Energy consumption, technological innovation, and environmental degradation in SADC countries

Abstract

This study investigates the relationship between energy consumption, technological innovation, and environmental degradation in Southern African Development Community (SADC) countries. With a focus on understanding the driving forces behind environmental degradation amidst rapid economic growth and industrialization, the research aims to provide crucial insights for informed policymaking and strategic interventions to promote sustainable development. Employing the Feasible Generalized Least Squares (FGLS) method to address panel data challenges including heteroscedasticity, serial correlation, and cross-sectional dependence, the analysis is conducted using data spanning from 1994 to 2021 for the SADC region. Findings reveal a direct, positive correlation between energy consumption and environmental degradation, particularly evident in increased carbon emissions. Additionally, the study finds that certain aspects of technological innovation, such as Internet usage and mobile phone subscriptions, mitigate the positive impact of energy consumption on carbon emissions, aligning with previous research emphasising the role of ICT in improving energy efficiency and resource management. Insights from this research provide valuable guidance for policymakers in steering SADC countries towards sustainable development trajectories. By leveraging technological innovation to optimize energy consumption and reduce carbon emissions, policymakers can effectively address environmental concerns while promoting economic growth and technological advancement in the region.

Keywords:

• Energy consumption
• technological innovation
• environmental degradation
• FGLS
• SADC

REVIEWING EDITOR:

• George Mudimu, Development Sciences, Marondera University of Agricultural Sciences and Technology, Marondera, Zimbabwe

SUBJECTS:

• Economics
• Macroeconomics
• Econometrics
• Microeconomics
• Economic Theory & Philosophy
• Environmental Economics
• Power & Energy

1. Introduction and background

Energy consumption, technological innovation, and environmental degradation are integral components of the complex relationship between human activities and the natural environment (Ulucak et al., 2020; Mughal et al., 2022). This nexus holds particular significance within the Southern African Development Community (SADC) countries, where diverse economies and ecosystems intersect with the challenges of meeting growing energy demands, fostering technological progress, and mitigating environmental harm (Adenle, 2020). SADC countries are lagging behind the global trend in technology adoption, with a significant dependence on traditional energy sources compared to their counterparts worldwide, who are at the forefront of embracing cutting-edge technologies (Bwalya & Healy, 2010). The environmental repercussions of these diverse approaches underscore the urgency of understanding a region’s sustainability dynamics.

Recent global initiatives, epitomized by the 28th Conference of the Parties (COP28), echo the urgency of addressing environmental challenges, particularly in rapidly industrialising regions like the Southern African Development Community (SADC) countries (Chopra et al., 2024; Arora & Mishra, 2024; Jiang et al., 2024). COP28, convened to tackle pressing environmental issues, aligns with the goals of achieving net-zero emissions by 2050 outlined in the United Nations Framework Convention on Climate Change and the Paris Agreement, which pledges to leave no one behind in the 2030 Agenda for Sustainable Development (Browning et al., 2023). Acknowledging the challenges posed by environmental degradation, there are calls for increased financial resources, institutional capacity, and technical support to be immediately scaled up, particularly in Least Developed and developing countries, including SADC (Khan et al., 2021). Technological innovation and accessibility to renewable energy technology emerge as pivotal factors recognized at the COP28 summit for mitigating emissions (Adenle, 2020). These outcomes underscore the imperative for policymakers to grasp the relationship between energy consumption, technological innovation, and environmental degradation, as emphasized in this study. By embedding our research within the broader context of international efforts to combat climate change, we endeavour to furnish insightful contributions that can inform policymaking and propel sustainable development agendas forward.

Energy consumption is an essential factor for facilitating a nation’s economic performance. Energy is consumed in different forms; one of the general forms of the Southern African Development Community (SADC) is electricity. Since electricity in SADC is generated from natural resources such as coal, which supplies approximately 62% of power, it is considered a significant source of environmental degradation (Madakufamba et al., 2017). In other words, the higher the coal consumption in electricity production, the greater the degradation of the environment. The top coal consumers in the SADC region between 1980 and 2021 are as follows: South Africa is seating at 195428.58 thousand tons, Zimbabwe at 2255.07, Botswana at 1538.97, Zambia at 1209.88, Mauritius at 811.97, Tanzania at 618.3, Madagascar at 471.95, Swaziland at 243.57, Malawi at 71.6, Mozambique at 59.77, Namibia at 45.02, and the DR Congo at 15.22 (The Global Economy, 2023).

According to Usman et al. (2022), the relationship between energy usage and environmental degradation is currently a top concern for investigation. Environmental deterioration continues to rise because coal is an energy source that accounts for a significant portion of total energy production, resulting in a substantial amount of carbon dioxide (CO2) emissions. Similarly, an increase in energy consumption leads to high degradation of the environment through CO2 emissions (Depren et al., 2022). Residential energy consumption and emissions have increased considerably in recent decades owing to rapid economic expansion and urbanization. The urban population is expected to reach 66% of the world’s population by 2050, resulting in an increase in energy consumption (Geng et al., 2017). This means that energy supply must be expanded to match the rising energy demand, resulting in significant CO2 emissions that degrade the environment. According to Xie et al. (2022) and Adebayo and Rjoub (2022), recent data shows that most countries cause CO2 emissions through energy consumption.

To mitigate environmental degradation, countries should consider technological innovation that can neutralize the impact of energy consumption on environmental degradation. Technological innovation is a crucial tool for encouraging integrated development of both the environment and economy (Liu et al., 2018). It is necessary to implement an energy strategy that applies technological innovation to attain energy security, while protecting the environment from pollution (Alam & Murad, 2020). Energy produced from technological innovations includes renewable energy and is considered the best alternative to fossil energy sources, such as coal, which are the largest contributors to CO2 emissions (Ganda, 2019; Ulucak et al., 2020; Altinoz & Dogan, 2021; Adebayo & Kirikkaleli, 2021; Destek & Aslan, 2020, Adebayo et al., 2023). In other words, one of the types of energy from technologies that plays a crucial role in promoting economic growth and reducing pollution, as well as societal improvement, is green energy (Ulucak et al., 2020).

