The Impact of Financial Efficiency and Renewable Energy Consumption on CO2 Emission Reduction in GCC Economies: A Panel Data Quantile Regression Approach

Abstract

As prominent oil producers, Gulf Cooperation Council (GCC) countries have played a significant role in the global energy market. However, as the world’s attention increasingly shifts towards environmental sustainability, understanding the implications of the GCC’s economic activities on CO2 emissions becomes indispensable. This research paper investigates the relationship between specific economic indicators and their impact on CO2 emissions in the GCC from 2001 to 2021. This study employs quantile regression, a robust statistical method that estimates the conditional quantiles of a response variable given a set of predictor variables. The findings reveal several essential insights: Financial institution efficiency is significant and negative at a 1% level at the lower (10th, −83,537.3) and higher quantiles (90th, −549,002.3). The relationship between the GDP per capita and CO2 emissions varies across quantiles, highlighting the complexity of the growth–environment nexus. Total patents exhibit a positive and significant relationship with emissions, underscoring the importance of directing innovation towards environmentally sustainable solutions. Renewable energy consumption displays a nuanced relationship with CO2 emissions, with a more substantial negative impact observed at higher consumption levels. This underscores the potential of renewable energy to mitigate emissions when integrated at scale. This study’s outcomes hold crucial policy implications for GCC countries as they seek to align economic growth with environmental sustainability. The findings emphasize the importance of fostering financial institution efficiency, promoting green innovation, and expanding renewable energy sources to reduce emissions.

1. Introduction

In an era marked by global concerns about environmental sustainability, it becomes increasingly essential to scrutinize the interplay between economic activities and their impact on the environment. The Gulf Cooperation Council (GCC), comprising Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates, is crucial in the global energy landscape. These nations, known for their prolific oil production, have witnessed substantial economic growth in recent decades. However, industrialization’s rapid expansion and reliance on hydrocarbons have raised concerns regarding carbon dioxide (CO2) emissions. This research explores the intricate relationships between various economic indicators and their influence on CO2 emissions within the GCC region.

GCC countries have experienced remarkable economic growth at the cost of environmental sustainability. Balancing economic prosperity with environmentally responsible practices is a critical challenge. However, considering diverse economic variables within the GCC context, the precise nature of this relationship still needs to be explored. This study seeks to address this gap by investigating the connection between key economic indicators and CO2 emissions in the GCC from 2000 to 2021.

The primary objective of this research is to examine the impact of specific economic factors, including financial development, financial institution efficiency, financial market efficiency, the GDP per capita, total trade in low-carbon technology products as a percentage of the GDP, all technologies (total patents), urban population growth (annual %), and renewable energy consumption (% of the total final energy consumption), on CO2 emissions in the GCC region.

Understanding this relationship holds substantial importance on several fronts. First, it aids in strategic planning as GCC countries look to diversify their economies beyond oil. Second, it encourages regional collaboration among GCC nations to adopt a unified approach to sustainable development. Third, it assists in aligning international commitments to reduce the carbon footprints made by many GCC countries with empirical evidence.

The majority of empirical studies in the field of environmental science share two fundamental flaws. One of these shortcomings is using OLS mean regression models to determine the dependent variable’s conditional mean in response to the independent variables [1]. The fact that the distribution of experimental data is heterogeneous is another shortcoming. As a result, OLS might offer inaccurate estimations [2]. To detect the relationship between CO2, the total trade in low-carbon technology products as a percentage of the GDP, all technologies (total patents), urban population growth (annual %), renewable energy consumption (% of the total final energy consumption), and the GDP per capita, this study is the first environmental experiment to apply the QR for multiple percentiles (i.e., 0.05, 0.25, 0.50, 0.75, and 0.95) against the OLS model. QR is a significant and well-established tool for planning and resource management, which could provide a meaningful explanation of this environmental relationship.

This study uses the panel data quantile regression (PDQR) approach to examine CO2 emissions and economic factors within GCC countries. This method provides a nuanced understanding of how economic factors influence CO2 emissions at various levels, a gap in the existing literature. This study reveals that economic factors like GDP, energy consumption, and trade openness have varying impacts on CO2 emissions, making it crucial for policymakers to design targeted interventions. It is the first to apply the PDQR approach to CO2 emissions/economic factors within GCC countries, providing empirical evidence that captures the differential impacts of economic determinants across the entire conditional distribution of CO2 emissions. This study also explores group differences across GCC countries, identifying and analyzing the heterogeneous impacts of economic factors on CO2 emissions across different quantiles. This provides comprehensive insights into region-specific dynamics.

This research paper is structured as follows: Following this Introduction, we delve into a literature review to provide context and insights from prior studies. Next, the Data and Methodology section illustrates the data sources, variable definitions, and statistical techniques employed. Descriptive statistics and normality tests present an overview of the dataset’s characteristics. Subsequently, the Quantile Regression section reveals the findings. Finally, we conclude with some recommendations based on our observations.

2. Literature Review

In this section, we delve into the complex interplay between climate change, economic growth, and the pivotal role of the financial sector in facilitating a transition from nonrenewable to renewable energy sources. The urgency of addressing climate change is universally acknowledged, and this review critically examines the evolution of global economies towards sustainability. The authors draw on a comprehensive analysis of several studies to provide a holistic understanding of the intricate dynamics that intertwine financial sector development with environmental conservation efforts while concurrently fostering economic growth.

2.1. The Nexus between Economic Growth, Financial Sector Development, and Climate Change

This section asserts that one of the most significant aims of the millennium is creating a clean environment for civilizations to reap the benefits of contemporary, clean technologies. If this is not the case, technologies that encourage pollution can negatively impact the environment while benefiting societies in the short and medium term. Clean technologies can contribute to a cleaner environment in the long run. In order to promote clean technology, financial sector loans should be classified; otherwise, ignorance would result in pollution and unsustainable development in this region and, by extension, the rest of the globe (Claessens and Feijen) [3].

