Skip to content
Frontiers
Search

Original Research

Frontiers in Environmental Economics ·

Volume 4 - 2026 | https://doi.org/10.3389/frevc.2025.1708624

An empirical insight into the validity of the Environmental Kuznets Curve hypothesis in the SADC region

  • Ngonidzashe Chiranga1*corresponding author
  • Stephen Zhanje2
  • Samuel Chingoiro3
  1. 1Department of Economics, Faculty of Economics and Finance, Tshwane University of Technology (TUT), Polokwane, South Africa
  2. 2Department of Economics, Faculty of Management and Law, School of Economics and Management, University of Limpopo (UL), Polokwane, South Africa
  3. 3Library, Teaching & Research Excellence Department, Botho University, Gaborone, Botswana
Article metrics0 viewsView details
PDFCiteMetrics

Abstract

Climate change has resulted in global warming (CO 2) extremes, leading to academic debates on how climate change is impacting economic growth. Additionally, how do regulatory environments address climate change issues from a developing country perspective? The novelty of the study stems from a contribution to the stock of empirical evidence on the SADC region (1990–2023), employing panel data and a robust econometric strategy that accounts for cross-sectional dependence, slope heterogeneity, and non-stationarity. The analysis utilizes the CS-ARDL and CCEMG estimators on the impact of climate change on economic growth in the short-term and long-term. The findings provide partial support for the Environmental Kuznets Curve (EKC) hypothesis. Although GDP and its squared term were not statistically significant across models, the expected inverted U-shaped relationship emerged in some specifications, indicating that environmental degradation initially increases with economic growth but may decline at higher income levels.

Keywords

Introduction and background

Climate change or rising global temperatures has resulted in global warming (CO2), which has meant extreme weather conditions that range from floods, pests, and heat waves (Yoro and Daramola, 2020). Livelihoods for the poor and marginalized are increasingly becoming vulnerable, and this has resulted in competition for scarce natural resources like water, grazing land, and cultivation land (Njiru, 2012). Furthermore, the competition for resources has attracted fierce competition and internal conflict (McGuirk and Nunn, 2024). As the world grapples with global warming extremes, academic debates continue on how climate change is impacting economic growth in both the short and long term. Additionally, how have political institutions addressed the issues surrounding climate change as they balance pressing developmental priorities with global climate change commitments? Also, what role does climate change plays in the control of the government? This comes from the realization that non-governmental organizations (NGOs) seem to take over roles traditionally meant by governments, as governments fail to meet basic UN SDGs.

Country-specific studies by Zhou et al. (2023) and Song et al. (2021) both covered data from China from 1995 to 2017. However, differences were registered in the findings by Zhou et al. (2023) in which they found that urbanization did not have any impact on the environmental quality, in contrast to the findings by Song et al. (2021) 's study where a significant positive correlation was registered between CO2 emissions and economic growth. However, Obobisa et al. (2022) found conflicting results from their study of 25 African countries using the Augmented Mean Group (AMG) estimation technique and the Common Correlated Effects Mean Group (CCCMG), where Green Technology impacted CO2 negatively, Institutional Quality, and Renewable Energy.

On the other hand, Acaroglu and Güllü (2022) found that temperature has a negative relationship with renewable energy and consumption. In contrast, Lindvall (2021) found that democracy leads to lower emissions using data from Japan, the USA, and the European Union. Despite mostly the study is mostly based on content analysis. In contrast, Mele et al. (2021) found that the ever-increasing use of renewable energy tends to sustain economic growth, as indicated by data from Brazil from 1980 to 2018, using a non-parametric modeling approach. Given the lack of consensus from the different regional perspectives, one wonders if the same can be said for the Southern African Development Community (SADC) region? This paper is an attempt to contribute to the academic debate by closing the empirical gap observed in the SADC region, as few studies exist on this area. Also, the reluctance by the biggest polluters (USA and China) to fully commit to the provisions of the Paris Accord, and the overemphasis on strict environmental compliance in the least polluted nations situated in Africa1 And this raises more questions than answers: is this not weaponizing these environmental agreements, making goods from less developed nations less price competitive, given the cost of fully committing to some of these set environmental requirements for exported goods? This effectively keeps the balance of wealth more favorable to countries in the global north, and eventually, those in Africa may become mere experimental areas to further enrich those in the North.

This paper specifically aims to ascertain if there is a dynamic relationship between climate change, institutional quality, and economic growth in the SADC region in line with the EKC framework, an area that has few studies, yet it remains a topical issue in policy debates. This paper novelly contributes to the academic scholarship by closing the empirical and time gaps in the SADC region in the period 1990 to 2023 by employing panel data and a robust econometric strategy that utilizes the CS-ARDL and CCEMG estimators. The rest of the paper is as follows: the review of literature follows. Thereafter, there are methods and results, and a discussion. Lastly, the paper focuses on the conclusion and policy implications.

Theoretical framework

This study is grounded in the EKC hypothesis, which provides the principal theoretical lens for examining the relationship between economic growth and environmental quality. The conceptual roots of the EKC originate from the work of Kuznets (2019), who proposed an inverted U-shaped relationship between income levels and income inequality. Although Kuznets' work did not address environmental considerations directly, the underlying logic was subsequently extended to the income–environment nexus, forming the foundation of the modern EKC hypothesis (Grossman and Krueger, 1991). Figure 1 illustrates the theorized relationship between income and environmental degradation, along with the key moderating variables.

ECK income-environmental cost relationship & key influencing factors. Source: Authors' Contribution.
Figure 1. ECK income-environmental cost relationship & key influencing factors. Source: Authors' Contribution.

As depicted in Figure 1, the EKC suggests a non-linear path for environmental degradation as income rises. In the early stages of economic development, countries usually concentrate on industrial expansion and economic growth, often with limited attention to environmental protection. This stage is associated with intensive fossil fuel use and resource dependent production, which in turn leads to increasing pollution. Once income reaches a certain critical threshold, often referred to as the turning point, further economic growth is expected to contribute to environmental improvement. This improvement is driven by gradual shifts toward service-oriented sectors, stronger public demand for a cleaner environment, increased investment in cleaner technologies, and improved institutional capacity as documented in recent studies (Jóéwik et al., 2021; Hayaloglu et al., 2025).

To empirically test for the inverted U-shape of the EKC, a common approach is to incorporate an income measure and its squared term within an econometric model. Departing from the frequent use of GDP per capita, this analysis employs the logarithm of Gross Domestic Product (GDP) as its core income variable, alongside the square of this logged value. The literature argues that population size is a key determinant of carbon emissions (Naveed et al., 2022) and is therefore introduced as a control variable in the EKC model. This is particularly important in Africa, where access to clean energy remains a major challenge due to widespread poverty (Munonye et al., 2025). Therefore, to avoid multicollinearity between population size and the dependent variable, GDP is used instead of GDP per capita. The use of log-transformed total GDP as an indicator for economic growth is well-established in the literature (Mar et al., 2023; Nica et al., 2025; Fumey et al., 2025; Okorontah et al., 2025). Applying the log transformation ensures consistency with the scaling of other covariates, mitigates heteroscedasticity, corrects for right-skewness, and facilitates the interpretation of results in terms of percentage changes. Importantly, using log GDP still aligns with EKC theory. The EKC pattern would be confirmed by a positive coefficient for log GDP and a negative one for its square, indicating that pollution first rises with income but later falls after a certain point. A positive coefficient on log GDP combined with a negative coefficient on its squared value would confirm the EKC pattern, where environmental degradation increases at lower income levels but declines once income surpasses the turning point.

This study extends the core EKC framework by incorporating additional factors that influence the relationship between income and emissions. As shown in the conceptual diagram, factors such as renewable energy consumption, fossil fuel consumption, institutional regulatory quality, population size, and trade openness are expected to shape both the slope of the curve and the positioning of the turning point. Strong regulatory institutions or a rapid shift toward renewable energy can flatten the curve or lower the income level at which improvement begins. In contrast, continued dependence on fossil fuels may steepen the curve and prolong environmental decline.

Therefore, the EKC provides a robust foundation for this analysis. It not only offers a testable hypothesis for the income-emission dynamic within the SADC region but also creates a structured rationale for examining how structural and institutional factors modify this core relationship. Ultimately, this framework allows the study to explore whether rising incomes in SADC are compatible with long-run environmental improvement and what complementary factors are necessary to facilitate a sustainable transition.

