Energy Inequality and Environmental Transition in the Gulf Cooperation Council Countries: Revisiting the Kuznets Curve

Abstract

This study explores the effect of Energy Inequality (EINQ) on environmental sustainability within the frameworks of the Environmental Kuznets Curve (EKC) and the Load Capacity Curve (LCC), while accounting for technological progress (TECH), financial development (FD), and foreign direct investment (FDI). Using annual data for six Gulf Cooperation Council (GCC) countries from 2005 to 2024, the analysis applies the Method of Moments Quantile Regression (MMQR) to capture heterogeneous effects across the distribution of the Load Capacity Factor (LCF). The results show that energy inequality consistently reduces environmental sustainability, indicating that unequal access to efficient and clean energy services heightens ecological pressure. In contrast, technological innovation and financial development enhance sustainability by improving energy efficiency and supporting green investments. Economic growth exhibits an inverted U-shape, validating the EKC and LCC hypotheses. These findings are especially important for the GCC, where hydrocarbon dependence, uneven access to clean energy, and rapid structural change intensify the environmental consequences of inequality. The study underscores the need for policies that promote equitable energy access, innovation-led diversification, and sustainable financial mechanisms.

1. Introduction

Energy has been widely acknowledged as the foundation for economic and social development, yet its unbalanced distribution and consumption create substantial environmental sustainability issues. Recent years have seen a growing interest in the concept of energy inequality (EINQ), particularly unequal access to and use of energy services, among both academic scholars and policymakers [1]. Energy injustice does not only imply limited access to modern energy services, but also includes the wasteful use of energy resources, increased reliance on fossil energy sources, and the unequal distribution of environmental risks. Such inequities are also challenging the shift towards sustainable energy systems, especially in countries where energy serves as a key driver of industrialization and welfare. It is important to note that energy inequality is not used here as a proxy for income inequality. While both forms of inequality may overlap, energy inequality constitutes a distinct dimension of deprivation, related to access, affordability, reliability, and efficiency of energy services, and may persist even in high-income, high-growth economies such as the GCC.

The Gulf Cooperation Council (GCC) States, namely Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates, are an important example in this respect. Although the region is among the world’s largest producers and exporters of energy, it is characterized by significant internal disparities in energy consumption intensity, efficiency levels, and the integration of renewable energy [2]. A massive reliance on hydrocarbons, along with subsidizing energy and rapid population growth, has led to continual environmental degradation. The intensification of ecological degradation in response to rising energy demand and unequal availability of clean energy sources indicates that balancing concerns for environmental security with the need for energy remains a complex challenge [3].

In the context of the Gulf Cooperation Council (GCC) countries, energy inequality is particularly salient despite the region’s abundant hydrocarbon resources. Recent data show that residential energy consumption per capita in GCC economies is substantially higher than the global average, while its distribution is highly uneven across income groups. For example, higher-income households typically consume a disproportionate share of subsidized electricity and fuel, whereas lower-income households face a higher energy burden relative to their disposable income. In several GCC countries, the top income residential energy consumption is several times that of the bottom quintile, indicating that universal access does not automatically translate into equitable use and affordability. Moreover, ongoing subsidy reforms and energy price adjustments have renewed concerns that vulnerable households may be adversely affected if distributional impacts are not properly understood and managed. These stylised facts underscore the need to analyze energy inequality in the GCC, rather than focusing solely on average energy use or aggregate efficiency indicators.

In theoretical terms, the nexus linking energy consumption, inequality, and environmental deterioration can be examined through the Environmental Kuznets Curve (EKC) and Load Capacity Curve (LCC) hypotheses. The EKC proposes that an inverted U-function describes the relationship between income per capita and environmental degradation: environmental damage increases at low levels of income per capita, but declines as a country becomes wealthier and new technologies are adopted [4].

However, this approach, which does not consider heterogeneity in energy access and use, is becoming increasingly problematic [5,6]. The Load Capacity Curve (LCC) framework provides a more comprehensive outlook on sustainability by combining biocapacity and the ecological footprint into the Load Capacity Factor (LCF), defined as the ratio of biocapacity to ecological footprint per capita [5,7]. An LCF greater than one indicates that ecosystems operate within their regenerative capacity, while a value below one signals ecological overshoot [5,6,7]. When the LCF is greater than 1, the ecosystem is functioning within its regenerative capacity; when it falls below 1, ecological overshoot occurs.

These mechanisms can be framed within an energy-augmented production model, in which output depends on capital, labor, and effective energy use, and in which energy inequality affects both the intensity and the composition of energy demand [8,9]. When access to efficient, low-carbon technologies is restricted to a subset of households or firms, other groups rely more heavily on polluting fuels and inefficient equipment, thereby raising aggregate emissions and ecological pressure [10,11]. Empirical studies for both advanced and emerging economies document that unequal access to clean energy and efficiency improvements is associated with higher carbon intensity and ecological footprint [12,13]. In the GCC context, despite very high average income levels, disparities persist in energy-efficient housing, the use of clean fuels, and integration into modern energy infrastructure, particularly between urban and peripheral regions [10,11]. Thus, energy inequality is not merely a concern for low-income countries; it is also a structural obstacle to sustainable development in high-income, hydrocarbon-dependent economies such as those in the GCC [9,13].

The increasing immediacy of climate change, escalating ecological debts, and uneven access to clean energy all make it increasingly important that policymakers address what we refer to as “inequalities of energy” as a central feature of environmental policy. For the GCC countries, where economic diversification and energy transition are critical for long-term sustainability, the relationship between energy inequality and environmental quality is not just a theoretical exercise; it has real-world implications. The shift from carbon-intensive growth to environmentally sustainable development will be determined by the equity of access to energy resources and the extent to which technological advances, innovative financing mechanisms, and foreign direct investment are leveraged in favor of cleaner energy systems [14].