Climate change is a pressing issue that has garnered attention worldwide, within developed and developing nations. The fundamental reason lies in the predominant dependency on nonrenewable energy sources across many economic spheres. Consequently, there is a global shift towards renewable resources. The financial sector’s capability to channel funds into sustainable energy projects is pivotal in this paradigm shift. A significant gap exists in the literature, where numerous studies have investigated the link between economic growth and CO2 emissions, often overlooking the critical role of financial sector evolution. The intertwined relationship of economic growth, environmental conservation, and renewable energy utilization is an overarching challenge for a prosperous and sustainable global economy.

These studies were bifurcated based on their focus: single-country-specific or multiple-country analyses. No singular hypothesis has emerged as dominant among the reviewed literature, especially in the context of GCC countries. However, the majority underscored the growth or feedback hypotheses, suggesting that GCC energy policies predominantly aim to ensure energy availability for industrial growth and development. Consolidating renewable energy solutions was posited as a strategic move to achieve sustainable development while addressing environmental challenges in the GCC region. Here are some highlighted research findings:

Al-Iriani utilized the Pedroni panel cointegration method, examining the nexus between energy consumption and GDP in GCC nations from 1971 to 2002. The study corroborated the conservation hypothesis [4].

Mehrara investigated the relationship between per capita energy consumption and GDP across 11 oil-exporting countries. The results endorsed the conservation hypothesis [5].

Squalli explored the link between electricity consumption and economic growth within OPEC countries. Findings suggested a bidirectional relationship for Saudi Arabia and Qatar and a one-directional one for Kuwait [6].

Mahadevan and Asafu-Adjaye focused on the energy–economic growth dynamic in net energy-exporting nations. Their findings resonated with the feedback hypothesis [7].

Hafeez et al. explored the impact of financial efficiency and environmental innovations on renewable energy consumption and CO2 emissions in highly polluted Asian economies. Using the ARDL-PMG model, they find that environmental innovations and financial institution efficiency positively influence renewable energy consumption and reduce carbon emissions [8].

Al-Mulali, Ozturk, and Lean [9] investigated the relationships between economic growth, urbanization, trade openness, financial development, and renewable energy on European pollution levels. Their findings highlight the complex interplay between these factors and suggest that economic policies must balance growth and environmental sustainability.

Hamdi, Sbia, and Shahbaz [10] explored the nexus between electricity consumption and economic growth in Bahrain, emphasizing the significant role of electricity as a driver for economic activities. They argue that policies to reduce electricity consumption should consider potential negative impacts on economic growth.

Salahuddin, Gow, and Ozturk [11] examined the long-term relationship between economic growth, electricity consumption, carbon dioxide emissions, and financial development in Gulf Cooperation Council (GCC) countries. Their study confirms the robustness of these relationships, suggesting that economic growth and financial development are crucial for sustainable energy policies in the region.

Global sentiment on climate change underlines the urgent need to understand and mitigate its impacts. The financial industry emerges as a catalyst in this scenario, offering innovative green financial solutions tailored to environmentally conscious entities. Consequently, the phrase “testing clean energy investment and financial development as determinants of environmental sustainability” has become a focal point in climate change discourse.

2.2. Recent Empirical Studies

In this subsection, we can outline some recent studies:

Akbulut implemented a panel threshold regression model to explore financial strength’s influence on carbon emissions among 11 CEE countries that EU member states which were part of the former Eastern bloc. The following countries are included Bulgaria, Czech Republic, Estonia, Hungary, Lithuania, Latvia, Poland, Romania, Slovenia, and Slovakia) from 1995 to 2018. Findings indicated financial depth’s dual role in enhancing and mitigating carbon emissions [12].

Mehmood analyzed the banking sector’s influence on CO2 emissions within the N-11 countries—Bangladesh, Egypt, Indonesia, Iran, Mexico, Nigeria, Pakistan, Philippines, South Korea, Turkey, and Vietnam—emphasizing the role of GDP, clean energy, and nonrenewable energy. The study highlighted the ambivalent impact of bank development on environmental sustainability [13].

Cheiu et al. probed the effect of financial development on the relationship between OECD RE and CO2 emissions. Their results depicted the dichotomous influence of stock market growth and banking sector evolution on CO2 emissions [14].

Duc Hong study evaluated the interconnected roles of trade liberalization, financial evolution, and urbanization. It unveiled multifaceted causality relationships within the examined parameters [15].

Anton et al. assessed the impact of financial development on renewable energy consumption among 28 EU countries. They discerned a positive correlation between financial development and renewable energy uptake [16].

Eregha et al. investigated the efficacy of environmental regulations on the ecological footprint within N11 economies. Their findings cast doubt on the effectiveness of current environmental policies [17].

Amin et al. examined a gamut of economic and environmental factors across South and East Asian nations. The study revealed the multifarious impacts of these parameters on CO2 emissions [18].

Ali et al. used advanced econometric techniques to establish long-term links. Their results indicated various positive impacts on green finance, except for inflation [19].

Zahoor delved into the interplay between renewable energy investment, financial growth, and environmental sustainability in China. The study unraveled the inverse relationship between clean energy investment and environmental degradation [20].

Dogan, Taskin, and Alharthi [21] provided a systematic review of the relationship between energy consumption and economic growth in GCC countries. They find that while energy consumption is a key driver of economic growth, it contributes significantly to environmental degradation. Their review underscores the need for policies that promote energy efficiency and renewable energy sources.

Eller and Medlock [22] discussed the implications of energy demand on economic development. They argue that as countries develop, their energy consumption patterns change, which can have significant implications for energy policy and environmental sustainability.

Nathaniel et al. [23] examined the impact of renewable energy consumption on environmental degradation in MENA countries. Their study shows that renewable energy consumption can significantly reduce environmental degradation, highlighting the importance of transitioning to sustainable energy sources.

