run due to the imported carbon-embedded goods. In other words, the technical level of imported goods may sometimes not meet the technological standards of well-developed economies. The time frame that is needed to correct the emissions contributed to affecting the environment in the short-run. However, within time, the emissions are contained, making them insignificant in the long-run. The carbon-inducing effects of mobile cellular subscriptions are plausible on the ground that the subscription could serve as a medium that encourages the use of energy-intensive gadgets, which by extension drives CO₂ emissions. The moderating effects of technology as evident in ICT exports are well supported in previous studies like Ahmad et al. (2023), Sahoor et al. (2023), Abbasi et al. (2022), Hu et al. (2021), Shan et al. (2021), and Shao et al. (2021b), which submitted that technology improves the environment towards carbon neutrality. The inverse relationship of the lag of CO₂ emissions per capita (PCCO2) can be clarified from a policy angle that regulations implemented to curtail previous environmental degradation can effectively moderate the surge in subsequent years. A graphical representation of the empirical outcomes is presented in Fig. 8a, b.

Fig. 8

a Graphical presentation of the empirical outcomes in the economic growth-population growth-energy consumption-CO₂ emissions nexuses. b Graphical presentation of the empirical outcomes in the natural resource abundance-technology-transport services-CO₂ emissions nexuses

Robustness checks: using different estimators (CCEMG and AMG)

The need to confirm the persistent nature of the estimates produced by the CS-ARDL estimator motivates the present study to conduct a robustness check by employing two different estimators to similar models. The results of the assessments, which are presented in Tables 6 and 7, provide some notable backings for the main outcomes. For example, the induced impacts of the agriculture and industrial sectors are supported by the long-run estimates of both CCEMG and AMG. This implies that all things being equal, the contributive effects of both agriculture and industrial sectors on CO₂ emissions in the G7 economies are not negligible. More so, the moderating impacts of the service sector on CO₂ emissions are supported by the results from both estimators. The sector-induced (agriculture, industry, and service) EKC is empirically supported. It should, however, be noted that the parameter estimates differ in terms of magnitude. Regarding the proxies of population growth, only rural-urban population growth is significant and robust for the main outcomes. The mitigating and facilitating impacts of renewable energy and nonrenewable energy are found to be robust. The effects of natural resource abundance are significant for oil, coal, gas, and forest rents in the estimates from both CCEMG and AMG. However, mineral rents are only significant in the results provided by AMG. The impacts of the components of technology are robust in terms of significance and signs with slight differences in magnitudes.

Table 6 Robustness checks-I

Table 7 Robustness checks-II

Conclusion and policy recommendations

The present study constitutes one of the emerging contributions to the environmental sustainability debates in recent periods. While the concept is not new, recent commitments by developed economies and international organizations reinvigorate the empirical interest in the concept. Consequently, this study examines how structural change, demographic mobility, natural resource abundance, technology, and energy consumption drive or deter the pathway to carbon neutrality in the G7 economies. This study conducts preliminary analyses comprising descriptive statistics, normality tests, cross-sectional dependence tests, first- and second-generation stationarity tests, and panel cointegration tests to achieve the set objectives. Upon confirming the fulfillment of the dataset to the basic tests, the study further estimated the empirical model using CS-ARDL as the main estimator. In addition, both CCEMG and AMG are employed for robustness checks. The following results are established from the estimated models. First, agriculture and industry sectors positively drive CO₂ emissions per capita (PCCO₂) while the service sector reduces it. In addition, the square of the three sectors significantly mitigates the surge in PCCO₂, which confirms the existence of EKC in the G7 economies. Second, population growth in rural has a negative effect on PCCO₂ in the short-run, respectively. Urban population growth, on the other hand, positively promotes PCCO₂ in both the short-run and long-run. Third, four of the components of natural resource abundance comprising oil rents, coal rents, mineral rents, and forest rents positively promote PCCO₂ while gas rents negatively influence it. The impacts of ICT indicators differ across the measures such as ICT exports (negative −), ICT imports (positive +), and mobile cellular subscriptions (positive +). Fourth, consequent upon the results, the following policy recommendations are suggested towards attaining 2050 net-zero CO₂ emissions in the G7 countries and other developed worlds with a similar level of advancement.

• i. To curtail agriculture-induced CO₂ emissions, the government of the various states of the G7 should implement policies that enhance the transition to conservation tillage and adopt sustainable and green livestock and manure management systems in the agriculture sector.

• ii. Agriculture-induced CO₂ emissions can also be mitigated by implementing initiatives that encourage carbon offsetting through payment incentives to farmers and rural landowners who adopt climate-friendly farming practices in the G7 economies. More so, farming equipment that enhances green farming practices should be subsidized and made readily available for farmers to access. This is important considering the levels of technological advancement in the G7 economies.

• iii. The government should implement laws to compel industries to adopt energy efficiency practices in their production and construction activities. More so, the adoption of renewable energy should be encouraged across firms and industries in the G7 economies.

• iv. The transition from carbon-induced energy (fossil fuel) to a clean and environmentally friendly energy source (renewable energy) should be holistically pursued from all fronts in line with the G7 communique of achieving 60% renewable energy electricity generation by 2030 and 100% by 2035. This could include subsidizing all products and services that promote renewable energy and tax waivers for industries and companies that adopt renewable energy as their primary energy source. A tax increment or embargo can be placed on energy-intensive (fossil fuel) products and services.

• v. The government should implement programs and initiatives that will promote diversification from natural resource dependence to other areas of the economy that are low-carbon intensive. A persistent concentration on developing the service sector could help the G7 economies achieve this suggestion.

• vi. The government should uphold its commitment to promoting and adopting technology in the different phases of economic and production activities.

In order to advance the knowledge frontier on the drive to net-zero emissions, future research can address the following caveats. First, future studies can consider the asymmetric effects of the key variables to explore a comprehensive evaluation of the positive and negative impacts of structural change, demographic change, and the energy transition on the G7 environment. Second, future studies can consider the role of technological innovation and population in the EKC and STIRPAT frameworks to provide robust policy implications towards accomplishing the carbon neutrality agenda. Third, the role of green finance which has gained increased attention in the literature can be considered in the G7 environmental sustainability model.