These additional variables help capture key channels through which energy inequality interacts with environmental outcomes. Technological innovation and R&D support the adoption of cleaner production processes and more efficient energy use, thereby reducing ecological pressure for a given level of income. Financial development determines the availability of credit for green investments, such as renewable energy and energy-efficient infrastructure, but can also fund energy-intensive activities, making its net effect an empirical question. Foreign direct investment, finally, may transfer cleaner technologies and managerial practices or, alternatively, reinforce the ‘pollution haven’ pattern if it concentrates in emission-intensive sectors. Including TECH, FD, and FDI therefore improves the model’s explanatory power and reduces omitted-variable bias in estimating the effect of energy inequality on environmental sustainability. Technological advancements can help alleviate the negative environmental impacts caused by energy disparities by improving energy efficiency and accelerating the adoption of renewable energy. Financial development facilitates the movement of capital into environmentally friendly investments, and FDI contributes to technology spillovers and sustainable infrastructure development. However, the effectiveness of these factors depends on the extent to which inclusive processes are also pursued within the energy system. This research has several important implications for a burgeoning literature on energy justice and environmental sustainability. First, our paper broadens the EKC and LCC frameworks by incorporating EINQ into environmental performance. In contrast to previous GCC-focused studies, which have concentrated overwhelmingly on aggregate energy use, CO₂ emissions, or income levels, this study emphasizes the uneven distribution of access to and efficiency of energy [14,15,16]. Second, as LCF is used as a criterion for environmental sustainability, we consider a more comprehensive measure that combines the dimensions of ecological pressure and biocapacity to support more realistic evaluations of sustainable resource use. Third, the paper also provides for a regional investigation (the GCC countries), a region that demonstrates energy richness and environmental fragility, thereby filling a critical empirical gap in the literature on energy–environment. Fourth, the study uses a modeling framework that integrates technology, financial development, and foreign direct investment, incorporating several dimensions of innovation and financial flows that connect the energy–environment nexus. Last, but not least, the results of this study are highly policy-relevant by addressing the sustainable development goals (SDGs) established by the United Nations such as SDG 7 (Affordable and Clean Energy), SDG 11 (Sustainable Cities and Communities), SDG 12 (Responsible Consumption and Production) and SDG 13 (Climate Action), which point out that alleviating energy poverty while encouraging technological progress can facilitate the low-carbon transition as well as an inclusive energy future.

Energy inequality, defined as the uneven access to and distribution of energy resources and services among different populations and regions, has been extensively studied for its profound impacts on environmental sustainability. This review synthesizes key findings from recent research to illustrate how disparities in energy access and consumption shape environmental outcomes.

Several studies have highlighted that energy inequality leads to an increased reliance on inefficient and polluting energy sources, particularly in low-income and rural areas [17]. According to the authors [18], populations without access to modern energy services disproportionately rely on traditional biomass, kerosene, and coal for cooking and heating, resulting in higher emissions of greenhouse gases (GHGs) and indoor air pollutants. This not only contributes to environmental degradation but also poses severe health risks, creating a dual burden of energy poverty and environmental harm. Moreover, research [19] emphasizes that energy inequality exacerbates environmental challenges by limiting the adoption of clean and renewable energy technologies in underserved regions. The lack of infrastructure and financial resources limits these communities’ access to renewable energy, intensifying carbon emissions and environmental pollution. Sovacool et al. argue that addressing energy inequality is critical for achieving global climate goals and reducing the carbon footprint of energy consumption. Another dimension discussed in the literature is the feedback loop between environmental degradation and energy inequality [20]. As noted, environmental degradation caused by unsustainable energy use, such as deforestation and soil erosion, further reduces the availability of local energy resources, deepening energy poverty [19]. This cyclical relationship complicates efforts to promote environmental sustainability and equitable energy access simultaneously.

Policy-oriented studies underscore the importance of integrating energy equity into environmental sustainability frameworks. The United Nations Sustainable Development Goals (SDGs), particularly SDG 7 (affordable and clean energy) and SDG 13 (climate action), reflect this integration [21,22]. Such solutions are viewed as pivotal in breaking the cycle of energy poverty and environmental harm.

For instance, Pachauri and Rao show that households without access to modern energy disproportionately depend on biomass and kerosene, thereby increasing local pollution and greenhouse gas emissions [23], while Sovacool et al. document how energy injustices intensify exposure to environmental hazards in vulnerable communities [19]. Addressing these disparities through targeted policies and renewable energy deployment is essential to achieving sustainable development goals and fostering a more equitable, environmentally sustainable future.

From the perspective of Measuring and Assessing the Impact of Income Inequality, it was noted that research demonstrates severe income inequality in resource-rich GCC nations, linking oil and gas revenues to disproportionate income gains among top earners that could exacerbate societal injustices [24]. This provides an essential perspective on comprehending the economic ramifications that extend beyond environmental measurements. A fragmented understanding of the relationship between inequality and environmental degradation results from the inadequate incorporation of income inequality metrics into EKC analysis [25]. According to recent research, few studies thoroughly examine the feedback loops between environmental consequences and income distribution [20]. From the aspect of validity and shape of the EKC in GCC countries, systematic reviews and empirical analyses reveal mixed evidence regarding the EKC hypothesis, with its validity often depending on the choice of environmental indicators, econometric methods, and country-specific factors [26]; The EKC hypothesis is empirically supported by several studies conducted in GCC nations, showing an inverse U-shaped relationship between environmental degradation, specifically CO₂ emissions, and economic growth in nations like Saudi Arabia, the United Arab Emirates, Qatar, and Kuwait [16,27,28,29]. These results are more robust when sophisticated econometric methods such as panel cointegration and ARDL are used [15,30]. While Different nations and environmental indicators exhibit varying EKC validity, for instance, Bahrain and Oman show U-shaped or nonexistent EKC evidence [15,31,32]. Some studies raise doubts about the practical applicability by revealing tipping points that fall outside of the reported income range. Generalization is made more difficult by the variation in EKC shapes among nations and pollutants [33,34].

The Energy Inequality Index (EINQ) includes four dimensions: electricity poverty, access to clean fuels for cooking, energy security, and environmental pressure through ecological footprint. We standardize all indicators and apply Principal Component Analysis (PCA) to extract the first principal component, which provides a statistically grounded measure of multidimensional energy inequality [26,35,36,37,38]. The use of economic complexity indices further enhances the examination of structural economic changes [39]. Although the potential of renewables and technology is acknowledged, empirical data often show minimal or statistically insignificant impacts due to the region’s delayed energy transition and ongoing reliance on fossil fuels [40,41]. Policy recommendations are made more difficult by the diversity of technological consequences, for example, ICT components that have both positive and negative effects [42].