This study posits that it is essential for the global community to comprehend the relationship among energy consumption, technological advancement, and environmental degradation. Most economies focus on enhancing their economic growth without observing the effects of energy use on the environment. This study also contributes to the body of knowledge on the relationship between energy consumption, technological innovation, and environmental degradation in several ways. First, it extends empirical literature to the current period. Second, the existing literature shows that many studies on energy consumption, technological innovation, and environmental degradation were done in nations such as Japan, South Asia, South Korea, Spain, BRICS, Belt and Road countries (Adebayo & Kirikkaleli, 2021; Adebayo et al., 2021; Mughal et al., 2022; Zang et al., 2023; Khan et al., 2023).

Previous studies (Udeagha & Ngepah, 2022; Kivyiro, 2023; Sibanda et al., 2023) have in the SADC region focused on the direct effects of energy consumption on environmental degradation. Udeagha and Ngepah (2022) focus only on the impact of technological innovation on environmental quality in South Africa. The Quantile autoregressive distributive lag (QARDL) indicates that technological innovation directly reduces carbon emissions while energy consumption degrades enviromental quality in South Africa between 1960 and 2020.

Again, the pooled mean group (PMG) analysis of environmental degradation effects of human capital in SADC countries between 1980 and 2020 by Sibanda et al. (2023) indicates that even if human capital improves environmental quality, energy consumption exacerbates environmental degradation in SADC countries. The causal link between urbanization, energy use, and carbon emissions in the SADC region by Kivyiro (2023) indicates a one-way Granger causality from energy consumption to carbon emission from to 1988-2020. The current study contributes to the literature by assessing the direct and indirect impacts of energy consumption on environmental degradation in the SADC region. Thus, we consider the effectiveness of technological innovation in the environmental degradation-energy consumption nexus. The rationale behind the selection of technological innovation is that it helps produce clean energy, boost economic growth without damaging the environment, and save natural resources, such as coal (Xie et al., 2022; Udeagha & Ngepah, 2022).

Current research stems from the pressing need to address the challenges posed by the intersection of energy consumption, technological innovation, and environmental degradation. As the global community grapples with the consequences of climate change and strives to achieve Sustainable Development Goals (SDGs), understanding regional contexts becomes paramount (Nizami et al., 2023). This study seeks to contribute empirical insights and analytical depth to the existing body of knowledge, providing an understanding of the relationships between energy, technology, and the environment in SADC nations. By doing so, this study aims to offer evidence-based recommendations for policymakers, businesses, and civil society to foster sustainable practices and guide the region toward a more resilient and ecologically balanced future. This study’s main contribution lies in its comprehensive examination of the factors influencing the environmental impact of energy consumption and technological innovation in SADC countries, thereby informing targeted interventions and policies.

The degradation of the environment has been harmful and affects all nations (Kartal et al., 2023), and this study strives to find a way to neutralise the impact of one of the sources of degradation, which is the consumption of energy. Hence, this paper examines the long-term equilibrium relationship between energy consumption, technological innovation, and environmental degradation in SADC countries, examining short-term dynamics and the mediation role of technological innovation. This paper uses FGLS techniques to effectively achieve its objective and rectify limitations and challenges in fundamental panel data techniques (Ikpesu et al., 2019). The reliability of panel data analyses is significantly impacted by concerns such as heteroscedasticity and serial correlation (Kumawat & Patel, 2022). The current paper utilizes FGLS to provide accurate and unbiased estimations, enhancing the quality and trustworthiness of empirical findings by effectively controlling for issues. This paper chose this method due to its explicit accounting for heteroskedasticity, cross-sectional and serial correlations in estimation (Bai et al., 2021). In other words, FGLS offers potential efficiency advantages in the presence of heteroscedasticity, without knowing its functional form (Miller & Startz, 2019).

2. Theoretical literature review

The literature review explores three key theories regarding the relationship between energy prices and economic growth. The Environmental Kuznets Curve theory (EKC) posits that initial economic growth may exacerbate environmental degradation but suggests that cleaner production technologies can mitigate this effect, promoting economic growth (Egli & Steger, 2005; Kaika & Zervas, 2013). EKC theory suggests that firms must adopt cleaner production technologies to reduce environmental degradation and promote economic growth (Khan et al., 2023). The second theory is the innovation diffusion theory developed by Rogers (1962). Innovation Diffusion Theory underscores the importance of adopting energy-efficient and clean technologies, emphasizing the stages of knowledge acquisition, decision-making, and implementation (Peansupap & Walker, 2006; Sahin, 2006). Finally, Energy Transition Theories discuss the global shift from fossil-based to renewable energy sources and highlight the role of technology in environmental problem solving (Geerts, 2017, 2018). This transition theory provides insights into the factors influencing the adoption of cleaner energy sources in regions, such as the Southern African Development Community (SADC) (Sovacool et al., 2021).

2.1. Empirical review

There is little empirical evidence on energy consumption, technological innovation, and environmental degradation. Numerous studies have focused on the relationship between the energy consumption and environmental degradation. For example, Depren et al. (2022) examined the nexus between energy consumption and environmental degradation on a global scale using bibliometric analysis from 2000 to 2008. Their findings indicated that disaggregated energy consumption worsens environmental degradation. These results align with the findings of Sharif et al. (2020), who found a negative connection between energy consumption and environmental degradation in China, the USA, Japan, Canada, Brazil, South Korea, and Germany between 1990 and 2017. However, the results contrast with the cases of India, Russia, and Indonesia. Additionally, the results of Granger causality in quantiles show a bidirectional causal relationship between renewable energy consumption and environmental degradation.

Similarly, the study conducted by Altinoz and Dogan (2021) examined renewable energy consumption and environmental degradation using a panel of 82 countries and quantile regressions between 1990 and 2014. The study found that renewable energy consumption reduced CO2 emissions, and its effect increased in higher quantiles. On the other hand, Destek and Aslan (2020) assessed the nexus between renewable energy consumption and environmental pollution in G-7 countries between 1991 and 2014 using the augmented mean group estimator and panel bootstrap causality method. Their results suggested that increasing biomass energy consumption was efficient in reducing carbon emissions in France, Germany, Japan, and the United States, while increasing hydroelectricity usage was efficient in reducing carbon emissions in Italy and the United Kingdom, wind energy consumption reduced emissions in Canada, and solar energy usage was efficient in reducing emissions in France and Italy.

