Rethinking Carbon Neutrality Pathways in MENAT: Unveiling the Roles of Social Globalization, Energy Intensity, and Human Capital Through the Environmental Kuznets Curve and STIRPAT Framework

Abstract

As the world races toward carbon neutrality, the true test lies not in ambition but in implementation, particularly in regions such as the Middle East, North Africa, and Türkiye (MENAT), where energy demand is accelerating and emissions trajectories remain uncertain. Despite increasing global focus on decarbonization, the MENAT region remains empirically underexplored, with limited and often inconclusive evidence on the environmental impacts of structural factors such as energy intensity, human capital, social globalization, and financial globalization. This study addresses these gaps by integrating the Environmental Kuznets Curve (EKC) hypothesis with the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) framework, employing an empirical strategy using panel data from MENAT countries covering the period from 2000 to 2021. Utilizing a suite of robust panel estimators, our results suggest that there is a U-shaped connection between income and CO₂ emissions, which invalidates the EKC hypothesis. Additionally, energy intensity, human capital, and urbanization are found to increase emissions, whereas technological innovation, social globalization, and financial globalization contribute to CO₂ emissions reduction. The panel heterogeneous causality tests give insights on the inference causality between CO₂ emissions and its drivers. These results highlight the urgent need for MENAT economies to embed renewable energy, low-carbon technologies, and sustainability-focused policies into the core of their development agendas to prevent the intensification of emissions alongside rising income levels.

1. Introduction

Achieving net zero will depend largely on areas where economic growth is rapid and energy production is highly carbon intensive, and the Middle East, North Africa, and Türkiye (MENAT) economies sit at this frontier. Carbon emissions from industry, transport, and energy production remain the dominant driver of global warming [1,2,3], so credible mitigation hinges on sustained reductions in CO₂ emissions as a core policy priority [4,5]. However, the pathways to decarbonization differ across countries because economic structure, energy use, and technological capability vary substantially [6], with energy intensity emerging as a pivotal mechanism, increasing the elasticity of CO₂ emissions with respect to output growth and trade openness [7,8]. According to [9], high energy intensity reflects inefficient energy use, often due to outdated technologies, heavy reliance on carbon-intensive fuels, and a lack of energy-saving practices. In MENAT, hydrocarbon dependence, high and heterogeneous energy intensity, rapid urbanization, and the expansion of energy-intensive manufacturing intersect with evolving institutions, subsidy reform, commodity price volatility, and geopolitical risk [10]. At the same time, the region is scaling renewables, modernizing grids, and exploring technologies such as green hydrogen, making targeted reductions in energy intensity a high-impact and immediately actionable lever [11,12,13,14]. These features make MENAT a decisive testing ground for evidence-based policy, where improvements in energy efficiency, technology adoption, and structural change can deliver outsized gains in emissions mitigation.

Human capital is a pivotal determinant of environmental outcomes, operating through opposing channels whereby rising education and skills can spur industrial expansion and urbanization that increase energy use and emissions [15], while simultaneously strengthening innovation capacity, absorptive capability, and environmental awareness that facilitate the diffusion of cleaner technologies and efficiency improvements [16]. For the MENAT regions, the ongoing education and skills reforms are expanding the region’s innovation and absorptive capacity, positioning human capital to lower emissions via faster diffusion of efficient technologies, even as the same capabilities can accelerate energy-intensive industrial expansion without strong standards [17]. Furthermore, as MENAT nations deepen their integration into global markets, they gain increased access to foreign investment for low-carbon infrastructure, enhanced opportunities for clean technology transfer, and greater exposure to international environmental standards that can reinforce domestic decarbonization efforts [18,19]. This growing interconnectedness positions the region to leverage globalization as a catalyst for sustainable development, provided that supportive regulatory and institutional frameworks are in place. Meanwhile, globalization can also drive higher consumption patterns and industrial growth [20], particularly in MENAT regions, which may increase energy demand and emissions. The effect is therefore an empirical matter that depends on domestic policy, institutional quality, and sectoral composition, emphasizing the need to identify how human capital and globalization sharpen the region’s environmental trajectory.

In addition to these factors, ref. [1] identified technological innovation (TI) as a critical driver of CO₂ emissions. In MENAT countries, where the fossil fuel industries dominate the energy landscape [21], technological advancement towards the improvement in energy efficiency and the development in renewable energy could offer a potential solution to decouple economic growth from CO₂ emissions [22]. Additionally, technological advancements must be coupled with strategic investments and policy support to fully realize their potential in reducing CO₂ emissions. Countries like Saudi Arabia and the United Arab Emirates have already begun investing in large-scale solar projects, signaling a shift towards cleaner energy sources [23,24]. Additionally, several major cities within the region, particularly in cities like Dubai, Istanbul, and Cairo, continue to urbanize rapidly, and the demand for energy-intensive infrastructure and transportation systems continues to rise [25,26]. Without smart urban planning, urbanization can lead to increased emissions, but if managed sustainably, it offers a unique opportunity to reduce CO₂ emissions via the implementation of green building codes, energy-efficient public transportation, and renewable energy integration [27,28]. Meanwhile, understanding the role of TI and urbanization on CO₂ emissions in MENAT nations remains a major concern and gap in the current literature.

The objective of this study is to probe how energy intensity, human capital, social globalization, technological innovation, urbanization, and financial globalization affect CO₂ emissions in MENAT nations, within the broader context of the Environmental Kuznets Curve (EKC) hypothesis. Specifically, this research seeks to answer the question of to what extent do energy intensity, human capital, globalization, technological innovation, and urbanization moderate the income–emissions relationship?

The contribution of this study is both conceptual and empirical. Conceptually, we integrate the EKC hypothesis with the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) framework in a single, coherent specification that moves beyond income-centric formulations to embed key socioeconomic mechanisms. This integrated design departs from conventional EKC models by explicitly incorporating (i) energy intensity as a proxy for technique and efficiency effects, (ii) human capital as a channel for innovation, absorptive capacity, and environmental awareness, and (iii) globalization as a composition and scale force shaping production structures and trade-mediated emissions. Empirically, and to the best of our knowledge, while prior studies such as [10] adopted the load capacity curve framework whilst neglecting the STIRPAT framework, they seldom model energy intensity, human capital, and globalization jointly. This is the first study to integrate EKC and the STIRPAT framework while jointly quantifying the roles of energy intensity, human capital, and globalization to CO₂ emissions across MENAT economies. The MENAT region’s pronounced structural heterogeneity, spanning hydrocarbon exporters and diversified importers, varied institutional capacities, and rapid demographic transitions, provides a demanding test bed that typical income-only EKC specifications cannot accommodate. By identifying these distinct channels, our approach yields sharper diagnostics of the scale, composition, technique, and capability effects at work, thereby generating policy-relevant evidence on how efficiency upgrades, human capital investment, and the governance of globalization can accelerate decarbonization in MEANT.