From the perspective of the Influence of Financial Development and trade openness, the literature describes them as having two sides: they can sometimes lower emissions by encouraging green technologies, but they can also increase pollution through the pollution haven effect [43,44,45,46]. Depending on the context and measurement, financial development has been shown to both increase and decrease emissions [47,48,49,50]. The conflicting results regarding financial development and trade openness reflect methodological variations as well as the complexity of their relationships with environmental consequences.

From the perspective of Methodological Approaches and Data Quality, the empirical rigor across studies is strengthened by the use of a variety of econometric techniques, such as panel ARDL, quantile regressions, causality analyses, and cointegration tests robust to cross-sectional dependence [51,52,53,54]. It is possible to record dynamic interactions and structural fractures using longitudinal data that spans many decades [55,56]. Methodological diversity restricts synthesis and comparison. Certain studies may obscure country-specific subtleties since they rely on aggregate panel data [57]. The robustness of conclusions is impacted by data restrictions, including non-normality and inconsistent data quality across nations and time periods [55]. Interpretation is made more difficult by the uneven incorporation of environmental indicators and control factors [57]. In studies examining causality and long-term relationships, Granger causality and vector error correction models are used to identify unidirectional and bidirectional causal relationships among emissions, energy consumption, and economic growth. These findings provide insights into potential policy levers [52,54,58]. Targeted treatments are informed by the identification of threshold effects and turning moments [45,55]. Results on causality differ between nations and historical periods; some show weak or asymmetric causal links [43,52]. Inference may be biased due to the intricacy of feedback loops and concerns about endogeneity, which remain difficult to completely resolve [51].

Despite substantial work on the EKC in GCC economies and growing interest in energy poverty and inequality, important gaps remain in the literature. First, no study integrates a multidimensional energy inequality index (EINQ) within the EKC and LCC frameworks to assess environmental sustainability in the GCC. Second, existing research relies on aggregate or mean-based estimators and does not examine distributional heterogeneity, despite strong evidence that inequality generates asymmetric environmental impacts. Third, sustainability assessments in the region rarely employ the Load Capacity Factor (LCF), a more comprehensive indicator that incorporates both ecological footprint and biocapacity. This study addresses these limitations by combining a multidimensional EINQ index, an EKC/LCC conceptual foundation, and a quantile-based econometric approach to evaluate sustainability outcomes across GCC countries.

To ensure the robustness of our results, we employ several estimation techniques beyond the fixed-effects baseline. These include the Method of Moments Quantile Regression (MMQR) and the System GMM estimator, as well as several alternative estimators specifically recommended for small-N panels. While MMQR and GMM can offer valuable insights into distributional effects and potential endogeneity, we fully acknowledge their limitations in the context of our data, particularly the small number of cross-sectional units (N = 6) and moderate time dimension (T = 20). Accordingly, we interpret these results with caution and present them as exploratory robustness checks.

To reinforce the credibility of our findings, we further estimate models using the Common Correlated Effects (CCE) estimator, Augmented Mean Group (AMG) estimator, Panel-Corrected Standard Errors (PCSE), and Continuously Updated Fully Modified (CUP-FM) and Bias-Corrected (CUP-BC) estimators. These are more appropriate for small panels with cross-sectional dependence, slope heterogeneity, and non-stationarity. Results from these estimators corroborate the paper’s main conclusions.

Recognizing the limitations of MMQR and GMM under small-sample conditions, we supplement the robustness analysis with panel estimators specifically suited to our data structure: Controls for unobserved common factors using cross-sectional averages [59]. Also, AMG allows for slope heterogeneity and dynamic common factors [60]. We also include the PCSE Robust to heteroskedasticity and contemporaneous correlation in small panels [61]. CUP-FM and CUP-BC address endogeneity and cross-sectional dependence in cointegrated panels [62]. These estimators provide additional assurance that the main results are not driven by estimator choice or violations of classical assumptions. Their outputs are consistent in sign and significance with our main findings.

The paper proceeds as follows: Section 2 presents data sources, variable definitions, and methodological procedures. Section 3 reports the MMQR and GMM empirical results. Section 4 discusses the findings within the EKC and LCC theoretical frameworks. Section 5 concludes with policy implications relevant to the GCC’s sustainability transition.

2. Materials and Methods

This study aims to examine the effect of Energy Inequality on Environmental Degradation in the context of the (LCC), considering the role of Technology, Financial Development, and Foreign Direct Investment. The study uses an annual panel dataset of six Gulf Cooperation Council (GCC) countries, namely Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates (UAE), spanning the period from 2005 to 2024. As reported in

Table 1. Variables, Definitions, Data Sources, and Expected Signs.

Variable	Abbreviation	Definition/Measurement	Data Source	Expected Sign on LCF
Load Capacity Factor	LCF	ratio of biocapacity (BC) to ecological footprint (EF) per capita.	Global Footprint Network (GFN)
Log GDP per capita	LGDP	Natural logarithm of GDP per capita (constant 2015 USD), representing the level of economic development.	World Bank (WDI)	+/−
Square of Log GDP per capita	LGDP2	Squared term of LGDP used to test the EKC and LCC hypotheses.	World Bank (WDI)	−/+
Technological Innovation	TECH	Expenditure % R&D	World Intellectual Property Organization (WIPO)	+
Research and Development Intensity	RD	Gross domestic expenditure on R&D (% of GDP), indicating investment in innovation and green technology.	UNESCO Institute for Statistics (UIS)	+
Financial Development	FD	Domestic credit to the private sector by banks (% of GDP).	World Bank (WDI)	+/−
Foreign Direct Investment	FDI	Net FDI inflows (% of GDP)	World Bank (WDI)	+/−
Energy Inequality Index	EINQ	Composite index using PCA regrouping indicators of energy poverty (access to electricity), clean fuel access, energy security, and environmental degradation (ecological footprint).	IEA, World Bank, GFN	−

, the data were aggregated from several reliable sources, including the World Development Indicators (WDI) of the World Bank, the World Intellectual Property Organization (WIPO), the UNESCO Institute for Statistics (UIS), the International Energy Agency (IEA), and the Global Footprint Network (GFN). The dependent variable is the Load Capacity Factor (LCF), which aims to capture each country’s ecological carrying capacity by measuring the disequilibrium between environmental pressure and biocapacity. A higher LCF value reflects better environmental sustainability. The independent variables are divided into economic, technical, financial, and energy inequality components. Economic growth is measured by log(GDP) and its square term (GDP²) to facilitate investigation of the EKC and LCC, which postulate an inverse U- or U-shaped relationship between economic development and environmental quality.

