Natural resource use, industrialization and climate change in Africa: Blueprints for sustainable regional development

Ngangnchi, Forbe Hodu; Aquilas, Nkwetta Ajong; Mbella, Mukete Emmanuel Article Natural resource use, industrialization and climate change in Africa: Blueprints for sustainable regional development Research in Globalization Provided in Cooperation with: Elsevier Suggested Citation: Ngangnchi, Forbe Hodu; Aquilas, Nkwetta Ajong; Mbella, Mukete Emmanuel (2024) : Natural resource use, industrialization and climate change in Africa: Blueprints for sustainable regional development, Research in Globalization, ISSN 2590-051X, Elsevier, Amsterdam, Vol. 9, pp. 1-10, https://doi.org/10.1016/j.resglo.2024.100245 This Version is available at: https://hdl.handle.net/10419/331171 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc/4.0/ Natural resource use, industrialization and climate change in Africa: Blueprints for sustainable regional development Forbe Hodu Ngangnchi a , Nkwetta Ajong Aquilas b , * , Mukete Emmanuel Mbella c a Department of Organizational Sciences, Higher Institute of Commerce and Management, The University of Bamenda, Cameroon b Department of Economics, University of Buea, South West Region, Cameroon c Department of Management Science, Higher Technical Teachers’Training College, Kumba, University of Buea, South West Region, Cameroon ARTICLE INFO Keywords: Climate change Environmental regulation Environmental Kuznets Curve Industrialization Natural resource use Sustainable development ABSTRACT Climate crisis continue to animate environmental policy debate and remains a fundamental global concern. While natural resources use and industrialization are main drivers of climate change especially in developing countries like Africa, previous empirical studies have failed to analyze the simultaneous effect of both on climate change. To close this gap, the current study investigates the relationship between natural resources use, industrialization and climate change in Africa. Specifically, this study analyzes the effect of natural resources rents and value added in manufacturing on carbon dioxide emissions. Addressing these issues is important, as the implications of natural resources extraction and industrial development on environmental sustainability cannot be underestimated. This paper is very relevant because of the limited research on comprehensive frameworks that integrate natural resources use with industrialization policies specific to Africa. Using data from 2005 to 2022 for 45 African countries obtained from the African Infrastructural and World Bank databases, and employing the Cross-Sectional Feasible Generalized Least Squares, Driscoll-Kraay effects, Dynamic Driscoll-Kraay effects and Panel Quantile on Quantile regression techniques, the empirical findings revealed that natural resources rents and industrialization adversely contribute to climate change in Africa. Quantitative results show that a 1% increase in natural resource rents increased carbon dioxide emissions by 0.000103% for both Feasible Generalized Least Squares and Driscoll-Kraay estimates and by 0.00813% for the Dynamic Driscoll-Kraay estimates. Furthermore, a 1% increase in value added in manufacturing led to a 0.377% increase in carbon dioxide emissions for both Feasible Generalized Least Squares and Driscoll-Kraay estimates and a 0.157% increase for Dynamic Driscoll-Kraay estimates. The result also shows that the effect of natural resource extraction and industrialization on climate change become negative at both the 25% quantile and 50% quantile. However, this turning point becomes nullified at the 75% and at 90% quantiles. Africa should adopt a green path to natural resource use and industrialization. 1. Introduction Since the start of the Industrial Revolution around 1760, human activities have contributed significantly to atmospheric emissions of carbon dioxide and other greenhouse gases, altering the planet’s climatic composition. The earth’s climate is also influenced by natural processes including variations in land use patterns. This explains why empirical debates focus on the role of natural resources use and industrialization on climate change. For instance, studies such as Adebayo and ¨ Ozkan (2024), Adebayo et al. (2024), Haseeb et al. (2020); Fan et al. (2023) and Aladejare (2023) have recorded that the amount of GHGs in the environment are further boosted by the expanding scales of economic activities particularly manufacturing. Minerals, energy, food, fiber and other natural resources from the environment are utilized by man for economic purposes, which degrades the ecological system. This is through chemical agents and energy particles and releasing of poisonous particles, producing palpable unbalanced environmental situations (Zubair et al., 2017). Increase in greenhouse gas (GHG) emissions is a significant factor contributing to this serious ecological crisis. Harmful components enter the environment through industrial sites that release harmful smoke, vapour and fumes, damaging the ecosystem and aggravating climate change. Growth in carbon dioxide (CO 2 ) emissions is strongly linked with the increased consumption of non-renewable energy consumption in economic activities relative to renewable energies, which reduce environmental degradation (Adebayo & ¨ Ozkan, * Corresponding author. E-mail address: [email protected] (N. Ajong Aquilas). Contents lists available at ScienceDirect Research in Globalization journal homepage: www.sciencedirect.com/journal/research-in-globalization https://doi.org/10.1016/j.resglo.2024.100245 Received 17 January 2024; Received in revised form 26 July 2024; Accepted 4 August 2024 Research in Globalization 9 (2024) 100245 Available online 12 August 2024 2590-051X/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/bync-nd/4.0/ ). 