The Relationship between Economic Growth, Energy Consumption and CO2 Emission in the Middle East and North Africa (MENA)


 
 Research background: CO2 emissions are considered to be the main reason for global warming, and for this reason, their regulation is a very important issue for governments. Due to the increasing use of energy, carbon dioxide emissions have increased dramatically over the past century, with a direct link to economic growth and development. The relationship between CO2 emissions, growth and energy consumption is therefore at the heart of current economic issues.
 
 Purpose: This study aimed at examining the relationship among economic growth, carbon dioxide (CO2) emissions and energy consumption in selected MENA countries, in the period 1995–2017.
 
 Research methodology: To prove these relations, a stationary data panel methodology is used supported by unitary root and cointegration tests.
 
 Results: The results indicated that there is a long-term relationship between CO2 emissions, energy consumption and GDP. In addition, it is found that the elasticity of CO2 emissions with respect to energy consumption is less than one (inelastic), and the elasticity of CO2 emissions with respect to GDP suggests the existence of an Environmental Kuznets Curve. An important finding is that energy consumption has a positive but relatively low effect on CO2 emissions. To reduce CO2 emissions, the countries of the MENA region are being called upon to increase significantly the use of renewable energies and the establishment of a more efficient energy policy.


Introduction
Environmental issues are the subject of a great deal of scientific research and are of constant concern to governments at the highest level around the world (Alkhathlan, Javid, 2013;Wang, Zhou, Wang, 2011). Indeed, the destruction of the ozone layer by the release into the atmosphere of greenhouse gases, including carbon dioxide (CO 2 ), constitutes the main threat to humanity.
According to Stern (2007), it could cost the entire world economy up to $ 550 billion if real action is not taken. Due to the increasing use of energy, carbon dioxide emissions have increased dramatically over the past century, with a direct link to economic growth and development (Lau, Tan, Lee, Mohamed, 2009;Sotyas, Sari, 2009;Soytas, Sari, Ewing, 2007). The relationship between CO 2 emissions, growth and energy consumption is therefore at the heart of current economic issues.
CO 2 emissions are considered to be the main reason for global warming, and for this reason, their regulation is a very important issue for governments. For example, the Kyoto protocol in 1997 set the goal of reducing greenhouse gases. As MENA countries is a group of countries that have recently experienced rapid economic growth, it is important to study the effect that GDP has on CO 2 emissions. Additionally, the extensive study of the relationship between energy consumption and GDP indicates that energy consumption patterns also have a positive impact on CO 2 emissions. In this sense, if the relationship is positive, mechanisms should be proposed for the conservation and efficient use of energy.
Economic literature refers to economic growth as the accumulation of factors of production, such as physical capital, labour, human capital and technical progress (Solow, 1956;Romer, 1986;Lucas, 1988). However, many economists, specialists in the economics of energy, support the idea that these factors of production can only be productive in the presence of energy, then the latter is considered to be the engine of economic growth (Asafu-Adjaye, 2000;Yang, 2000;Stern, 2000;Ghali, Khalifa, El-Sakka, 2004). Therefore, the causes of environmental degradation are not reduced to just one factor, namely economic growth, but other factors are also involved such as energy consumption. Indeed, the most widely used energy sources in the MENA countries are fossil fuels. The latter are the most polluting, because the use of gas or oil in economic activity, for example the production of electricity, emits CO2 in the air and contributes to the greenhouse effect. This air pollution is causing climate change. According to a World Bank report published in 2011, countries in the region will be hit hard by climate change. Because the average rate of warming will be higher than the world average, there is also the phenomenon of drought and water stress.
It should be noted that statistics from the World Bank (WDI, 2017) show that CO 2 emissions in MENA countries were 3.6 metric tons per capita in 1990 and 6.2 metric tons per capita in 2014, an increase of 172% in the space of just two decades. While the global emission averages for the same periods (1990 and 2014) are only 4.2 metric tons per capita and 5 metric tons per capita. The upward trend in emissions will be a real problem that the countries of the MENA region will have to face, since their international commitments vis-à-vis climate change (Paris Agreement of 2016) will oblige them to work to reduce emissions from greenhouse gas. This requires, of course, a reduction in the consumption of fossil fuels, bearing in mind that clean, socalled renewable energies are not sufficiently developed to make the energy transition a success, which risks jeopardizing the economic development of the countries in question.
The objective of this work is to study the relationship between energy consumption, GDP and CO 2 emissions. Through a non-stationary macroeconomic panel data model, the long-term elasticity's of CO 2 emissions to energy consumption and GDP are estimated. The above, in order to provide empirical evidence about the relationship between CO 2 emissions and some of its determinants, in this particular case, energy consumption and GDP. In addition, evidence is presented on the hypothesis of the existence of an Environmental Kuznets Curve (EKC) for the MENA group of countries.
Likewise, the contribution of this work to the existing literature lies in the use of nonstationary panel data methodologies, with unit root tests for panel data and panel cointegration tests, and empirical evidence is presented for a group of countries (economic region) little studied as in the case of the MENA.
Households and firms determine the growth of economies, in part, since through the production of goods and services by firms, the consumption of goods and services are provided so that households, can satisfy their needs. Now, if household consumption increases, firms will increase their production to meet household demand, and the problem here is that the said increase in production will generate possible negative and positive externalities added to possible excesses compared to the use of environmental resources, through increases in energy consumption and income. Referring to this context, the construction of economic policies, in most cases, does not link the factors of pollution and degradation of the environment, to try to reduce the negative impact of these on the economy or considerably limit the use of these resources. Here, then, is the importance of knowing the effects that increases in energy consumption and GDP have on the CO 2 emissions of an economy. This paper is divided into five sections, including an introduction, and it is organized as follows. The first section presents a brief review of international empirical literature on The relationship between Economic Growth,Energy Consumption and CO 2 Emission... 135 the relationship between CO 2 emissions, energy consumption, and GDP. In the second, the methodology and data are exposed. In the third section, the results of the unit root and panel cointegration tests are presented, in addition to the estimates. Finally, the fourth and last section presents the conclusions.