Secondly, although a sizable body of the literature examines the drivers of CO₂ emissions in advanced economies and major emerging blocs such as Brazil, Russia, India, China, and South Africa (BRICS), Mexico, Indonesia, Nigeria, and Türkiye (MINT), and the Association of Southeast Asian Nations (ASEAN), the MENAT region remains comparatively underexamined despite its distinct production structures, energy endowments, and institutional contexts. MENAT economies are at a pivotal juncture marked by rapid urbanization, industrial upgrading, and deeper integration into global value chains, alongside exposure to commodity price volatility and geopolitical risk. A major variable for the region is energy intensity, which captures the efficiency of the capital stock and production processes and is shaped by technology adoption, pricing regimes, and sectoral composition. High and heterogeneous energy intensity across MENAT countries amplifies the emissions response to growth and trade, making it a critical mechanism through which socioeconomic and globalization forces translate into CO₂ emissions outcomes. At the same time, these countries face an urgent need to diversify away from hydrocarbons, expand renewable capacity, and lower energy intensity through efficiency improvements to meet mitigation targets. Against this backdrop, a robust assessment of how socioeconomic fundamentals, energy intensity, and globalization dynamics jointly shape CO₂ emissions in MENAT is both timely and essential for credible, region-specific policy design.

Thirdly, we employ the Panel-Corrected Standard Errors (PCSE) method as our baseline estimator. To assess robustness, we replicate the analysis using the Driscoll–Kraay standard errors (DKSE) and Feasible Generalized Least Squares (FGLS) methods. These estimators address complementary features common in macro-panels, including heteroscedasticity, serial correlation, and cross-sectional dependence, and our main results remain materially unchanged across specifications. Based on the empirical findings, the EKC hypothesis is invalid in MEANT nations, while energy intensity, urbanization, and human capital positively impact CO₂ emissions. Conversely, technological innovation and social and financial globalization mitigate CO₂ emissions. By leveraging on the MENAT economies, the evidence has practical relevance beyond the region. It offers externally valid benchmarks for developing countries with comparable structural characteristics, such as high energy intensity, rapid urbanization, and increasing global integration. Accordingly, the findings contribute to the broader discourse on sustainable development and provide an evidence base for policies that prioritize efficiency gains, innovation diffusion, and integration strategies consistent with decarbonization goals.

The paper is organized as follows: Section 2 situates our contribution within the extant literature. Section 3 details the data sources, variable construction, and econometric strategy. Section 4 reports the empirical results and robustness analyses. Section 5 concludes by distilling policy implications and outlining avenues for future research.

2. Literature Review

An extensive literature investigates the determinants of environmental pollution, typically proxied by CO₂ emissions, with particular attention to economic growth, energy consumption, urbanization, and globalization, while emphasizing heterogeneity across countries. However, the association between these factors and ecological deterioration has been highlighted to be complex in nature due to difference in context with respect to nation(s), period of study, and econometric approach.

2.1. Nexus Between CO₂ Emissions and Economic Growth

A major important determinant is economic growth, wherein the EKC hypothesis depicts that EG contributes to ecological deterioration at an initial stage before the improvement in ecological quality occurs due to economic advancement [29]. Moreover, the growth–environment nexus has been examined across nations and regions. For instance, ref. [30] was focused on MENA nations, and it exposes the positive effect of economic growth on CO₂ emissions in Middle East and North Africa (MENA) countries. Ref. [31] tested the validity of EKC in Algeria and validated the evidence of EKC but confirmed that an income turning point has not been reached, so emissions rise with income. Ref. [10] conducted both time series and panel analysis for the case of MENAT nations. They reported that the load capacity curve (LCC) is invalid for panel data analysis, whereas, for time series analysis, the LCC is valid in Egypt and Iraq but invalid in Morocco. Meanwhile, ref. [32] reported that the EKC hypothesis holds in G20 nations. Similarly, ref. [33] also confirmed the validity of EKC in African nations over the period between 1990 and 2019. Furthermore, ref. [34] observed that that LCC and EKC hold in APEC nations over the period between 1991 and 2020. In contrast, ref. [35] observed that there is an inverted N-shaped EKC in ASEAN nations, suggesting that gains can reverse at higher income without sustained policy effort.

2.2. Nexus Between CO₂ Emissions and Globalization

The role of globalization has revealed both positive and negative impact on ecological deterioration. The work by [36] studied how globalization affects CO₂ emissions in 64 nations and found that globalization positively impacts CO₂ emissions. Meanwhile, ref. [37] highlighted a different finding, in which a mixed outcome was uncovered, showing a negative and positive effect of globalization on CO₂ emissions in Nigeria. Ref. [38] also delved into the role of economic growth, renewable energy, and economic globalization on CO₂ emissions, revealing that economic growth increases CO₂ emissions, while renewable energy (REC) and economic globalization reduce CO₂ emissions. For panel dataset settings, ref. [39] show that in Asia-Pacific Economic Cooperation (APEC) nations both economic growth and economic globalization increase CO₂ emissions, while [40] find for the G7 nations that globalization induces CO₂ emissions across the distribution, with the effect being stronger at higher quantiles. Consistently, ref. [41] reported that globalization, social globalization, political globalization, and economic globalization increase CO₂ emissions in ten countries. For time series analysis, for Saudi Arabia, ref. [42] showed that trade globalization and social globalization increase CO₂ emissions, whereas political globalization and financial globalization reduce CO₂ emissions. In contrast, ref. [43] finds that globalization lowers CO₂ emissions in Türkiye, suggesting technology transfer and regulatory convergence can outweigh scale effects.

2.3. Nexus Between CO₂ Emissions and Human Capital

Prior studies have focused on the nexus between economic growth, energy usage, and human capital. A critical investigation of how economic growth and human capital affect environmental impact in Italy was conducted by [44], while the role of trade globalization and renewable energy was also explored. Using the autoregressive distributed lag (ARDL) method to analyze the dataset, they uncovered that GDP induces environmental deterioration, while human capital, REC, and trade globalization reduce environmental deterioration. Expanding on the subject matter, ref. [45] observed the role of financial development on carbon footprint (CF) in top polluting nations while including other factors like economic growth, human capital, and GDP. They uncovered that GDP contributes to CF, and the reduction in CF could be realized through the rise in REC, human capital, and financial development. An investigation on a panel of 64 BRI nations by [46] established the negative impact of human capital, renewable energy, and ICT on CO₂ emissions while demonstrating that GDP induces CO₂ emissions.