In the context of the relationship between energy consumption and technological innovation, Adebayo and Kirikkaleli (2021) employed Wavelet analysis to investigate the relationship between renewable energy consumption, technical innovation, and CO2 emissions in Japan between 1990 and 2015. Based on these data, renewable energy consumption reduces CO2 emissions, but technological innovation contributes to CO2 emissions. Furthermore, the findings suggest that using renewable energy reduces CO2 emissions in the medium- and long-term. However, the autoregressive distributed lag (ARDL) bounds testing method and the dynamic ordinary least squares (DOLS) method during 1978 and 2018 by Li and Solaymani (2021) found contradictory results. Their study posits that technological innovation that increases energy efficiency can only reduce energy consumption in the industrial sector, thereby lowering emissions. Mughal et al. (2022) used annual time-series data from 1990 to 2019 to assess the link between technological innovation, environmental pollution, energy consumption, and sustainable economic growth in selected South Asian economies. The fully modified ordinary least squares (FMOLS) method shows that energy consumption increases CO2 emissions, resulting in environmental degradation.

The studies that have used panel data include Ganda (2019), who examines the link between financial development, natural resource rents, technological innovation, foreign direct investment, energy consumption, human capital, and trade on environmental degradation from 1990 to 2019 in the new BRICS economies. The panel data generalized least squares (GLS) and panel-corrected standard error (PCSE) methods in Ganda indicate that energy consumption has a positive connection with environmental degradation and is also favorable for technological innovation. However, Chen and Lei (2018) showed that increased energy consumption eventually led to higher CO2 emissions between 1980 and 2014.

The indecisive argument could result from differences in the study period, variables used, methodology, and country in which the study was conducted. For example Altinoz and Dogan (2021), Adebayo and Kirikkaleli (2021) and Destek and Aslan (2020) examine the relationship between the renewable energy and energy consumption, while others such as Li and Solaymani (2021), Mughal et al. (2022) and Ganda (2019), tried to check the nexus between technological innovation and energy consumption. On the other hand, the empirical evidence between energy consumption and pollution is investigated by the studies such as (Destek & Aslan, 2020; Chen & Lei, 2018). Very few studies have focused on energy consumption, technological innovation, or environmental degradation using panel data. Even studies that employed panel data did not use feasible generalized least squares (FGLS). Therefore, this study extends the literature by investigating energy consumption, technological innovation, and environmental degradation from 1994 to 2021 in 16 SADC countries using FGLS to overcome this drawback.

3. Methodology

3.1. Theoretical framework

The current paper adopted the theoretical framework from Zhang et al. (2022) which is based on the impact of the Environmental Kuznets Curve on environmental pollution accounting for the moderating role of technological innovation. It is further demonstrated by Figure 1:

Figure 1. Theoretical framework of EKC.

Source: Zhang et al., 2022.

Figure 1 describes that when there is a shock in one variable, the response of another variable is determined. C02 is a dependent variable, while technological innovation plays a crucial role in mitigating the effect of CO2 on the environment. Therefore, the theoretical framework helps us to examine the energy-technology- environment degradation nexus in this study.

3.2. Model specification

In analysing the relationship between energy consumption, technological innovation, and environmental degradation in 16 SADC countries, the econometric model applied by Sadiq et al. (2022) was modified: (1) $$\mathit{\text{LnCO}}{2}_{\mathit{\text{it}}}=\gamma 0+\gamma 1\mathit{\text{LnNE}}{C}_{\mathit{\text{it}}}+\gamma 2\mathit{\text{LnER}}{T}_{\mathit{\text{it}}}+\gamma 3\mathit{\text{LnGLO}}{B}_{\mathit{\text{it}}}+\gamma 4\mathit{\text{LnE}}{G}_{\mathit{\text{it}}}+\gamma 5\mathit{\text{LnPOP}}{D}_{\mathit{\text{it}}}+{\eta }_{\mathit{\text{it}}}$$(1)

Where CO2 denotes carbon emissions, NEC signifies nuclear energy consumption, ERT stands for environmental technology, GLOB represents globalization, EG denotes economic growth and, POPD indicates population density, and $\gamma 0$ is the intercept term, and the parameters $\gamma 1$ to $\gamma 5$ are the long-run elasticity coefficients, $t$ is the study time, $i$ is the number of countries and $\eta$ is an error term.

To account for the mediating effect of technological innovation on the environmental degradation-energy consumption nexus, this study introduces an interaction term between them. The estimated regression model for this study is presented in Equation 2(2) $${\mathit{\text{LC}}02\mathit{\text{PC}}}_{\mathit{\text{it}}}={\beta }_0+{\beta }_1{\mathit{\text{LAPEC}}}_{\mathit{\text{it}}}+{\beta }_2{\mathit{\text{TI}}}_{\mathit{\text{it}}}+{\beta }_3{\mathit{\text{LAPEC}}\mathrm{*}\mathit{\text{LTI}}}_{\mathit{\text{it}}}+{\beta }_4{X}_{\mathit{\text{it}}}+w_i+{\upsilon }_t+{\epsilon }_{\mathit{\text{it}}}$$(2) , as follows: (2) $${\mathit{\text{LC}}02\mathit{\text{PC}}}_{\mathit{\text{it}}}={\beta }_0+{\beta }_1{\mathit{\text{LAPEC}}}_{\mathit{\text{it}}}+{\beta }_2{\mathit{\text{TI}}}_{\mathit{\text{it}}}+{\beta }_3{\mathit{\text{LAPEC}}\mathrm{*}\mathit{\text{LTI}}}_{\mathit{\text{it}}}+{\beta }_4{X}_{\mathit{\text{it}}}+w_i+{\upsilon }_t+{\epsilon }_{\mathit{\text{it}}}$$(2)

Where ${\mathit{\text{LC}}02\mathit{\text{PC}}}_{\mathit{\text{it}}}$ is the dependent variable that measures the environmental degradation. t for country $i$ at time $t$. ${\mathit{\text{LAPEC}}}_{\mathit{\text{it}}}$ is the main regressor, which is energy consumption. ${\mathit{\text{TI}}}_{\mathit{\text{it}}}$ represents technology innovation measured by three indicators: fixed broadband subscription (fbs), Internet usage (itu), which refers to the Internet users per 100 individuals who have used the Internet from any particular location and through any source each year, and the mobile phone usage (mcs) rate, which measures the percentage of the population that has access to a mobile phone, either through ownership or regular use.