Magazzino and Cerulli [24] investigated the relationship between renewable energy consumption and economic growth in MENA countries using a bootstrap rolling window approach. They find a positive relationship, suggesting that renewable energy investments can drive economic growth and improve environmental outcomes.

Koenker [25], Graham et al. [26], and Candy [27] presented methodologies for applying quantile regression to panel data. These approaches are particularly relevant for analyzing the relationships between economic variables and environmental outcomes across different countries and over time. They emphasize the importance of considering the entire data distribution rather than focusing solely on the mean.

2.3. Research Gap

Despite the substantial research in the world and GCC context, a notable empirical research gap exists. Only some studies have deployed quantile regression techniques to analyze the relationship between economic variables and CO2 emissions in GCC countries, with some focus on financial efficiency and renewable energy consumption. This approach is crucial because it allows for a nuanced understanding of how economic factors impact environmental outcomes across different quantiles of CO2 emissions distribution. Additionally, comprehensive studies considering various economic indicators and their collective influence on CO2 emissions in the GCC are scarce.

2.4. Hypothesis

Building on the theoretical framework and insights from the existing literature, as well as the empirical results presented above, we propose the following hypotheses for this study:

Hypothesis 1 (H1). 

An inverted U-shaped relationship exists between per capita GDP (income) and carbon dioxide (CO2) emissions in Gulf Cooperation Council (GCC) countries based on the Environmental Kuznets Curve (EKC) hypothesis. This hypothesis suggests that in the early stages of economic growth, CO2 emissions tend to increase due to industrialization and higher energy consumption. However, as economies mature and per capita income rises further, there is a growing emphasis on environmental protection, adopting cleaner technologies, and implementing stricter environmental regulations, leading to decreased CO2 emissions.

Hypothesis 2 (H2). 

The turning point (income threshold) at which GCC countries transition from increasing to decreasing CO2 emissions levels will vary among the member states due to their differing economic structures, energy consumption patterns, and policy approaches.

Hypothesis 3 (H3). 

Factors beyond income, such as energy consumption, urbanization, and technological advancements, significantly influence CO2 emissions in GCC countries due to the complex and multifaceted nature of the factors beyond income, such energy consumption, urbanization, and technological advancements, which can influence CO2 emissions in varied ways. For instance, increased energy consumption might lead to higher emissions, whereas advancements in renewable energy technologies can offset this by reducing emissions. Similarly, urbanization can lead to higher emissions due to increased industrial and transportation activities, but it can also facilitate more efficient energy use and waste management practices. Given these diverse influences, it is prudent not to predict a single direction but to explore the relationships empirically through quantile regression analysis.

3. Data and Methodology

Quantile regression is a statistical method used to estimate the conditional quantiles of a response variable given a set of predictor variables. It was first introduced by Koenker and Bassett in 1978. The authors employed quantile regression with dynamic analysis. Since the models used in this study offer different outcomes for quantile levels and sequence-wise models, they are better than other methods. This technique has the potential to yield more diverse outcomes. Rather than estimating the mean of the studied variables, quantile regression estimates their median. Hence, quantile regression allows for estimating different quantiles, such as the median, lower, or upper quantiles.

Line [28] and Koenker [29] discuss the application of quantile regression in R, highlighting its usefulness in understanding the impact of different variables across various distribution quantiles. Quantile regression provides a more nuanced understanding than traditional mean regression approaches, which is particularly useful in environmental and economic studies where the impacts of variables can vary significantly across different levels of the outcome variable.

StataCorp [30] and the R Core Team [31] provide software tools that facilitate complex statistical analyses, including quantile regression. These tools are essential for conducting robust and reliable econometric environmental and economic research analyses.

The quantile regression methodology minimizes a loss function based on the absolute deviations between the observed and predicted quantiles. This approach is robust to outliers and does not assume a specific distributional form for the data. Estimating quantile regression coefficients can be achieved using various algorithms, such as linear programming or iterative reweighted least squares.

Quantile regression has several advantages over OLS regression. It provides a more comprehensive understanding of the relationship between the predictor variables and different parts of the conditional distribution of the response variable. This is particularly useful when the data have heteroscedasticity or the relationship between the variables is not linear [32]. Quantile regression also allows for analyzing distributional changes over time or across different groups. In recent years, quantile regression has gained popularity in various fields, including economics, finance, environmental studies, and healthcare. It has been used to analyze income inequality, asset pricing models, climate change impacts, and healthcare outcomes, among other applications. Quantile regression is a powerful tool providing valuable insights into the relationship between variables across different parts of the conditional distribution [33]. Its flexibility and robustness make it a valuable methodology for researchers in various disciplines.

Firstly, quantile regression allows for the estimation of different quantiles of the conditional distribution of the response variable, such as the median; this provides a more comprehensive understanding of the relationship between the predictor variables and different parts of the distribution [34].

Secondly, quantile regression is robust to outliers and does not assume a specific distributional form for the data. This makes it particularly useful when the data have heteroscedasticity or the relationship between the variables is not linear [34].

Thirdly, quantile regression allows for analyzing distributional changes over time or across different groups. It can capture how the relationship between the variables varies at different points in the distribution, providing insights into potential heterogeneity or conditional effects.

Lastly, quantile regression is flexible and can handle non-normal and skewed data, making it suitable for analyzing real-world datasets that may not meet the assumptions of OLS regression. Quantile regression provides a more comprehensive and robust analysis of the relationship between variables, allowing for a deeper understanding of the data’s conditional distribution and potential heterogeneity.

This paper employed quantile regression techniques to analyze the relationship between economic variables and CO2 emissions in GCC countries. It takes Carbon Dioxide emissions (CO2) as the dependent variable. The independent variables are listed in the

Table 1. Variable and data sources.