2024; Adebayo et al., 2024; Xu et al., 2022; Adebayo, 2022; Aquilas et al., 2024; Aquilas &Atemnkeng, 2022). Greenhouse gas concentrations have risen significantly since the industrial revolution, with increased industrialization and consequently extraction of natural resources. For example, atmospheric CO 2 concentrations reached 409.8 ppm (parts per million) in 2019, which is higher than at any other time in the last 800,000 years (Lindsey et al., 2020). Prior to the Industrial Revolution, the CO 2 concentration was 280 ppm, and for the last three centuries, it has varied between 180 and 280 ppm. This figure rose in the past 10 years, reaching 2.1 ppm yearly from the 1950s when it was higher by about 0.7 ppm yearly (Bergamaschi et al., 2013). Greenhouse gas emissions (GHG) in India and China significantly reduced between 2019 and 2020 due to a surge in the COVID 19 crisis which hampered economic activities globally. However, the emissions regained a positive trajectory as the temporary lockdowns were lifted (Ahmed et al., 2022). Environmental circumstances also worsened at same period. For instance, the average global temperature in 2021 was 1.11◦C which is higher than pre-industrial levels (Aladejare, 2022). In the wake of increased environmental pressures and stresses, the Paris climate agreement, often known as COP21 was born in 2015 with the aim of limiting the growth of greenhouse gas (GHG) emissions (Alola &Adebayo, 2023c; Irfan et al., 2022; Aquilas &Atemnkeng, 2022; Aquilas et al., 2022). This accord targets net-zero GHG emissions by achieving less than 2 ◦C average temperatures globally and making sure it does not bypass 1.5 ◦C. Successive United Nations (UN) accords after this (COP26, COP27 and COP28) have continued to emphasize the development and utilization of renewable energy technologies as the channel through which growth in carbon emissions can be curtailed. Most recently, building on previous agreements, COP28 emphasized four pillars global climate action needs to focus on in order to keep global average temperatures at bay. These include fast-tracking an energy transition described as just, orderly and equitable; increasing the commitments of the Parties towards climate finance mobilizations; dwelling on the people, their lives as well as livelihoods and fostering full inclusion in climate action. In fact, the issue of collaboration especially with non-members of the Party to the climate change convention (cities, businesses, investors, subnational regions and civil society) was re-echoed (United Nations Framework Convention on Climate Change [UNFCCC], 2024). The level of manufacturing activities, resource exploitation, globalization, urbanization, income growth, and the development of human capital are all crucial elements that affect industrial development on the African continent (Zhan, 2020). However, the use of these components in any nation’s industrial growth might raise questions about the sustainability of the environment. Through damaging extractive operations, natural resources required for industrial expansion led to environmental degradation. In order to industrialize, developing nations have extracted huge amounts of natural resources, which has had detrimental ecological repercussions (Aladejare, 2023). Examples include the degradation of the ecosystem caused by the extraction of non-renewable natural resources such as minerals, crude oil, and natural gas (Majeed et al., 2021; Aladejare, 2022; Ganda, 2022; Adekoya et al., 2022; Adebayo et al., 2022). The expanding paradigm shift, particularly in developing nations with social and economic opportunities from rural to urban areas, is another environmental concern brought on by industrialization. As urbanization in wealthy nations continues to increase, rural–urban migration has been increasing in developing nations (Awad &Mallek, 2023; Kundu, Venkat, &Babu, 2020; Quito, del RíoRama, ´ Alvarez-García, &Dur´ an-S´ anchez, 2022; Yang &Khan, 2022). If not properly managed, such rapid industrialization and extraction of natural resources can be damaging to the environment, creating more climatic problems. Most African countries are witnessing an increase in industrial production, which has become a development priority, coupled with a stable increase in resources extraction. For instance, manufacturing output in South Africa, Rwanda, and Nigeria increased by 39.3 %, 30.2 %, and 4.6 % respectively in 2021 (Moll de Alba &Todorov, 2022). Efforts by developing countries to win support for industrial transformation have therefore intensified in recent years and this has led to very rapid increases in the amount of energy used to produce products and services (Nazlioglu et al. 2011, Solarin et al. 2017). Future prosperity for the continent may be dim if these industrial expansions are not environmentally sound. Demand for natural resources is rising as a result of industrialization both within and outside of Africa. Around 30 % of the world’s mineral reserves are thought to be contained in Africa’s massive extractive industries (UNEP, 2024). Additionally, 40 % of the world’s gold deposits and 90 % of the chromium and platinum deposits of the world are located on the African continent (Aladejare, 2020). Again, roughly 12 % and 8 % of the world’s total reserves of oil and natural gas respectively are found in Africa (UNEP, 2024). In the last three decades, Africa has witnessed a continuous increase in CO 2 emissions, on the average, though the continent records the lowest emissions levels compared to Asia (Aquilas &Atemnkeng, 2022). According to World Bank (2024) data, CO 2 emissions for Sub-Saharan Africa increased from about 401805kt in 1990 to 760868kt in 2020 while for the Middle East and North Africa, it rose from 860,056 kt in 1990 to 2,416,065 kt in 2020. This increasing trend in CO 2 emissions reflects an increase in economic activities, highly fueled by consumption of energy from non-renewable sources. Africa’s ecological degradation is further aggravated by increased exploitation of its natural resources (Aladejare, 2022), leading to natural resources depletion. For instance, the natural resources depletion rate for Sub-Saharan Africa that stood at 5 % of GNI in 1990 settled at 6 % in 2020, though a rate of up to 11 % was witnessed within the period (World Bank, 2024). Consequently, the climate catastrophe remains a major concern across the African continent, as the continent is more vulnerable to climatic shocks. With the youngest population in the world and an average annual population growth rate of 2.4 % for the last 30 years (AfDB, 2024) and the rate of urbanization predicted to be the greatest in the world by 2050 (Walther, 2021), the climate problem remains a reality that must be dealt with. Nomenclature Acronyms and abbreviations AfDB African Development Bank AMG Augmented Mean Group ARDL Autoregressive Distributed Lag CD Cross Section Dependence CIPS Cross-sectionally Augmented Im-Pesaran-Shin CO 2 Carbon Dioxide Emissions DOLS Dynamic Ordinary Least Squares FDI Foreign Direct Investment EKC Environmental Kuznets Curve FGLS Feasible Generalized Least Squares FMOLS Fully Modified Ordinary Least Squares GHG Greenhouse Gas GHGs Greenhouse Gases GNI Gross National Income PMG Pooled Mean Group UN United Nations UNEP United Nations Environment Programme UNFCCC United Nations Framework Convention on Climate Change USA United States of America VECM Vector Error Correction Model WDI World Development Indicators F. Hodu Ngangnchi et al. Research in Globalization 9 (2024) 100245 2 Thus, the need for studies to guide policy formulation regarding the implications of natural resources use and industrial activities on climate change in Africa. On the basis of the above, this study sets out to realize two operational objectives; first is to investigate the effect of natural resource rents on CO 2 emissions; second is to analyze the effect of value added in manufacturing on CO 2 emissions. Haven gone through the introduction section, section two reviews the literature, section three presents the methodology, section four presents the findings and discussions while section five draws up a conclusion and provides policy suggestions. 