${\mathit{\text{LAPEC}}*\mathit{\text{LTI}}}_{\mathit{\text{it}}}$ is the interaction term, the multiplication of energy consumption and technological innovation. Thus, the impact of energy consumption on environmental degradation, conditional on technological innovation, was derived as follows: (3) $$\frac{\partial \mathit{\text{LC}}02\mathit{\text{PC}}}{\partial \mathit{\text{LAPEC}}}={\beta }_1+{\beta }_3\mathit{\text{LTI}}$$(3)

Thus, the overall effect of energy consumption on environmental degradation outcomes is the sum of the direct impact (${\beta }_1$) and indirect effect of technology innovation (${\beta }_3*\mathit{\text{LTI}}$). $X_{\mathit{\text{it}}}$ represents other explanatory variables, including institutional quality (inst), economic growth (rgdp), which measures national wealth, economic globalization (kof), human capital index (hci), access to electricity (acte), urbanization (urbn), foreign direct investment (fdi), corruption perception index (cp), political stability (pls), and civil liberties (cil). All variables were transformed into log form. ${\beta }_0$ is a constant and ${\beta }_1,{\beta }_2,{\beta }_3 \mathit{\text{and}} {\beta }_4$ are the coefficients of the predictor variables to be estimated, which can be the same in each period. ${\omega }_i$ is the unobserved country-specific time-varying effect. ${\upsilon }_t$ is the time trend. ${\epsilon }_{\mathit{\text{it}}}$ is a disturbance term.

3.3. Data sources

Table 1 below outlines that the data for this study was sourced from the Global Economy Database, the World Bank, World Integrated Trade Solutions, and World Penn Tables. To empirically evaluate the model, this study utilized annual panel data spanning 1994 to 2021, encompassing data from 16 SADC countries. This database was chosen because of its credibility and comprehensive coverage of variables for the specified period.

Table 1. Definition and source of variables.

Variable	Definition	Expected A priori	Source
C02pc	Carbon dioxide emissions are those stemming from the burning of fossil fuels and the manufacture of cement. They include carbon dioxide produced during consumption of solid, liquid, and gas fuels and gas flaring.		Global Economy Database (GED)
Apec	Energy consumption refers to primary energy before transformation to other end-use fuels, equal to indigenous production plus imports and stock changes, minus exports and fuels supplied to ships and aircraft engaged in international transport.	positive	GED
TI_Fbs	The broadband subscribers (fbs), which measures technology penetration or access to internet speed– refers to the number of households and organizations with fixed subscriptions per 100 individuals at a rate greater than or equal to 256 bits/second.	negative	GED
TI_itu	Internet usage refers to the number of individuals who have used the Internet (from any location) in the last three months. The Internet can be used via a computer, mobile phone, personal digital assistant, games machine, digital TV, etc.	negative	GED
TI_mcs	The mobile cellular subscription (mcs) refers to all subscriptions paid for public mobile telephone services per 100 individuals.	negative	GED
TI_ict	Is the technology index value generated through PCA computed using the PCA from fixed broadband subscription, internet use, and mobile phone usage data	negative	GED
Inst	Is the institutional index computed using PCA from six indicators of institutional quality, including the rule of law, government effectiveness, control of corruption, regulatory quality, voice accountability, and political stability data—the values of the six components range from −2.5 weak to 2.5 strong.	negative	Author’s calculations of institutional quality composites with data from the GED database
Rgdp	Real Gross domestic product (% annual)	negative	World Bank
Kof	Economic globalization has two dimensions: actual economic flows and restrictions to trade and capital. The sub-index on actual economic flows includes trade, FDI, and portfolio investment data. The sub-index on restrictions takes into account hidden import barriers, mean tariff rates, taxes on international trade (as a share of current revenue), and an index of capital controls.		World Integrated Trade Solutions (WITS)
Hci	Human capital index: based on years of schooling and return to education	negative	World Penn Tables
Acte	Access to electricity is the percentage of the population with access to electricity. Electrification data are collected from industry, national surveys, and international sources.	positive
urbn	Urbanisation: Percentage (%) of people living in urban areas	positive	GED
Fdi	Foreign direct investment, net inflows (% GDP)	positive/negative	International Monetary Fund
Cp	The Corruption Perceptions Index indicates perceptions of public sector corruption, that is, administrative and political corruption.	negative	GED
Pls	The Index of Political Stability and Absence of Violence/Terrorism	negative	GED
Cil	The Civil Liberties Index indicates freedom of expression and belief, associational and organizational rights, rule of law, personal autonomy, and individual rights. The rating ranges from 1 (strong liberties) to 7 (no liberties).	negative	GED

Source: Author’s compilation from STATA.

Energy consumption is expected to have a positive and significant effect on environmental degradation as demonstrated in Table 1, following Chein and Lei (2018) and Ganda (2019), who argue that energy consumption increases carbon emissions, leading to environmental degradation. Moyer and Hughes (2012) and Haini (2021) argue that technological innovation, particularly ICT, improves ecological stability. Thus, Table 1 indicates that technological innovation is expected to have an adverse effect on carbon emissions. The current study also considers economic growth, economic globalization, foreign direct investment, human capital development, and urbanization following Sehrawat and Giri (2014), Lin and Li (2020), and Gule (2021), who document evidence of the significance of these variables in environmental degradation.

Electricity has long been critical in ensuring energy security and environmental stability. However, its impact on environmental degradation is claimed to depend on its cleanliness (Lin & Li, 2020). Thus, Khan et al. (2021) suggest that access to electricity worsens carbon emissions. However, Lin and Li (2020) posited that electricity levels significantly negatively affect carbon emissions when clean energy-based electricity accounts for a larger portion. Thus, this study expects both positive and negative effects of electricity access on SADC countries’ carbon emissions. Karim et al. (2022) posited that strict enforcement of environmental regulations through robust and efficient institutions is imperative for carbon emission control. However, numerous studies (Gani, 2012; Halkos & Tzeremes, 2013; Mawutor et al., 2023) have documented evidence of the negative effect of institutional quality on carbon emissions. Therefore, this study expects institutional quality to have a negative impact on SADC countries’ carbon emissions.

3.4. Estimation techniques

This study uses fixed- and random-effects estimation techniques for baseline results. The Hausman ‘s(1978) test indicates that the fixed effect is an appropriate baseline estimation technique. However, even though these basic estimation techniques control for both time and individual heterogeneity, spherical errors, such as heteroscedasticity serial and cross-sectional dependence, are normally present within fixed and random effect models (Musau et al., 2015). Therefore, this study adopts robust panel estimation techniques to address the violation of the classical model assumption, including heteroscedasticity and serial correlation. The robust estimator consists of feasible generalized least squares (FGLS), panel standard error correction for heteroscedasticity and autocorrelation, and Driscoll-Kraay (SCC), which controls cross-sectional dependencies (Pham & Islam, 2022).

