Does environmental stress affect economic growth: evidence from the Gulf Cooperation Council countries?

The current paper empirically investigates the effect of environmental stress on economic growth in the Gulf Cooperation Council countries during 1995–2016. A panel cointegration analysis, specifically an autoregressive distributed lag model, is used to achieve the paper’s goal. The present work is motivated by the high carbon dioxide emissions per capita and environmental stress in these countries relative to other countries, and it assumes that the income per capita is a function of the natural resource’s rents and environmental stress. The findings show that environmental stress enhances economic growth, mainly in the long run. At the same time, the natural resources’ rents improve it in the short run and impede it in the long run. These results are significant because they tell that the Gulf Cooperation Council countries’ environmental stress did not reach critical levels that produce vast negative influences on the economy, and the resource curse hypothesis is valid in the long run. The current study’s policy implication states that economic policymakers should monitor and evaluate future environmental stress outcomes in these countries. There is no guarantee that the positive influence prevails. Therefore, the Gulf Cooperation Council countries should adopt genuine procedures to diversify their economies. Besides, it should continue its initial steps to expand renewable energy resources, i.e., nuclear, wind, and solar.


Introduction
The concept of sustainability affirms that the present community must manage and consume resources so as not to compromise future generation's needs and wants. It includes three pillars: environmental, economic, and social factors (Greer et al. 2020). Social sustainability is outside the scope of this paper. Environmental sustainability explores how to permanently maintain the human life support system by balancing the nation's production and consumption capacities. Thus, Goodland (1995) defined environmental sustainability as a set of restrictions on the four major activities regulating the scale of the human economic subsystem: the use of renewable and non-renewable resources on the production side, and pollution and waste assimilation on the consumption side. Meanwhile, sustainable economics is an ancient economic thought and has been attracting the attention of many economists. Its core idea concentrates on allocating the limited resources to satisfy the unlimited needs and desires or adjusting the lifestyle to comply with the finite resources (Sweidan and Alwaked 2016). Nowadays, understanding economic and environmental sustainability are growing very fast because of global environmental threats, i.e., global warming, pollution, and extensive exploitation of natural resources (Ahmed et al. 2019). Recently, Drupp et al. (2020) have surveyed the heterogeneous research area of sustainability economics to identify dominant research center, trends, and gaps. They reached many interesting conclusions: (1) of which research in the sustainability economics area is growing rapidly and (2) when the empirical work forms the largest share of this contribution. Baumgartner and Quaas (2010) defined economics sustainability as preserving the scarce resources while accomplishing the two normative goals of (1) the fulfilling of the desires of the current population and (2) justice, including justice between the present and future generations (intergenerational justice, justice between different humans of the same generation (intragenerational justice) and justice toward nature (physiocentric ethics). The definitions of environmental and economic sustainability are very close, and the latter focuses on justice while using natural resources.
The association between economic development, energy consumption, and the environment is within the sustainability domain concepts, and thus, it has been attracting the scholars' attentions of different drip lines. The literature has widespread applied studies on this significant topic. In an interesting work, Tiba and Omri (2017) briefly compiled around 400 papers in this research area. In a similar vein, Kais and Sami (2016) summarized 27 articles on the same research strand. Since the current study focuses on the Gulf Cooperation Council (GCC) countries, we noticed that the survey of Tiba and Omri (2017) reported two articles only with two different methodologies, i.e., the panel cointegration (Al-Iriani 2006), Granger causality (Salahuddin and Gow 2014), that explored the connection among these variables in the GCC countries as a group. Besides, they depicted two papers concerning Saudi Arabia (Alkhathlan et al. 2012;Alkhathlan and Javid 2013); both papers used the same methodology but employed different variables and time horizon. Additionally, they reported one work regarding the UAE .