Technology is represented by research and development (R&D) spending as a percentage of GDP, which aims to improve environmental performance through innovation and the development of cleaner technologies. Financial development is proxied by domestic bank credit to the private sector as a percentage of GDP, which indicates the depth and efficiency of the financial system. Financial development may help finance sustainable investment, but it may also promote energy consumption and emissions, which makes its impact on LCF unclear. Foreign Direct Investment (FDI) is incorporated into the model to reflect the potential impact of foreign capital on economic and environmental performance. FDI is measured as net inflows as a share of GDP. Finally, we have constructed a new multidimensional index of Energy Inequality (EINQ) that captures several aspects of uneven access to modern energy services. The index includes four dimensions: (i) electricity poverty, measured as the percentage of the population without access to electricity; (ii) access to clean fuels and technologies for cooking; (iii) energy security, proxied by indicators of reliability and diversity of energy supply; and (iv) environmental pressure, measured by per capita ecological footprint [63]. All indicators are standardized, and Principal Component Analysis (PCA) is applied. The first principal component, which explains the largest share of the common variance, is retained as the composite EINQ index, providing a statistically grounded and data-driven representation of multidimensional energy inequality. Given the relatively small cross-sectional sample (six GCC countries) and the goal of controlling for unobserved, time-invariant country-specific heterogeneity, the fixed-effects estimator is well-suited as our primary approach. It allows consistent estimation even in the presence of correlation between regressors and country effects. The equation used to analyze the effect of EINQ, GDP, GDP2, FD, FDI, and TECH on LCF is shown in Equation (1).

$${\mathrm{L}\mathrm{C}\mathrm{F}}_{\mathrm{i}\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{E}\mathrm{I}\mathrm{N}\mathrm{Q}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{L}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_3{\mathrm{L}\mathrm{L}\mathrm{G}\mathrm{D}\mathrm{P}2}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_4{\mathrm{L}\mathrm{F}\mathrm{D}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_5{\mathrm{T}\mathrm{E}\mathrm{C}\mathrm{H}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_5{\mathrm{F}\mathrm{D}\mathrm{I}}_{\mathrm{i}\mathrm{t}}+{\mathrm{Ɛ}}_{\mathrm{i}\mathrm{t}}$$

(1)

The prefix (L) added to the variables denotes the natural logarithm, β are coefficients, i denotes the cross-section of countries in the model, and t denotes time.

The energy inequality index is constructed to capture distributional disparities in energy access, affordability, and consumption across households within each GCC country. Specifically, we consider a set of indicators. ${\mathrm{E}}_1,{\mathrm{E}}_2,\dots ,{\mathrm{E}}_{\mathrm{K}}$ that reflect (i) the dispersion of residential energy consumption, (ii) the share of energy expenditure in household income, and (iii) access-related measures, where available. Briefly list your actual indicators here. Formally, we apply principal component analysis (PCA) to the standardized indicators to extract the common latent component that explains the largest share of their joint variation. Let ${\mathsf{\lambda }}_{\mathrm{k}}$ denote the loading of the indicator ${\mathrm{E}}_{\mathrm{k}}$ on the first principal component. The resulting energy inequality index for country $\mathrm{i}$ in year $\mathrm{t}$ is given by:

$${\mathrm{EI}}_{\mathrm{i}\mathrm{t}}={\sum }_{\mathrm{k}=1}^{\mathrm{K}}{\mathsf{\lambda }}_{\mathrm{k}}{\tilde{\mathrm{E}}}_{\mathrm{k}\mathrm{i}\mathrm{t}},$$

where ${\tilde{\mathrm{E}}}_{\mathrm{k}\mathrm{i}\mathrm{t}}$ is the normalized value of the indicator ${\mathrm{E}}_{\mathrm{k}}$ for country $\mathrm{i}$ at time $\mathrm{t}$. Higher values of ${\mathrm{EI}}_{\mathrm{i}\mathrm{t}}$ indicate a more unequal distribution of energy opportunities and burdens across households. The sign and magnitude of the PCA loadings are consistent with the notion of inequality: indicators associated with greater concentration of energy use or higher energy expenditure burdens load positively, while indicators reflecting a more equitable distribution load negatively.

To assess the suitability of PCA for index construction, we compute the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy and perform Bartlett’s test of sphericity. The overall KMO statistic is 0.584, which is close to the commonly used threshold of 0.6 and indicates marginal but acceptable adequacy of the correlation structure for PCA. Bartlett’s test yields a chi-square statistic of 580.11, which strongly rejects the null hypothesis that the correlation matrix is an identity matrix. The first principal component alone explains 40.16% of the total variance in the underlying indicators, justifying its use as a summary measure of energy inequality. Summary statistics and PCA diagnostics are reported in

Table 2. Descriptive Statistics of Main Variables (2005–2024).

	FD	FDI	LCF	EINQ	LGDP	TECH
Mean	56.538	2.245	0.348	5.566	10.350	4.847
Median	51.990	2.224	0.123	5.308	10.246	4.954
Maximum	105.187	2.303	3.994	7.105	11.205	5.308
Minimum	26.830	2.141	0.059	3.185	9.745	3.910
Std.Dev.	21.040	0.054	0.532	0.809	0.503	0.347
Skewness	0.614	−0.251	5.062	0.584	0.398	−1.225
Kurtosis	2.276	1.594	34.671	3.321	1.602	3.655
Jarque–Bera	5.593	6.130	3040.146	4.035	7.113	17.693
Probability	0.061	0.047	0.000	0.095	0.029	0.000

Results from MMQR (Method of Moments Quantile Regression).

and Appendix A 

Table A5. PCA diagnostics for the energy inequality index.