Empirical literature

From a developed country perspective, (Greiner 2024), (Khoualfia and Bardi 2025), (Tzeremes et al. 2023), and (Höhler et al. 2021) focused primarily on European countries from 1990 to 2020. However, they applied different estimation techniques. More specifically, Greiner (2024) applied the General Methods of Moments (GMM) on a panel of 24 European countries and found that higher temperatures had a positive impact on economic growth, but were insignificant in other models. In contrast, Karimi Alavijeh et al. (2023) employed the novel method of moments regression using FMOLS, DOLS, and fixed effects (FE) estimation, which revealed that institutional quality improves environmental quality. On the other hand, Höhler et al. (2021) used a panel corrected standard error time series and cross-sectional regression analysis, where the populist government movement positively influenced greenhouse gas emissions. Whilst left-wing populist governments in Southern Europe had an opposite effect.

In other studies from Europe and America, Islam (2024) found that the US economy had rising occurrences of storms, average temperature, and CO2 emissions. Further, CO2 emissions had a negative influence on GDP per capita using data from 1970 to 2023, estimated using linear multiple regression. Yet Chou et al. (2023), estimated the USA economy from 1990 to 2013 using quantile regression and found that CO2 emissions were reduced with enhanced democracy. In contrast, Christoforidis and Katrakilidis (2025). The operational quality of institutions was found to be added to ecological sustainability, with estimates based on a panel of 29 OECD countries from 1984 to 2016, estimated using the ARDL method.

Contextually, there is a general lack of consensus on how CO2 affects economic growth from a developing country perspective. Although most studies used precipitation or temperature to measure climate change, no consensus exists on its impact on economic growth. Some scholars used some storm events or CO2 as the dependent variable. However, the impact on economic growth is less clear and continues to draw mixed reviews. The question that lingers on is whether the same results could be replicated from a regional perspective. More specifically, in the SADC region, this remains unclear due to an empirical void. However, there is a worrying push by the larger polluting countries in the North to force the least polluting countries in Sub-Saharan Africa to adhere to the provisions of the Paris Agreement, some of which call for a switch to renewables. Yet, they are the backbone of export earnings for some small SADC economies and are available in abundance as a source of energy to fuel electricity generation, citing the case of coal.

One wonders if there is wisdom in rapidly switching to renewables in the SADC region when it's expensive and out of reach to use renewable energy that can vary with changing weather? These are conversations that policymakers engage in when small open economies plunge into darkness as they rapidly adopt policies that are insensitive to the plight of the majority that still depend on fossil fuels. The question is, is it significant to justify closing some coal-powered power plants on the advice and funding of the World Bank, yet fossil-related pollution accounts from Africa account for less than 5% of global emissions? Also, you question the agenda that the World Bank is pushing in Africa, is it not perpetuating poverty, yet pretending to be solving it? You can't close a power station and plunge an existing town into a ghost town, then push renewable energy at a time when there is no infrastructure in place? What is the agenda behind this? Is it not creating business for other nations at the expense of creating poverty in Africa?

In studies from SADC and the rest of Africa, Sanderson and Jeke (2025) focused on 4 SADC countries, namely Angola, Botswana, Tanzania, and Zimbabwe, estimated using ARDL, yet Udeagha and Ngepah (2022) focused on South Africa only. What is most similar about these studies is that they were primarily conducted between 1990 and 2021, except for the study by Khurshid et al. (2022), which spanned the period from 1980 to 2021. On the other hand, Destiartono (2025) used ARDL bounds testing and the Toda-Yamamoto (TY) causality testing method. In contrast, Giovanis and Ozdamar (2022) used OLS and Fixed effects. On the other hand, Udeagha and Ngepah (2022) applied a system of GMM. In contrast, Rusiadi et al. (2024) applied Static Random Effects and Fixed Effects. Furthermore, Ayombe et al. (2021) applied the Kaya Framework on a panel of 27 African countries for a time framework of 1990 to 2017. This lack of a consistent estimation technique opens the scope to probe further if, in the SADC context, the same or different results can be obtained.

On the contrary, Khurshid et al. (2022)'s study findings from Pakistan suggest that CO2 emissions and mean temperature pose asymmetric results concerning economic growth, both in the short and long term. On the other hand, Imran et al. (2022) found that average changes in most of the climate indicators had a detrimental effect on economic growth, yet sea level pressure contributes positively to economic growth. Could this result be contextualized to the SADC region, and get a better understanding of whether countries with a coastline (seaside) experience increasing economic growth, or landlocked SADC regions have a different result?

In other studies from the ASEAN region and other emerging economies, Destiartono (2025) found energy intensity to have a positive impact on CO2, yet renewable energy had a vigorous role in reducing emissions and was linked negatively to CO2. In contrast, Destiartono (2025) found a significant decrease in the fiscal balance at 7.3%. Yet, Destiartono (2025) found from study projections of South Africa that it could lose about $1.82 billion due to climate change.

In summary, several studies have shown that without further empirical evidence in this area and triangulation with several methodologies, one can dispute that environmental cost sharing has become an accepted practice that naturally attracts policy debates. Furthermore, other factors that could influence climate change, yet less developed countries, are excluded from the protocols that guide climate change guidelines, yet play a critical role in coming up with such numbers. Careful considerations need to be made when recommending policies regarding the impact of climate change on economic growth, as there exist some empirical voids. However, policy may be driven by learning from other regions. When considering the SADC case, there are parallels between reality and recommendations. Figure 2 presents a conceptual framework that will be adopted in this study.

Conceptual framework diagram. Source: Authors' Contribution.
Figure 2. Conceptual framework diagram. Source: Authors' Contribution.

Methods

Variables and data

The study employs an unbalanced panel dataset comprising 10 SADC countries over the period 1990 to 2023. The countries included in the panel are Botswana, Democratic Republic of Congo, Madagascar, Mauritius, Mozambique, Namibia, South Africa, Tanzania, Zambia, and Zimbabwe. Country selection was based on the availability and consistency of data across the relevant variables. While the panel is unbalanced, this structure remains suitable for empirical analysis as it allows for the inclusion of countries with incomplete time-series data, thereby increasing the sample size and preserving regional representation (Baltagi, 2008). The dependent variable is total emission of CO2 in natural logs (LNCO2), used as a proxy for environmental degradation. The core explanatory variables include total GDP in natural logs (LNGDP) and its squared term (LNGDPSQD), included to test the EKC hypothesis. The rationale for using the logarithm of total GDP as a proxy for economic growth has already been explained in the theoretical framework. Additional control variables are incorporated to capture the effects of energy consumption through fossil fuel consumption (Fosco) and renewable energy consumption (RenCo), as well as institutional regulatory quality (RegQual), population size (LnPPN), and trade openness (Tropen). All variables are transformed into natural logarithms where appropriate, in order to linearize the relationships and allow the coefficients to be interpreted as elasticities. However, trade openness, renewable energy consumption, and fossil fuel consumption are expressed as percentages and therefore cannot be log-transformed. Similarly, institutional regulatory quality is an index that is already standardized on a normal distribution ranging from positive to negative values on a scale of 2.5. Log transformation would distort this scale, and it is therefore left in its original form. The use of governance indicators in their original values is also supported by other scholars, including Asongu and Odhiambo (2020), (Liu et al. 2020), and (Wu et al. 2025). The data were obtained from the World Development Indicators published by the World Bank (2025). Definitions of each variable, aligned with the World Bank (2025), are provided as follows:

  • CO2 total emissions: Total annual emissions of carbon dioxide, one of the six Kyoto greenhouse gases (GHG), from the agriculture, energy, waste, and industrial sectors, excluding Land Use, Land Use Change, and Forestry (LULUCF). LULUCF emissions are excluded due to their higher level of uncertainty. The concentration of carbon dioxide, as a major contributor to environmental degradation, has been widely used as a proxy indicator in empirical research (Barišauskaite and Mikalauskiene, 2025; Greiner, 2024; Shahani and Khaneja, 2024).
  • Total GDP at current US Dollar: GDP at purchaser's prices is the sum of gross value added by all resident producers in the economy, plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. Data are in current U.S. Dollars. Dollar figures for GDP are converted from domestic currencies using single-year official exchange rates. For a few countries where the official exchange rate does not reflect the rate effectively applied to actual foreign exchange transactions, an alternative conversion factor is used.
  • Total GDP at current US Dollar squared: This variable captures the non-linear, inverted U-shaped relationship proposed by Simon Kuznets (see theoretical framework). At low levels of GDP, environmental degradation tends to increase. However, beyond a certain income threshold, degradation begins to decline as societies can afford cleaner technologies and implement stronger environmental regulations. The inclusion of this squared term is a standard methodology for testing the EKC hypothesis (Mougenot et al., 2022).
  • RegQual: Captures perceptions of the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. The estimate reflects the country's score on the aggregate indicator, expressed in units of a standard normal distribution (approximately ranging from −2.5 to 2.5). It is included in the model due to its established, fundamental influence on environmental outcomes (ul Hasan et al., 2024).
  • PPN: Total population, based on the de facto definition, includes all residents regardless of legal status or citizenship. Values are midyear estimates. In line with the literature, Consistent with the established literature, it is included as a control variable to account for the scale of human activity. This helps to mitigate omitted variable bias, preventing the conflation of demographic effects with those of economic affluence (Hanif and Gago-de-Santos, 2017).
  • Tropen: Trade openness is the sum of exports and imports of goods and services measured as a share of gross domestic product. It is included to account for the ambiguous relationship between international trade and the environment. Trade can exacerbate pollution by scaling up economic activity, but may also facilitate emissions reduction via the transfer of cleaner technologies (Pham and Nguyen, 2024).