2. Review of past studies From a theoretical perspective, the Environmental Kuznets Curve (EKC) hypothesis, named after the American economist Simon Kuznets provides the theoretical basis for understanding the implications of economic activities on climate change. The idea of the relationship between economic aggregates and environmental quality was first reported in 1952 and since then, has been discussed in technical conversations about environmental policy (Grossman &Krueger, 1991). Following this hypothesis, environmental quality invariably improves with increased clean energy production and consumption as it lowers carbon emissions. The EKC hypothesis has attracted great attention from both academics and policy makers, mainly because of its implications for sustainable development (Zhang et al., 2022; Aquilas et al., 2022). The EKC framework argues that at early stages of development, improvement in economic activities can only be achieved while increasing environmental degradation, but this tendency changes at higher incomes. This is because as industrialization and economic development advances, environmental damage increases due to the greater use of natural resources, more emission of pollutants, operation of less efficient and relatively dirty technologies, high priority given to increase in material output, and disregard for or ignorance of the environmental consequences of growth. However, as growth continues, cleaner water, improved air quality and a generally cleaner habitat become more valuable as people make choices at the margin about how to spend their incomes. In the postindustrial stage, cleaner technologies and a shift to information and service-based activities combined with a growing willingness to enhance environmental quality becomes necessary. The argument has moved beyond this today as the consequences of climate change and global warming have become more glaring (higher global temperatures, rising sea levels, desertification of previously forested areas, drying off of river beds, amongst others), casting doubts on the possibility of an improving environment in the presence of an increase in economic activities. Several empirical studies have examined how natural resources extraction and industrialization affect climate change, with most concluding that natural resources extraction and industrial development is detrimental to the environment. According to Ding et al. (2022), natural resources extraction is one of the key elements affecting a nation’s progress in industrial development. However, they observed that unplanned and unmanaged industrialization has a significant negative influence on the environment, particularly on the fundamental ecology, the habitats of species and the diversity of life on Earth. Raheem and Ogebe (2017) studied the effect of industrial development and urbanization on CO 2 emissions using a panel of 20 African countries with data from 1980 to 2013. The estimates from pooled mean group (PMG) showed that both industrialization and urbanization directly contribute to CO 2 emissions, but indirectly diminish it. Li and Lin (2015) however shows, with data from 1971 to 2010 for 73 countries that the effect of industrialization on CO2 emissions depends on income level. Panel threshold regression estimates reveal that in the lowand middleincome group, industrialization increases carbon emissions but at high income levels, the impact becomes insignificant. Using data from 1970 to 2015 and employing the Autoregressive Distributed Lag (ARDL) model and vector error correction model (VECM), Liu and Bae (2018) concluded for China that industrialization increases CO 2 emissions. Similar evidence exists for Bangladesh (Shahbaz et al., 2014), Tunisia (Ghazouani, 2022), developed countries (Dong et al., 2019), group of developing countries (Sikder et al., 2022), Saudi Arabia (Mahmood et al, 2020), Turkey (Pata, 2018). Steinbach (2019) examined the EKC hypotheses for climate change using data that covered the years 1986 through 2014, noting strongly that climate change is a real problem in many countries. Fodha and Zaghdoud (2010) found evidence to support the EKC hypothesis of climate change in Tunisia. Furthermore, Shahbaz and Sinha (2019) showed that estimation for CO 2 emissions remained speculative. However, Song et al. (2013) tested the EKC hypothesis in China and found inconsistent outcomes across the country. According to Liu et al. (2017), the relation between economic growth and CO 2 emissions that causes climate change was proven in some ASEAN countries over the long term but was not supported in others. Aquilas et al. (2022) showed that agriculture and gross domestic product significantly cause deforestation in the Congo Basin, with a U-shaped pattern observed. Moreover, an inverted U-shaped relationship was found between manufacturing and deforestation, buttressing the fact that though economic activities can create environmental stresses, the environment can still be improved while not reducing the activities. Ahmed et al. (2022) used panel data from 55 countries in the AsiaPacific region from 1995 to 2020 and applied an autoregressive distributed lag (ARDL) model to study how industrialization and foreign direct investment affect the environment. The findings demonstrated that industrialization has a beneficial and considerable influence on the environment, whereas FDI generally has a major negative impact and increases methane and CO 2 emissions, leading to the acceptance of pollution haven (PH) and the environment Kuznets curve (EKC) hypotheses in the Asia-Pacific region. Evidence suggests that much of the literature concentrates on developed economies (Yuan et al., 2020; Khan et al., 2021; Rahman &Lamsal, 2021; Nasir et al., 2021; Sarkodie & Owusu, 2021), demonstrating that industrialization