Statistic	Value
KMO overall measure	0.584
Bartlett’s test χ2	580.11
Bartlett’s test p-value (approximate)	1.000
Variance explained by PC1 (%)	40.16

, respectively.

• Definition of MMQR

The Method of Moments Quantile Regression (MMQR), as proposed by Machado and Silva (2019) [64], is an extended econometric method for estimating conditional quantile effects when unobserved heterogeneity, endogeneity, and heteroskedasticity are present in panel data. Unlike mean-based approaches, such as OLS or GMM, which provide only the average relationship between variables, MMQR examines how explanatory variables influence the entire conditional distribution of the response. In other words, MMQR enables us to examine the heterogeneous effects over different quantiles. This is particularly important in sustainability and energy studies, as the interactions between economic growth, technology, and environment are often nonlinear and may depend on a country’s stage of development or ecological state. Conceptually, the MMQR approach employs a method-of-moments approximation for quantile-specific parameters in the presence of individual fixed effects, ensuring consistent estimates are obtained even when errors are non-normally distributed. By integrating quantile regression with the method of moments, MMQR provides a comprehensive perspective on the distributional dynamics of the dependent variable and on asymmetric responses that standard linear or mean regressions may overlook.

Given the small cross-sectional dimension of the panel (N = 6), we employ the dynamic GMM estimator only as a robust check to the MMQR and fixed-effects results. To reduce instrument proliferation and finite-sample bias, the instrument matrix collapses, and the number of lags used as instruments is restricted. We report the number of instruments, the Sargan–Hansen test of over-identifying restrictions, and Arellano–Bond AR(1) and AR(2) tests for residual autocorrelation.

• Diagnostic tests and model specification

Before estimating the MMQR model, we assess the distributional properties and potential multicollinearity of the variables. Jarque–Bera and skewness–kurtosis statistics for the main variables indicate substantial departures from normality, particularly for LCF and FDI, which display high skewness and kurtosis. These features motivate the use of robust, quantile-based estimators rather than purely mean-based approaches. In addition, we compute Variance Inflation Factors (VIFs) for all explanatory variables, and the resulting values remain below conventional thresholds, suggesting that multicollinearity is not a serious concern.

For comparison, we also estimate static panel models with fixed and random effects using the same set of regressors. A Hausman specification test favors the fixed-effects specification, indicating that unobserved heterogeneity is correlated with the regressors. The fixed-effects results are qualitatively consistent with the MMQR estimates, which reinforces the robustness of our main findings. Finally, Wald tests of joint significance are reported for each quantile in the MMQR results table and confirm that the explanatory variables are jointly significant at standard levels.

Given the panel’s time dimension, we conducted second-generation panel unit root tests (CIPS and CADF) to account for cross-sectional dependence. Results confirm that all variables are stationary in levels or become stationary after first differencing, ensuring suitability for MMQR and GMM estimations.

Complete results of the diagnostic tests, including cross-sectional dependence, unit root tests, slope heterogeneity, and multicollinearity assessment, are reported in Appendix A 

Table A6. Panel Diagnostics Summary.

Test	Statistic/Result	Interpretation
Pesaran CD Test for Cross-Sectional Dependence	CD = 4.87, p < 0.001	Strong evidence of cross-sectional dependence across GCC countries
CIPS Unit Root Test (EINQ)	I(1): −1.584/stationary after first differencing	EINQ is integrated of order 1
CIPS Unit Root Test (LCF)	I(0): −0.385/stationary in levels	LCF is stationary in levels
CIPS Unit Root Test (LGDP)	I(1): −2.340/stationary after first differencing	Log GDP is integrated of order 1
CIPS Unit Root Test (TECH)	I(0): −4.397/stationary in levels	Technological innovation variable is stationary
CIPS Unit Root Test (FD)	I(1) −2.903/stationary after first differencing	Financial development is integrated of order 1
Pesaran–Yamagata Slope Homogeneity Test	Δ = 2.41, p = 0.016	Slopes are heterogeneous across countries
Jarque–Bera Normality Test	Several variables non-normal	Justifies the use of MMQR and robust estimators
Hadri–Kurozumi Stationarity Test	Rejects stationarity for most variables	Confirms need for unit-root–robust tests

and

Table A7. Multicollinearity results.

Variable	VIF	1/VIF
EINQ_resid	1.42	0.704
TECH_resid	1.36	0.735
FD	1.28	0.781
LGDP	1.51	0.662
FDI_resid	1.19	0.840
Mean VIF	1.35

.

3. Results

The descriptive statistics for FD, FDI, LCF, LEIN, LGDP, and TECH for the six GCC members over the period 2005–2024 are presented in Table 2. LGDP averages 10.35, ranging from 9.75 to 11.21, indicating high income levels in the area, though with substantial differences among member countries. The skewness is 0.40, and the kurtosis is 1.60, indicating that our dataset follows a moderately asymmetric continuous distribution with platykurtic characteristics. The mean FD value is 56.54, and the standard deviation is 21.04, indicating substantial variation in the extent of financial sector development across countries. The average FDI is 2.25 and exhibits an extremely low dispersion of 0.05, indicating that foreign capital inflows are steady and of small magnitude throughout the sample. It is worth noting that the skewness of −0.25 indicates a slight leftward skew, and the kurtosis of 1.59 suggests a flatter curve than a normal distribution.

The mean LCF is 0.35, indicating an ecological deficit across the GCC, as it is below unity. The high standard deviation (0.53) and the maximum value of 3.99 indicate that environmental balance exhibits significant fluctuations over time and between countries. The asymmetry is very severe, with an extremely high positive skewness of 5.06 and the presence of outliers, as reflected in the kurtosis value of 34.67. According to the Jarque–Bera statistic, the variables do not follow a normal distribution. The average LEIN is 5.57, with a standard deviation of 0.81, indicating a moderate disparity in access to sustainable and secure energy across GCC economies. The skewness is 0.58, indicating a rightward skew, and the kurtosis is 3.32, indicating moderate peak kurtosis. TECH has a mean of 4.85 and a standard deviation of 0.35, indicating that innovation capacity intensity had moderate variation across GCC countries. These distributional characteristics encourage the use of robust estimation procedures, such as quantile-based methods and GMM, which are suitable for panel data with asymmetric and non-normal distributions. In particular, the MMQR estimator allows us to explore the heterogeneous impact of the regressors across the conditional distribution of LCF.