Further, a new promising research strand of environmental sustainability is the ecological intensity of well-being (EIWB). It refers to the pressure or stress placed on the environment per unit of human well-being. Empirically, it is tested by many scholars by using different. For instances, Dietz et al. (2009) examined the effect of affluence, impacts, and human capital on EIWB in 135 countries. Dietz et al. (2012) tested the environmental Kuznets curve hypothesis by using EIWB in 58 countries. Moreover, Knight and Rosa (2011) investigated the factors that affect EIWB in a sample of 105 nations. Further,  studied how the impact of economic development on EIWB has changed since 1970 for 106 countries. Tiba and Omri's (2017) review did not report any study that used the EIWB indicator in its model. However, the literature shows that Sweidan (2018) and Sweidan and Alwaked (2016) explored the effect of economic growth and economic performance environmental sustainability. The former investigated the relationship in the Middle East and North Africa (MENA) region, while the latter in the GCC countries. Both studies showed that economic development and economic performance rose the environmental stress or economic unsustainability in these group of countries.
Within the framework of the EIWB concept, researchers, i.e., Jorgenson et al. 2014, defined environmental sustainability (stress) as the ratio of ecological footprint, i.e., CO 2 emissions per capita divided by life expectancy at birth. Technically, it represents the ratio of the stress on the environment to the human well-being. Dietz et al. (2012) presented additional definitions such as the ratio of a country's total energy consumption or CO 2 emissions divided by its gross domestic product. According to the economic philosophy, economic development seeks to improve human well-being or the quality of life. Thus, economic agents exploit the environment to achieve their goals. To produce goods and services, it converts the environment from a balance condition into a stress status. Alternatively, economic growth generates environmental stress, i.e., pollution and climate change, that negatively affects human well-being. As a general rule, the efficient nation is the one that can generate the largest well-being with less stress on the environment. As for EIWB indicator, more efficient or sustainable countries have small ratios, while less efficient nations have large ratios (Dietz et al. 2012). It is a general rule to compare between nations without a specific benchmark. The current paper uses this definition of environmental stress because many empirical works use it in the literature, and it measures the net effect of economic development on environmental stress and human well-being. Besides, the data to compute this indicator are available for the GCC countries.
Within the context of this crucial debate, the current paper is trying to move the focus on the GCC countries. Therefore, the present paper investigates empirically if environmental stress influences the GCC countries' economic growth. It is well known that the GCC countries are rentier states, which means that they permanently rely on oil and natural gas revenues to generate economic business cycles (Kabbani and Ben Mimoune 2021). The paper is motivated by the fact that CO 2 emissions per capita and the environmental stress are high in the GCC countries compared to other nations. Technically, we raise the opposite question of Sweidan (2018) and Sweidan and Alwaked (2016). Based on our knowledge, the literature lacks such a critical study, mainly in the GCC countries. We claim it is the first study that explores this question in the GCC countries, mainly using the structure of environmental stress variable. The GCC countries have a severe lack of environmental studies. Thus, our paper contributes to the literature by filling this gap.
We model environmental stress (unsustainability) as an input in the production function. We implicitly assume it has an impact on economic growth. The environmental stress is a proxy of exhausting (use) the nation's resources to generate more welfare. Alternatively, environmental stress is initially generated by economic development. Hence, we explore its influence on economic growth as a feedback effect. If environmental stress has a positive impact on economic growth, it indicates that the positive side of draining resources outweigh its negative ones. On the contrary, if the influence is negative, then the adverse impacts are larger than the positive ones, and the nation should adopt emergency actions, i.e., efficient allocation of resources, to alter the current situation. This paper is organized as follows. In "Literature review" section reviews the relevant literature about the current paper's question. "GCC countries: CO 2 emissions" section presents some facts about the GCC countries' CO 2 emissions and policy actions. "Data, model, and methodology" section introduces the data and methodology of the current study. "Empirical estimation" section analyzes and discusses the empirical estimations and the results. Conclusions and policy implications are made in "Conclusions and policy implications" section.