Symbol	Variable	Definition	Sources
CO2	CO2 emissions	CO2 emissions (kt)2	https://databank.worldbank.org/
FI	Financial Institutions Efficiency	Financial Institutions Efficiency Index	https://data.imf.org/?sk=f8032e80-b36c-43b1-ac26-493c5b1cd33b
FM	Financial Markets Efficiency	Financial Markets Efficiency Index	https://data.imf.org/?sk=f8032e80-b36c-43b1-ac26-493c5b1cd33b
FD	Financial development	Financial Development Index	https://data.imf.org/?sk=f8032e80-b36c-43b1-ac26-493c5b1cd33b
GDP	GDP per capita	GDP per capita (constant 2015 US$)	https://databank.worldbank.org/
AT	All technologies	All technologies (total patents)	https://stats.oecd.org/Index.aspx?
TR	Total trade in low-carbon technology products	Total trade in low-carbon technology products as a per cent of GDP	https://stats.oecd.org/Index.aspx?
pop	Urban population growth	Urban population growth (annual %)	https://databank.worldbank.org/
ECN2	Renewable energy consumption	Total energy consumption from nuclear renewables and other	https://databank.worldbank.org/

Note: own construction.

below. Each independent variable has been selected based on its potential impact on carbon dioxide emissions, supported by the literature and empirical studies.

Equations illustrate the study variables.

$$\mathrm{E}(\mathrm{yit}|\mathrm{xit},{\alpha }_{\mathrm{i}})=\beta {\mathrm{X}}_{\mathrm{i}\mathrm{t}}^{\mathrm{T}}+{\alpha }_{\mathrm{i}}$$

(1)

where it denotes the logarithm of the number for country yi at year t, it = (${\mathrm{X}}_{\mathrm{i}\mathrm{t},1}^{\mathrm{T}}$, …, ${\mathrm{X}}_{\mathrm{i}\mathrm{t},\mathrm{p}}^{\mathrm{T}}$), ^T is ap ×1vector of independent variables. Ai denotes the (unobserved) county effect, which controls time-invariant sources of unobserved heterogeneity, such as geography and formal institutions.

$$\mathrm{QCO}2{}_{\mathrm{it}}(\tau |\mathrm{xit},{\alpha }_i)={\beta }_{1\tau }{\mathrm{FD}}_{\mathrm{it}}+{\beta }_{2\tau }{\mathrm{FI}}_{\mathrm{it}}+{\beta }_{3\tau }{\mathrm{FM}}_{\mathrm{it}}+{\beta }_{4\tau }{\mathrm{GDP}}_{\mathrm{it}}+{\beta }_{5\tau }{\mathrm{ECN}}_{\mathrm{it}}+{\beta }_{6\tau }{\mathrm{POP}}_{\mathrm{it}}+{\beta }_{7\tau }{\mathrm{TR}}_{\mathrm{it}}+{\beta }_{8\tau }{\mathrm{AT}}_{\mathrm{it}}+{\alpha }_i$$

(2)

Here, the authors represent individual countries in the GCC, and t represents time. ${\alpha }_i$ is the error term and at nine quantiles, namely τ = 0.10, 0.20, 0.30, 0.40, 0.50, 0.60, 0.70, 0.80 and 0.90. The expected relationships between various economic indicators and CO2 emissions in the GCC (Gulf Cooperation Council) countries from 2001 to 2021 provide a brief analysis of each indicator’s potential relationship with CO2 emissions based on the existing literature and empirical findings.

$\beta 1>0,\mathrm{o}\mathrm{r}\beta 1<0$: For financial development, the coefficient is projected to be either negative or positive, indicating that it can lead to more efficient resource allocation and the promotion of green technology and practices. However, it can also increase industrial activity, increasing emissions [35].

$\beta 2>0,\mathrm{o}\mathrm{r}\beta 2<0$: Efficient financial institutions predict the coefficient as negative or positive, implying that it might promote sustainable initiatives while simultaneously financing carbon-intensive sectors [36].

$\beta 3>0,\mathrm{o}\mathrm{r}\beta 3<0$: In inefficient financial markets, the coefficient is expected to be negative or positive and could facilitate funds transfer to cleaner technologies. However, it can also promote industries, increasing emissions depending on market priorities [37].

β4 > 0: A higher per capita GDP coefficient is expected to be positively associated with increased CO2 emissions; as countries grow wealthier, they initially have increasing emissions due to industrialization. However, after reaching a certain income level, countries might invest more in clean technology and enforce stricter environmental regulations, thus potentially reducing emissions.

β5 < 0: Renewable energy consumption (% of total final energy consumption) contributes to reducing CO2 emissions through efficient energy resource allocation. The coefficient is expected to be negative. This implies that increasing renewable energy consumption reduces CO2 emissions through more efficient energy resource allocation.

β6 > 0: The coefficient is expected to be positive for urban population growth (annual %). This suggests that increased urban population growth is associated with increased CO2 emissions.

β7 $<0$: Total trade in low carbon technology. If a country imports a higher share of environmental goods, the coefficient is expected to be negative. It suggests a transition towards cleaner technology and practices, potentially lowering emissions.

β8 $<0$: The coefficient is expected to be negative for all technologies (total patents). This indicates that an increase in total patents, reflecting technological innovation, is associated with a decrease in CO2 emissions

Originally presented in the seminal study by Koenker and Bassett [34], quantile regression is an expansion of the classic least-squares estimation of the conditional mean to a group of models for various conditional quantile functions. The ${\tau }^{th}$ regression quantile estimate βˆ(τ) is the solution to the following minimization problem, where t is a parameter (0 < τ < 1), which represents quantile size. According to Bao, Lee, and Saltoglu [38], the primary benefit of quantile regression over ordinary least-squares (OLS) analysis is its ability to examine the whole distribution. On the other hand, conditional distributions can only be partially described by conventional OLS regression, which only allows researchers to estimate the conditional mean and median located at the center of the distribution [39]. Additionally, quantile regression is more forgiving than OLS due to its greater insensitivity to outliers and distributions with heavy tails. In order to construct the regression model, we use penalized panel quantile regression with a fixed effect. The penalized version of the estimator is then solved using the following equation:

$$\beta {}^{^}(\tau )={arg}_{\beta \in R^K}\min {\sum }_{i\in \left\{i:yi\ge x_i\beta \right\}}\tau \left|y_i-x_i\beta \right|+{\sum }_{i\in \left\{i:yi<x_i\beta \right\}}(1-\tau )\left|y_i-x_i\beta \right|$$