2.4. Nexus Between CO₂ Emissions and Energy Intensity

The impact of energy intensity on the environment has been examined in prior studies such as [47,48], which highlighted different outcomes. The work of [47] was focused on Iceland, and it confirmed that energy intensity negatively impacts environmental deterioration. Meanwhile, ref. [48] uncovered that energy intensity contributes to CO₂ emissions across regions and income levels. A similar study on the nexus between energy intensity and CO₂ emissions conducted in 30 Chinese provinces by [49] provided a similar perspective. The work of [50] highlighted that a higher energy intensity level positively impacts CO₂ emissions in 30 nations. Ref [47] emphasized the need to transition energy sources towards green energy sources, presenting a viable pathway for energy intensity to mitigate carbon emissions.

2.5. Nexus Between CO₂ Emissions and Technological Innovation

Technological innovation (TI) has emerged as a central determinant of CO₂ emissions by shaping their trajectory through its effects on both energy and carbon intensities. A set of prior studies highlighted that a higher level in terms of TI exerts a pivot role in decreasing CO₂ emissions, suggesting that higher technological advancement results in a less detrimental influence of human activities on the environment. Ref. [51], for instance, conducted an investigation in South Africa from 1975–2020 and discovered that TI negatively impacts CO₂ emissions, thereby revealing that technological advancement exerts a positive impact on ecological quality. This perspective can be corroborated by the work of [52], who, in their investigation in a panel of 58 selected nations, revealed that TI helps in decreasing CO₂ emissions. Ref [53] inspected a panel of G7 nations from 1990–2020, highlighting that TI improves ecological quality, whereas EG has a different impact. Research by [54], on a panel of 35 Belt-and-Road nations, confirmed that a higher level of TI induces economic growth and contributes to ecological quality. Meanwhile, this viewpoint is not without conflict. For instance, the investigation of [28] on a panel of BRICS nations highlighted that TI, EG, and urbanization increase CO₂ emissions, comparing them with the negative role of REC on CO₂ emissions.

2.6. Gap in the Literature

Prior empirical work on the environmental impact of energy intensity, technological innovation (TI), globalization, human capital, and urbanization yields mixed and sometimes contradictory results, reflecting differences in econometric design, country coverage, sample periods, and proxy selection, as well as inconsistent treatment of endogeneity, cross-sectional dependence, dynamic adjustment, and parameter heterogeneity. In the Middle East, North Africa, and Türkiye (MENAT), these challenges are amplified by structural features such as hydrocarbon dependence, rapid urban expansion, and uneven institutional capacity, which can produce context-specific responses that are not well captured by single-equation or narrowly specified models. This study addresses these gaps by providing a systematic, integrated assessment of the joint roles of energy intensity, TI, globalization, human capital, and URB in shaping CO₂ emissions across MENAT economies. We employ a harmonized empirical framework that nests Environmental Kuznets Curve and STIRPAT perspectives, incorporates interaction and nonlinearity to identify effects, and uses panel estimators. This analysis helps reconcile uneven findings in the literature and delivers context-sensitive evidence to guide sustainability pathways for MENAT economies undergoing structural transformation.

3. Data and Methods

Given that this research is centered on the MENAT region, its coverage spans between 2000 and 2021. Meanwhile, we end at 2021 because consistent post-2021 data on energy intensity and human capital are not yet available for MENAT countries. In particular, we adopted CO₂ emissions as the response variable for this study to capture the level of environmental impact. Moreover, the data were obtained from the World Bank database indicator (WBDI). Human capital is one of the core regressors in CO₂ emissions functions, and the data are sourced from the [55] database. In addition, GDP per capita is also one of the major regressors, which captures the impact of income on CO₂ emissions, and the data were sourced from the WBDI. Meanwhile, the square of GDP per capita is included in the CO₂ emissions functions to capture the effect of additional growth in income on CO₂ emissions. We used energy intensity as another primary regressor, which captures the amount of energy required per unit output in the CO₂ emissions functions, and the data were acquired from the WBDI. In addition, we adopted social globalization, financial globalization, urbanization, and technological innovation as control variables. Social globalization and financial globalization help to capture the effect of social and financial integration on CO₂ emissions, and the dataset was sourced from database of [56]. However, the WBDI provided the dataset for technological innovation and urbanization.

Table 1. Variables of the study.

Symbol	Description	Unit of Measurement	Source
CO2	Carbon emissions	Metric tons per capita	[57]
HC	Human capital	Index	[55]
GDP	GDP per capita	GDP per capita (constant 2015 USD)	[57]
EI	Energy intensity	MJ/USD2017 PPP GDP
SGLO	Social globalization	Index	[56]
FGLO	Financial globalization	Index
TI	Technological innovation	Total patent applications (sum of residents and non-residents)	[57]
URB	Urbanization	% of total population

shows a detailed description of the studied variables.

3.1. Theoretical Framework and Model Specification

The empirical foundation of this study draws on two influential theoretical constructs in environmental economics, which are the Environmental Kuznets Curve (EKC) hypothesis and STIRPAT framework. Meanwhile, we begin with the IPAT identity, which was originally proposed by [58]. The IPAT identity is denoted as:

$$I=P\times A\times T$$

(1)

where environmental impact $I$ is population $\left(P\right)$ times affluence $(A$) times technology $(T$). IPAT is an accounting identity with unit elasticities, so it is not directly estimable as a stochastic model.

Next, the Environmental Kuznets Curve (EKC) treats environmental impact as a nonlinear function of affluence. A common reduced form is:

$$\ln I_{it}={\theta }_0+{\theta }_1\ln A_{it}+{\theta }_2{\left(\ln A_{it}\right)}^2+{\epsilon }_{it},$$

(2)

Hence, it permits inverted-U shaped profiles of impact with income but does not explicitly incorporate population or technology.