Literature review
The literature on the connections between energy, environment, and economic growth can be divided into three different directions: first, exploring the relationship between energy indicators and economic growth. It assumes a bidirectional causality link between economic performance and energy consumption; second, investigating the existence of the environmental Kuznets curve (EKC). The EKC hypothesis states that as the gross domestic product per capita of a nation rises, the environmental stress placed by that nation boosts until it reaches a turning point, after which any further growth reduces the environmental stress; and third, examining the relationship between economic growth, energy consumption, and CO 2 emissions. Tiba and Omri (2017) survey confirmed the link between economic growth, energy consumption, and CO 2 emissions. Technically, the current work contributes to the third direction of empirical research. Nevertheless, it differs from the previous studies by the structure of the primary variable that measures economic sustainability. Table 1 presents the previous studies of this research strand in the GCC and MENA regions. The information of Table 1 raises the following notes. First, the number of studies in the area of economic growth, energy consumption, and the environment of the GCC and MENA regions is limited. It may be explained by data availability. The current studies' trend reveals a definite increase. It seems scholars are interested to understand the relationship between the region's above-mentioned variables. So, it is the proper time to move forward and explore the effect of environmental stress on economic growth to complete the knowledge. Second, the previous studies confirmed a link between economic growth, energy consumption, and CO 2 emissions in the GCC countries and the MENA region. Furthermore, economic development and performance in the area boost environmental stress.
On the international level, many studies investigated the influence of different economic variables on environmental sustainability. These variables include natural resources such as coal, oil, and natural gas (Sun and Wang 2021), globalization, financial development, economic growth, and energy consumption (Sethi et al. 2020), disaggregated energy consumption, and natural resources rents (Ulucak and Danish 2020), information and communications technology (ICT) trade openness (Murshed 2020), trade policy, monetary policy, and migration index (Alola 2019), income inequality, climate change damage, and depletion of nonrenewables (Long and Ji 2019). All the empirical studies in this research area concluded that the different economic variables significantly impact environmental sustainability. Usually, environmental sustainability is measured by many different proxies such as CO 2 emissions, ecological footprint, environmental pollution, solid waste emissions, and green products. However, the current work defines environmental sustainability or stress by dividing CO 2 emissions per capita over life expectancy at birth, as stated above.
Recently, some studies have started to focus on understating the effects of environmental stress on various economic and social aspects by scholars from different disciplines. Wang et al. (2020) introduced a modeling framework for urban studies to capture spatial connectivity and teleconnection among the US cities in reaction to different environmental stressors such as extreme heat and air pollution. Blampied (2021) modeled a connection between economic growth and environmental stress (ecological footprint). The theoretical base for this work is that future permanent economic growth would not be attainable because of the environmental stress such as increasing pollution and exploiting the natural resources. He used Lucas (2000) model and assumed that nations faced an environmental constraint after 1970. Matthew et al. (2021) explored the potential future economic burden of workplace heat risk because of global warming and climate change. Further, a new significant area of research in environmental economics offers supervision to the governments and assessment models on CO 2 emissions reduction strategies to motivate low-carbon residential buildings. For example, Ma et al. (2020) explored the factors that can reduce CO 2 intensity in China's residential buildings. In addition, Ma et al. (2019) identified a roadmap to mitigate CO 2 emissions in China's residential buildings.
In sum, the literature has extensive studies on understating environmental sustainability due to the economic development process. Recently, some scholars from different disciples have sought to understand the effect of environmental stress or unsustainability on economic growth. The current work contributes to the literature on this spectrum. It is the first work that explores the influence of environmental stress on economic growth in the GCC countries. Environmental stress has many different proxies or indicators in the literature. We use the EIWB definition to achieve the goal of the current work.