According to Bai et al. (2021), the FGLS is consistent and efficient when the time dimension exceeds cross-sectional dimensions. The PSCE is efficient and effective when T is smaller than N (Zakaria & Bibi, 2019). Again, the Driscoll-Kraay estimator can be used when cross-sectional dependence is present in the data across groups. Cross-sectional dependence tests showed no cross-sectional correlation across SADC countries. The study results are based on the FGLS, given that the time dimension is greater than the cross-sectional dimension. The study presents only the FGLS results for the analysis of the paper (the fixed effects and Driscoll Kraay) coefficients are available on request). Before presenting the primary results, this paper presents the preliminary results, which include descriptive statistics, correlation coefficients, stationarity tests, cross-sectional dependence tests, heteroscedasticity tests, and serial correlation tests. The following table presents descriptive statistics, which show the nature of the data used in the study.

4. Empirical results

4.1. Descriptive statistics of variables

Before estimating the econometric model, summary statistics of the variables must be determined. Table 2 summarizes these findings.

Table 2. Summary statistics of variables.

Variable	Obs	Mean	Std. dev	Min	Max
lC02pc	400	.8679822	.8359811	.0199987	2.903334
lapec	417	1.01184	2.23749	−4.217369	5.653589
lfbs	276	.7328891	1.042271	0	4.35096
litu	444	2.083059	1.620681	0	5.094892
lmcs	447	3.305107	2.016757	0	5.916533
lict	274	.2586813	.9710939	−.9822264	2.72238
linst	352	.0001648	1.368111	−2.438916	2.124998
lrgdp	448	1.570556	1.512713	−3.565815	3.759415
lkof	430	4.556746	.2554543	3.868716	5.138066
lhci	364	1.383679	.2116514	.9595914	1.798919
lacte	383	4.074506	.8967651	1.060026	5.298342
lurbn	432	4.243794	.3892779	3.252646	4.954185
lfdi	416	.3930039	.7074817	−2.699162	3.001204
lcp	288	4.196915	.3520928	3.402307	4.882859
lpls	352	−.0700293	.6736369	−1.766598	1.066198
lcil	432	1.964759	.355857	.8813736	2.644121

Note. C02pc is the carbon emissions per capita; apec-annual primary energy consumption; fbs- fixed broad subscription; itu-internet usage; mcs -mobile phone subscription; ict-the information communication technology index; inst -institutional quality index; rgdp-real gross domestic product; kof-economic globalisation; hci-human capital index; acte-access to electricity; urbn-urbanisation; fdi-foreign direct investment; cp-corruption perception index; pls-political stability; cil-civil liberties.

Source: Author’s compilation from STATA.

The significant difference between the maximum and minimum values of the series shows significant variation in the trends of the variables. The standard deviation is also sufficiently large to explore the variations in the series under consideration.

4.2. Correlation coefficient tests

Table 3 outlines the correlation coefficient outcomes for the variables of interest.

Table 3. Correlation coefficient test results.

	lC02pc	lapec	lfbs	litu	lmcs	lict	linst
lC02pc	1.0000
lapec	−0.0569*	1.0000
lfbs	0.6582*	−0.0590*	1.0000
litu	0.5465*	−0.0713*	0.7072*	1.0000
lmcs	0.3991*	−0.0066	0.5752*	0.9158*	1.0000
lict	0.7158*	−0.1077*	0.8724*	0.9266*	0.7878*	1.0000
linst	0.6619*	0.0129	0.5490*	0.3448*	0.2813*	0.5215*	1.0000
lrgdp	−0.0877*	0.2370*	−0.1015*	−0.1307*	−0.0168	−0.1852*	0.0544*
lkof	0.5626*	0.0554*	0.6080*	0.4143*	0.3816*	0.5326*	0.6942*
lhci	0.2303*	−0.0871*	−0.1442*	0.2449*	0.2761*	0.0857*	0.1898*
lacte	0.7212*	−0.0651*	0.6443*	0.6622*	0.5437*	0.7074*	0.3872*
lurbn	0.6459*	−0.0324	0.4090*	0.4207*	0.3897*	0.4686*	0.3078*
lfdi	0.1486*	0.0705*	0.0100	0.1541*	0.2482*	0.0332	0.0907*
lcp	0.6696*	−0.0251	0.6007*	0.5305*	0.5070*	0.6254*	0.9064*
lpls	0.4744*	0.0814*	0.4736*	0.3064*	0.2540*	0.4045*	0.8541*
lcil	−0.5883*	−0.0762*	−0.4011*	−0.3543*	−0.3214*	−0.4076*	−0.9084*
	lrgdp	lkof	lhci	lacte	lurbn	lfdi	lcp
lrgdp	1.0000
lkof	0.0856*	1.0000
lhci	−0.0589*	0.0590*	1.0000
lacte	−0.1054*	0.3428*	0.0948*	1.0000
lurbn	−0.0884*	0.3370*	0.2170*	0.6510*	1.0000
lfdi	0.1600*	0.1435*	0.2685*	0.0602*	0.2329*	1.0000
lcp	−0.0689*	0.6764*	0.2666*	0.5244*	0.3063*	0.0599*	1.0000
lpls	0.0847*	0.6082*	0.1782*	0.3784*	0.2041*	−0.1058*	0.7678*
lcil	−0.0656*	−0.5402*	−0.1890*	−0.3765*	−0.3270*	−0.1879*	−0.8043*
	lpls	lcil
lpls	1.0000
lcil	−0.7461*	1.0000

Source: Author’s compilation from STATA.

4.3. Stationary tests

The stationary tests employed include the IM Pesaran and ADF. The results are presented in Table 4.

Table 4. Stationary tests test results.