Finally, the STIRPAT model is the stochastic generalization of IPAT that relaxes unit elasticities and is statistically testable and extendable:

$$I_{it}=C{P}_{it}^{\alpha }{A}_{it}^{\beta }{T}_{it}^{\gamma }{\epsilon }_{it}$$

(3)

where $I_{it}$ is environmental impact, $P_{it}$ is population, $A_{it}$ is affluence (income per capita), $T_{it}$ is technology, $C>0$ is a constant, $\alpha ,\beta ,\gamma$ are elasticities, ${\epsilon }_{it}$ is a multiplicative error term, and $i,t$ are indexes of country and year. Moreover, STIRPAT can incorporate additional variables; it nests the EKC by augmenting the specification with a quadratic in affluence.

$$\ln I_{it}=\ln a+\alpha \ln P_{it}+{\beta }_1\ln A_{it}+{\beta }_2{\left(\ln A_{it}\right)}^2+\gamma \ln T_{it}+{\epsilon }_{it}$$

(4)

For estimation, CO₂ emissions serve as the proxy for environmental impact, GDP per capita reflects affluence, technological innovation captures the technological dimension, and the population is captured though the lens of urbanization.

$${CO}_{2it}={\beta }_0+{\beta }_1{URB}_{it}+{\beta }_2{GDP}_{it}+{{\beta }_{3}GDPSQ}_{it}+{\beta }_4{TI}_{it}+{\epsilon }_{it}$$

(5)

Aligning with prior works such as [47,59,60,61], we improve this model by including human capital (HC), energy intensity (EI), and globalization (GLO) into the CO₂ function, and it is expressed as follows:

$${CO}_{2it}={\beta }_0+{\beta }_1{URB}_{it}+{\beta }_2{GDP}_{it}+{{\beta }_{3}GDPSQ}_{it}+{\beta }_4{TI}_{it}+{\beta }_5{EI}_{it}+{\beta }_6{HC}_{it}+{\beta }_7{GLO}_{it}+{\epsilon }_{it}$$

(6)

We further advance this model by including social globalization (SGLO) and financial globalization (FGLO) to develop three different models, which are expressed as follows:

$$\mathrm{M}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l} 1: {CO}_{2it}={\beta }_0+{\beta }_1{GDP}_{it}+{{\beta }_{2}GDPSQ}_{it}+{\beta }_3{TI}_{it}+{\beta }_4{URB}_{it}+{\beta }_5{EI}_{it}+{\beta }_6{HC}_{it}+{\beta }_7{SGLO}_{it}+{\epsilon }_{it}$$

(7)

$$\mathrm{M}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l} 2:{CO}_{2it}={\beta }_0+{\beta }_1{GDP}_{it}+{{\beta }_{2}GDPSQ}_{it}+{\beta }_3{TI}_{it}+{\beta }_4{URB}_{it}+{\beta }_5{EI}_{it}+{\beta }_6{HC}_{it}+{\beta }_7{FGLO}_{it}+{\epsilon }_{it}$$

(8)

$$\mathrm{M}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l} 3: {CO}_{2it}={\beta }_0+{\beta }_1{GDP}_{it}+{{\beta }_{2}GDPSQ}_{it}+{\epsilon }_{it}$$

(9)

where CO₂ indicates carbon emissions, GDPSQ, TI, URB, HC, EI, SGLO, and FGLO denote the square of GDP, technological innovation, urbanization, human capital, energy intensity, social globalization, and financial globalization, respectively. Meanwhile, GDP represents GDP per capita. Meanwhile, prior to analyzing the three CO₂ emissions functions, a logarithmic transformation was applied to each variable in the dataset to stabilize variance and reduce data heterogeneity, thereby enhancing the reliability of the regression estimates. Figure 1 shows the plot of the research framework.

3.2. Econometric Strategy

3.2.1. Cross-Sectional Dependence (CSD) Tests

It is critical to accurately identify and resolve the CSD in panel datasets to ascertain the validity and trustworthiness of the findings. It arises when there is no degree of independence among the observations made within the same cross-sections (for example, nations). This problem is often attributed to the effect of some unidentified common features affecting all units, which may differ. For this study, we adopt the Pesaran CD (PCSD) test and the Pesaran Scaled LM (PSLM) test to identify the CSD among the variables.

3.2.2. Slope Heterogeneity Test

When there are significant changes or variations among the units in a panel dataset, this is called panel heterogeneity. It shows how the traits, actions, or connections of the cross-sectional units vary. Unidentified individual-specific factors impact the response variable, which may give rise to heterogeneity. Thus, the [62] heterogeneity test was employed in our current investigation. This method is more resilient in the face of heteroscedasticity and autocorrelation. It is applicable in cases where the assumptions of homoscedasticity and serial correlation in the error terms may not be met, as it does not make these assumptions. Furthermore, the test statistics become reliable as the size of the sample expands. This testing procedure enhances the reliability of the outcome of the panel data analysis by including CSD and individual-specific effects. This problem aids the statistical analysis in identifying associations while accounting for panel heterogeneity.

3.2.3. Unit Root Tests

To evaluate the stationarity characteristics of the panel dataset, there is a need to conduct unit root tests. Precisely, these help to ascertain whether the studied series have unit root issues (i.e., non-stationarity), resulting in a convergence towards a stable mean over time. In this research, the Cross-sectional Augmented IPS (CIPS) test and the Cross-sectional Augmented Dickey–Fuller (CADF) test were adopted to assess whether the series had unit root issues. The CIPS test integrates cross-sectional data by augmenting the IPS test statistic with supplementary words that account for CSD. It is deemed superior to first-generation unit roots, which neglect to explicitly address CSD, resulting in biased outcomes. Conversely, the CADF test not only addresses cross-sectional dependency but also facilitates lag length and the coefficient constraints, thereby addressing the variation in unit root behavior across units. The CADF and CIPS tests are favored for identifying unit roots in panel data.

3.2.4. Cointegration Tests

To establish a long-term connection among variables within panel data, the cointegration test needed to be conducted. For the purpose of the cointegration test, this study applied two different panel cointegration tests, which were the cointegration testing approaches of [63,64]. The Kao panel cointegration test is an adaptation of Engle and Granger cointegration test, specifically designed for panel data analysis. It considers both cross-sectional and temporal dependency in the data. Similar to the Kao test, the Pedroni test takes into account both CSD and the time series characteristics of panel data. Furthermore, the T-statistic for the Pedroni cointegration test could be classified into two: individual-specific cointegration and group-mean cointegration. The group-mean cointegration indicates that the cointegrating link is evident for all individuals within the panel. The null hypothesis for each tests posits that no cointegration is present among the variables, while its alternate hypothesis posits the existence of a cointegrating association.