Econometric modeling

To assess the relationships proposed in the theoretical framework, this section sets out an econometric specification that extends the conventional EKC formulation. The extension incorporates the principal intermediary factors of renewable energy use, reliance on fossil fuels, trade openness and regulatory quality. By doing so, the modeling approach serves two main purposes. Firstly, it reflects more accurately the diverse influences on emissions within the SADC region. Secondly, it enables an examination of the extent to which institutional and structural conditions shape the EKC trajectory, including the point at which the curve begins to decline. To test this hypothesis and evaluate the influence of institutional quality on economic growth, the following three variations model is estimated to test alternative specifications with(out) renewable energy consumption:

LNCO2it=β0+β1LNGDPit+β2LNGDPSQDit+β3FossilConsit +β4RenewaConit+β5LNPPNit+β6Tropenit

+β3RegQualit+εit(1)

Model 1 includes both renewable and non-renewable energy consumption variables.

LNCO2it=α0+α1LNGDPit+α2LNGDPSQDit+α3FossilConsit

+α4LNPPNit+α5Tropenit+α6RegQualit+μit(2)

Model 2 excludes the renewable energy consumption variable.

LNCO2it=ω0+ω1LNGDPit+ω2LNGDPSQDit +ω3RenewaConit+ω4LNPPNit+ω5Tropenit

+ω6RegQualit+δit(3)

Model 3 excludes the non-renewable energy consumption variable.

Model 1 serves as the baseline specification, while the other two models are designed to isolate the individual effects of renewable energy consumption and non-renewable energy consumption on environmental degradation.

Results and discussions

Descriptive statistics

The descriptive statistics are presented in Table 1 and provide a preliminary overview of the distribution, central tendencies, and variability of the key variables used in the analysis. These statistics inform the underlying assumptions of the empirical modeling that follows. The variables capture core components essential to the study, as discussed in the previous sections. As stated in section Empirical literature, they include environmental indicators (CO2 emissions and energy consumption patterns), economic performance (GDP and its squared term), institutional quality (regulatory quality), trade openness, and population. Understanding these summary statistics is crucial before proceeding to the empirical estimation of the relationships among climate change, institutions, and economic growth in the SADC region.

Table 1. Summary of descriptive statistics.

VariableObsMeanStd DevMinMax
LN CO23401.79041.5619−0.09676.1901
LNGDP34023.37461.222921.712526.8506
LNGDPSQD340547.864358.9895471.4333720.9531
FossilCons34039.118130.40991.6493.65
RenewaCon34062.279730.69767.698.3
LNPPN34016.25641.391913.872618.4770
Tropen34071.966926.802817.2251135.2905
RegQual340−0.35230.7092−2.20151.1969

Authors' computations.

The mean value of 23.37, the standard deviation of 1.22, and the range of 21.71 to 26.85 indicate moderate variation in GDP output across the SADC region, which roughly reflects the presence of both lower-income and middle-income economies within the region. In contrast, the data show significant variability in CO2 emission levels across countries, and the presence of a negative minimum (−0.097) suggests extremely low emissions (slightly above zero without the LN) in some cases. Similarly, fossil fuel consumption reflects wide variation in the energy mix across SADC countries, with some relying more heavily on renewable energy instead. In terms of population, a range of 13.87 to 18.48 indicates notable differences in population sizes across the region. Trade openness also exhibits considerable variation, ranging from 17.23 to 135.29, with some countries being highly open while others are more closed. Finally, the negative mean value for regulatory quality suggests generally weak institutional environments across the SADC region, though there are countries with relatively high institutional quality, as indicated by positive values. The next section displays the pairwise correlations among these variables.

Correlation matrix

Table 2 presents the correlation matrix, providing insights into the direction and strength of linear relationships between the variables. The matrix is essential for understanding potential multicollinearity before proceeding to the regression analysis.

Table 2. Correlation analysis results.

VariablesLNCO<sub>2</sub>LNGDPLNGD PSQDFossil ConsRenewa ConLNPPN
LNCO2R1.0000
p-value
LNGDPR0.8667**1.0000
p-value0.0000
LNGDPSQDR0.8745**0.9995**1.0000
p-value0.00000.0000
FossilConsR0.5616**0.3406**0.3512**1.0000
p-value0.00000.00000.0000
RenewaConsR−0.4785**−0.3147**−0.3247**−0.9538**1.0000
p-value0.00000.00000.00000.0000
LNPPNR0.3934**0.5569**0.5537**−0.4867**0.5169**1.0000
p-value0.00000.00000.00000.00000.0000
TropenR−0.2452**−0.3148**−0.3160**0.4040**−0.4412**0.7138**
p-value0.00000.00000.00000.00000.00000.0000
RegQualR0.1410**0.07160.07860.5754**−0.6893**−0.5207**
p-value0.00920.18760.14830.00000.00000.0000

Authors' computations. ** indicate the significant at the 1% level.

The results in Table 2 reveal that most variables are statistically significant in the correlation matrix, suggesting robust associations relevant for examining the impacts of climate change and institutions on growth. There are strong positive correlations between CO2 emissions (LN CO2) and GDP (LNGDP) (r = 0.87; p = 0.00), as well as with its square (LNGDPSQD) (r = 0.87; p = 0.00), indicating a link between growth and emissions. Fossil fuel consumption is moderately positively correlated with LN CO2 (r = 0.56; p = 0.00), while renewable consumption shows a moderate negative correlation with CO2 emissions (r = −0.48; p = 0.00) and a strong negative correlation with fossil fuel consumption (r = −0.95; p = 0.00), highlighting a substitution effect. Trade openness (Tropen) shows weak negative correlations with both emissions (r = −0.25; p = 0.00) and GDP (r = −0.31; p = 0.00). Regulation quality (RegQual) shows a weak positive correlation with LN CO2 (r = 0.14; p = 0.01), a moderate positive correlation with fossil fuel consumption (r = 0.58; p = 0.00), and a strong negative correlation with renewable consumption (r = −0.69; p = 0.00), indicating potential institutional challenges in fostering green growth.

Table 2 also shows an almost perfect correlation between LNGDP and LNGDPSQD (0.9995), indicating potential multicollinearity. Likewise, the strong negative correlation between fossil fuel consumption and renewable consumption (−0.9538) suggests possible instability if both variables are included in the same regression. While other correlations are moderate, these correlations do not imply cross-sectional dependence or heterogeneity. Cross-sectional dependence and heterogeneity tests are generally necessary when using panel data (Baltagi and Hashem Pesaran, 2007). Tables 3, 4 provide these results.

Table 3. Cross-sectional dependence test results.

VariableCD-test*p-value*Average joint TMean ρMean abs(ρ)
LNCO220.6590.00034.000.530.77
LNGDP34.0270.00034.000.870.87
LNGDPSQD34.0320.00034.000.870.87
FossilCons1.5730.11634.000.040.52
RenewaCon0.6510.51534.000.020.59
LNPPN38.3250.00034.000.980.98
TrOpen6.130.00034.000.160.33
RegQual−1.2810.20034.00−0.030.58

Ho, Cross-sectional independence. Authors' computations.

Table 4. Homogeneity/heterogeneity test results.

Slope homogeneity testModel 1Model 2Model 3
Statistic*p-value*Statistic*p-value*Statistic*p-value*
Delta_tilde14.3980.00016.1180.00015.9330.000
Delta_tilde_adj16.7910.00018.4320.00018.2200.000

Cross-sectional dependence test results

The findings in Table 3 show a significant CD-test statistic, which supports the rejection of the null hypothesis of cross-sectional independence. Thus, cross-sectional dependence is present and needs to be accounted for in the subsequent analysis.

Based on the p-values in Table 3, the null hypothesis of cross-sectional independence is rejected at the 1% significance level for LN CO2, LNGDP, LNGDPSQD, LNPPN, and Tropen (p-value = 0.000). This indicates significant cross-sectional dependence for these variables. Conversely, the p-values for FossilCons (0.116), RenewaCon (0.515), and RegQual (0.200) suggest that the null hypothesis for these variables is not rejected, indicating no significant cross-sectional dependence in those variables. Overall, since the majority of the variables exhibit cross-sectional dependence, the use of first-generation estimators would be inefficient in this case.