and natural resource exploitation have important bearing on the climate change through CO 2 emissions. However, there is also evidence that developing economies have not been neglected (Sharma et al., 2020; Zafar et al., 2020; Haseeb et al., 2020; Zahoor et al., 2022; Gallego-Bono &TapiaBaranda, 2022; Usman &Balsalobre-Lorente, 2022). These studies have shown that industrialization is positive on increasing CO 2 emissions, hence deteriorating the climate system. The impact of resource extraction and industrialization on climate variability through CO 2 emissions is real, necessitating the business community and legislators to factor climate change in business decision making (Yang &Khan, 2022; Quito et al., 2022). Liu (2023) analyzed the nexus between natural resources exploitation and CO 2 emissions in the USA. Employing data from 1984 to 2021 and the quantile regression technique, the study showed that natural resource extraction impacts negatively on climate change, necessitating the fast expansion and utilization of renewable energy and technological innovations. Kwakwa et al. (2020) investigated the effect of natural resources extraction on CO 2 emissions and energy consumption in Ghana from 1971 to 2013 using the ARDL technique. The result revealed that resources extraction aggravates CO 2 emissions in the long run. Alhassan and Kwakwa (2023) also reached a similar conclusion for Ghana with an extended dataset from 1971 to 2018 based on the Fully Modified Least Squares method within the Environmental Kuznets Curve (EKC) framework. Cai et al. (2023) also argues, based on the Fully Modified Ordinary Least Squares (FMOLS), Dynamic Ordinary Least Squares (DOLS), and Canonical Cointegration Regression (CCR) with data from 1989 to 2021 for China that total natural resources extraction deteriorates the environment through rising CO 2 emissions and compromises the COP26 objectives. Again, utilizing panel data for ASEAN countries between 2006 and 2020, Nguyen et al. (2023) queues up with other studies to confirm that increased natural resources rents leads to high emissions of CO 2 but that renewable energy consumption use will increase environmental sustainability. Balsalobre-Lorente et al. (2021) F. Hodu Ngangnchi et al. Research in Globalization 9 (2024) 100245 3 however conclude that an inverted U-shaped relationship exist between natural resources extraction and CO 2 emissions for European countries, insinuating that at a certain level, environmental sustainability can be guaranteed. The contribution and novelty of the current study to the literature is evident. To the best of our knowledge, this study is the first to investigate the simultaneous effect of natural resources use and industrialization on climate change in Africa. While some studies (Liu, 2023; Alhassan & Kwakwa, 2023; Cai et al., 2023; Nguyen et al., 2023; Ni, 2023; Balsalobre-Lorente et al., 2021; Kwakwa et al., 2020; Baloch et al., 2019) have concentrated on the effect of natural resources extraction on climate change, others (such as Sikder et al., 2022; Dong et al., 2019; Liu &Bae, 2018; Raheem &Ogebe, 2017; Li &Lin, 2015; Mahmood et al, 2020; Pata, 2018) focused on the effect of industrialization. No study so far appears to have analyzed the combined effect of resource extraction and industrialization on CO 2 emissions. This study therefore represents an addition to the existing literature on this subject, which is still scanty but growing. By studying the combined effect of resources extraction and industrialization on climate change, this study becomes more comprehensive as integrates two major drivers of climate change (exploitation of natural resources and industrialization) in Africa. Furthermore, this study uses an updated dataset spanning from 2005 to 2022 while also extending the number of African countries to 45 unlike most previous studies. Again, most past studies explored have employed econometric techniques such as FMOLS, DOLS, CCR, ARDL, VECM, PMG, panel data methods (OLS and fixed effects), augmented mean group (AMG), cross-sectional ARDL (CS-ARDL), panel ARDL, threshold regression with a few employing Panel Quantile on Quantile regression and none using the feasible generalized least squares (FGLS) and the Driscoll-Kraay and Dynamic Driscoll-Kraay models. Following Adebayo (2024) and Adebayo and ¨ Ozkan (2024), adoption of panel quantile regression is advantageous, as it makes possible the estimation of heterogenous effects that may characterize the conditional distribution of the dependent variable, while also taking care of individual and timespecific extraneous variables. The panel quantile on quantile regression methodology allows us to understand the relationships between the variables outside of the mean of the dataset (25 % quantile, 50 % quantile, 75 % quantile and at the 90 % quantile). Additionally, the employment of the novel feasible generalized least squares (FGLS), Driscoll-Kraay and Dynamic Driscoll-Kraay models are appropriate and consistent in the presence of cross-sectional dependence and heteroskedasticity, as the case is with the current study. Thus, this study fills both the empirical and the methodological gaps. 3. Materials and methods The thematic scope of this study is on the nexus between natural resource use, industrialization and climate change in Africa. The study utilizes a global panel data spanning from 2005 to 2022 inclusive for 45 African countries collected from African Development Bank (2022) and the World Development Indicators (2023). These countries whose list appears on the appendix were selected on the basis of availability of complete and non-truncated data. The empirical model for this study is informed by the empirical studies of Aladejare and Nyiputen (2023), Liu (2023), Alhassan and Kwakwa (2023), Cai et al. (2023), Nguyen et al. (2023), Balsalobre-Lorente et al. (2021), Ni (2023), Kwakwa et al. (2020), Baloch et al. (2019), Sikder et al. (2022), Dong et al. (2019), Liu and Bae (2018), Raheem and Ogebe (2017), Li and Lin (2015), Mahmood et al. (2020), Pata (2018) and Ahmed et al. (2022) and the EKC hypothesis. Thus; CO2it =β0+β1NRRit +β2INDUSit +γi χ it +ni+ τ +ξit (1) From Eq. (1),CO2it represents a measure of climate change, NRRit is natural resource rents, INDUSit is industrialization as measured by share of manufacturing to GDP, χ it is a vector of control variables such as aggregate infrastructural development index, gross domestic product (GDP), trade openness, foreign direct investment (FDI), tertiary school enrolment (TER) and population (POP). The symbol β0is the constant term, β1is the coefficient