(3)

where τ is a parameter (0 < τ < 1) representing quantile size. According to Bao, Lee, and Saltoglu [38], the main advantage of quantile regression over OLS is its ability to analyze the entire distribution. In contrast, traditional OLS regression only allows researchers to approximate the conditional mean and conditional median located in the center of the distribution, providing only an incomplete description of a conditional distribution. Mosteller and Tukey [39]. Furthermore, quantile regression is more forgiving than OLS because it is relatively insensitive to outliers and distributions with heavy tails. Thus, we employ the penalized panel quantile regression with a fixed effect to build up the regression model and solve the penalized version of the estimator by the following equation.

$$(\beta {}^{^}({\tau }_{\mathrm{K}},\lambda ),\left\{{\alpha }_i(\lambda )\right\}{n}_{i=1}=\mathit{arg}\min ({\sum }_{k=1}^K{\sum }_{t=1}^T{\sum }_{i=1}^n{\omega }_k{P}_{\tau \mathrm{K}}(y_{it}-x_{it}^T\beta ({\tau }_k)-{\alpha }_i)+\lambda {\sum }_{i=1}^n\left|{\alpha }_i\right|$$

(4)

where $\beta {\tau }_k$ (u) = u(tk − I(u ≤ 0)) is the standard quantile loss function (Koenker, 2004 [25]) and ${\omega }_k$ is the relative weight given to the kth quantile, which controls the contribution of the kth quantile to the estimation of the fixed-effects panel data mode.

4. Results and Analysis

4.1. Descriptive Statistics

Descriptive statistics summarize the central tendency, dispersion, and shape of a dataset’s distribution, presented in the

Table 2. Descriptive statistics.

Stats	CO2	FD	GDP	FI	FM	AT	TR	ECN2	POP
Mean	134,507.7	0.432779	33,441.54	0.614043	0.372024	60.11748	1.403897	0.00186	4.305412
Median	79,189.9	0.432155	23,671.85	0.676653	0.240786	6.60625	1.317068	3.71 × 10−5	3.13081
Maximum	565,190	0.58593	73,493.3	0.780721	1	758.08	3.282995	0.076731	19.61203
Minimum	15,876.5	0.264384	15,512.7	0.285995	0.011617	0.33	0.022589	−0.002029	−2.605206
Std. Dev.	145,738.1	0.074635	16,637.95	0.13143	0.30993	145.7936	0.691809	0.009169	4.048381
Skewness	1.770123	−0.174216	0.854789	−0.954099	0.609715	3.565122	0.743584	6.870983	1.664043
Kurtosis	5.034254	2.489945	2.35958	2.746825	2.030884	15.35824	3.061748	51.40331	6.509573

Source: Authors’ calculation.

below.

In the given dataset, the mean CO2 emission is approximately 134,507.7 kt, characterized by a pronounced rightward skewness and elevated kurtosis, signaling the preponderance of lower values with a spread of higher values. The Financial Institutions Efficiency Index exhibits a leftward skewness, suggesting a prevalence of higher values with a mean of 0.614. The Financial Markets Efficiency Index possesses a mild rightward skewness with a mean of 0.372, and the GDP per capita has a slightly rightward skewness with an average of 33,441.54. Remarkably, the variables ‘Total Patents’ and ‘Renewable Energy Consumption’ display extremely high rightward skewness and kurtosis, highlighting numerous extreme values.

The results from normality tests, predominantly relying on assessments of skewness and kurtosis, unveil that most variables diverge from normal distributions, with p-values below the 0.05 threshold. Specifically, the variables, including CO2, fi, fm, GDP, at, tr, pop, and ecn2, display substantial deviations from normality in their distributions, entailing potential implications for subsequent analyses utilizing methods premised on the assumption of normality.

The detailed results of Levin, Lin, and Chu t*, Breitung t-stat, Im, Pesaran, and Shin W-stat, and ADF—Fisher chi-square tests are shown in

Table 3. Panel unit root tests.

Variables	Level				First Difference
	Levin, Lin and Chu t *	Breitung t-Stat	Im, Pesaran and Shin W-Stat	ADF—Fisher Chi-Square	Levin, Lin and Chu t *	Breitung t-Stat	Im, Pesaran and Shin W-Stat	ADF—Fisher Chi-Square
CO2	3.7363 (0.9999)	−4.0657 *** (0.0000)	4.81104 (1.0000)	0.32347 (1.0000)	−3.98180 *** (0.0000)	1.27221 (0.8984)	−2.25793 * (0.0120)	33.5467 *** (0.0008)
Fi	−0.7393 (0.2299)	−2.34471 *** (0.0095)	−0.63224 (0.636)	12.419 (0.4126)	−10.296 *** (0.0000)	−6.97417 *** (0.0000)	−9.01634 *** (0.0000)	78.2641 *** (0.0000)
Fd	2.98683 (0.9986)	−1.36721 (0.0858)	1.07577 (0.8590)	5.58413 (0.9356)	−7.56227 *** (0.0000)	−7.10773 *** (0.0000)	−8.38458 *** (0.0000)	70.490 *** (0.0000)
Fm	0.82168 (0.7944)	1.48981 (0.9319)	1.0255 (0.8475)	7.8403 (0.7975)	−8.9639 *** (0.0000)	−6.9493 *** (0.0000)	−7.3531 *** (0.0000)	64.4648 *** (0.0000)
Gdp	1.76007 (0.9608)	1.37743 (0.9158)	2.0642 (0.9805)	3.63452 (0.9892)	−3.13688 *** (0.0009)	−2.64493 *** (0.0041)	−1.9424 (0.0260)	22.136 ** (0.0360)
Tr	0.6865 (0.7538)	−0.50047 (0.3084)	0.8230 (0.7948)	31.9863 *** (0.0014)	−11.098 *** (0.0000)	−6.25717 *** (0.0000)	−10.4500 *** (0.0000)	88.2078 *** (0.0000)
At	0.1310 (0.5521)	−1.85462 ** (0.0318)	0.4355 (0.6684)	7.7786 (0.8022)	1.4324 (0.9240)	−2.31317 ** (0.0104)	−0.8973 (0.1848)	91.8424 *** (0.0000)
Pop	2.42895 (0.9924)	3.48707 (0.9998)	3.1279 (0.9991)	1.7767 (0.9997)	−3.3257 *** (0.0004)	−3.9807 *** (0.0000)	−1.7989 ** (0.0360)	20.497 ** (0.0582)
ecn2	4.4477 (1.0000)	6.16539 (1.0000)	6.0540 (1.0000)	0.7796 (1.0000)	−3.82540 *** (0.0001)	0.5807 (0.7193)	−2.0383 ** (0.0208)	28.747 *** (0.0043)