Heterogeneity test results

Table 4 presents the results of the slope heterogeneity test for the three models, assessing whether the relationship between CO2 emissions and the explanatory variables differs significantly across countries in the panel.

The results in Table 4 provide strong evidence of slope heterogeneity across all three models. This is demonstrated by the high Adjusted-Deltã statistics and p-values of 0.000. The findings confirm that the relationship between CO2 emissions and the explanatory variables varies significantly across countries in the panel. These variables encompass GDP, energy consumption, and control variables such as institutional quality, population size, and trade openness. Among the models, Model 2 shows the highest level of heterogeneity, indicating that differences in fossil fuel use and economic structures play a key role in cross-country variation. Based on these results, the application of heterogeneous panel estimators is appropriate for capturing country-specific dynamics more accurately.

Stationarity test results

To examine the time series properties of the panel data, this study applies two complementary unit root testing approaches known as first-generation and second-generation panel tests. The first-generation tests, represented by the Im, Pesaran, and Shin (IPS) test, assume cross-sectional independence across panel units. However, this assumption may be too restrictive, especially in macroeconomic panels where regional countries are likely to be interconnected. To address potential cross-sectional dependence, the second-generation approach incorporates the Cross-sectionally Augmented IPS (CIPS) test developed by Pesaran (2007), which allows for common factors across units (Menegaki, 2021).

Table 5 presents the results of both the first- and second-generation panel unit root tests, applied to the variables in both trend and non-trend forms. This dual approach ensures a robust assessment of the stationarity properties of the series, providing a reliable foundation for selecting appropriate panel estimation techniques in subsequent analysis.

Table 5. First- and second-generation panel unit root tests results.

Variables/Unit root tests1stGeneration2ndGeneration
IPS w/o trendIPS w trendCIPS w/o trendCIPS w trend
LNCO20.585 I(0)−0.220 I(0)−2.188 I(0)−2,375 I(0)
−9.070***I(1)−8.323***I(1)−5.274***I(1)−5.529***I(1)
LNGDP2.565 I(0)−0.583 I(0)−3.044***I(0)−3.535***I(0)
−11.360***I(1)−9.764***I(1)−5.718***I(1)−5.843***I(1)
LNGDPSQD2.712 I(0)−0.706 I(0)−3.024***I(0)−3.532***I(0)
−11.390***I(1)−9.770***I(1)−5.717***I(1)−5.845***I(1)
FossilCons5.223 I(0)8.340 I(0)−0.601 I(0)−0.336 I(0)
1.5177 I(1)1.821 I(1)−2.468**I(1)−3.211***I(1)
RenewaCon3.984 I(0)8.968 I(0)0.500 I(0)−0.973 I(0)
−0.2397 I(1)0.2397 I(1)−4.316***I(1)−4.363***I(1)
LNPPN2.602 I(0)−4.313***I(0)−2.217*I(0)−2.539 I(0)
−7.1027***I(1)−8.425***I(1)−2.943***I(1)−4.108***I(1)
Tropen−1.544 I(0)−2.421***I(0)−2.305*I(0)−3.090***I(0)
−11.802<sup></sup>I*(1)−10.189***I(1)−5.524***I(1)−5.543***I(1)
RegQual−0.426 I(0)1.797 I(0)−1.737 I(0)−2.580 I(0)
−11.689***I(1)−10.719***I(1)−6.190***I(1)−6.413***I(1)

***p < 0.01, **p < 0.05, and *p < 0.1, Authors' computations. 1(0) is at levels and 1(1) at first difference.

The results presented in Table 5 provide consistent evidence that variables under consideration are integrated either of order zero or one, I (0) or I (1), across both first- and second-generation panel unit root tests. In the first-generation framework, the Im, Pesaran, and Shin (IPS) test results indicate that, although no variables appear to be stationary without a trend at the level, but stationary at 1st difference. The inclusion of a trend component reveals significance at 5% level and hence stationarity of some variables (LNPPN, Tropen) at levels, and the majority of the variables (LNCO2, LNGDP, LNGDPSQD, RegQual) confirm stationarity after first differencing. Similarly, the second-generation CIPS test, which accounts for potential cross-sectional dependence, reinforces these findings by showing that most of the variables (LNCO2, FossilCons, RenewaCon, LNPPN, RegQual) become stationary only after first differencing, both with and without trend specifications. The CIPS test results also revealed that other variables (LNGDP and LNGDPSQD—without and with trend; Tropen-with trend) are stationary at the level since the null hypothesis of the existence of a unit root is rejected at 5% level. These results confirm that other series are stationary in levels, but others are stationary in first differences, justifying the treatment of variables as either I (0) or I (1) processes in the panel estimations. The mixed unit root behavior across both generations of tests supports the use of estimation techniques suitable for non-stationary panel data, as determined by most model variables, such as panel cointegration approaches.

Cointegration test results

(Pedroni 1999), (2004) cointegration methodology is applied using the seven parametric and non-parametric test statistics to the LNCO2 Models 1, 2 and 3 accordingly. The Pedroni cointegration technique tests the hypothesis of no cointegration using a single tailed test which follows a normal distribution. Hence, Table 6 presents the results of Pedroni's panel cointegration test. The test was carried out to examine whether a long-run relationship exists between CO2 emissions, economic growth, and other explanatory variables. This analysis forms part of the investigation into the EKC hypothesis within the SADC region.

Table 6. Pedroni's panel cointegration test results.

StatisticModel 1 (LNCO2 as dependent variable)Model 2 (LNCO2 as dependent variable)Model 3 (LNCO2 as dependent variable)
PanelGroupPanelGroupPanelGroup
V−1.273–−0.5452–0.6449–
Rho1.1912.3090.01571.1210.53021.58
T−2.803−2.523−3.566−3.533−2.757−2.534
Adf−3.088−2.431−3.244−3.356−3.695−3.328

Authors' computations.

Ho, no cointegration.

The results of Pedroni's panel cointegration test, shown in Table 6, indicate a clear long-run relationship among the variables across all three models. Although the panel V-statistics are mostly insignificant, the panel t-statistics and panel ADF statistics are consistently negative and statistically significant across all models. These statistics reject the null hypothesis of no cointegration, indicating that CO2 emissions, economic growth, and the included control variables are cointegrated in the long run. Additionally, the group statistics, particularly the group t-statistics and group ADF statistics, also support the presence of cointegration, reinforcing the panel-based results. This suggests the existence of a long-run relationship where CO2 emissions, GDP, squared GDP, fossil fuel consumption, renewable energy consumption, population, trade openness, and regulatory quality share a common trend.

Westerlund (2005) cointegration methodology applies the variance ratio (VR) ratio test statistic to determine the existence or non-existence of cointegration in some of the panels. The panel-specific cointegration parameters are α1*i* to α6*i* and ω1*i*to ω6*i*for Models 2 and 3, respectively. The null hypothesis of no cointegration is tested against the alternative hypothesis that some of the panel-specific cointegration parameters produce cointegrated series, that is, some panels are cointegrated. In the same realm, the null hypothesis of no cointegration is also rejected as depicted by Westerlund's cointegration test results in Table 7 for the disaggregated models 2 and 3 using the 10% significance level threshold applied by other studies (Selmey et al., 2025; Ximei et al., 2025). The consistency of these findings across models lends empirical support to the existence of a stable long-run equilibrium (Neal, 2014), a condition critical for testing the EKC hypothesis in the context of the SADC region. This justifies the subsequent use of long-run panel estimation techniques such as Cross-Sectional Autoregressive Distributed Lag (CS-ARDL) and the Common Correlated Effects Mean Group (CCEMG) (Sheikh and Hassan, 2023).

Table 7. Westerlund panel cointegration test results.

Variance ratio testModel 2 (LNCO2 as dependent variable)Model 3 (LNCO2 as dependent variable)
Statistic*p-value*Statistic*p-value*
Variance ratio−1.41350.0788−1.37450.0846

Ho, No cointegration.

H1, panels are cointegrated.