of natural resource use, β2is the coefficient of industrialization, γiis a vector of coefficients for the vector of control variables, niis the country fixed effect, τ is the time fixed effect, and ξit is the stochastic error term. Table 1 shows the variables, their descriptions and sources from which the data was obtained. Literature suggests that panel data frequently have the issue of crosssectional dependence. Due to the financial or economic interdependence of nations, a number of unobserved factors and shocks happen simultaneously (De Hoyos &Sarafidis, 2006; Dogan et al., 2017; Latif et al., 2018). Therefore, it is important to determine if shocks affect all crosssectional units equally. In the field of macroeconomic research, contemporary estimation techniques have attracted a lot of interest from around the world. However, in this age of globalization, when changes in other nations have a significant impact on others, it is impossible to continue using conventional techniques. The most recent dynamic Driscoll-Kraay model of Driscoll and Kraay (1998) is used for this purpose. Thus, considering the linear Eq. (2); CO2it =β0+β1NRRit +β2INDUSit +γi χ it +ni+ τ +ξit (2) Stacking all observations as shown by Eqs. (3) and (4) is common: CO2= [CO21t11⋯⋯..CO21T1,CO22t21⋯⋯..CO2NTN,]ʹ(3) NRR = [NRR1t11⋯⋯..NRR1T1,NRR2t21⋯⋯..NRRNTN,]ʹ(4) This formula allows the panel to be unbalanced since for individual i, only a subset ti1, …,Ti with 1 ≤ti ≤Ti ≤T of all T observations may be available. It is assumed that the regressors FIti are uncorrelated with the scalar disturbance term ζis for all s, t (strong exogeneity). However, the disturbances ζti themselves are allowed to be autocorrelated, heteroskedastic and cross-sectionally dependent. Under these presumptions, λi can consistently be estimated using the OLS regression, as shown by Eq. (5).  λi= (NRR*NRRʹ)−1NRRʹ* (5) By relying on cross-sectional averages, standard errors estimated by this approach are consistent and independent of the panel’s cross-sectional dimension. Driscoll and Kraay (1998) showed that this consistency result holds even for the limiting case where N → ∞. Furthermore, estimating the covariance matrix with this approach yields standard errors that are robust to general forms of cross-sectional and temporal dependence. This method can also be applied when dealing with unbalanced panel data (Ditzen, 2016) and data that has structural fractures Table 1 Description and sources of data. Variable Description/measurement Source CO 2 proxy for climate change through CO 2 emissions World Development Indicators (2023) NRR it NRR is natural resource rents as a percentage of GDP African Development Bank (2022) INDUS it the log of manufacturing value added in USD World Development Indicators (2023) AIDI Aggregate infrastructural development index African Infrastructure Database, 2022 GDP GDP in USD African Development Bank (2022) OPEN trade openness, measured as African Development Bank (2022) FDI foreign direct investments in USD African Development Bank (2022) TER tertiary school enrolment rate in percentage World Development Indicators (2023) POP Population World Development Indicators (2023) F. Hodu Ngangnchi et al. Research in Globalization 9 (2024) 100245 4 (Kapetanios et al., 2011). 4. Results and discussion 4.1. Descriptive statistics The main objective of this study was to investigate into the effects of natural resource extraction and industrialization on climate change in Africa. The summary statistics of variables (Table 2) show that the average level of CO 2 emissions that cause climate change in Africa stands at 7.9 percent, natural resource extraction has an average value of 10.87 percent, industrialization as proxy by the log of manufacturing value added is 22.39 percent, and the aggregate infrastructural development index stands at 22.035 percent in Africa. We note that the mean of control variables such as foreign direct investments, GDP, population, tertiary school enrolment, and trade openness are 19.45 percent, 22.71 percent, 15.83 percent, 14.32 percent, and −0.0567 percentage points respectively. 4.2. Pairwise correlations Econometric principles suggest that, collinearity issues pose a fundamental problem in regression analysis. To this respect, we assertain the quantitative relationship between the explanatory variables with the aid of the pairwise correlation matrix, as reported on Table 3. It can be observed from the pairwise correlation matrix (Table 3) that all correlation coeffeicients among the explanaory variables are less than 0.75. This indicates that there are no problems of collinearity among the independent variables captured in the model. Therefore, the variables could be included in regression models separately as different explanatory variables that have the ability to explain the changes in climate through CO 2 emissions in Africa. 4.3. Unit root tests Econometric principles indicate that non-stationary data constitutes a fundamental problem in regression analysis. We test for stationary of data in this study via the CIPS Second-Generation test which has the ability to control for the problem of cross-sectional dependence in panel data. The test results reported on Table 4 show that CO 2 emissions, natural resource extraction, industrialization, population growth and trade openness are stationary at levels. At the same time, aggregate infrastructural development index, foreign direct investment, economic growth, and tertiary school enrolment are stationary after first difference. As a result, we conclude that all panels are stationary. The fitted scatter plots presented on Fig. 1 show that the expected relationship between natural resource extraction, industrialisation, foreign direct investment, economic growth, and population growth with CO 2 emission that cause climate change are apparently positive and linear. Furthermore, we observe that the expected relationship between tertiary school enrolement rate, trade openness, aggregate infrastructural development, with CO 2 emission that cause climate change are apparently constant. In this respect, the following section presents a suitable regression methodology in tackling the quantitative extent through the linear models specified. 