GCC countries: CO 2 emissions
The GCC countries, Bahrain (BA), Kuwait (KW), Oman (OM), Qatar (QA), Saudi Arabia (SA), and the United Arab Emirates (UA), are positioned at the heart of the world and located in the Arabian Peninsula. Their total population number is 58.0 million in 2019, which is around 0.8% of the world's population. They are abundant in natural resources, including massive supplies of natural gas and crude oil. The British petroleum statistical review of the world energy for June 2020 reported that these countries own around 30.5% and 19.6% of the world's proven oil and gas reserves, respectively. Consequently, these countries rely heavily on oil and gas sales revenues to enhance economic development, amend social and health indicators, and build a first-class infrastructure. There is no doubt that the oil revenues contribute positively to the public well-being in the GCC countries. For example, the average annual income per capita during the period (1995-2018) is around $40 thousands. Moreover, the average life expectancy at birth increased from 72.8 y in 1995 to 77.2 y in 2018.
The World Bank development indicators display that the GCC countries' average contribution to the world CO 2 emissions reached 2.3% during the period (1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014). Table 2 presents and compares the CO 2 emissions per capita between the GCC countries and three industrial countries (France, Japan, and the UK) for the period (1995-2016). On average, Table 2 tells that the CO 2 emission per capita in these countries is higher than that of the industrial countries by more than three times. The extensive CO 2 emissions cause global climate change and pollution. Figure 1 presents the CO 2 emissions per capita for the GCC countries during the same period. It displays that the CO 2 emission of these countries is fluctuating but has a descending trend. It is also higher than that of some industrial countries with a higher number of populations. The global ecological footprint can also be a useful indicator for understanding and comparing the environmental threats and challenges on the regional and global levels. Table 3 introduces and compares the ecological footprint of consumption per person of the GCC countries and other countries in 2021 based on data from 2017. The ecological footprint of all the GCC countries is higher than that of the world average (2.8 ha/person). Further, the individual GCC countries have a higher ecological footprint of consumption compared with individual developed and emerging countries. Furthermore, Qatar (14.6 ha/person) and Luxembourg (12.8 ha/person) have the highest ecological footprint of consumption worldwide. The Intergovernmental Panel on Climate Change (IPCC) has published many reports about climate change over many years. It assured that global warming is an observable fact and a severe problem. They also warned that there is rising in the temperature of lands and oceans and more frequent heat waves in most land regions. These dangerous developments will increase poverty and starvation levels, and human health suffers. The reports also warned the Middle East region of a high risk of drought if the region's temperature increased by more than 1.5 degrees Celsius. Further, the Arab Forum for environment and development (AFED) confirmed the GCC countries' low air quality. Air pollution's primary sources are industrial and vehicle emissions, electric power generations, and inappropriate disposal of solid and hazardous waste. Furthermore, Heesterman (2020) confirmed that many of the world's wealthy coastal cities which historically developed as ports will disappear below the sea level. The quick solution to avoid the climate change catastrophic effects is to limit the CO 2 emissions. It requires a unique degree of international cooperation, a high level of taxation on the extraction of coal and crude oil, and the use of pressurized liquid petrol gas as aviation fuel.

The governments' actions
What did the GCC countries do to face regional and global environmental threats and challenges? We can highlight the answer via three dimensions. First, on the institutional level, they have well-established institutional arrangements to achieve sustainable development and protect the environment. The GCC governments' policies and action plans have the following levels: individual country actions, the GCC countries arrangements via the GCC general secretariat, and the League of Arab States decisions via the economic and social council (Abdel Gelil 2017). Second, the GCC countries relied on economic diversification as an extra strategic tool to confront the effects of climate change and boost economic growth. They realized that global warming harms non-oil economic sectors, i.e., agriculture, infrastructure, and tourism. Accordingly, the GCC countries' governments intervened via their strategic visions to support these sectors, preserve the environment, and enhance economic growth (Al-Sarihi 2018). Further, these countries recognized that the oil sector is finite and experience considerable fluctuations in the oil demand. So, it creates unstable oil prices and weak government revenues. Besides, oil revenues crowd out the other economic sectors. All the GCC countries have currently included an environmental dimension, i.e., energy efficiency, in their long-term development plans (Al-Sarihi 2018). Third, the GCC countries added a marketbased instrument (energy price reforms) starting from 2015 to protect the environment and ration energy consumption. Thus, these countries' energy subsidies have fallen from $116 billion in 2014 to $47 billion in 2016 (IMF 2017).