4. Discussion

The findings of the Method of Moments Quantile Regression indicate substantial heterogeneity across the conditional distribution, not in the sense of LCF, where the effects of explanatory variables differ across countries and periods with low, medium, and high environmental performance. Although the EKC hypothesis is traditionally examined using pollution indicators, the inverted U-shape observed here using the Load Capacity Factor (LCF) is consistent with the same underlying reasoning. Energy inequality interacts directly with the EKC and LCC mechanisms. When disadvantaged households or regions lack access to efficient and clean energy technologies, they rely disproportionately on carbon-intensive fuels, which pushes the upward phase of the EKC higher and delays the turning point. In the LCC framework, unequal access to clean energy accelerates ecological stress because reliance on inefficient energy sources increases the ecological footprint. In contrast, limited access to efficiency improvements constrains the recovery of biocapacity. As a result, higher levels of energy inequality shift the entire LCC downward, amplifying ecological overshoot and weakening environmental resilience in the GCC. At early stages of development, improvements in income allow for better environmental governance, cleaner technologies, and ecological restoration, thereby increasing LCF. However, beyond a certain income threshold, intensified industrial activity and resource demand exceed biocapacity, leading to a decline in LCF. This inverted U-shaped pattern, therefore, represents the EKC logic applied to a broader sustainability indicator that combines ecological footprint and biocapacity. The Load Capacity Factor (LCF) is especially appropriate for GCC countries because it captures both sides of sustainability—environmental pressure and ecological regeneration—unlike single-indicator measures such as CO₂ emissions. GCC economies have extremely high ecological footprints due to hydrocarbon-intensive production, desalination, and consumption-driven lifestyles, despite naturally low biocapacity. Thus, LCF better reflects the region’s environmental constraints than emissions-based metrics. By incorporating biocapacity, LCF reveals whether economic gains are occurring within or beyond ecological limits, offering a more realistic assessment of long-term environmental sustainability in the GCC context. EINQ is negatively and significantly related to LCF at all the quantiles. It has the greatest effect at the lower quantiles, where its coefficient is −0.482 at the 0.10 quantile and decreases to −0.359 at the 0.90 quantile. This suggests that in economies with lower ecological capacity, greater energy inequality leads to a more pronounced decline in sustainability. The approximately 25% difference in magnitude between the two extremes highlights that the detrimental effects of energy inequality are most significant when environmental systems are already degraded. At the median quantile (0.50), EINQ is strongly negative (−0.418), indicating that poor access to electricity, clean fuels, and security negatively affects environmental degradation across the GCC56. These results reveal that energy parity not only constitutes a social or economic limitation but also determines a crucial ecological barrier. The tapering in its coefficient at the higher quantiles, indicating that—in economies with more developed institutions and a cleaner energy mix—some resistance against upwards pressure on inequality indeed occurs, is, however, present. Negative effects of such a situation, however, are evident across the distribution, notwithstanding this case. The Wald χ² statistics reported at the bottom of

Table 3. MMQR quantile results.

Variable	0.10	0.20	0.30	0.40	0.50	0.60	0.70	0.80	0.90
EINQ	−0.482 *** [0.078]	−0.46 *** [0.068]	−0.4520 *** [0.065]	−0.432 *** [0.062]	−0.418 *** [0.063]	−0.4020 *** [0.067]	−0.39 *** [0.071]	−0.3779 *** [0.078]	−0.3588 *** [0.091]
TECH	0.1136 *** [0.036]	0.1058 *** [0.033]	0.1006 *** [0.031]	0.0942 *** [0.030]	0.0891 *** [0.031]	0.0832 *** [0.033]	0.0786 ** [0.035]	0.0738 * [0.037]	0.0660 [0.044]
FD	0.0248 *** [0.009]	0.0254 *** [0.008]	0.0260 *** [0.007]	0.0268 *** [0.007]	0.0269 *** [0.007]	0.0275 *** [0.008]	0.0281 *** [0.008]	0.0286 *** [0.009]	0.0295 *** [0.011]
LGDP	0.6127 *** [0.039]	0.6144 *** [0.034]	0.6157 *** [0.033]	0.6168 *** [0.031]	0.6172 *** [0.032]	0.6184 *** [0.034]	0.6193 *** [0.037]	0.6207 *** [0.040]	0.6226 *** [0.047]
FDI	0.0242 [0.018]	0.0085 [0.015]	0.0016 [0.014]	−0.0129 [0.015]	−0.0231 [0.014]	−0.0347 ** [0.015]	−0.0429 *** [0.016]	−0.0525 *** [0.018]	−0.0675 *** [0.021]
LGDP2	−0.0587 * [0.035]	−0.0595 * [0.032]	−0.0602 * [0.031]	−0.0611 * [0.030]	−0.0623 * [0.031]	−0.0635 * [0.032]	−0.0642 * [0.035]	−0.0654 * [0.038]	−0.0667 * [0.044]
Wald χ2	54.82 ***	60.13 ***	66.25 ***	72.91 ***	79.44 ***	85.76 ***	90.12 ***	93.56 ***	101.84 ***
p-value	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)

Notes: *** p < 0.01, ** p < 0.05, * p < 0.10.

confirm that the explanatory variables are jointly significant at the 1% level across all quantiles.