4.4. Econometric estimates We note from Table 5 that the Hausman test statistics is significant at the 1 percent level. This invalidates the reliability of the random effect model in favour of the fixed effects model results for inference. Furthermore, tests such as the Breusch-Pagan LM test of independence and the Modified Wald test for groupwise heteroscedasticity with significant tests statistics at 1 percent levels confirms the presence of heteroskedasticity, thereby invalidating the fixed effects model for inference. In this study, we adopt the Pesaran (2004) and Pesaran (2015) cross sectional dependence test to check for the econometric problem of crosssectional dependence within the panel. Chudik et al. (2013) explained that the presence of the problem of cross-sectional dependence is due to the interaction among the countries and other unobserved factors. Therefore, failure to address the problem of cross-sectional dependence (CD) produces biased and inconsistent estimates. From Table 5, the CD test strongly rejects the null hypothesis of no cross-sectional dependence in our regression model. Although it is not the case here, a possible drawback of the CD test is that adding up positive and negative correlations may result in failing to reject the null hypothesis even if there is evidence of cross-sectional dependence in the errors. The average absolute correlation is 0.243, which is a very high value. Hence, there is enough evidence suggesting the presence of crosssectional dependence under a cross-sectional fixed effect specification. We therefore properly estimate reliable parameters based on the Feasible Generalized Least Squares (FGLS) and the Drisco-Kraay standard errors estimators (Table 6) that have the ability to eliminate such problems in econometric analysis. The results are validated by the adjusted R squared of 0.978, which suggests that about 98 percent of changes in climate from environmental degradation is due to joint variations in natural resource extraction, industrialization, infrastructural development and other control variables used in this study. The results are further validated by the F-statistics of 6430.06 with p-value of 0.0000 indicating that the DriscoKraay model is significant at the 1 percent level and thus, the coefficients estimated are at least, 99 percent reliable for policy inference. The regression coefficients suggest that previous year’s CO 2 emission has a 1 percent significant and positive effect on the current year CO 2 emission in Africa. This is possibly in the absence of environmental cleaning which is absent in Africa. Furthermore, a percentage increase in CO 2 emission in the current year increases the current year level CO 2 level in year 2 by at least 0.630 percentage points. In this respect, it is necessary to keep the CO 2 emission level low, as an increase in a given period will affect concentration and implications in the next period positively. Thus, CO 2 emissions must be regulated given its far reaching consequences on the environment and human life. The findings across the different models (Cross-Sectional FGLS, Table 2 Summary descriptive statistics. VARIABLE Obs. Mean Std. Dev. Min Max CO 2 810 7.90043 0.2274886 7.536691 8.273545 NRR 805 10.87095 11.09167 −1.19976 67.9176 LINDUS 773 20.39026 3.122002 3.800957 25.15552 AIDI 810 22.03491 18.34845 1.12 98.88 LFDI 769 19.45464 1.805362 10.36072 23.02867 LGDP 810 22.71743 3.140711 4.438643 27.03318 LGDP 2 810 525.9335 104.7409 19.70156 730.793 LPOP 792 15.83075 1.617791 11.32488 19.16262 TER 810 14.32596 3.072544 8.673049 19.52035 LTOPEN 784 −0.0567314 2.339061 −1.573949 15.46033 F. Hodu Ngangnchi et al. Research in Globalization 9 (2024) 100245 5 Driscoll-Kraay, and Dynamic Driscoll-Kraay) have established that natural resource extraction as proxy by total natural resource rent, industrialization as proxy by manufacturing value added and the aggregate infrastructural development index all have positive effects on climate change, captured by CO 2 emissions in Africa. The quantitative results indicate that a percentage increase in natural resource extraction, industrialization, and infrastructural development all have the ability to increase CO 2 emission (climate change) by 0.00813, 0.157 and 0.0129 percentage points respectively. These coefficients are significant at 5 percent, 10 percent and 1 percent respectively. These findings are consistent with Kwakwa et al. (2020), who noted that natural resources in Africa have been depleted due to abusive use and poor management leading to climatic problems. Furthermore, Alola and Adebayo (2023a) showed that the consumption of domestic metallic ores aggravates environmental degradation by increasing GHG emissions while fossil fuels domestic material consumption and biomass reduce GHG emissions in Iceland. Adebayo and Kirikkaleli (2021) however argues that renewable energy consumption prevents the growth of CO 2 emissions in both the short and medium run while Adebayo et al. (2023), confirms that it worsens ecological footprint. Ahmed et al. (2022) and Aladejare and Nyiputen (2023) also found consistent results as rooted in the greenhouse theory of climate change, the Wobble, roll and stretch theory. These studies hold that though there are natural factors that contribute to climate change, there is also the possibility to restore the climate system to equilibrium through proper and timely intervention. Table 3 Pairwise correlation matrix. NRR LMVA AIDI LFDI LGDP LPOP SETT LTOPEN NRR 1.0000 LMVA 0.1248 1.0000 AIDI −0.0753 0.2012 1.0000 LFDI 0.1237 0.2945 0.1204 1.0000 LGDP 0.1652 0.6764 0.1523 0.2880 1.0000 LPOP 0.1091 0.6750 −0.3203 0.5345 0.5889 1.0000 TER −0.1756 0.0867 0.2349 0.1402 0.0827 0.0754 1.0000 LTOPEN −0.0526 −0.6453 −0.0274 0.0946 −0.6891 −0.5311 −0.0179 1.0000 Table 4 CIPS second-generation panel unit root test. VARIABLE Levels First Difference Order of Integration Statistics P-value Statistics P-value CO2 −3.387 0.000 −-−- I(0) NRR −5.311 0.000 −-−- I(0) LMVA −6.297 0.000 −-−- I(0) AIDI −0.281 0.390 −10.518 0.000 I(1) LFDI −0.409 0.134 −9.405 0.000 I(1) LGDP −2.250 0.341 −7.345 0.000 I(1) LGDP2 −0.271 0.393 −7.450 0.000 I(1) LPOP −8.113 0.000 −-−- I(0) TER −0.220 0.410 −9.261 0.000 I(1) OPEN −2.597 0.005 −-−- I(0) Fig. 1. Fitted scatter plots. F. Hodu Ngangnchi et al. Research in Globalization 9 (2024) 100245 6 Alola and Adebayo (2023b) further revealed that efficiency in the use of natural resources mitigates carbon emissions. The findings further indicate that foreign direct investments and tertiary school enrolment used as a measure of technological input have negative implications on CO 2 emission in Africa. This indicates that a percentage increase in foreign direct investments and technological inputs reduces CO 2 emission by 0.124 and 0.0273 (significant at 1 percent) percentage points, hence a reduction in climate change issues in Africa. While the coefficient of foreign direct investment is consistently negative across all models