Method	IM Pesaran		ADF
Order of integration	1(0)	1(1)	1(0)	1(1)
Variable	t-stat	t-stat	t-stat	t-stat
lC02pc	1.637	−9.663***	2.198	−14.070***
lapec	−11.078	−13.661***	−15.530	−24.169***
lfbs	−3.507	−5.698***	2.493	−7.481***
litu	8.676	−6.536***	2.014	−7.599***
lmcs	−2.574***	−6.792***	7.561	−6.876***
lict	6.894	−3.589	4.442	−3.681***
linst	0.330	−8.219***	−1.659**	−11.974***
lrgdp	−9.494***	−14.062***	−13.231***	−24.095***
lkof	−2.597***	−10.820***	−0.667	−15.320***
lhci	4.297	−2.303***	−4.376***	−2.807***
lacte	−1.061	12.195***	−7.394***	−21.589***
lurbn	5.937	0.635	−4.460***	0.532*
lfdi	−4.614***	−12.429***	−5.611***	−21.114***
lcp	−1.450*	−7.167***	−0.847	−9.050***
lpls	−3.549***	−9.454***	−3.293***	−16.123***
lcil	−1.990***	−2.610**	−0.729	−13.557**

***p < 0.01, **p < 0.05, *p < 0.1.

Source: Author’s compilation from STATA unit root test results.

4.4. Diagnostic results

The other empirical analysis conducted in this study involved diagnostic tests. The results of the Hausman test, Pesaran cross-sectional dependence test, Friedman cross-sectional dependence test, Wald heteroscedasticity test, Wooldridge serial correlation test, Wooldridge serial correlation test, and Jaque-Bera normality tests are shown in Table 5.

Table 5. Diagnostic results.

Hausman test results
	fbs	itu	mcs	ict	apecfbs	apecitu	apecmcs	apecict
Chi2(12)	145.37	23.38	562.09	266.85	288.34	189.85	171.02	261.19
P-value	0.0000	0.0247	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
Cross-sectional dependance test results
Pesaran cross-sectional dependence test
	fbs	itu	mcs	ict	apecfbs	apecitu	apecmcs	apecict
P-test	−1.269	1.387	−1.696	−0.707	−1.007	−1.393	−1.761	−0.668
Absolute val	0.365	0.377	0.381	0.360	0.353	0.377	0.378	0.359
P value	0.2045	0.1655	0.0898	0.4798	0.3138	0.1637	0.0783	0.5044
Friedman cross-sectional dependence test
	fbs	itu	mcs	ict	apecfbs	apecitu	apecmcs	apecict
F-test	0.429	9.229	5.269	0.429	0.5149	9.486	5.843	0.343
Absolute val	0.530	0.367	0.396	0.523	0.522	0.367	0.398	0.518
P value	1.000	0.7555	0.9586	1.000	1.000	0.7354	0.9517	1.000
Wald heteroscedasticity test
Chi2	36188.24	346.03	633.44	5492.01	41171.34	345.42	604.74	6870.40
P-value	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000	0.0000
Wooldridge serial correlation test
F(1,13)	2.052	2.326	2.251	1.886	2.037	3.147	2.571	2.335
P-value	0.756	0.1521	0.1574	0.1929	0.1771	0.0995	0.1329	0.1504
Jaque-Bera normality test
JB test	3.174	4.290	4.555	2.794	2.622	4.525	4.545	2.513
P-value	0.2046	0.1195	0.1025	0.2473	0.2696	0.1041	0.1031	0.2846

Source: Author’s compilation from STATA.

The Hausman test had p-values of less than 0.05. Thus, the p-value was sufficiently small to reject the null hypothesis. Therefore, the null hypothesis was rejected, indicating that the fixed-effect estimator is the appropriate method. The null hypothesis of the cross-sectional dependence test (CD Pesaran and Friedman) deduces that the residuals are cross-sectionally uncorrelated and the alternative posits that the residuals are cross-sectionally correlated. Thus, if the p-value is less than 0.05, the null hypothesis must be rejected, meaning that the residuals are cross-sectionally dependent. The Pesaran (2004) and Friedman (1937) cross-sectional dependence test (see Table 5) p-values are greater than 0.05%. Thus, neither test rejects the null hypothesis of no cross-sectional dependence.

4.5. Empirical results

The empirical results in Table 6 are based on FGLS and are presented in eight models. Models 1–4 present the results of the direct effect of energy consumption on environmental degradation. The three components of ICT include fixed broadband subscription, internet usage, and mobile phone subscription. These variables are highly correlated; therefore, they were treated separately. Thus, Models 1–3 consider each ICT component. Model 4 includes the ICT index, which was generated from the three components (fbs, itu, and mcs) using the Principal Component Analysis (PCA). Models 5 to 8 present the indirect effect of energy consumption on environmental degradation, which examines the impact of each ICT indicator and ICT index on the ecological degradation and energy consumption nexus. Thus, our main concern for Models 5–8 is the mediating variable coefficient.

Table 6. The impact of energy consumption on environment degradation.