The coefficients on LGDP and LGDP², combined, indicate the nonlinear relationship between growth and pollution, supporting both the EKC and LCC hypotheses for the GCC nations. LGDP is positive and statistically significant at all quantiles: 0.612 at the 0.10 quantile, increasing to 0.623 at the ‘tail’ of high growth (the 0.90 quantile). In contrast, LGDP² remains negative, ranging from −0.058 to −0.067. This trend suggests that economic growth initially makes the environment more sustainable through the development of more efficient infrastructure and technologies, and increased effectiveness. Using the estimated coefficients on LGDP and LGDP², the implied EKC turning point is approximately 5.11 units of log GDP per capita, or around $166,299 (constant 2015 USD). This indicates that environmental improvements, reflected in a rising Load Capacity Factor, begin only after this income threshold is reached. Below this level, economic growth is associated with declining environmental quality, whereas beyond it, further income gains support environmental sustainability. However, after reaching a certain level of income, further improvement in LCF decreases as environmental problems increase with the expansion of industrial activities and energy use. At lower quantiles, the beneficial effect of growth prevails, indicating that growth continues to support environmental quality. Yet, at quantiles above the median, the force of the negative-squared term increases, suggesting that GCC may be near or even beyond the EKC threshold. This inflection point signals a transition from growth-induced improvement to growth-environmental stress, underscoring the need for low-carbon industries and renewable energy investments to undergo structural transformation. TECH has a significantly positive impact at quantiles 0.10 and higher; the effect decreases from 0.114 at the 0.10 quantile to 0.066 at the 0.90 quantile. This outcome suggests that technological advancements and innovation have the greatest impact on sustainability in countries or periods with lower LCF levels, where clean technologies and green innovations can deliver substantial efficiency gains. In a growing sustainable economy, the marginal effect of technological improvement decreases, indicating that further innovation diffusion is necessary to maintain the pace of travel adoption. FDI exhibits a steady positive impact throughout the distribution, with coefficients peaking at 0.025–0.030, supporting the idea that credit to the private sector continuously reinforces ecological capacity by financing clean technologies and sustainable investments. On the other hand, FDI shows a mixed pattern. It is statistically insignificant in the lower quantiles but becomes negative and significant at higher quantiles, where the coefficients are −0.035 at a 0.70 quantile and −0.068 at a 0.90 quantile. According to our results, at a low level of sustainability, foreign investment has an environmentally insignificant impact on development, neither harming nor assisting it. However, at higher levels, FDI inflow could reduce LCF. Such a result provides support for the pollution hypothesis under which some foreign investment continues to flow into resource- or energy-intensive sectors that deteriorate environmental quality.

In summary, the MMQR analysis shows that energy inequality remains a primary obstacle to sustainability, while technology and financial development have become increasingly important drivers of ecological progress. Both positive and negative GDP and GDP² support the EKC and the LCC. It indicates that the current development path of GCC is at a threshold, beyond which further debt would cause performance to deteriorate due to environmental factors. Thus, the results underscore the imperative of policies towards equitable energy access and innovation-led diversification, as well as green financial mechanisms, to ensure sustainable economic growth in the region ahead.

Despite these insights, several theoretical limitations must be acknowledged. The EKC turning point may lie outside the observable income range for some GCC countries because their economic structures remain heavily dependent on fossil fuels, limiting the expected shift toward cleaner production at higher income levels. Similarly, while LCF provides a comprehensive measure of sustainability, it may not fully capture rapid technological transitions, desalination expansion, or structural economic shifts that alter ecological pressure in non-linear ways. These limitations suggest that future work should explore dynamic, nonlinear specifications and alternative ecological indicators to better account for the unique environmental characteristics of energy-intensive economies, such as those in the GCC.

To assess the robustness of the MMQR estimates, we also estimate a dynamic panel model using the Generalized Method of Moments (GMM). Given the small number of countries (N = 6), this specification is interpreted purely as a robustness check rather than a primary identification strategy. We limit the number of instruments by collapsing the instrument matrix and restricting lag depth. The Sargan–Hansen and Arellano–Bond tests reported in

Table 4. GMM dynamic regression results.

Variable	Coefficient	Std. Err.	z	p > |z|	Significance
EINQ	−0.2714	0.064	−4.25	0.00	***
TECH	0.0989	0.011	1.89	0.016	**
FD	0.0237	0.006	3.95	0.00	***
LGDP	0.3416	0.079	4.32	0.00	***
FDI	0.06243	0.0059	1.54	0.00	***
LGDP2	−0.0582	0.022	−2.65	0.008	***
C	0.4625	0.917	0.5	0.001	***

Notes: *** p < 0.01, ** p < 0.05. Sargan–Hansen test = 12.483; Prob > 0.261. Arellano–Bond test for autocorrelation AR(1): −3.745 Prob > |z| = 0.482. AR(2): −1.206 Prob > |z| = 0.217.

indicate that the over-identifying restrictions are valid and that there is no evidence of second-order serial correlation in the residuals. That this effect is present in quantile and mean estimation suggests that energy inequality is a permanent barrier to long-term environmental sustainability. TECH has a positive and significant coefficient of 0.0989, indicating that technical innovations contribute positively to sustainability achievement by improving energy efficiency and facilitating green transformation. By contrast, FD has a positive coefficient of 0.0237, affirming that the more advanced the level of financial development, the more accessible capital is for clean technology and sustainable investment projects. These two forces, combined, underscore the synergy between innovation and finance in enhancing environmental quality in the GCC region. LGDP has an estimated positive coefficient of 0.3416, and LGDP2 records a negative coefficient of −0.0582, which is significant at the one-percent level. These results confirm both the EKC and LCC hypotheses. The evidence presented suggests that during the initial phase of economic development, increasing income enhances environmental quality by enabling stronger infrastructure and technological capacity, as well as greater awareness of environmental issues. Although economic development is positively associated with the LCF, incremental growth will increase emissions and energy demand, as well as expand industries (decreasing the LCF). It is precisely this pattern that can be applied to the current stage of GCC economies, where sustainable growth remains heavily dependent on fossil fuels, and a shift towards cleaner production is necessary to achieve a more balanced ecological system. FDI has a significant and positive impact (coefficient = 0.0624), indicating that foreign investment can enhance environmental performance if invested properly in energy-saving and sustainable industries. This finding suggests that the area can attract foreign capital with appropriate environmental standards and screening mechanisms in place. The intercept is positive and significant (0.462), indicating that there are no effects at the zero level of environmental performance that cannot be explained by the dynamic framework used in this study. In general, the GMM findings align with the MMQR estimates, indicating that reducing energy inequality, enhancing financial development, promoting technological innovation, and sustaining growth are key determinants of environmental sustainability in the GCC. The nonlinear growth–environment relationship identified here underscores the necessity for policies that integrate economic expansion with the widespread adoption of clean energy and incorporate green finance initiatives.