3.2.5. Long Term Estimators

Due to slope heterogeneity and CSD, this research used the PCSE regression approach of [65], which exhibits the capacity to resolve the problem of correlated errors within the panel dataset. Additionally, it reduces biasness in variables by efficiently managing unexplained heterogeneity particular to individual panels. This approach is particularly suitable for panel datasets when the cross-sections surpass the period of study. For the robustness analysis of the PCSE result, this study adopted the FGLS method and DKSE devised by [66,67], respectively. The FGLS, like the PCSE, is a modification of the OLS approach that considers panel-data-specific properties like heteroscedasticity and serial correlation. Furthermore, it accounts for individual-specific effects with either random-effects or fixed-effects models. Conversely, the DKSE computes the standard errors in regression models where there is a possibility of heteroscedasticity or correlation in the error components. Furthermore, it offers unbiased and consistent estimations of the standard errors, considering the presence of heteroscedasticity or correlation in the error terms. Lastly, the causal connection between CO₂ emissions and its regressors was examined in this current work by applying the panel heterogeneous causality test devised by [68]. Figure 2 shows the step-by-step procedure of the econometric analysis.

4. Presentation of Results and Discussion

4.1. Presentation of Results

Table 2. Descriptive statistics.

	CO2	GDP	HC	SGLO	TI	URB	EI	FGLO
Mean	0.680	3.883	0.313	1.751	2.136	1.833	0.631	1.754
Median	0.504	3.626	0.345	1.747	2.725	1.862	0.602	1.794
Maximum	1.678	4.866	0.513	1.933	3.932	2.000	0.999	1.910
Minimum	−0.510	0.000	0.000	1.486	0.000	1.419	0.243	1.302
Std. Dev.	0.556	0.672	0.130	0.106	1.318	0.146	0.150	0.121
Skewness	0.258	−2.071	−1.213	−0.368	−0.739	−1.026	0.233	−1.373
Kurtosis	1.913	14.553	3.656	2.567	2.001	3.286	2.483	4.846

reports the descriptive statistics for all variables. GDP has the largest mean and standard deviation, whereas HC records the lowest mean, and SGLO has the smallest dispersion. All series display substantial variability around their central tendencies. Skewness is negative for most variables, but CO₂ and EI are positively skewed, reflecting occasional high realizations. Kurtosis values show that GDP, HC, URB, and FGLO are leptokurtic (peaked with heavy tails), while the remaining variables are platykurtic (flatter than normal).

We assess CSD for each series using the Pesaran CD (PCSD) and Pesaran scaled LM (PSLM) statistics. The results of both tests are presented in

Table 3. Outcome of the CSD test.

Variable	Pesaran CD	Pesaran Scaled LM
CO2	21.399 *	60.034 *
GDP	6.389 *	42.761 *
GDPSQ	6.200 *	43.299 *
EI	−0.685	19.144 *
HC	32.975 *	99.162 *
SGLO	1.325	25.365 *
FGLO	32.320 *	94.547 *
TI	5.153 *	9.502 *
URB	22.137 *	76.166 *

Note: * depicts <1%.

. The PCSD test rejects the null of no CSD for all variables except EI and SGLO, whereas the PSLM test rejects the null for all variables, indicating pervasive dependence across cross-sections. These outcomes confirm the presence of CSD in our panel, which motivated subsequent tests for heterogeneity and stationarity of the series. Furthermore, we examined parameter heterogeneity using the [62] test. The findings are shown in

Table 4. Outcomes of the heterogeneity test.

	Model 1	Model 2	Model 3
Delta Tilde	−2.706 *	−3.271 *	−1.667 ***
Delta Tilde Adjusted	−3.520 *	−4.225 *	−1.853 ***

Note: * and *** depict <0.01 and 0.1.

. The outcomes confirm the rejection of the null hypothesis of homogeneity for the three models, validating the evidence of heterogeneity in the two models.

We assess the order of integration using cross-sectionally augmented panel unit-root tests (CIPS and CADF), the results of which are reported in

Table 5. Unit root test.

	CIPS		CADF
	I(0)	I(1)	I(0)	I(1)
CO2	−0.947	−3.444 *	−1.952	−2.954 **
GDP	−1.392	−3.953 *	−2.205	2.831 **
GDPSQ	−1.270	−4.101 *	−2.466	−4.140 *
HC	−1.698	−1.833	−1.211	−2.468 *
SGLO	−1.687	−3.664 *	−2.229	−3.421 *
FGLO	−2.391	−4.793 *	−2.595	−3.104 *
EI	−2.443	−4.375 *	−2.242	−3.361 *
TI	−2.543	−5.045 *	−1.948	−3.679 *
URB	−1.512	−2.472	−1.207	−30.551 *

Note: * and ** depict <1% and <5%.

. Based on the result from the CIPS unit roots test, it is evident that most variables are non-stationary in levels but stationary after the first difference (I(1)). Meanwhile, the CADF unit root test results corroborate this pattern by rejecting the null of non-stationarity for all series in first differences. Thus, the observed variables exhibit a stationarity nature.

The outcome of the [63,64] cointegration testing approaches are presented in

Table 6. Cointegration test.

(A): Kao cointegration
Panel A: Model 1
		T-statics		Prob
ADF		−2.984 *		0.001
HAC variance		0.026
Residual variance		0.037
Panel B: Model 2
		T-statics		Prob
ADF		−3.122 *		0.000
Residual variance		0.037
HAC variance		0.026
Panel C: Model 3
		T-statics		Prob
ADF		2.045 **		0.020
Residual variance		0.037
HAC variance		0.038
(B): Pedroni cointegration
Panel A: Model 1
	Within panel		Entire panel
	Statistic	Prob.	Statistic	Prob.
Modified Phillips–Perron t	0.673	0.250	−0.200	0.420
Phillips–Perron t	−12.183 *	0.000	−11.229 *	0.000
Augmented Dickey–Fuller t	−12.084 *	0.000	−11.151 *	0.000
Panel B: Model 2
Modified Phillips–Perron t	0.716	0.237	−0.647	0.259
Phillips–Perron t	−12.758 *	0.000	−14.209 *	0.000
Augmented Dickey–Fuller t	−12.624	0.000	−13.622 *	0.000
Panel C: Model 3
Modified Phillips–Perron t	−3.038 *	0.001	−1.299 ***	0.096
Phillips–Perron t	1.701 **	0.044	2.007 **	0.022
Augmented Dickey–Fuller t	1.098	0.1361	1.443 ***	0.074

Note: *, ** and *** depict <1%, <5% and <10%.