Note: Standard errors are in parenthesis. *** p < 0.001, ** p < 0.05, and * p < 0.1.

to demonstrate that every factor in our model has an I(0) or I(1) process, indicating that every series is stationary at its levels or the first difference. The authors use the previously stated panel unit root test results to see if there is a long-term relationship across the variable. Recent econometrics research has introduced generation unit root tests to control the issue of cross-sectional dependence throughout panel units [40,41].

The authors performed dynamic analysis on collected data using nine models in different periods.

Table 4. Renewable energy and CO2 relationship.

Variable	M-1	M-2	M-3	M-4	M-5	M-6	M-7	M-8	M-9
Period	2021–2020	2020–2019	2019–2018	2018–2017	2017–2016	2016–2015	2015–2014	2014–2013	2013–2012
Fi	−1702.121 (7524.118)	1092.601 (7905.362)	−854.3858 (9325.619)	506.3395 (10,684.5)	980.3714 (12,397.37)	8223.686 (14,392.32)	14,013.52 (16,559.74)	19,411.11 (17,816.25)	−22,282.43 (19,389.39)
Fd	29,007.87 (18,493.67)	24,213.11 (19,351.37)	24,649.48 (21,012.08)	15,685.95 (23,921.19)	20,049.24 (27,537.05)	5992.077 (31,752.06)	−4935.701 (35,844.46)	−13,976.4 (38,349.44)	3544.443 (41,549.47)
Fm	899.6757 (4570.033)	572.3488 (4950.002)	2184.765 (5379.091)	6122.058 (5795.661)	3512.103 (6590.809)	7464.449 (7437.67)	9921.395 (8428.047)	14,345.19 (9034.014)	14,031.75 (9970.689)
Gdp	0.2515758 (0.1032343)	−0.2211375 ** (0.1101319)	−0.2451776 ** (0.1175154)	−0.2540144 ** (0.1230287)	−0.2167689 (−0.2167689)	−0.1713489 * (0.1362349)	−0.1618858 (1418558)	0.1650761 (0.1420702)	−0.247454 (0.1593102)
Tr	−1239.236 (1235.452)	−1374.116 (1295.465)	−1564.843 (1373.09)	−1404.534 (1470.246)	−1338.478 (1575.444)	−2488.661 (1609.569)	−2659.711 (1730.392)	−2961.522 (1894.782)	−3405.381 (2082.315)
At	−54.66656 *** (9.194927)	−53.51432 *** (11.34117)	−64.37327 *** (14.502)	−83.85143 *** (16.65998)	−73.13967 *** (20.25201)	−89.80826 *** (22.61398)	−100.169 *** (25.2357)	−131.269 *** (27.13217)	−106.5161 *** (30.77096)
Pop	361.7188 ** (152.6018)	368.0806 ** (161.7725)	403.8418 ** (174.9024)	408.8782 ** (0.646)	366.0271 ** (196.675)	395.4598 ** (199.1936)	400.5054 ** (224.0622)	285.4512 (262.3137)	437.2312 (333.5236)
ecn2	2567.129 (6428.038)	2403.447 (6705.235)	2443.179 (7066.791)	325.0101 (7319.28)	3459.337 (8019.85)	6656.735 (8092.902)	6543.821 (8453.006)	9076.003 (8485.313)	10,601.04 (9062.235)
0 bs.	126	120	114	108	102	96	90	84	78

Note: This table shows the beta coefficients and standard error in parenthesis. *** p < 0.001, ** p < 0.05, and * p < 0.1.

shows the findings for the year. Standard deviations are given in parentheses. Models 1 to 6 indicate a significant negative effect on the GDP per capita and all technologies (total patent) on CO2 with −0.2515758 ** and −54.66656 ***, respectively. The result may be attributed to the high-income level of GCC countries, which may enable them to increase investment in clean technologies, enforce stricter environmental regulations, and reduce emissions. Other variables regarding technologies have negative effects, demonstrating that an increment in them reflects innovative development and is related to a diminishing CO2 outflow. On the other hand, models 1 to 9 indicated an expected significant positive impact of urban population growth on CO2 emissions with a value of a coefficient of 361.7188. CO2 is shown to have a considerable and detrimental impact on GCC countries’ investment. The following models in diverse time zones showed an insignificant but positive relationship between FD, FM, ECN2, and CO2, implying that financial institutions and the market efficiency environment need more progress to reduce CO2 emissions. Different beta values are shown in the case of FI in different models concerning periods. If GCC countries want to reduce their CO2 emissions, they must invest more in renewable energy projects. The economic indicator, GDP per capita, also exerted a negative pressure on CO2 emissions with a −0.2515758 *** beta value in model 1. Following this, the authors examined the data through the quantile regression technique and examined the impact of Fd, Fi, Fm, GDP, at, tr, pop, and ecn2 on CO2 in GCC countries.