Model selection

The diagnostic tests presented in Tables 3–5 highlight key features of the panel data. Table 3 confirms the presence of cross-sectional dependence, while Table 4 shows significant slope heterogeneity across countries. Table 5 further indicates that all variables are integrated of order zero or one, I (0) or I (1), based on both first- and second-generation unit root tests. In addition, the Pedroni cointegration test results and Westerlund cointegration results in Tables 6, 7, respectively confirm the existence of a long-run relationship among the variables. Together, these findings necessitate the use of estimation techniques that can account for cross-sectional dependence, non-stationarity, slope heterogeneity, and cointegration. Based on this, the CS-ARDL model was adopted to examine both the short- and long-run relationships among the variables. This model controls for cross-sectional dependence and allows for country-specific dynamics. To ensure the robustness of the results, CCEMG estimator was also applied. The CCEMG approach further addresses cross-sectional dependence by incorporating cross-sectional averages of the regressors and residuals and is suitable for panels with heterogeneous slopes. These estimators provide more reliable insights into the relationship between climate change, institutions, governance, and economic growth in the SADC region.

CS-ARDL model results

Table 8 reports the estimation results from the CS-ARDL model applied to the three specified models. The CS-ARDL approach accounts for both cross-sectional dependence and heterogeneity across countries, making it well-suited for panel data with integrated variables and potential long-run relationships. This estimation technique enables the identification of both short-run dynamics and long-run equilibrium effects between CO2 emissions, economic growth, institutional quality, and other explanatory variables within the SADC region. The results presented in Table 8 offer important insights into the nature and direction of these relationships, particularly in the context of the EKC hypothesis.

Table 8. CS-ARDL model results.

Variables(1)(2)(3)
LNCO2LNCO2LNCO2
Group short run
L.LNCO20.04730.1010.165**
(0.803)(1.583)(2.398)
LNGDP−0.828−1.434−1.608
(−0.342)(−0.707)(−0.624)
LNGDPSQD0.01850.03090.0357
(0.353)(0.704)(0.635)
FossilCons−0.00002840.00631**
(−0.00625)(2.121)
RenewaCon−0.0126***−0.00847**
(−2.771)(−2.578)
LNPPN−0.2380.9090.214
(−0.116)(0.458)(0.128)
TrOpen0.001560.001140.00188*
(1.365)(1.046)(1.960)
RegQual−0.0797***−0.0388−0.0779**
(−2.610)(−1.566)(−2.285)
ECT−0.953***−0.899***−0.835***
(−16.15)(−14.05)(−12.14)
Group long run
lr_FossilCons0.0001270.00716**
(0.0298)(2.218)
lr_LNGDP−0.438−1.554−2.803
(−0.174)(−0.697)(−0.838)
lr_LNGDPSQD0.01050.03400.0627
(0.193)(0.706)(0.860)
lr_LNPPN0.2181.2391.111
(0.106)(0.574)(0.523)
lr_RegQual−0.0873**−0.0497−0.0810*
(−2.191)(−1.301)(−1.689)
lr_RenewaCon−0.0125***−0.00988***
(−2.960)(−2.770)
lr_Tropen0.001700.001360.00265*
(1.217)(1.060)(1.934)
Observations330330330
R-squared0.4040.4600.446
Number of groups101010
F1.387**1.408***1.489***

Authors' computations.

Standard errors are in parentheses, ***p < 0.01, **p < 0.05, and *p < 0.1.

The short-run and long-run parameter estimates derived from the CS-ARDL methodology for all three model specifications are presented in Table 8. A stable long-run cointegrating relationship is validated by the error correction terms (ECTs), which are consistently negative and statistically significant at the 1% level across all models (Androniceanu and Georgescu, 2023). This consistency with prior research that utilizes CS ARDL and similar cointegration techniques reinforces the common finding of rapid convergence toward equilibrium in environmental-economic systems (Mata et al., 2024).

Short-run dynamics reveal that renewable energy consumption significantly curbs CO2 emissions in Models 1 and 3. This indicates that scaling up renewables is an effective strategy for lowering emissions within the SADC region. Such a conclusion mirrors earlier empirical work attributing environmental gains to the displacement of fossil fuels by renewable sources (Xu et al., 2023). Furthermore, the significant negative coefficient for regulatory quality in the same models underscores the importance of robust institutions in promoting adherence to environmental policies and strengthening mitigation initiatives. This aligns with analyses of the regulatory-environment link in developing economic contexts (Abbas et al., 2025).

The results for trade openness show only weak significance in Model 3. This ambiguity reflects the lack of consensus in the existing literature, where some studies highlight the pollution-intensifying scale effects of trade (Wang et al., 2023), while others point to beneficial outcomes from technological spill overs and cleaner production methods (Gava et al., 2025). A similarly limited effect is observed for fossil fuel consumption, which shows a positive and significant impact on emissions solely in Model 2. This anticipated result points to the ongoing role of carbon-based energy, a trend noted in previous studies (Androniceanu and Georgescu, 2023).

Over the long run, the negative and significant relationship between renewable energy consumption and CO2 emissions persists in Models 1 and 3, confirming the vital role of clean energy in achieving decarbonization. This finding is supported by a substantial body of international evidence on the long-term carbon-reducing effects of renewables (Busu and Nedelcu, 2021; Alam et al., 2025). Regulatory quality also maintains its significant negative influence in Model 1, lending further weight to the proposition that effective governance is crucial for improving environmental performance, as it encourages sustainable investment and allocative efficiency (Abdelkawy and Alzuwaidi, 2025; Abbas et al., 2025). On the contrary, institutional quality in other studies appears to be without any substantial long-term influence on carbon emissions (Xu et al., 2023).

While the coefficients for LNGDP and its squared term lack statistical significance, their positive and negative signs, respectively, correspond with the theoretical EKC pattern, suggesting a potential inverted U-shaped relationship. Such weakly defined EKC results are not uncommon in analyses of developing regions, where factors like structural economic limitations, technological gaps, or inconsistent policy enforcement can obscure the relationship (Voumik et al., 2023). The presence of the expected signs nonetheless implies that income growth may still influence emissions through scale and composition channels, even if this effect is not statistically robust in the present SADC analysis.

Models 1 to 3 reveal that a positive but insignificant relationship exists between the population and CO2 emissions. It can be noted that a possible increase in population density may result in environmental distress (Mata et al., 2024). Moreover, the trade- CO2 emissions relationship proved to be positive for all CS-ARDL models and only Model 3 showed a significant outcome in the long run (Wang et al., 2023). On the contrary and increase in trade growth has a potential of decreasing the CO2 emissions growth due to expansion of service trade (Zhou et al., 2025). Also, countries and regions are intertwined in such a way that new developments and opportunities are enhanced in the process (Wang et al., 2023).

In summary, this analysis identifies renewable energy deployment, the quality of institutions, and, to a lesser extent, trade openness as pivotal factors influencing environmental quality in the SADC region. By applying the CS-ARDL approach, this study offers fresh regional insights, illustrating that advancements in environmental quality are achievable through clean energy and enhanced regulation, even in contexts where the link with economic growth remains statistically ambiguous.

CCEMG model results

Table 9 reports the estimation results from the CCEMG model applied to the three specifications. The CCEMG estimator addresses both cross-sectional dependence and slope heterogeneity by incorporating cross-sectional averages of the dependent and independent variables. This makes it particularly suitable for panels where unobserved common factors and country-specific dynamics may bias traditional estimators. The model provides robust long-run coefficients, complementing the CS-ARDL results and serving as a robustness check for the relationships between CO2 emissions, economic growth, institutional quality, and other explanatory variables in the SADC region.

Table 9. CCEMG model results.

Variables(1)(2)(3)
LNCO2LNCO2LNCO2
Group long run
LNGDP0.0818−1.276−1.658
(0.0304)(−0.532)(−0.511)
LNGDPSQD0.00005120.02910.0382
(0.000888)(0.569)(0.544)
FossilCons−0.0008500.00619**
(−0.196)(2.069)
RenewaCon−0.0130***−0.00918**
(−2.580)(−2.476)
LNPPN0.4791.6770.786
(0.221)(0.842)(0.407)
Tropen0.001270.0007830.00141
(0.800)(0.614)(1.041)
RegQual−0.0983**−0.0663*−0.0986**
Group cross-section averaged regressors
__00000M_LNCO2(−2.273)(−1.664)(−2.262)
0.924***0.972***0.956***
__00000L_LNGDP(3.232)(3.408)(3.833)
6.4716.5855.557
__00000L_LNGDPSQD(1.133)(1.393)(0.981)
−0.137−0.141−0.118
__00000L_FossilCons(−1.126)(−1.400)(−0.983)
−0.0203−0.00874
__00000L_RenewaCon(−1.280)(−0.570)
−0.0252**−0.00169
__00000L_LNPPN(−2.078)(−0.110)
−0.646−1.297−0.643
__00000L_Tropen(−0.231)(−0.566)(−0.283)
−0.0005000.000638−0.00182
__00000L_RegQual(−0.209)(0.374)(−0.966)
0.1680.2300.363
Constant(0.669)(0.966)(1.282)
−71.69−67.96−46.97
(−1.135)(−1.307)(−0.829)
Observations340340340
Number of countries101010
chi2150.35***42.30***12.62**

Authors' computations. Standard errors are in parentheses, ***p < 0.01, **p < 0.05, and *p < 0.1.