Did the above-mentioned policy actions were enough to generate adequate environmental protection? The answer can be highlighted by the Arab Forum for Environment and Development (AFED) public opinion surveys. They explored the Arab public awareness of environmental threats because of climate change. The AFED performed three surveys in 2006, 2009, and 2017 that covered all the Arab countries. The same main conclusions were reached by the surveys.
The surveys found that 98 percent of the respondent believed that the climate is changing, and 89 percent considered this was due to human activities, including excessive use of energy. Survey respondents expressed their willingness to participate in environmental actions to guard the environment. More importantly, it found that practically no work is being implemented to prepare the Arab countries for climate change threats. Technically, there were no research efforts to investigate the effects of climate change on health, infrastructure, biodiversity, tourism, water, and food production. Moreover, the Arab public alleged that the environment has continued to deteriorate over the last ten y. The public said that some governments had initiated specific policies that promoted green and sustainable economy, but these procedures are not enough to tackle the threats.

Data and model
Our paper uses perfectly balanced data of the six GCC countries covering the period (1995-2016) and consists of 132 observations. The data's source is the World Bank World Development Indicators (WBWDI) via Data Stream. This paper uses the idea of the Cobb-Douglas production function to achieve its goal (Mankiw et al. 1992). Nevertheless, we adjust the model to be consistent with the GCC countries' case. The GCC countries are endowed with natural resources, mainly oil and gas. Their incomes are generated based on two items: the rents of natural resources and these resources' use (the stress). Alternatively, economic growth is generated basically from extracting and consuming the natural resources in these countries. As a result, the income per capita is a function of the rent of the total natural resource and environmental stress. The specification of the current paper's model is as follows: where Y it is the income per capita in the country i at time t, A 0 indicates the parameter of the technology progress indicator or total factor productivity, and R it represents the rent of the natural resources in the country i at time t. It includes the sum of oil rents, natural gas rents, coal rents, mineral rents, and forest rents. S it is the environmental stress in the country i at time t, and θ and γ are the model's parameters or the factors of production shares. Based on economic theory, the expected signs of θ could be positive or negative, and it depends on the validity of the resource curse hypothesis. For example, Adekoya (2020) found that the economic growth of the resource-rich countries adversely responds to oil consumption in the long run, while the short-run impact is positive. Likewise, Dell'Anno (2020)  (1) answer is that γ could be positive or negative too. What is the meaning of a positive impact of environmental stress on the GCC countries' economy? We think the interpretation is related to how we design the environmental stress indicator. In our paper, we defined environmental stress as CO 2 emissions divided by life expectancy. Environmental stress can be interpreted as a sign to boost economic activity. If the economic development process succeeded in increasing human well-being more than wasting the resources (the stress), we should expect a positive sign. However, if the economic activities enhanced wasting the resources more than human well-being, we should expect a negative sign. There are no previous warning lights in the GCC countries tell that the pollution reached a very high level. Thus, the governments did not adopt emergency actions or procedures. As a result, the governments did not implement any sudden environmental policy that negatively affected economic growth. It indicates that the GCC governments' actions in the environmental area is not sufficient.
Researchers prefer to use the natural logarithm to improve the assumptions of normality and linearity. Further, Hahs-Vaughn et al. (2013) proposed a rule of thumb to decide whether to use the data in its raw form or transform it by using the natural logarithm. They stated if the distribution of skewness and kurtosis for all the variables are within the acceptable ranges, then there is no need to transform the data. The acceptable ranges for skewness and kurtosis are ± 3 of the standard error of each one. The standard error for skewness is √ 6∕132 = 0.213, while the standard error for kurtosis is √ 24∕132 = 0.426. As a result, the acceptable arranges are from + 0.640 to − 0.640 and from + 1.279 to − 1.279, respectively. The results show that the distribution of skewness and kurtosis are outside the acceptable ranges for some variables, and using the natural logarithm improves the distribution. Accordingly, we transform Eq. (1) into the log-linear equation. The new form of Eq. (1) is:

Equation (3) can be expressed as a general linear equation as follows:
where a 0 is the intercept and U it is the error term.