4.2. Quantile Regression

Indeed, in the quantile regression analysis for the given dataset, the influence of several independent variables on CO2 emissions is investigated across different quantiles, which provides a more comprehensive understanding of their relationships than standard linear regression models. Below is an elaboration on the quantile regression analysis, focusing on each independent variable.

The quantile regression model was specified considering the various quantiles, ranging from 0.10 to 0.90, and the potential impact of each independent variable on carbon dioxide emissions was analyzed at these specified quantiles (

Table 5. Quantile regression.

	CO2	CO2	CO2	CO2	CO2	CO2	CO2	CO2	CO2
	0.10	0.20	0.30	0.40	0.50	0.60	0.70	0.80	0.90
Fi	−83,537.3 ** (0.027)	−358,232.6 *** (0.000)	−443,214.9 *** (0.000)	−486,985.89 *** (0.000)	−516,025.3 *** (0.000)	−563,621.7 *** (0.000)	−572,810.8*** (0.000)	−478,746.6 *** (0.000)	−549,002.3 *** (0.000)
Fd	−87,153.89 * (0.187)	−197,111 * (0.172)	−116,505 (0.404)	−55,075.87 (0.596)	−30,645.2 (0.771)	−33,379.6 (0.767)	74,308.56 (0.542)	102,151.2 (0.356)	−93,298.7 (0.534)
Fm	77,642.75 *** (0.000)	107,923.4 *** (0.001)	120,278.2 *** (0.000)	104,810.7 *** (0.000)	97,727.15 *** (0.000)	84,795.45 *** (0.001)	56,327.45 ** (0.047)	68,475.65 *** (0.008)	76,673.69 ** (0.028)
Gdp	−0.55983 ** (0.029)	−1.14338 ** (0.041)	−1.34533 * (0.014)	−1.64755 *** (0.000)	−1.36072 *** (0.001)	−1.04476 * (0.018)	−1.07388 ** (0.024)	−0.917857 ** (0.033)	−0.0077746 (0.989)
Tr	3116.676 (0.572)	7919.605 (0.66)	8693.199 (0.456)	1859.903 (0.830)	−557.288 (0.950)	−7109.72 (0.452)	−20,458.8 ** (0.047)	−17246.5 (0.064)	−16,393.4 (0.193)
At	628.4973 *** (0.000)	510.4718 *** (0.000)	471.8405 *** (0.000)	451.273 *** (0.000)	426.6955 *** (0.000)	467.2045 *** (0.000)	476.6955 *** (0.000)	631.5421 *** (0.000)	612.0191 *** (0.000)
Pop	2135.022 *** (0.007)	1420.594 (0.407)	454.5198 (0.784)	566.5764 (0.646)	−365.927 (0.770)	−1094.23 (0.415)	−717.229 (0.621)	−1867.74 (0.157)	−1704.15 (0.340)
ecn2	1,397,177 *** (0.000)	940,622.4 (0.245)	565,664.2 (0.470)	651,639.4 (0.265)	354,910 (0.549)	71,639.37 (0.910)	53,860.52 (0.937)	−216,598 (0.727)	−660,405 (0.433)
0 bs.	132	132	132	132	132	132	132	132	132
R-squared	0.4749	0.5081	0.5737	0.6166	0.6549	0.6934	0.7441	0.7843	0.8200

Standard errors are in parenthesis. *** p < 0.001, ** p < 0.05, and * p < 0.1.

and Figure 1).

Across all quantiles, an increase in the Financial Institutions Efficiency Index is consistently associated with decreased CO2 emissions. This negative relationship is significant, suggesting that as the efficiency of financial institutions increases, CO2 emissions tend to decrease, a significant and negative relationship with CO2 emissions across all quantiles. These results indicate that the effect is statistically significant and negative at a 1% level at the lower (10th) −83,537.3 and the higher quantile (90th) −549,002.3. This could be attributed to more efficient financial institutions fostering investments in cleaner technologies or enforcing stricter environmental standards.

Conversely, the Financial Markets Efficiency Index shows a significant and positive relationship with CO2 emissions across all quantiles; these results indicate that the effect is statistically significant and positive at a 1% level at the lower (10th) 77,642.75 and the higher quantile (90th) −549,002.3. This suggests that as the efficiency of financial markets increases, CO2 emissions tend to rise by 76,673.69. This may imply that more efficient financial markets while fostering economic activity, may also lead to higher emissions without corresponding environmental regulations.

The relationship between GDP per capita and CO2 emissions varies across the quantiles. A decrease in CO2 emissions is observed in most quantiles with increased GDP per capita, signifying a possibility of cleaner production processes or higher environmental consciousness in higher-income economies. However, there is a slight and insignificant increase in CO2 emissions at the 0.90 quantile. This is evident from the significant and negative coefficient of GDP per capita in the quantile regression results, particularly at higher quantiles. These results support hypotheses H1 and H2 regarding the findings indicating variation in the impact of GDP per capita on CO2 emissions across different quantiles, suggesting that the turning point varies among GCC countries. This highlights the member states’ diverse economic and energy profiles and suggests that increased consumption and production may offset the benefits of cleaner technologies or environmental consciousness beyond a certain income level. Total patents directly and significantly correlate with CO2 emissions across all quantiles. These results indicate that the effect is statistically significant and positive at a 1% level at the lower (10th) 628.4973 and the higher quantile (90th) 612.0191. This can imply that the regions with a higher number of patents, possibly indicating higher innovation and industrial activity, tend to have higher CO2 emissions. It is crucial to understand the nature of these patents to determine whether they are related to cleaner technologies or industries with high emissions. The variable representing renewable energy consumption portrays a nuanced relationship with CO2 emissions. It has positive coefficients at lower quantiles, switching to negative at higher quantiles. This can imply that an increase in renewable energy consumption does not significantly decrease CO2 emissions at lower levels of renewable energy consumption. However, at higher levels, it does have a substantial impact on reducing emissions. The variables ‘Total trade in low carbon technology products as a per cent of GDP’ (tr), ‘Urban Population Growth’(pop), and ‘Financial development’ (FI) show inconsistent relationships with CO2 emissions across different quantiles.