Table 9 reports the long-run CCEMG estimates for the three model specifications. Essentially, the results are broadly consistent with those obtained using the CS-ARDL estimator, reinforcing the robustness of the key relationships identified.

Renewable energy consumption remains negative and statistically significant in Models 1 and 3, corroborating the earlier findings and confirming that increased deployment of clean energy contributes to long-term reductions in CO2 emissions in the SADC region. This outcome is in line with empirical evidence from both developed and developing countries that attributes lower emissions to the substitution of carbon-intensive fuels with renewable sources (Voumik et al., 2023).

Fossil fuel consumption exhibits a positive and statistically significant effect on emissions in Model 2, underscoring its central role in driving environmental degradation. This finding aligns with a substantial body of literature demonstrating the strong contribution of fossil-based energy to emission growth, particularly in regions with limited renewable energy penetration (Voumik et al., 2022; Rai et al., 2025).

Regulatory quality continues to display a negative and significant relationship with emissions across all model specifications, further highlighting the importance of effective governance in promoting environmentally responsible behavior and enforcing mitigation policies. These results are consistent with studies emphasizing that well-functioning institutions facilitate environmental regulation, stimulate green investment and enhance compliance (Sarfraz et al., 2021).

The coefficients for LNGDP and LNGDP squared remain statistically insignificant. However, their signs in Models 2 and 3 reflect the anticipated inverted U-shape implied by the EKC. Similar findings have been reported in studies where income and environment relationships in developing regions fail to reach statistical significance, often due to structural inefficiencies and uneven policy implementation (Rai et al., 2025). Nonetheless, the preserved signs indicate that scale, composition and technique effects may still be operating, albeit weakly, within the SADC context.

The inclusion of cross-section averages in the CCEMG specification effectively accounts for unobserved common factors, strengthening the reliability of these estimates. The alignment of CCEMG and CS-ARDL results underscores the importance of renewable energy adoption and governance quality as core determinants of environmental performance in the SADC region, while confirming the continuation of fossil-fuel intensity as a major constraint.

In a nutshell, the CCEMG findings provide additional empirical support for the role of clean-energy transitions and institutional capacity in reducing emissions, thereby extending the literature by offering region-specific evidence based on techniques that address cross-sectional dependence and heterogeneous dynamics.

Conclusion and policy implications

This study investigates the long-run relationship between climate change, economic performance, institutional quality, and other control variables within the SADC region, with a particular focus on testing the EKC hypothesis. Employing panel data and a robust econometric strategy that accounts for cross-sectional dependence, slope heterogeneity, and non-stationarity, the analysis utilizes the CS-ARDL and CCEMG estimators. These methods were chosen based on supporting evidence from cross-sectional dependence tests, panel unit root tests, and Pedroni's cointegration test. The findings provide partial support for the EKC hypothesis. Although GDP and its squared term were not statistically significant across models, the expected inverted U-shaped relationship emerged in some specifications, indicating that environmental degradation initially increases with economic growth but may decline at higher income levels. This delicate finding aligns with existing literature from both developed and emerging economies, where the EKC relationship is often found to be context-dependent and sensitive to regional dynamics (Greiner, 2024; Karimi Alavijeh et al., 2023; Khurshid et al., 2022).

More robust and consistent evidence was found regarding the impact of renewable energy consumption and institutional quality. Both estimators revealed that renewable energy use significantly reduces CO2 emissions in the short and long run, reinforcing conclusions from studies in the ASEAN and African contexts that highlight renewables as a vital tool in mitigating environmental degradation (Destiartono, 2025; Sanderson and Jeke, 2025). Similarly, regulatory quality exerted a significant negative effect on emissions, underscoring the critical role that strong institutions and effective governance play in promoting environmental sustainability, consistent with findings from OECD and European studies (Christoforidis and Katrakilidis, 2025; Karimi Alavijeh et al., 2023).

Fossil fuel consumption remains a primary driver of emissions in the SADC region, while trade openness showed weak or mixed effects depending on the model specification. This reflects ongoing policy debates in the region, where the transition from fossil fuels to renewables faces challenges related to energy access, economic dependency, and infrastructure constraints (Sanderson and Jeke, 2025). The literature highlights concern about the feasibility and socio-economic implications of rapid shifts to renewable energy, especially given the energy needs of developing countries and their relatively low contributions to global emissions (World Bank, 2025).

In summary, while the EKC relationship is not conclusively confirmed, this analysis highlights the importance of targeted policy efforts in renewable energy adoption and institutional strengthening as critical pathways for balancing economic growth with environmental protection in the SADC region. Given the empirical gaps and regional particularities identified in the literature, future research should focus on disaggregating country-level dynamics, incorporating additional institutional and environmental performance indicators, and exploring the socio-economic trade-offs of energy transitions to provide more context-sensitive policy recommendations.

This paper offers several policy insights for SADC policymakers. Foremost among them is the need for a gradual yet strategic transition toward renewable energy. Although renewable energy has demonstrated effectiveness in reducing emissions, its adoption must be guided by considerations of cost, infrastructure availability, and broader socio-economic conditions. Also, expanding renewable energy technologies should go hand in hand with developing a supportive regional infrastructure and implementing targeted subsidies to make clean energy more accessible and affordable, especially in remote and disadvantaged communities. In addition, reinforcing institutional frameworks is essential. This implies governments should improve regulatory quality and strengthen institutional capacity to ensure the effective implementation and enforcement of environmental regulations. Furthermore, instead of pursuing a blanket reduction in fossil fuel use, SADC countries should embrace a balanced energy mix strategy that harnesses their natural resource advantages while maintaining a commitment to long-term decarbonization objectives. Lastly, climate-related policies should be tailored to the regional context, avoiding generic, externally imposed solutions. A bottom-up, inclusive approach to policy design that addresses energy poverty, engages local communities, and aligns with national development priorities is essential for promoting sustainable growth without compromising economic performance.

Nonetheless, this study is not without limitations. Its geographical focus is restricted to the SADC region. Undertaking a comparative analysis that includes other African blocs, such as ECOWAS, would enhance the robustness and generalizability of the findings. Extending the study period could also strengthen the empirical rigor.

Future research may benefit from broadening the analytical scope to incorporate the shadow pricing of renewable energy sources. In addition, complementing secondary data with primary, bottom-up approaches such as survey-based investigations would provide deeper insights and further reinforce the validity of the conclusions.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frevc.2025.1708624/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Phoebe Koundouri, Athens University of Economics and Business, Greece

Reviewed by: Theodoros Daglis, University of Patras, Greece

Mohammad Jibran Gul Wani, Jazan University, Saudi Arabia

Footnotes

According to Statistica.com. Africa contributes between 3%−4% of global emissions.

Statements

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by Botho University towards the publication fees.

Data availability

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

NC: Conceptualization, Writing – review & editing, Validation, Supervision, Formal analysis, Writing – original draft. SZ: Formal analysis, Writing – original draft, Conceptualization, Writing – review & editing, Investigation, Methodology. SC: Funding acquisition, Resources, Conceptualization, Writing – review & editing, Formal analysis, Writing – original draft, Methodology.

Acknowledgements

The authors would like to thank their respective institutions for providing an environment that enabled this research. These institutions are the Tshwane University of Technology (South Africa), University of Limpopo (South Africa), and Botho University (Botswana).