A,B, respectively. Table 6A reveals that there is evidence of a long-run interaction between CO₂ emissions and their response variables for the three models. The findings of the [63] cointegration in Table 6B show that for Model 1 (Panel A), four out of the six statistics are statistically significant. On the other hand, for Model 2 (Panel B), only three out of the six statistics are statistically significant. Moreover, for Model 3 (Panel C), only five out of the six statistics are statistically significant. Thus, a cointegrating interaction between CO₂ emissions and their response variables is evident in the three models.

The results of the PCSE method for Models 1, 2, and 3 are presented in

Table 7. Results of the PCSE estimator.

	Model 1		Model 2		Model 3
Variable	Coefficient	SE	Coefficient	SE	Coefficient	SE
GDP	−0.355 *	0.063	−0.340 *	0.064	−0.401 *	0.081
GDPSQ	0.124 *	0.009	0.119 *	0.009	0.168 *	0.012
HC	1.001 *	0.145	1.034 *	0.151	-	-
SGLO	−0.282 *	0.073	-	-	-	-
FGLO	-	-	−0.423 **	0.186	-	-
EI	0.669 *	0.062	0.593 *	0.072	-	-
TI	−0.018 **	0.008	−0.015 **	0.007	-	-
URB	0.473 *	0.108	0.747 *	0.134	-	-
Constant	−0.923	0.0294	−1.137	0.410	−0.380	0.155

Note: * and ** depict the rejection of the null hypothesis at 0.01 and 0.05, SE: standard error.

. The results show that GDP negatively impacts CO₂ emissions, wherein a 1% surge in GDP leads to a decline in CO₂ emissions by 0.355% (Model 1), 0.340% (Model 2), and −0.401% (Model 3). Moreover, GDPSQ positively affects CO₂ emissions. A rise in GDPSQ by 1% increases CO₂ emissions by 0.124% (Model 1), 0.119% (Model 2), and 0.168% (Model 3). Thus, the validity of the EKC method does not hold in MENAT nations. Furthermore, the results revealed that HC positively impacts CO₂ emissions in both Model 1 and 2, wherein every 1% rise in HC induces an increase in CO₂ emissions by 1.001% (Model 1) and 1.034% (Model 2). Thus, human capital contributes to CO₂ emissions in MENAT nations. Meanwhile, SGLO has a negative influence on CO₂ emissions, wherein a 1% rise in SGLO reduces CO₂ emissions by 0.282% (Model 1). Thus, SGLO mitigates environmental impact in MENAT nations. Likewise, we discovered that the role of FGLO in CO₂ emissions is negative, wherein a 1% increase in FGLO leads to a decline in CO₂ emissions by 0.423% (Model 2). Thus, FGLO mitigates ecological degradation in MENAT nations. Meanwhile, we noticed that EI positively impacts CO₂ emissions in Model 1 and 2. As a result, a rise in EI by 1% induces CO₂ emissions by 0.669% (Model 1) and 0.593% (Model 2). Thus, EI contributes to harmful emissions in MENAT nations. However, our findings reveal that TI negatively impacts CO₂ emissions, wherein a 1% increase in TI leads to a decline in CO₂ emissions by 0.018% (Model 1) and 0.015% (Model 2). Hence, ecological quality in MENAT nation could be enhanced by increasing the level of TI. Lastly, the role of URB on CO₂ emissions is positive. This result suggests that URB increases CO₂ emissions in MENAT nations, in which a percentage increase in URB leads to a decrease in CO₂ emissions by 0.473% (Model 1) and 0.747% (Model 2).

Furthermore, the verification of the results for the PCSE method for the three models was conducted in this study using the FGLS and DKSE approaches (see

Table 8. Results of the robustness analysis.

	Model 1				Model 2				Model 3
	FGLS		DKSE		FGLS		DKSE		FGLS		DKSE
	Coeffic	SE	Coeffic	SE	Coeffic	SE	Coeffic	SE	Coeffic	SE	Coeffic	SE
GDP	−0.385 *	0.014	−0.421 *	0.063	−0.368 *	0.017	−0.415 *	0.069	−0.406 *	0.017	−0.401 *	0.036
GDPSQ	0.133 *	0.002	0.140 *	0.006	0.129 *	0.002	0.139 *	0.009	0.168 *	0.002	0.168 *	0.004
HC	0.899 *	0.030	0.979 **	0.424	0.904 *	0.029	0.989 **	0.419	-	-	-	-
SGLO	−0.378 *	0.014	−0.329 *	0.093	-	-	-	-	-	-	-	-
FGLO	-	-	-	-	−0.322 *	0.020	−0.410 ***	0.199	-	-	-	-
EI	0.705 *	0.019	0.692 *	0.212	0.671 *	0.015	0.608 **	0.228	-	-	-	-
TI	−0.013 *	0.001	−0.019	0.017	−0.005 *	0.001	−0.012	0.015	-	-	-	-
URB	0.283 *	0.019	0.248 ***	0.134	0.441	0.023	0.489	0.117	-	-	-	-
Cons	−0.458	0.043	−0.437	0.455	−0.854	0.052	−0.707	0.419	−0.364	0.035	−0.380	0.116

Note: *, **, and *** depict the rejection of the null hypothesis at 0.01, 0.05, and 0.1.

). The results of these three models indicate that GDP negatively impacts CO₂ emissions, while GDPSQ positively impacts CO₂ emissions. This implies that at low-income levels, incremental growth is associated with lower CO₂ emissions, but beyond a threshold, growth induces CO₂ emissions again in MENAT nations. Thus, the EKC hypothesis does not hold in MENAT nations. Furthermore, the results of the FGLS and DKSE methods show that HC exerts a positive impact on CO₂ emissions in MENAT nations. However, the findings of the FGLS and DKSE methods reveal that SGLO reduces CO₂ emissions in MENAT. Likewise, these estimators disclose that FGLO reduces CO₂ emissions in MENAT nations. On the other hand, the results of the FGLS and DKSE methods confirm that EI has a negative impact on CO₂ emissions in MENAT nations. Likewise, the two methods disclose that URB increases CO₂ emissions in MENAT nations. Lastly, the two estimators confirm the negative role of URB on CO₂ emissions in MENAT nations. Thus, we observed that the results of both the FGLS and DKSE estimators corroborate the results obtained from the PCSE estimator.

Furthermore, we used the panel heterogeneous causality test of [68] to examine the causal connection between CO₂ emissions and their explanatory variables in MENAT nations.

Table 9. Panel heterogeneous causality test.