estimated, the coefficient of technological inputs became positive and significant in the more robust Dynamic DriscollKraay model. This is due to the weak level of technology in Africa. This is in line with Njimanted et al. (2023) who noted that Africa needs to finance education and develop skills through increasing annual budgetary allocations to education to a minimum of 30 % of the state budget. The effect of FDI is in contrast to the findings of Seker et al. (2015), yet consistent with Demena and Afesorgbor (2020). The estimates show that the openness of the economy to foreign trade and population growth rate has positive and significant effects on the volume of CO 2 emission in Africa. As such, a percentage increase in the openness of the economy to foreign trade and population growth rate will increase the volume of CO 2 emission by 0.506 and 0.205 percentage points respectively, leading to climate change complications in the continent. Both coefficients are significant at the 10 percent level. We note that in the more robust Dynamic Driscoll-Kraay estimates, trade openness is negative but insignificant on the volume of CO 2 emission, in line with Al-Mulali et al. (2015), Yu et al. (2019), Ahmed et al. (2017) who had positive impact of trade openness on the environment for both low-, middleand high-income countries. The result indicates that all the coefficients of GDP are positive in all models estimated. This is consistent with the works of Alola and Adebayo (2023a), Adebayo and Kirikkaleli (2021), Alola and Adebayo (2023b) and Adebayo et al. (2023) which found that economic growth deteriorates the environment. However, the coefficients of GDP 2 across different regressions are negative, showing that the environmental U sharped Kuznets curve hypothesis is verified for Africa. This is in line with Grossman and Krueger (1991) that the first or initial stages of economic growth are associated with increases in environmental degradation up to a certain peak and then starts decreasing, and this relation depicts an inverted U-shaped curve. The inverted U-shaped stems principally from the scale, technique, and composition effect. The scale effect shows how economic growth can increase pollution through the scaling up of economic activities within economies. We carried out further evidence of the findings through Panel Quantile on Quantile regression methodology which allows us to understand the linkages and complexities between the variables outside of the mean of the dataset. These results are shown on Table 7. It can be noted from the quantile regression coefficients (Table 7) that outside the mean of the variables, the impact of natural resource extraction on climate change becomes negative. This result suggests that natural resource extraction initially increases CO 2 emission in line with the EKC hypothesis up to a certain level beyond which emission reduces. This is in line with Kwakwa (2021) who examined the effects of natural resource extraction and renewable energy consumption on carbon dioxide emissions in Sub-Saharan Africa. The results for industrialization Table 5 Baseline estimates based on pooled, fixed effects and random effects estimations. (Pooled OLS) (Fixed Effect) (Random Effect) VARIABLES CO 2 CO 2 CO 2 Natural resources rents 0.000103 −0.00207 −0.00297 (0.00708) (0.00182) (0.00181) Log of manufacturing 0.377*** −0.118*** −0.105*** (0.137) (0.0391) (0.0391) Aggregate infrastructure development index 0.0145** −0.00849*** −0.00895*** (0.00698) (0.00206) (0.00203) Log of foreign direct investment inflows −0.124* −0.00489 −0.00410 (0.0667) (0.00828) (0.00832) Log of population 0.205* 0.865*** 0.479*** (0.124) (0.180) (0.133) School enrolment, tertiary −0.0273 −0.0244** −0.00953 (0.0327) (0.00967) (0.00814) Log of trade openness 0.506** 0.307*** 0.319*** (0.239) (0.0427) (0.0427) Log of gross domestic product 0.924 −0.212 −0.185 (1.340) (0.431) (0.431) Log of gross domestic product squared −0.0280 0.00493 0.00432 (0.0280) (0.00926) (0.00927) Constant −16.40 −9.978* −4.626 (15.30) (5.716) (5.432) Observations 688 688 688 R-squared 0.979 0.982 Number of id 45 45 45 Hausman chi2(9) =133.09; Prob >chi2 =0.0000Breusch-Pagan LM test of independence: chi2 (820) =2385.138, Pr =0.0000 Modified Wald test for groupwise heteroskedasticity chi2 (41) =26.95 Prob >chi2 =0.9554 Pesaran’s test of cross-sectional independence = − 1.006, Pr =1.6771 Average absolute value of the off-diagonal elements =0.243 Standard errors in parentheses. *** p <0.01, ** p <0.05, * p <0.1. Table 6 Feasible generalized least squares, Driscoll-Kraay and dynamic Driscoll-Kraay estimates. (CrossSectional FGLS) (Driscoll-Kraay) (Dynamic Driscoll-Kraay) VARIABLES CO 2 CO 2 CO 2 Lag of CO 2 emissions 0.630*** (0.153) Natural resources rents 0.000103 0.000103 0.00813** (0.00701) (0.00225) (0.00376) Log of manufacturing 0.377*** 0.377*** 0.157* (0.136) (0.0477) (0.0914) Aggregate infrastructure development index 0.0145** 0.0145*** 0.0129*** (0.00691) (0.000898) (0.00376) Log of foreign direct investment inflows −0.124* −0.124 −0.0913** (0.0661) (0.0758) (0.0439) Log of population 0.205* 0.205* 0.155*** (0.123) (0.110) (0.0328) School enrolment, tertiary −0.0273 −0.0273*** 3.518** (0.0324) (0.00643) (1.493) Log of trade openness 0.506** 0.506* −0.0148 (0.237) (0.286) (0.223) Log of gross domestic product 0.924 0.924** 2.156* (1.328) (0.364) (1.251) Log of gross domestic product squared −0.0280 −0.0280*** −0.0436* (0.0278) (0.00503) (0.0248) Constant −16.40 −16.40*** −45.76** (15.16) (3.168) (19.18) Observations 688 688 654 R-squared 0.979 0.987 Number of groups 45 45 45 Wald chi2(9) =31332.79 Prob >chi2 =0.0000 F(9, 40) =877.98 Prob >F=0.0000 R-squared = 0.9785 F(10, 40) = 6430.06 Prob >F=0.0000 R-squared = 0.9866 Standard errors in parentheses. *** p <0.01, ** p <0.05, * p <0.1. F. Hodu Ngangnchi et al. Research in Globalization 9 (2024) 100245 7 reveal that industrialization continuously increases CO 2 emission, leading to worsening climate change. This is so because even out of the mean of industrialization, the effect on CO 2 emission as seen across the different quantile remains positive due to accumulation effects resulting from lack of environmental cleaning. This is consistent with what we observed in the Cross-Sectional FGLS, Driscoll-Kraay and Dynamic Driscoll-Kraay models estimated. The aggregate infrastructural development index that was positive in the Cross-Sectional FGLS, Driscoll-Kraay and Dynamic Driscoll-Kraay models estimated become negative out of the mean of infrastructural development and CO 2 emission at the 25 % quantile and the 50 % quantile. As