References

  1. Abbas F., Malik U. I., Abdul Rahman A. (2025). Nexus between institutional quality, economic openness, and environmental quality: new evidence from the MENA region. Fut. Bus. J. 11: 128. doi: doi:10.1186/s43093-025-00552-4
  2. Abdelkawy N. A., Alzuwaidi L. (2025). Institutional quality, energy efficiency, and natural gas: explaining CO2 emissions in the GCC. Sustainability 17: 6746. doi: doi:10.3390/su17156746
  3. Acaroglu H., Güllü M. (2022). Climate change caused by renewable and non-renewable energy consumption and economic growth: a time series ARDL analysis for Turkey. Renew. Energy 193, 434 – 447. doi: doi:10.1016/j.renene.2022.04.138
  4. Alam M. F., Singh A., Tripathi A., Singh R., Kumar R. (2025). Linkages among income, renewable energy consumption and CO 2 emissions and their implications for sustainable development in the OECD countries. SAGE Open 13: 21582440251378810. doi: doi:10.1177/21582440251378810
  5. Androniceanu A., Georgescu I. (2023). The impact of CO2 emissions and energy consumption on economic growth: a panel data analysis. Energies 16: 1342. doi: doi:10.3390/en16031342
  6. Asongu S. A., Odhiambo N. M. (2020). Governance, CO2 emissions and inclusive human development in sub-Saharan Africa. Energy Explor. Exploit. 38, 18 – 36. doi: doi:10.1177/0144598719835594
  7. Ayombe C. E., Amenhwi A. G., Ngala M. I. (2021). Forest cover change due to large scale plantation agriculture and community perception of its impact on climate variability in Nguti sub-division, South West Cameroon Region. Int. J. Biodivers. Conserv. 13, 109 – 119. doi: doi:10.5897/IJBC2021.1481
  8. Baltagi B. H. (2008). Econometric Analysis of Panel Data, Vol. 4. Chichester: John wiley and sons. p. 135 – 145.
  9. Baltagi B. H., Hashem Pesaran M. (2007). Heterogeneity and cross section dependence in panel data models: theory and applications introduction. J. Appl. Econometr. 22, 229 – 232. doi: doi:10.1002/jae.955
  10. Barišauskaite G., Mikalauskiene A. (2025). Economic and environmental factors determining carbon dioxide emissions in Lithuania. Transform. Sustain., 1, 1 – 11. doi: doi:10.63775/t67saw86
  11. Busu M., Nedelcu A. C. (2021). Analyzing the renewable energy and CO2 emission levels nexus at an EU level: a panel data regression approach. Processes 9: 130. doi: doi:10.3390/pr9010130
  12. Chou C., Clark R., Kimbrough S. O. (2023). What do firms say in reporting on impacts of climate change? An approach to monitoring ESG actions and environmental policy. Corp. Soc. Responsib. Environ. Manag. 30, 2664 – 2678. doi: doi:10.1002/csr.2509
  13. Christoforidis T., Katrakilidis C. (2025). Assessing the impacts of economic complexity and economic freedom on the energy-induced environmental performance: new evidence from a panel of EU countries. J. Knowl. Econ. doi: doi:10.1007/s13132-025-02672-2
  14. Destiartono M. E. (2025). Do economic growth and urbanization drive CO2 emissions? Role of renewable energy and energy efficiency in Indonesia. Int. J. Energy Econ. Policy 15: 647. doi: doi:10.32479/ijeep.18354
  15. Fumey M. P., Sackitey G. M., Boateng S. A., Wiredu J., Moro K. D. (2025). Economic growth, renewable energy consumption patterns, and environmental sustainability in Ghana: testing the environmental Kuznets curve. Sustain. Econ. 3: 1320. doi: doi:10.62617/se1320
  16. Gava E., Seabela M., Ogujiuba K. (2025). The effect of trade openness on environmental quality in Southern African Customs Union (SACU) countries: the CS-ARDL approach. Economies 13: 233. doi: doi:10.3390/economies13080233
  17. Giovanis E., Ozdamar O. (2022). The nexus between business–investment climate and firm performance in the Middle East and North Africa (MENA) region. Rev. Econ. Polit. Sci. 7, 257 – 285. doi: doi:10.1108/REPS-09-2020-0138
  18. Greiner A. (2024). Uncertainty of climate models and policy implications: a European perspective. Paper presented at the List Forum für Wirtschafts-und Finanzpolitik. doi: doi:10.1007/s41025-024-00266-5
  19. Grossman G. M., Krueger A. B. (1991). Environmental Impacts of a North American Free Trade Agreement. Available online at:
  20. Hanif I., Gago-de-Santos P. (2017). The importance of population control and macroeconomic stability to reducing environmental degradation: an empirical test of the environmental Kuznets curve for developing countries. Environ. Dev. 23, 1 – 9. doi: doi:10.1016/j.envdev.2016.12.003
  21. Hayaloglu P., Recepoglu M., Cakir M. A. (2025). Institutional structure and environmental pollution: an application within the framework of the environmental Kuznets curve hypothesis. Int. J. Energy Econ. Policy 15: 230. doi: doi:10.32479/ijeep.19337
  22. Höhler C., Rasamoel N. D., Rohrbach N., Hansen J. P., Jahn K., Hermsdörfer J.. (2021). The impact of visuospatial perception on distance judgment and depth perception in an augmented reality environment in patients after stroke: an exploratory study. J. Neuroeng. Rehab. 18: 127. doi: doi:10.1186/s12984-021-00920-5
  23. Imran M., Khan S., Zaman K., ur Rashid Khan H., Rashid A. (2022). Assessing green solutions for indoor and outdoor environmental quality: sustainable development needs renewable energy technology. Atmosphere 13: 1904. doi: doi:10.3390/atmos13111904
  24. Islam M. K. (2024). Impact of climate change on the economic growth of the USA. Int. J. Adv. Res. Econ. Financ. 6, 56 – 68.
  25. Jóéwik B., Gavryshkiv A. V., Kyophilavong P., Gruszecki L. E. (2021). Revisiting the environmental Kuznets curve hypothesis: a case of Central Europe. Energies 14: 3415.
  26. Karimi Alavijeh N., Salehnia N., Salehnia N., Koengkan M. (2023). The effects of agricultural development on CO2 emissions: empirical evidence from the most populous developing countries. Environ. Dev. Sustain. 25, 12011 – 12031.
  27. Khoualfia A., Bardi W. (2025). The effect of climate change on economic growth in European countries: an empirical evidence from Panel ARDL approach. Environ. Sci. Pollut. Res. 32, 736 – 748. doi: doi:10.1007/s11356-024-35788-0
  28. Khurshid N., Khurshid J., Shakoor U., Ali K. (2022). Asymmetric effect of agriculture value added on CO2 emission: does globalization and energy consumption matter for pakistan. Front. Energy Res. 10: 1053234. doi: doi:10.3389/fenrg.2022.1053234
  29. Kuznets S. (2019). “Economic growth and income inequality,” in The Gap Between Rich and Poor (Routledge), 25 – 37.
  30. Lindvall D. (2021). Democracy and the Challenge of Climate Change. Stockholm: IDEA Publications.
  31. Liu X., Latif K., Latif Z., Li N. (2020). Relationship between economic growth and CO2 emissions: Does governance matter? Environ. Sci. Pollut. Res. 27, 7221 – 17228. doi: doi:10.1007/s11356-020-08142-3
  32. Mar M., Seraj M., Tursoy T. (2023). The role of fiscal policy in G20 countries in the context of the environmental Kuznets curve hypothesis. Enegies 16, 1 – 16. doi: doi:10.3390/en16052215
  33. Mata J. P. V., Bautista M. G. G., Granda L. E. S., Zurita Moreano E. (2024). Evaluating the environmental Kuznets Curve: the role of renewable energy, economic growth, urban density and trade openness on CO2 emissions. An analysis for high-income countries using the CS-ARDL Model. Int. J. Energy Econ. Policy 14, 580 – 596. doi: doi:10.32479/ijeep.16740
  34. McGuirk E. F., Nunn N. (2024). Transhumant pastoralism, climate change, and conflict in Africa. Rev. Econ. Stud. 92/:rdae 027. doi: doi:10.1093/restud/rdae027
  35. Mele M., Gurrieri A. R., Morelli G., Magazzino C. (2021). Nature and climate change effects on economic growth: an LSTM experiment on renewable energy resources. Environ. Sci. Pollut. Res. 28, 41127 – 41134. doi: doi:10.1007/s11356-021-13337-3
  36. Menegaki A. (2021). “Chapter 2: stationarity and an alphabetical directory of unit roots often used in the energy-growth nexus,” in A Guide to Econometrics Methods for the Energy-Growth Nexus (Cambridge, MA), 31 – 61.
  37. Mougenot B., Durand Santa María R. P., Koc Olcese C. L. (2022). Testing the environmental Kuznets curve hypothesis: an empirical study for Peru. Int. J. Energy Econ. Policy 12, 193 – 199. doi: doi:10.32479/ijeep.11895