Null Hypothesis	W Stat.	Zbar Stat.	Prob.
EI → CO2	2.812	0.621	0.534
CO2 → EI	5.469 *	3.891	0.000
GDP → CO2	13.580 *	13.874	0.000
CO2 → GDP	4.138 **	2.252	0.024
HC → CO2	2377.40 *	2923.14	0.000
CO2 → HC	6.681*	5.382	0.000
TI → CO2	32.162 *	36.743	0.000
CO2 → TI	4.971 *	3.278	0.001
URB → CO2	4.648 *	2.880	0.004
CO2 → URB	1.781	−0.648	0.516
GDPSQ → CO2	13.587 *	13.883	0.0000
CO2 → GDPSQ	4.049 **	2.143	0.032
SGLO → CO2	1.198	−1.365	0.172
CO2 → SGLO	3.007	0.860	0.389
FGLO → CO2	2.551	0.299	0.764
CO2 → FGLO	2.895	0.723	0.469

Note: *, and ** depict <1%, and <5%.

shows the results of the panel heterogeneous causality test. We observed that there is a unidirectional causal association from CO₂ emissions to energy intensity, while a one-way causality is observed from human capital and URB to CO₂ emissions. Meanwhile, there is a bidirectional causality between CO₂ emissions and the variables GDP, GDPSQ, and TI. Lastly, there is no causal association between CO₂ emissions and the variables SGLO and FGLO.

4.2. Discussions

The results from the three estimators for the three models show that there is a U-shaped association between income level and CO₂ emissions. This contradicts the inverted-U EKC, which posits that CO₂ emissions rise with income up to a peak and then decline thereafter. This indicates that at the early stages of development, technique and composition effects such as efficiency improvements, modest fuel switching, and technology catch-up contribute to lower CO₂ emissions. At higher income levels, scale effects, rapid urbanization, and fossil-intensive diversification dominate and drive emissions upward, reversing the trajectory. This further reflects the fact that MENAT nations are associated with structural features such as energy mix, pricing distortions, and uneven policy stringency. To address this trend, MENAT countries must expedite their transition to renewable energy and implement more stringent environmental policies in order to maintain development and decrease emissions. This study’s outcomes offer a contrasting insight from prior studies such as [44] in Italy, [46] in 64 BRI nations, [69] in the US, and [70] in African nations. These studies established the validity of the EKC hypothesis.

Furthermore, the results revealed that HC positively impacts CO₂ emissions in both Model 1 and 2. Thus, human capital contributes to CO₂ emissions in MENAT nations. This result could be triggered by the region’s industrial sector, particularly energy-intensive industries such as oil and gas production companies. Additionally, improvements in human capital will accelerate industrialization and urbanization, which will drive higher energy usage and, in turn, lead to an increased level of harmful emissions. Moreover, improved human capital increases income levels, driving more consumption of energy-dependent products. Meanwhile, the region’s dependence on fossil fuels intensifies these effects, instigating the need to balance the development of human capital with sustainable energy practices. This result aligns with prior studies such as [15] in Latin American nations and [71] in ASEAN nations. Other studies such as [46] in 64 BRI nations and [45] in top polluting nations provide a contrasting insight by highlight the negative role of HC on CO₂ emissions.

Furthermore, SGLO has a negative influence on CO₂ emissions. Thus, SGLO mitigates harmful emissions in MENAT nations. Ref. [6] emphasized that SGLO facilitates increasing global interconnectedness, which fosters the exchange of knowledge about the usage of green energy and technologies, thereby fostering the transition away from fossil fuel dependency. Furthermore, SGLO offers exposure to environmental norms and policies, which place pressure on policymakers to implement stringent environmental regulations [72]. Additionally, social globalization boosts public consciousness about environmental issues, triggering changes in consumer behavior towards more sustainable practices. This study aligns with the study conducted using a panel analysis of 124 nations by [73], who concluded that an increasing level of SGLO induces ecological quality. Likewise, the study of [74] in China also corroborates our findings. However, ref. [6] disapproved our result by emphasizing that SGLO increases CO₂ emissions in ten selected nations.

We discovered that the role of FGLO on CO₂ emissions is negative. Thus, FGLO mitigates ecological degradation in MENAT nations. Moreover, ref. [75] argued that one of the potential benefits of FGLO entails enhancing the availability of international green financing and facilitating investment in renewable energy initiatives, thereby contributing to a sustainable environment. Furthermore, the negative role of FGLO on CO₂ emissions could be caused by the role of FGLO towards facilitating the transfer of funds from environmentally aware global investors who emphasize sustainable initiatives, promoting the development of greener energy infrastructure. Moreover, FGLO could spur countries to implement more stringent environmental restrictions to attract international investments prioritizing low-carbon businesses. Moreover, ref. [76] also provide backing for this result by revealing that FGLO contributes to mitigating the CO₂ emissions level in the United States. Likewise, ref. [77] also established the positive role of FGLO towards environmental quality in G-11 nations. Meanwhile, ref. [78] concluded that the decreasing ecological quality level in NIC nations is caused by FGLO.

Meanwhile, we noticed that EI positively impacts CO₂ emissions in Model 1 and 2. Thus, EI contributes to harmful emissions in MENAT nations. This results aligns with the work of [48] for African nations, who concluded that a rise in energy intensity contributes to a surge in CO₂ emissions. Likewise, a set of prior studies, such as [49] in 30 Chinese provinces and [50] in 30 selected nations, concluded that EI increases the level of CO₂ emissions. Meanwhile, an opposing view was provided in the study of [47] in Iceland. The authors suggested that EI negatively impacts CO₂ emissions. The energy intensity of MENAT countries increases CO₂ emissions due to their economies’ substantial dependence on fossil-fuel-based energy sources, especially oil and natural gas, which significantly contribute to carbon emissions. Major industries in the region, such as petrochemicals, construction, and heavy manufacturing, are energy-intensive and operate with relatively low efficiency, leading to high energy consumption per unit of output. The insufficient adoption of renewable energy technology and energy-efficient practices results in a proportionate increase in CO₂ emissions as these economies expand and use more energy, owing to their continued dependence on carbon-intensive energy systems [79].