The GCC countries have a constraint on data availability. Thus, we employ the best available choices to estimate the model of Eq. (3). It justifies why our study covers the period . Researchers in environmental economics paid attention to an important issue when using a ratio as an indicator in the regression analysis. This issue is related to the construction of the ratio. This complication can be described as follows: The variability and the range of the numerator and denominator of the ratio can differ significantly. Thus, the indicator can be dominated by either the former or the latter. In our data illustration, the coefficient of variation (standard deviation divided by the mean) of the CO 2 emissions per capita is 0.52, and its range is 6.7-70.0 t per capita. As for life expectancy, the variation coefficient is 0.03, and its range is 69.8-79.9 y. The two variation coefficients tell that the relative movements in the CO 2 emissions per capita, the numerator, are much larger than the variation in life expectancy, the denominator. It indicates that variation in the numerator is driving the variation of the indicator. As a result, we should resolve this complication before using the ratio in our regression analysis. We follow researchers from the New Economic Foundation (2009) to solve the problem. The solution method consists of forcing the numerator and denominator's variation coefficient to be equal by adding a constant number to one of them. It shifts the mean without altering the variance. It means the variation coefficient of the two variables will be equal. This technique was used by many researchers in the environmental economics area (Sweidan 2018). We equate the coefficients of variation for the two variables in our data sample by adding the correction factor 462.47 to the CO 2 emissions per capita. The correction factor is estimated based on the following formula, CF = S CO 2 * M LE ∕S LE − M CO 2 , where S stands for the standard deviation, M denotes the mean, CO 2 is the emissions per capita, and LE indicates the life expectancy. As a result, the adjusted S it is computed by the following formula: Figure 2 presents the adjusted S it of the GCC countries over the period . It reveals that the adjusted S it has a declining trend, which means that human well-being is larger than environmental stress. Despite the massive CO 2 (4) S it = CO 2 emissions per capita it + 462.47 Life expectancy it × 100 emissions in the GCC countries, they can mitigate its negative impacts on the economy by expanding public welfare. Table 4 presents the descriptive statistics of the variables included in the current paper model.

Methodology
Our paper employs the panel autoregressive distributed lagged (ARDL) model introduced by Pesaran et al. (1999) to extract the empirical evidence. It detects the presence of a long-run relationship between the variables in the model. It computes short-run and long-run coefficients and estimates the speed of adjustment (error correction term) toward the long-run equilibrium. This approach is valid regardless of the variables' integration orders. It is applicable if the orders of the variables are I(0) or I(1) or I(0) and I(1), but not I(2). The general specification of the ARDL(p, q) model is as follows: where Y it denotes the independent variable for a group of countries i, X it indicates a list of explanatory variables for a group of countries i, α it stands for the fixed effect, β it and ′ ij represent the estimated coefficients, and ε it is the random disturbance. Pesaran et al. (1999) rewrite Eq. (5) to match the panel error correction form: where φ it and i are the long-run parameters, and we employ them to estimate the speed of adjustment (error correction term) to long-run steady state, and γ ij and ′ ij are the short-run coefficients lagged dependent and independent variables, respectively. We write the paper's model by using the variables' notations in Eq. (6) as follows:

Panel unit root tests
The first step in the estimation process is to check the stationarity of our model's dependent and independent variables. It guarantees that our variables satisfy the condition of the variables' order. We conduct three panel unit root tests: Levin et al. (2002) (LLC) unit root test, Breitung (2000) test, and Pesaran (2003) test. The last test is classified among the second-generation unit root tests, which allow for crosssectional dependence. Nevertheless, the first two tests do not have this feature. The H 0 for the LLC (2002) and Breitung (2000) test tells that the panels contain unit roots, against the H 1 , i.e., the panels are stationary. The H 0 for Pesaran (2003) test states that all panels are non-stationary. We report the results of the three panel unit root tests in Table 5. The results display that some variables are stationary at the level and/or at the first difference. Thus, we can use the panel ARDL approach.