Variables	(Model_1)	(Model_2)	(Model_3)	(Model_4)	(Model_5)	(Model_6)	(Model_7)	(Model_8)
	(FGLS)	(FGLS)	(FGLS)	(FGLS)	(FGLS)	(FGLS)	(FGLS)	(FGLS)
	lC02pc	lC02pc	lC02pc	lC02pc	lC02pc	lC02pc	lC02pc	lC02pc
lapec	0.0173	0.0127	0.0121	0.0180	0.0287**	0.0510**	0.0917**	0.0167
	(0.0117)	(0.0114)	(0.0113)	(0.0117)	(0.0137)	(0.0231)	(0.0389)	(0.0115)
linst	0.732***	0.498***	0.480***	0.742***	0.743***	0.515***	0.504***	0.756***
	(0.0887)	(0.0803)	(0.0825)	(0.0884)	(0.0884)	(0.0801)	(0.0824)	(0.0874)
lrgdp	−0.0389**	−0.0255	−0.0235	−0.0400**	−0.0421**	−0.0306*	−0.0340*	−0.0473***
	(0.0178)	(0.0176)	(0.0178)	(0.0178)	(0.0178)	(0.0176)	(0.0183)	(0.0178)
lkof	−0.235	−0.0583	−0.0651	−0.215	−0.213	−0.0467	−0.0536	−0.204
	(0.168)	(0.162)	(0.162)	(0.165)	(0.167)	(0.161)	(0.160)	(0.163)
lhci	0.471***	0.363***	0.371***	0.444***	0.442***	0.348***	0.363***	0.412***
	(0.124)	(0.121)	(0.121)	(0.121)	(0.125)	(0.121)	(0.120)	(0.120)
lacte	0.653***	0.628***	0.649***	0.639***	0.657***	0.637***	0.651***	0.661***
	(0.0593)	(0.0637)	(0.0541)	(0.0605)	(0.0590)	(0.0634)	(0.0535)	(0.0604)
lurbn	0.216**	0.243***	0.239***	0.211**	0.197**	0.222***	0.224***	0.171**
	(0.0863)	(0.0814)	(0.0793)	(0.0856)	(0.0866)	(0.0815)	(0.0789)	(0.0862)
lfdi	−0.0919***	−0.0734**	−0.0700**	−0.0902***	−0.0911***	−0.0711**	−0.0714**	−0.0852***
	(0.0316)	(0.0311)	(0.0315)	(0.0314)	(0.0314)	(0.0309)	(0.0312)	(0.0310)
lcp	−0.814***	−0.372**	−0.324*	−0.869***	−0.831***	−0.390**	−0.373**	−0.888***
	(0.191)	(0.174)	(0.181)	(0.196)	(0.191)	(0.172)	(0.180)	(0.194)
lpls	−0.730***	−0.606***	−0.597***	−0.728***	−0.727***	−0.610***	−0.601***	−0.723***
	(0.0831)	(0.0790)	(0.0798)	(0.0829)	(0.0826)	(0.0784)	(0.0791)	(0.0818)
lcil	0.127	−0.111	−0.130	0.133	0.163	−0.0753	−0.0834	0.178
	(0.183)	(0.176)	(0.177)	(0.182)	(0.184)	(0.176)	(0.176)	(0.181)
lfbs	0.0402				0.0834*
	(0.0404)				(0.0486)
litu		0.00406				0.0175
		(0.0289)				(0.0295)
lmcs			−0.0171				0.00781
			(0.0259)				(0.0282)
lict				0.0571				0.0849**
				(0.0409)				(0.0421)
Interaction Terms
lapeclfbs					−0.0298
					(0.0189)
lapeclitu						−0.0160*
						(0.00844)
lapeclmcs							−0.0189**
							(0.00882)
lapeclict								−0.033**
								(0.0144)
Constant	0.838	−1.228	−1.368	1.106	0.833	−1.233	−1.321	1.182
	(1.146)	(1.059)	(1.065)	(1.175)	(1.139)	(1.051)	(1.055)	(1.159)
Observations	191	226	226	191	191	226	226	191
Number of countrynum	14	14	14	14	14	14	14	14

Standard errors in parentheses.

***p < 0.01, **p < 0.05, *p < 0.1.

Source: Author’s compilation from STATA.

Table 6 shows the direct (see Models 1 to 4) and indirect (see Models 5 to 8) effects of energy consumption on environmental degradation in the SADC countries. The results indicate that the direct impact of energy consumption is insignificant, as shown in Models 1–4. These results concur with those of Tahir et al. (2022), who asserted that energy consumption has an insignificant influence on environmental degradation. However, the effect became positive and significant after controlling for the mediating variables in Models 5, 6, and 7. Thus, a 1% increase in energy consumption is associated with a 0.03%-0.09% increase in carbon emissions at the 5% significance level. These findings are in line with those of Ali et al. (2021), Khan et al. (2021), Udeagha and Ngepah (2022), and Kivyiro (2023), who documented evidence of the positive impact of energy use on carbon emissions. This positive impact suggests that SADC countries rely heavily on fossil fuels such as coal, oil, and natural gas for their carbon-intensive energy sources, leading to higher carbon emissions. Again, limited access to and investment in renewable resources and inefficient energy infrastructure precede the use of carbon-intensive sources. As a result, (Dogan & Seker, 2016; Khan et al., 2021) suggest that the minimum consumption of energy leads to lower carbon emissions and better environmental quality.

The empirical results also indicate that institutional quality, human capital development, access to electricity, and urbanization are positively related. A percentage increase in institutional quality increases carbon emissions by 0.5%–0.7% at the 1% significance level. Several studies have found a positive (Salman et al., 2019; Khan et al., 2021) connection between quality institutions and carbon emissions, whereas others have found a negative relationship (Azam et al., 2022) relationship. The positive effect of institutions on environmental degradation could be due to poor institutional quality, which spurs weak environmental regulation, lax enforcement of emissions standards, permitting violations, and inadequate monitoring of industrial and commercial activities that contribute to carbon emissions.

A 1% increase in human capital was associated with a 0.4% increase in carbon emissions at the 1% level. This result aligns with that of Bano et al. (2018), who find a unidirectional causal link between human capital and carbon emissions. These findings are consistent with those of Wiredu et al. (2023), who found that human capital causes carbon emissions. Other studies (Umar et al., 2022; Adikari et al., 2023) have found that human capital reduces carbon emissions. Thus, a positive effect is attributed to the fact that high levels of human capital can improve living standards and urbanization, leading to increased energy and resource consumption, potentially resulting in higher carbon emissions, primarily if the energy is derived from carbon-intensive sources. Even if education can raise awareness about environmental issues, it does not guarantee that individuals prioritize sustainability. In some cases, well-educated individuals may still engage in unsustainable practices if they prioritize economic benefits over environmental concerns.

Again, a percentage increase in access to electricity and urbanization increases carbon emissions by 0.7% and 0.2%, respectively, at the 1% significance level. The positive effect of access to electricity could be attributable to the SADC’s reliance on fossil fuels for electrification, the region’s primary source of carbon emissions. A positive impact of urbanization could be that the increasing population in urban areas increases the demand for energy and use (such as fossil fuels for transport, heating, and cooling), leading to higher carbon emissions. These results align with those of Kwakwa et al. (2021) and Nayaga et al. (2022), who argued that the relationship between electricity, urbanization, and carbon emissions is significant and positive.

The empirical results indicate that economic growth, perceptions of corruption, and political stability are carbon emission reduction factors in SADC countries. That is, a percentage increase in economic growth reduces carbon emissions by 0.03%–0.04% at the 5% significance level. The negative effect of economic growth on environmental degradation corroborates the findings of Liu et al. (2021) and Rahman and Alam (2022), who documented evidence of the negative impact of economic growth on carbon emissions, leading to a reduction in environmental degradation. This implies that increasing economic growth enhances innovation and technological advancements, which improve the use of cleaner energy-efficient technologies, such as renewable energy technologies, which reduce carbon emissions.

The negative sign of the corruption perception index implies that a 1% increase in corruption perception reduces carbon emissions by 0.3%-0.8% at a 5% level in SADC countries. When corruption is reduced, public resources are allocated more efficiently and environmental projects are less likely to be misappropriated or diverted for personal gain. As a result, funds can be directed toward clean energy projects, energy-efficient technologies, and emission reduction initiatives. These results are in accordance with those of Muhammad and Long (2021) and Yang et al. (2023), who suggested that corruption reduces carbon emissions.