The results confirm hypothesis H3, which states that the efficiency of financial institutions, financial markets, and technological innovation (patents) significantly impact CO2 emissions. Specifically, the efficiency of financial institutions shows a significant negative relationship with CO2 emissions, while financial market efficiency has a positive impact. Additionally, the number of patents is positively associated with CO2 emissions, underscoring the importance of green innovation. Renewable energy consumption also exhibits a nuanced relationship with CO2 emissions, showing a more substantial negative impact at higher consumption levels.

To visualize the results of a quantile regression by plotting a series of line graphs, each line representing a different quantile (0.10, 0.20, 0.30, etc.) for each independent variable (fi, fd, fm, etc.). Each independent variable will have a graph showing how its relationship with CO2 emissions varies across the different quantiles. Connecting these points, the coefficient of financial institutions’ efficiency index “fi” changes across different quantiles, reflecting the varying impact of “fi” on CO2 emissions across the distribution of the dependent variable.

This test refers to a process in statistical analysis where the researcher checks the reliability and validity of their results under different conditions, assumptions, or parameters. This can involve altering the dataset, using different statistical models, or changing key variables to see if the results remain true. Robustness testing is crucial in research to ensure that the findings are not just specific to a particular set of circumstances but can be generalized to a broader context. Probit and Tobit regression tests (see

Table 6. Testing the robustness.

	Tobit		Probit
Fd	29,998.35 ***	(0.007)	2.240671 ***	(0.024)
Fi	23,746.12.***	(0.000)	1.575652 ***	(0.000)
Fm	10,874.26 ***	(0.000)	0.5193258 ***	(0.000)
Gdp	0.1935762	(0.774)	0.00000974	(0.883)
At	2052.171 **	(0.000)	0.0018809 ***	(0.000)
Tr	2090.038	(0.488)	0.1788664	(0.935)
ecn2	5,798,047 *	(0.110)	11.78696	(0.280)
Pop	317.3253	(0.255)	0.0246225	(0.156)

Standard errors are in parenthesis. *** p < 0.001, ** p < 0.05, and * p < 0.1.

). Both tests indicated that the same significant results had been found in the case of dynamic analysis and quantile regression.

5. Conclusions

The Gulf Cooperation Council (GCC) nations have contributed significantly to the global energy market as major oil producers. However, with the global focus increasing towards environmental sustainability, it is imperative to comprehend how the GCC’s economic activities affect CO2 emissions. This study comprehensively analyses the relationship between various economic factors and CO2 emissions in the GCC region using quantile regression. This approach allows for a nuanced understanding of how these relationships vary across different levels of CO2 emissions. The empirical findings indicated that FI is a significant negative predictor, with 549,002.3 beta values in model 0.09. Thus, green energy channels can be improved by raising a consistent relationship between financial institutions’ efficiency and CO2 emissions, suggesting that efficient financial institutions can contribute to reducing emissions by fostering investments in cleaner technologies, respectively; policymakers should encourage financial institutions to provide capital and credit at reasonable costs to deploy renewable energy projects and promote green activities, this finding also accepts the H2.

However, FM has a significant positive relationship with CO2, with a beta value of 76,673.69, a positive correlation highlighting the potential environmental cost of increased economic activities facilitated by efficient financial markets.

A variable relationship between GDP per capita and CO2 emissions indicates the complex interplay between economic growth and environmental impact.

A direct and significant relationship between technological innovation (measured through patents) and CO2 emissions, emphasizing the need for green innovations. There is a nuanced relationship between renewable energy consumption and CO2 emissions, with its impact varying across different emissions levels.

The study utilizes a panel data quantile regression approach, which provides a comprehensive understanding of the relationship between economic variables and CO2 emissions across different quantiles of the emissions distribution. This method offers a significant advantage over traditional OLS regression models, which only estimate the mean effect by allowing the analysis to capture the varying impacts of predictor variables across the entire response variable distribution. This approach is beneficial in the GCC economies, where the relationship between economic activities and environmental outcomes can be highly heterogeneous.

Additionally, while this study finds significant relationships between financial efficiency, renewable energy consumption, and CO2 emissions, it is essential to acknowledge that the data used span from 2000 to 2021, and recent developments might influence these relationships. The authors suggest that ongoing data collection and analysis are necessary to update and validate the findings

6. Recommendations

Promoting Financial Institutions’ Efficiency: Policymakers should encourage financial institutions to invest in cleaner technologies and enforce stricter environmental standards. This can be achieved through policy incentives and regulatory frameworks that prioritize sustainable investments.

Balancing Financial Market Efficiency and Environmental Regulations: While efficient financial markets are crucial for economic growth, stringent environmental regulations to mitigate the associated increase in CO2 emissions must counterbalance this.

Fostering Sustainable Economic Growth: Efforts should be made to ensure that economic growth, especially in higher-income GCC economies, is accompanied by sustainable practices and technologies to limit the increase in CO2 emissions.

Innovation and Green Technology: There is a need for a strategic shift in innovation towards green and sustainable solutions. Governments and private sectors should invest in research and development of low-carbon technologies and incentivize patents in environmentally friendly technologies.

Renewable Energy Development: Given the complex relationship between renewable energy consumption and CO2 emissions, GCC countries should invest more in renewable energy infrastructure, particularly at levels that can significantly impact emission reductions.

Further Research and Policy Implementation: Ongoing research into the specific dynamics of the energy–growth relationship in the GCC countries is crucial. Policymakers should use these insights to implement balanced strategies that promote energy efficiency, economic growth, and environmental sustainability.

Regional Collaboration and Knowledge Sharing: GCC countries should collaborate in sharing best practices, technological advancements, and policy experiences to address environmental challenges and achieve sustainable development goals collectively.