Panel cointegration tests
We explore the presence of a long-run relationship among the variables using Pedroni (1999Pedroni ( , 2004 and Westerlund (2005) panel cointegration tests. These two tests have the same null hypothesis of no cointegration. Pedroni (1999, 2004) alternative hypothesis states that all panels are cointegrated, whether the alternative hypothesis of Westerlund (2005) is that some panels are cointegrated. We present the panel cointegration tests' results in Table 6. We perform the two cointegration tests by including fixed effect, trend and remove the cross-sectional means. The two tests' results reject the null hypothesis of no cointegration in favor of the existence of a long-run relationship between our variables.

The results
The presence of a long-run relationship between the variables indicates we can estimate the ARDL model of Eq. (7). There are three estimators: pooled mean group (PMG), mean group (MG), and dynamic fixed effect (DFE). The PMG estimator levies homogeneity restriction on the long-run coefficients across counties while maintaining heterogeneity for the intercept and short-run parameters. In contrast, the MG does not place any restrictions. It allows all parameters to change and to be heterogeneous in the short and long run. The DFE estimator imposes homogeneity restriction on the slopes, and the intercepts are allowed to alter across countries. Under the long-run homogeneity assumption, the PMG estimator is more efficient than the MG and DFE estimators (Pesaran et al. 1999). Accordingly, we should find out the most efficient estimator for our data among the three estimators. The direct method is to employ the Hausman test to check whether there is a significant difference between PMG and/or MG or PMG and DFE estimators. The PMG estimator is efficient and consistent under the null hypothesis of slope homogeneity. The PMG estimator is recommended if the null hypothesis is not rejected.   results [0.41 (Prob. 0.813) and 0.26 (Prob. 0.880),respectively] are statistically insignificant in both cases. Thus, the long-run PMG estimator is recommended. Accordingly, we compute the ARDL (2, 0, 0) model coefficients by the PMG estimator. The numbers inside the parentheses (2, 0, 0) are the computed optimum lags that minimizing the log likelihood by using the FORVAL programming command.
For robustness check, we approximate three ARDL model scenarios based on the generated environmental stress as follows. Scenario 1 (original model) uses the environmental stress produced during the consumption of solid, liquid, and gas fuels and gas flaring (ln S it ). Scenario 2 utilizes the environmental stress generated from natural gas (ln SG it    present the environmental stress of Scenario 2 and Scenario 3, respectively. Table 7 documents the log-term parameters, short-term coefficients, and the error correction terms (ECT) of the three scenarios. These scenarios produced almost very similar results. The ECT in the three scenarios is statistically significant and negative. It approves a converging and stable long-run relationship between the model's variables. The deviations from long-run equilibrium are corrected by an average adjustment speed of 20.3 percent in the current period.
All the short-run parameters of the natural resources' rents are statistically significant and positive, with an average value of (0.078). In the long run, they continue to be statistically significant, in two scenarios, but with a negative sign. Moreover, the average size of the long-run parameters (− 0.017) is smaller than the short-run coefficients. This evidence supports the resource curse hypothesis in the GCC countries.
The short-run parameters of environmental stress are statistically insignificant in two scenarios. However, in the long run, all the parameters switched to be statistically significant and positive. It means the GCC countries' environmental stress succeeded in producing positive effects more than negative ones on the economic activities. Hence, the final observable influence is positive. The environmental stress can be viewed in terms of more consumption, investment, economic activity, and higher economic welfare. It may also indicate that these countries' environmental stress did not reach critical levels that produce significant adverse effects. As a result, we think the GCC countries are lucky because the positive impacts prevail, and thus, they should start giving more attention to the environmental stress before any possible enlargement to the negative influences.