Moreover, the results indicate that a percentage increase in political stability reduces carbon dioxide emissions by 0.6% -0.7% at the 1% significance level. This result suggests that governments can develop and implement coherent and sustained climate and energy policies in a stable political environment, which is essential for fostering investment in low-carbon technologies and infrastructure. The result is consistent with Muhammad and Long (2021) and Benlemlih et al. (2022), who argue that high political stability reduces CO2 emissions.

The indirect effect (see Models 5 to 8) shows the impact of each ICT indicator and ICT index on the environmental degradation-energy consumption nexus. Thus, Model 5 shows that the interaction term of energy consumption and fixed broadband is negative but insignificant, yet the energy consumption coefficient is significant and positive. This equates to the positive marginal effect of energy consumption on carbon emissions. This implies that, as more energy is consumed, it directly contributes to higher carbon emissions. This suggests that energy consumption and carbon emissions are positively related. This result aligns with those of Mobeen and Rashid (2017), Munir and Riaz (2020), Ali et al. (2021), and Khan et al. (2021), who argued that energy consumption enhances carbon emissions, leading to environmental degradation. The positive marginal effect of energy consumption on carbon emissions highlights the importance of managing and reducing carbon emissions in SADC countries.

The interaction of energy consumption with Internet usage and mobile phone subscription was significant and negative. This implies that internet usage and the use of mobile phones in SADC countries reduce the positive effect of energy consumption on carbon emissions. This result is consistent with those of Moyer and Hughes (2012) and Jin et al. (2018), who suggest that advances in technological innovations, mainly ICT, increase economic productivity, energy efficiency, and renewable energy production and use, which minimize carbon emissions, thereby reducing environmental degradation. The negative sign of the mediating variable (Internet usage) suggests that Internet use enables more efficient resource management in both cities and industries. For example, smart grids, connected transportation systems, and energy-efficient building management systems can optimize energy use and reduce carbon emissions (Erol-Kantarci & Mouftah, 2015). Internet use also enhances mobile apps that help users track their carbon footprint and encourages environmentally friendly choices, such as reducing energy consumption. Mobile phones provide access to information on sustainable practices, environmental awareness, and climate change (Zhao et al., 2022).

A significant and negative sign of the interaction term between energy consumption and the ICT index implies that ICT worsens the impact of energy consumption on SADC countries’ carbon emissions. This is attributable to the fact that many SADC countries may have older and less energy-efficient ICT infrastructures. Hence, outdated equipment and inefficient data centers can consume more energy than modern energy-efficient alternatives (Hilty et al., 2009; Adom, 2016). This inefficiency can exacerbate’s the carbon footprint.

5. Conclusions, summary, and policy implications

A multifaceted approach is recommended to mitigate carbon emissions within the Southern African Development Community (SADC) region by combining theory, practice, and policy elements. First, there is a pressing need to encourage SADC countries to diversify their energy sources away from carbon-intensive fuels such as coal and oil, and instead embrace cleaner and renewable energy options. This strategic shift could play a pivotal role in curbing the environmental impacts of energy consumption.

Simultaneously, strengthening environmental regulations and institutional quality should be of paramount focus. SADC countries can effectively combat the detrimental effects of weak governance on carbon emissions by enhancing governance standards, implementing stricter emission regulations, and bolstering monitoring mechanisms. Moreover, incorporating environmental education and sustainability principles into educational curricula is essential to ensure that human capital development aligns with eco-friendly practices.

Therefore, promoting investment in clean and renewable energy technologies is crucial. Governments can incentivize such investments through subsidies and support, facilitating economic growth and reducing emissions. Addressing corruption is vital to ensure that funds allocated for environmental projects and clean energy initiatives are used judiciously.

Political stability serves as a foundational element of consistent climate and energy policies. Governments that prioritize stability are more likely to attract investment in low-carbon technologies and infrastructure.

On the policy front, implementing energy-efficiency measures in urban areas is imperative to counterbalance the impact of urbanization on carbon emissions. These measures encompass energy-efficient transportation systems and building codes.

Promoting Internet access and adopting Information and Communication Technology (ICT) can facilitate resource management and environmental awareness. The development of smart grids, connected transportation systems, and energy-efficient technologies can make a significant contribution. Investing in upgrading ICT infrastructure to enhance energy efficiency is another crucial step.

6. Areas for future research

As for areas of future research, investigations should delve into the specific mechanisms through which ICT and Internet access influence carbon emissions, including the roles of intelligent technologies, data centers, and online behavior. Strategies for expediting the transition from carbon-intensive energy sources to cleaner options in SADC countries merit further exploration. Behavioral analysis, especially concerning education and awareness, should be the focus of this study. Furthermore, considering economic, social, and environmental factors, assessing the long-term sustainability of emission reduction policies is vital. Comparative studies with other regions can provide valuable insights and best practices applicable to SADC countries regarding their carbon emission reduction efforts.

Author contribution

The authorship contributions for this manuscript are as follows: Kin Sibanda conceptualized the study, conducted formal analysis, and contributed to the methodology, original draft preparation, review, editing, and visualization. Siyakudumisa Takentsi participated in investigation, original draft preparation, review, editing, and visualization. Dorcas Gonese contributed to methodology, formal analysis, investigation, original draft preparation, review, editing, and visualization. All authors have read and agreed to the final version of the manuscript.

Abbreviations
ARDL	=	Autoregressive distributed lag
CO2	=	Carbon dioxide
DOLS	=	Dynamic ordinary least squares
EKC	=	Environmental Kuznets Curve theory
FGLS	=	Feasible generalized least squares
FMOLS	=	Fully modified ordinary least squares
ICT	=	Information and Communication Technology
PCSE	=	Panel-corrected standard error
GLS	=	Panel data generalized least squares
PMG	=	Pooled mean group
PCA	=	Principal Component Analysis
QARDL	=	Quantile autoregressive distributive lag
SADC	=	Southern African Development Community
SDGs	=	Sustainable Development Goals

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data that supports the findings of this study is available on request from the corresponding author. The data was sourced from theglobaleconomy database which is a credible data source.

Additional information

Funding

The researchers received funding to attend a writing retreat to finalise the article form Walter Sisulu University.