However, our findings reveal that TI negatively impacts CO₂ emissions. Hence, ecological quality in MENAT nations could be enhanced by increasing the level of TI. This finding contradicts prior studies such as [28,80], which concluded that TI positively impacts CO₂ emissions in BRICS nations. However, prior studies such as [52] give backing to these findings by showing that TI reduces CO₂ emissions in 58 nations. Likewise, a study conducted in 35 BRI nations also agrees with our findings. Moreover, TI facilitates the implementation of renewable energy sources, such as solar energy [81], thereby reducing harmful emissions in MENAT nations. TI fosters improvements in energy efficiency, allowing enterprises to decrease their energy consumption per unit of production, thus reducing emissions produced by conventional fossil-fuel-based energy systems. Moreover, advancements in carbon capture and storage technology and smart grids facilitate the reduction in emissions from the current energy infrastructure [82], enabling MENAT nations to progress towards a low-carbon economy while sustaining industrial development.

Lastly, the role of URB in CO₂ emissions is positive in the two models. This reveals that the expansion of cities induces more energy demand for cooling and heating electricity and transportation, thereby driving up CO₂ emissions in the region. Moreover, the concentration of people in urban areas intensifies infrastructure development, leading to the proliferation of automobiles and the construction of energy-intensive buildings, which depend heavily on fossil fuels [83]. Furthermore, ref. [84] emphasized that the rapid acceleration of urban expansion often surpasses the implementation of sustainable urban planning, leading to inefficient energy usage, thereby intensifying harmful emissions as cities continue to grow. Moreover, the results are backed by prior works carried outby [85] in Somalia and [86] in South Asian nations, who found that URB induces CO₂ emissions. Likewise, refs. [27,87] also confirmed similar outcomes by reporting that URB induces CO₂ emissions in China. Meanwhile, the works of [88] in Sweden and [89] in G7 nations concluded that URB reduces CO₂ emissions.

5. Conclusions and Policy Remarks

Drawing on a panel of MENAT economies over 2000 to 2021, we examined how energy intensity (EI), human capital (HC), social globalization (SGLO), and financial globalization (FGLO) shape CO₂ emissions within an EKC framework while controlling for urbanization (URB) and technological innovation (TI). A series of pre-estimation test, such as a cross sectional dependence test, slope heterogeneity test, and panel unit roots, Were conducted, followed by a series of cointegration methods, namely the [63,64] cointegration testing approaches, which jointly indicate a stable long-running relationship between CO₂ emissions and the covariates. The estimations of the PCSE methods confirmed that the EKC hypothesis is invalid, indicating a U-shaped income–emissions relationship, in which emissions decline at lower income levels but increase beyond an estimated turning point, consistent with the joint significance of the quadratic income terms. Furthermore, the positive role of energy intensity, human capital, and urbanization on CO₂ emissions is observed. On the other hand, technological innovation, social globalization, and financial globalization mitigate CO₂ emissions in MENAT nations. This study also used the FGLS and DKSE estimators, which corroborate the findings of the PCSE method. Lastly, the panel heterogeneous causality test shows that there is a one-way causality flow from CO₂ emissions to energy intensity and from human capital and URB to CO₂ emissions. Conversely, there is a two-way causality pathway between CO₂ emissions and the variables GDP, GDPSQ, and TI.

5.1. Policy Remarks

Given the U-shaped relationship between GDP and CO₂ in MENAT nations, mitigation must begin before the EKC turning point through a single, capacity-calibrated policy mix. A “low-carbon-first” budget rule should be institutionalized that earmarks a predictable share of resource rents, carbon revenues, or concessional finance for competitive renewable auctions, grid flexibility (storage, demand response), and industrial decarbonization, all under a transparent MRV by strengthened regulators. Time-of-use tariffs, enforceable building codes and minimum energy performance standards, utility efficiency obligations, and concessional on-lending via national development banks for SME electrification and process upgrades should be rolled out. Simplified procurement and results-based grants should be used to deploy PAYG solar and mini-grids where administrative capacity is thin. Fossil fuel subsidy reform should be phased-in alongside lifeline tariffs and digital cash transfers to protect vulnerable households. Green public procurement should be activated to create demand for low-carbon cement, steel, and public buildings. Hard-to-abate sectors should be targeted with CCUS and high-temperature electrification, and district cooling should be integrated in new urban zones. Power should be diversified through regional interconnectors, and green hydrogen should be piloted only where logistics and offtake are credible. Financing should be blended via green bonds/sukuk, partial-risk guarantees (IsDB/EBRD/AfDB), and standardized PPP pipelines. Sequencing near-term audits and tariff reform, medium-term auctions and standards, and longer-term industrial transformation provides concrete, institution-aware steps to decouple growth from emissions before the turning point.

The prominence of human capital in both models implies that education and skills development bolster long-running sustainability only when directed toward green capabilities, such as energy auditing, efficient HVAC operation, grid management, and industrial process optimization; otherwise, rising skills risk amplifying energy demand. Because urbanization and energy intensity significantly increase CO₂ emissions, skills policy should be coupled with capacity-aware but integrated actions, including adopting and enforcing building and industrial efficiency codes, expanding mass transit with district-cooling performance standards, upgrading power plant heat rates and cutting distribution losses, tightening appliance and lighting standards, and pursuing transit-oriented development linked to affordable public transport, emphasizing minimum building and appliance standards, bus rapid transit and fleet renewal, and standardized energy service contracts for public facilities. To significantly reduce energy intensity, targeted concessional credit should be paired with performance-based incentives for industrial retrofits, and public procurement should be used to pull high-efficiency equipment into the market. The negative association between financial globalization and emissions indicates that deeper integration with global capital can finance these priorities by adopting clear green finance taxonomies, streamlining permitting clean projects, deploying partial-risk guarantees to entice private investment, and structuring partnerships that transfer technology and project-delivery expertise. Aligning financial reforms with investments in green technology and human capital while tailoring urban and industrial measures to institutional capacity enables MENAT economies to sustain growth while bending the CO₂ emissions path downward.

5.2. Limitations of the Study and Future Direction

This study is limited to MENAT economies, so future research should broaden the scope to include both developed and emerging regions such as Latin America, ASEAN, and the European Union to test external validity and uncover context-specific heterogeneity. Comparative cross-regional panels could examine whether institutional quality, energy mix, and market structure condition the income–emissions nexus, while country-specific case studies using higher-frequency data can trace policy shocks and structural breaks more precisely. Subsequent work should also diversify environmental outcomes beyond CO₂ emissions to include ecological footprint, aggregate greenhouse gases, and load capacity factor. Methodologically, employing heterogeneous panel estimators, spatial models to account for cross-border spillovers, and causal designs around policy reforms would strengthen inference. Sectoral and urban–rural disaggregation, coupled with robustness checks across alternative indicators and data vintages, would provide more actionable and generalizable policy guidance.