We think our results making sense and consistent with the previous studies' results concerning the UAE (Sbia et al. 2017) and Saudi Arabia (Alkhathlan and Javid 2013). These previous works confirmed a bidirectional relationship between CO 2 emissions and economic growth in the GCC countries. However, recall that the current paper utilizes a different or a new proxy for environmental stress other than CO 2 emissions. From the other direction, Sweidan (2018) and Sweidan and Alwaked (2016) concluded that economic development and performance created more pressure on the environment in the GCC countries and MENA region. Hence, the environmental stress is a proxy to the dynamics of the economic activities, and logically its effect could be positive unless the government took some procedures to reduce the environmental stress, which might adversely affect the economic growth. Overall, our paper confirms that the GCC countries' environmental stress contributes positively to economic growth, mainly in the long run.

Conclusions and policy implications
Environmental issues and sustainable development have been attracting the world economic and environmental organizations and scholars alike. The literature has extensive empirical works focused on exploring the relationship between economic development, energy consumption, and the environment. In contrast, there are limited studies focused on the link between environmental stress and economic development. The GCC countries have limited empirical works in this research area. They are oil and gas exporting countries and have around 29.3% and 21.5% of the world's proven oil and gas reserves, respectively. They also use oil and gas sales revenues to enhance economic growth, amend social and health indicators, and build a first-class infrastructure. The different economic and health indicators showed clear improvements during the past four decades. However, many reports displayed high CO 2 emissions per capita and environmental stress in GCC countries.
Our paper empirically investigates the effect of environmental stress or environmental unsustainability on the GCC countries' economic growth. We model the environmental stress as an input in the production function. Our paper's primary motivation is the high CO 2 emissions and the environmental stress in the GCC countries relative to other nations. These countries lack such critical study. The literature has two studies, Sweidan (2018) and Sweidan and Alwaked (2016), that explored the opposite question, which is the effect of economic development on the environmental stress in the MENA region and GCC countries. The reports showed that these countries had established a promising institutional framework to deal with environmental and sustainable development. Unfortunately, the region lacks genuine interest in environmental issues, which prevent establishing useful work, sound policies, and sustainable development evaluations.
The current paper uses the panel ARDL methodology to estimate the models' parameters. The results reveal that environmental stress has a positive and statistically significant effect on the GCC countries' economic growth, mainly in the long run. To check the robustness of our results, we employ three proxies of environmental stress. The three scenarios generated almost similar results. Our conclusion tells that the GCC countries' environmental stress produced more positive impacts than adverse effects on the economy. The environmental stress is a signal for more consumption, investment, and higher economic welfare in the GCC countries. If the economic development mechanics achieved more human well-being than damaging the resources (the stress), then a positive feedback effect is expected. The GCC countries' environmental stress did not reach critical levels that produce vast negative influences on the economy. Historically, the governments did not adopt harsh decisions that hindered economic activities. On the contrary, the GCC governments did not give serious attention to environmental and sustainable development problems. The GCC countries are lucky because the positive impacts prevail. However, economic and environmental policymakers should monitor and evaluate the future environmental stress outcomes in these countries. The positive impacts are not guaranteed to dominate over the long time. The GCC countries should adopt an efficient energy environmental policy by shifting to the renewable energy, i.e., wind and solar. This move helps these countries to diversify their energy resources and reduce the environmental stress. These countries have initiated some reforms, i.e., the establishment of research and development centers, to enhance renewable energy deployment. This includes Masdar City in the United Arab Emirates and King Abdullah City for Atomic and Renewable Energy in Saudi Arabia. The UAE has initiated its nuclear energy program and built its first power plant in the region (Al-Saidi and Haghirian 2020). Furthermore, our results confirm that the resource curse hypothesis is valid in the GCC countries in the long run. This finding opens up the debate on the importance of diversifying the GCC economies.