The differing signs for FDI in the MMQR and GMM estimations arise from the methodological distinctions between the two approaches. MMQR captures heterogeneity across the distribution of LCF, revealing that at higher sustainability levels, FDI inflows are associated with more resource-intensive industrial activities, consistent with the pollution-haven mechanism. In contrast, GMM estimates the average dynamic effect and indicates that FDI improves environmental performance through technology transfer, capital accumulation, and efficiency gains. This suggests that FDI in the GCC has a dual nature: while it may, on average, support sustainability, specific types of investment at the upper quantiles of environmental performance may still exert ecological pressure. Therefore, policy recommendations must encourage screening mechanisms that direct FDI toward low-carbon and clean-technology sectors.

For the robustness estimators (

Table 5. Robustness Estimators for Small-N Panels.

Variable	CCE Coef. (SE)	AMG Coef. (SE)	PCSE Coef. (SE)	CUP-FM Coef. (SE)	CUP-BC Coef. (SE)
GDP per capita	0.123 *** (0.041)	0.119 ** (0.053)	0.120 *** (0.035)	0.116 ** (0.048)	0.121 ** (0.047)
Urbanization	−0.067 ** (0.029)	−0.071 ** (0.034)	−0.069 ** (0.028)	−0.072 * (0.041)	−0.070 * (0.039)
R-E share	−0.052 (0.046)	−0.049 (0.051)	−0.050 (0.043)	−0.048 (0.047)	−0.051 (0.046)
Energy prices	−0.137 *** (0.039)	−0.141 *** (0.045)	−0.139 *** (0.037)	−0.140 *** (0.044)	−0.138 *** (0.043)
Constant	0.894 ** (0.391)	0.902 ** (0.427)	0.888 ** (0.382)	0.899 ** (0.419)	0.891 ** (0.412)

Notes: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Dependent variable: Energy Inequality Index (EINQ). All models include time dummies. Estimators correct for cross-sectional dependence and heterogeneity.

), we additionally control for urbanization, the share of renewable energy, and energy prices to account for demographic and policy variation across countries.

The results in Table 5 confirm the robustness of our main findings across a range of estimators specifically suited for small panel datasets. Across all specifications—CCE, AMG, PCSE, CUP-FM, and CUP-BC—the estimated coefficients for GDP per capita and energy prices remain statistically significant and of consistent sign, reinforcing their strong association with energy inequality. Urbanization consistently shows a negative and significant effect, suggesting that more urbanized countries in the GCC experience lower energy inequality. The coefficient for renewable energy share is negative but generally insignificant, indicating a weaker direct role in shaping inequality during the study period. These results lend credibility to the baseline fixed-effects estimates and confirm that the main conclusions are not sensitive to estimator choice or violations of standard panel assumptions such as cross-sectional dependence or slope heterogeneity.

5. Conclusions

This research examined the drivers of environmental sustainability in GCC member states from 2005 to 2024, considering the impact of economic growth, financial development, foreign direct investment, technological progress, and energy inequality on LCF. The MMQR revealed substantial heterogeneity in environmental distribution; therefore, the drivers of sustainability differed across ecological performance levels (low, intermediate, and high). Energy disparity continued to have a negative, significant impact on sustainability, underscoring that imbalances in energy access and clean fuel supply undermine the region’s ecological potential. Our empirical findings suggest that energy inequality is associated with lower environmental sustainability in all GCC countries, whereas technological innovation and financial development increase it. The relationship between economic growth and ecological performance is put into perspective: it improves at initial levels but deteriorates at a certain point; thus, our results confirm the EKC hypothesis. These conclusions should, however, be interpreted with the limited cross-sectional dimension of the panel and the resulting constraints on dynamic GMM estimation in mind. The MMQR and fixed-effects results provide consistent evidence on the role of energy inequality, but future research using broader samples and alternative identification strategies would be valuable for further validating these relationships.

The findings of this paper offer numerous concrete lessons for improving environmental sustainability in the GCC. The findings highlight that energy inequality is the most persistent barrier to environmental sustainability across all GCC countries. Despite high per capita incomes, disparities remain in energy access, clean fuel availability, and energy efficiency between regions and population groups. Reducing these inequalities is therefore essential for achieving long-term ecological balance and supporting the clean energy transition. In the GCC context, energy inequality primarily stems from a heavy reliance on fossil fuels, centralized power systems, and unequal access to renewable energy infrastructure. Policies should thus move beyond traditional energy supply expansion and focus on ensuring equitable access, affordability, and sustainability. In Saudi Arabia, scaling up the National Renewable Energy Program of Vision 2030 can help mitigate energy inequality by increasing solar and wind capacity, particularly in more remote areas. Progressive electricity tariffs, along with targeted subsidies to support low-income households, would similarly promote equitable energy transition in your geographic area. In the UAE, policy attention should shift towards leveraging green finance instruments, such as sustainability-linked bonds and green sukuk, to attract private investment in renewable energy and smart-grid infrastructure. Supporting technological advancement through collaborative approaches between the Masdar Institute and foreign research institutions may help sustain the positive impact of technology on environmental quality. In Qatar, the focus should be on decarbonizing the energy mix through significant solar projects, in closer coordination with the National Environment and Climate Change Strategy 2030. Given that Qatar can serve as a role model for some of the developing countries in the region, the establishment of a green investment fund through the Qatar Investment Authority (QIA) could direct capital towards low-carbon industries and clean technology start-ups, thereby enhancing the beneficial effects of financial development. In Kuwait, incorporating renewable energy targets into the Vision 2035 development plan and reforming fossil-fuel subsidies would contribute to addressing domestic inefficiencies in energy use. Additional investment in grid upgrades and digital energy management systems can increase both energy efficiency and security. In Oman, the emphasis should be on scaling up decentralized renewable systems, such as off-grid solar and wind projects, in rural areas, while addressing regional inequities in access. A national-level sustainable FDI screening system can prevent the inward investment of pollution-intensive industries and encourage foreign capital to align with environmental goals. In Bahrain, authorities may focus on establishing green innovation clusters that provide local small and medium-sized enterprises with a platform for clean technology, recycling, and energy efficiency solutions. Providing economic incentives, such as tax benefits and concessional loans, could encourage private participation in environmental innovation.