  38. Munonye J. O., Onyeneke R. U., Ankrah D. A., Agyarko F. F., Onyeneke C. J., Nejad J. G.. (2025). “Do climate change, access to electricity and renewable energy consumption matter in aquaculture production in Africa?,” in Natural Resources Forum, Vol. 49, No. 2 (Oxford: Blackwell Publishing Ltd), 986 – 1009.
  39. Naveed A., Ahmad N., Aghdam R. F., Menegaki A. N. (2022). What have we learned from Environmental Kuznets Curve hypothesis? A citation-based systematic literature review and content analysis. Energy Strat. Rev. 44: 100946. doi: doi:10.1016/j.esr.2022.100946
  40. Neal T. (2014). Panel cointegration analysis with xtpedroni. Stata J. 14, 684 – 692. doi: doi:10.1177/1536867X1401400312
  41. Nica I., Georgescu I., Kinnunen J. (2025). Economic growth, innovation, and CO2 emissions: analyzing the environmental Kuznets curve and the innovation Claudia curve in BRICS countries. Sustainability 17, 1 – 37. doi: doi:10.3390/su17083507
  42. Njiru B. N. (2012). “Climate change, resource competition, and conflict amongst pastoral communities in Kenya,” in Climate Change, Human Security and Violent Conflict: Challenges for Societal Stability, eds. J. Scheffran, M. Brzoska, H. Brauch, P. Link, and J. Schilling (Berlin: Springer), 513 – 527.
  43. Obobisa E. S., Chen H., Mensah I. A. (2022). The impact of green technological innovation and institutional quality on CO2 emissions in African countries. Technol. Forecast. Soc. Change 180: 121670. doi: doi:10.1016/j.techfore.2022.121670
  44. Okorontah C. F., Uruakpa Chiagoziem G., Obisike N. E., John O. O., Uzoma K. P. (2025). Assessment of the validity of the environmental Kuznets Curve (EKC) hypothesis in the Nigerian economy. J. Econ. Financ. Manag. Stud. 8, 6562 – 6575. doi: doi:10.47191/jefms/v8-i9-63
  45. Pedroni P. (1999). Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxf. Bull. Econ. Stat. 61, 653 – 670. doi: doi:10.1111/1468-0084.61.s1.14
  46. Pedroni P. (2004). Panel cointegration: asymptotic and finite sample properties of pooled time series tests with an application to the PPP hypothesis. Econ. Theory 20: 597 –625. doi: doi:10.1017/S0266466604203073
  47. Pesaran M. H. (2007). A simple panel unit root test in the presence of cross-section dependence. J. Appl. Econom. 22, 265 – 312.
  48. Pham T. Q. A., Nguyen Q. L. H. T. T. (2024). “Climate change and firm performance: the case of the top 100 largest companies in Vietnam,” in The International Conference on Economics, Law, and Government (ELG 2024), p. 177.
  49. Rai P., Gupta P., Saini N., Tiwari A. K. (2025). Assessing the impact of renewable energy and non-renewable energy use on carbon emissions: evidence from select developing and developed countries. Environ. Dev. Sustain. 27, 3059 – 3080. doi: doi:10.1007/s10668-023-04001-6
  50. Rusiadi R., Adivia A., Yusuf M. (2024). The impact of the green economy on sustainable development in the ABRIC countries. Paper presented at the International Conference on Humanity Education and Society (ICHES).
  51. Sanderson A., Jeke L. (2025). Impact of rainfall and temperature on economic growth in Angola, Botswana, Tanzania and Zimbabwe. Int. J. Energy Econ. Policy 15: 719. doi: doi:10.32479/ijeep.19168
  52. Sarfraz M., Ivascu L., Cioca L. I. (2021). Environmental regulations and CO 2 mitigation for sustainability: panel data analysis (PMG, CCEMG) for BRICS nations. Sustainability 14: 72. doi: doi:10.3390/su14010072
  53. Selmey M. G., Elbialy B. A., Mohamed I. A., Alhumud A. A., Sheta I. H. (2025). The Impact of primary energy consumption, trade openness and foreign direct investment on economic growth: evidence from MENA countries. Int. J. Energy Econ. Policy 15, 398 – 409. doi: doi:10.32479/ijeep.20794
  54. Shahani R., Khaneja N. (2024). Is environment degradation a precondition to economic growth? A case of India and China. Vision 28, 98 – 110.
  55. Sheikh A., Hassan O. I. (2023). Impact of economic growth, energy use and research and development expenditure on carbon emissions: an analysis of 29 OECD countries. Int. J. Energy Econ. Policy 13, 454 – 464. doi: doi:10.32479/ijeep.13647
  56. Song Y., Zhang X., Zhang M. (2021). The influence of environmental regulation on industrial structure upgrading: based on the strategic interaction behavior of environmental regulation among local governments. Technol. Forecast. Soc. Change 170: 120930. doi: doi:10.1016/j.techfore.2021.120930
  57. Tzeremes P., Dogan E., Alavijeh N. K. (2023). Analyzing the nexus between energy transition, environment and ICT: a step towards COP26 targets. J. Environ. Manag. 326: 116598. doi: doi:10.1016/j.jenvman.2022.116598
  58. Udeagha M. C., Ngepah N. (2022). Does trade openness mitigate the environmental degradation in South Africa? Environ. Sci. Pollut. Res. 29, 19352 – 19377. doi: doi:10.1007/s11356-021-17193-z
  59. ul Hasan M., Ullah S., Sultan S. (2024). Environmental quality, economic growth and role of institutional quality: panel data analysis from selected Asian countries. Ann. Soc. Sci. Perspect. 5, 351 – 361. doi: doi:10.52700/assap.v5i2.427
  60. Voumik L. C., Rahman M., Akter S. (2022). Investigating the EKC hypothesis with renewable energy, nuclear energy, and RandD for EU: fresh panel evidence. Heliyon 8, 1 – 10. doi: doi:10.1016/j.heliyon.2022.e12447
  61. Voumik L. C., Rahman M. H., Nafi S. M., Hossain M. A., Ridzuan A. R., Mohamed Yusoff N. Y. (2023). Modelling sustainable non-renewable and renewable energy based on the EKC hypothesis for Africa's ten most popular tourist destinations. Sustainability 15, 1 – 19. doi: doi:10.3390/su15054029
  62. Wang Q., Wang L., Li R. (2023). Trade openness helps move towards carbon neutrality—Insight from 114 countries. Sustain. Dev. 32, 1081 – 1095. doi: doi:10.1002/sd.2720
  63. Westerlund J. (2005). New simple tests for panel cointegration. Econ. Rev. 24, 297 – 316. doi: doi:10.1080/07474930500243019
  64. World Bank (2025). World Bank Data. Available online at: (Accessed May 11, 2025).
  65. Wu H., Du J., Rasool Y. (2025). Navigating environmental concerns: unveiling the role of economic governance, energy transition, and population aging on transport-based CO 2 emissions in China. Enegies 18, 1 – 26. doi: doi:10.3390/en18071748
  66. Ximei K., Javaid M. Q., Shams T., Sibt-e-Ali M. (2025). Information and communication technology, economic globalization, and environmental sustainability in APEC nations: insights from LCC and EKC hypotheses testing. Sustain. Dev. 33, 878 – 903. doi: doi:10.1002/sd.3163
  67. Xu X., Dai W., Muhammad T., Zhang T. (2023). The dynamic relationship between carbon emissions, financial development, and renewable energy: a study of the N-5 Asian countries. Sustainability 15: 13888. doi: doi:10.3390/su151813888
  68. Yoro K. O., Daramola M. O. (2020). “CO2 emission sources, greenhouse gases, and the global warming effect,” in Advances in Carbon Capture (Cambridge: Elsevier), 1 – 28.
  69. Zhou R., Guan S., He B. (2025). The impact of trade openness on carbon emissions: empirical evidence from emerging countries. Energies 18: 697. doi: doi:10.3390/en18030697
  70. Zhou Y., Luo H., Tang J., Zhang L., Zhu H., Sun S. (2023). Study on ecological environment damage compensation in China. J. Nat. Conserv. 76: 126503. doi: doi:10.1016/j.jnc.2023.126503

Citation

Chiranga, N., et al. (2026). An empirical insight into the validity of the Environmental Kuznets Curve hypothesis in the SADC region. Frontiers in Environmental Economics, Volume 4. https://doi.org/
Received
Accepted
Published
Volume
4 - 2026

Copyright

Copyright © 2026 Chiranga, Zhanje and Chingoiro.

License: https://creativecommons.org/licenses/by/4.0/

*Correspondence: Ngonidzashe Chiranga

Article metrics

Views
0
Downloads
0
Citations
0
Shares
0

Related articles

Hypothesis and TheoryClimate policies and adaptive capacities: a cross-analysis of sectoral, intersectoral, and international actionsJean-Marie Gankou Fowagap, Magali Duchel Mandio AbomoOriginal ResearchEnergy and resource sustainability patterns in the European Union through clustering and convergence dynamicsNicoleta Barbuta-Misu, Puiu Lucian Georgescu, Monica Laura Zlati, Costinela Fortea, Valentin Marian AntohiOriginal ResearchIntegrating climate adaptation initiatives for coastal resilience: governance lessons from the Catalan coastJoana Díaz-Pont, Queralt Plana, Sandra Casas, Xavier Martinez