such, we observed a turning point for the effect of infrastructural development on CO 2 emission in Africa at the 25 % quantile and the 50 % quantile, but this turning point is nullified at the 75 % and at the 90 % quantiles out of the mean of the variables. Though studies suggest that building of road infrastructure generates CO 2 emissions (Liu et al., 2015) as cited by Xu et al. (2022), other studies (Kumar et al., 2019) indicate that CO 2 emissions vary inversely with highway mileage and mobility of traffic. The development of road infrastructure also cuts down atmospheric gas emissions through agglomeration which save energy (Shao et al., 2019 as cited by Xu et al., 2022). The study observed that, even out of the mean of foreign direct investment and CO 2 emissions, the relation foreign direct investment and climate change through CO 2 emission remains negative, in line with Demena and Afesorgbor (2020). We observe population growth rate that was positive in within the period of the data becomes negative out of the mean of population growth rate and climate change through CO 2 emission in Africa at the 25 % and 50 % quantiles (consistent with the results of Aye &Edoja, 2017). Nonetheless, this negative effect of population growth rate on climate change through CO 2 emissions is nullified at the 75 % and 90 % quantiles. The study finds that out of the mean of the data, the effect of tertiary school enrolment is negative while that trade openness is positive on climate change through CO 2 emission at the 25 %, 50 %, 75 % and 90 % quantiles. Again, out of the mean of the dataset, economic growth becomes negative on climate change, while the quadratic term of economic growth becomes positive on climate change. This is a contradiction to the Environmental Kuznets Curve (EKC) hypothesis explained by Grossman and Krueger (1991) but in line with Miti´ c et al. (2023). 5. Conclusion and policy implications This study explores the effect of natural resource use and industrialization on climate change in African using data from 2005 to 2022 for 45 African countries. Using contemporary methods and procedures, the study found that the future of climate change in Africa is directly related to the previous climatic conditions in terms of CO 2 emissions. It was established that natural resource extraction, industrialization and infrastructural development all have positive effects on climate change in Africa. However, the study found that outside the mean of the variables, the effect of natural resource extraction on climate change becomes negative. Furthermore, the effect of infrastructural development that was positive becomes negative out of the mean on climate change at both the 25 % quantile and the 50 % quantile, but this turning point becomes nullified at the 75 % and the 90 % quantiles out of the mean of the variables. The results of this study imply that sustainable regional development will remain a far-fetched objective for the African continent. First, the African continent greatly relies on the exploitation of its natural resource potentials as a main source of revenue mobilization which should propel growth. The rapid extraction of these resources consequently results in their degradation, with attendant consequences for the environment, thereby jeopardizing sustainable development. Moreover, industrialization accelerates depletion of natural resources and fosters the emission of GHGs in to the atmosphere, further worsening the environmental problems. With both the extraction of natural resources and industrial development in Africa worsening climate change, it is obvious that sustainable development will be greatly hampered, as sustainability requires that the economy should grow while not seriously destroying the environment. The quantitative findings of this study yield some useful policy guides for African government in their drive towards industrialisation in the face of global warming; First, following a green path to natural resource extraction and industrialization through resource-efficient industrial productivity, the use of sustainable materials in production, using production resources in an eco-friendly way, the use of green technologies that reduces energy consumption and CO 2 emission, and the adoption of a pragmatic approach to environmental governance is necessary as the world battles the ravaging implications of global warming. Second, there is also the necessity for strong environmental policies to regulate the emission and more importantly, the cleaning of the environment for sustainable management of the consequences of climate change. Lastly, proper environmental cost-benefit analysis is therefore necessary to guide natural resource extraction, industrialization and infrastructural development in Africa so as to mitigate the impact of such investment opportunities on the global environment. Future studies could investigate the disaggregated effects of natural resources and manufacturing activities on more comprehensive indicators of environmental degradation. CRediT authorship contribution statement Forbe Hodu Ngangnchi: Writing –original draft, Methodology, Formal analysis, Data curation, Conceptualization. Nkwetta Ajong Aquilas: Writing –review &editing, Visualization, Validation, Methodology, Formal analysis, Data curation. Mukete Emmanuel Mbella: Table 7 Quantile regression estimates. (25 % Quantile) (50 % Quantile) (75 % Quantile) (90 % Quantile) VARIABLES CO 2 CO 2 CO 2 CO 2 Natural resources rents −0.00391* −0.000425 −0.00142 −0.00523 (0.00217) (0.00207) (0.00331) (0.00725) Log of manufacturing 0.158*** 0.153*** 0.283*** 0.289** (0.0420) (0.0401) (0.0640) (0.140) Aggregate infrastructure development index −0.00432** −0.00182 0.00748** 0.0103 (0.00214) (0.00204) (0.00326) (0.00715) Log of foreign direct investment inflows −0.00426 −0.00213 −0.0908*** −0.0961 (0.0204) (0.0195) (0.0311) (0.0683) Log of population −0.165*** −0.112*** 0.0819 0.197 (0.0380) (0.0363) (0.0579) (0.127) School enrolment, tertiary −0.0141 −0.0187* −0.0357** −0.0606* (0.0100) (0.00958) (0.0153) (0.0335) Log of trade openness 0.687*** 0.640*** 0.921*** 0.984*** (0.0733) (0.0700) (0.112) (0.245) Log of gross domestic product −2.355*** −2.275*** −2.720*** −0.740 (0.411) (0.392) (0.626) (1.373) Log of gross domestic product squared 0.0516*** 0.0485*** 0.0545*** 0.0112 (0.00859) (0.00820) (0.0131) (0.0287) Constant 24.55*** 23.72*** 27.30*** 3.544 (4.689) (4.480) (7.148) (15.68) Observations 688 688 688 688 Standard errors in parentheses. *** p <0.01, ** p <0.05, * p <0.1. F. Hodu Ngangnchi et al. Research in Globalization 9 (2024) 100245 8