Literature review
Numerous researchers have studied the relationship between energy consumption and GDP over the last thirty years, 1990-2020; however, it is necessary to include CO 2 emissions in this relationship, given the growing environmental concern, due to the problem of climate change that is occurring worldwide. Some of the most relevant works on this topic, during the first decade of the 21st century, are reviewed below.
Since the study by (Panayotou, 1993), empirical work on the question of the relationship between economic growth and the environment continues to abound. Indeed, (Panayotou, 1993) had attempted, through both theoretical and empirical analysis, to verify the hypothesis of the Kuznets environmental curve (EKC). This assumption assumes that the economic growthenvironment relationship takes the form of an inverted "U". Indeed, the empirical study consisted of highlighting the relationship between the effects of economic development on deforestation and air pollution, for a sample comprising developed and developing countries. The results obtained confirm the EKC hypothesis. According to the author, the confirmation of the EKC involves important political consequences. Because during the development period, environmental degradation is inevitable. However, when the economy reaches an advanced level of growth, the environmental framework begins to improve significantly.

Methodology
This section describes the methodology used to determine the effects that GDP and energy consumption have on CO 2 emissions. In addition, graphs of the data with which this methodology is applied are presented.

Panel data
During the last two decades, 2010-2020, panel data have been used as an analysis tool by researchers from various areas to study the relationships between different variables. The main reason is that this methodology combines a time dimension (time series) with a cross-sectional one (cross section).
When working with macroeconomic panel data, which contemplate a time series greater than the number of individuals, the existence of a long-term relationship between the variables that are analyzed for the group of individuals must be taken into account. In other words, it must be ensured that a cointegration relationship exists to avoid the problem of obtaining spurious results as argued by (Engle, Granger, 1987;Entorf, 1997;Kao, 1999;Phillips, Moon, 1999), who introduced the term of spurious relationships in the use of panel data, when the time observations are greater than the number of individuals in a panel. For this, it is necessary to apply unit root tests that allow determining the order of integration of the time series, and cointegration tests to determine whether or not the time series of each country are cointegrated, and thus avoiding spurious results.
The implementation of this panel data methodology links the fact that individuals, firms, banks or countries, as in this case (the MENA) are heterogeneous. Cross-sectional and time series analyzes do not attempt to control for this heterogeneity at the risk of biased results. An additional advantage is that they allow a better study of the dynamics of the adjustment processes of the data taken for the MENA countries.
According to the international literature on energy economics, the following equation can be formulated in order to determine the long-term relationship between CO 2 emissions, energy consumption and GDP. Other authors such as (Pao, Tsai, 2010;Apergis, Payne, 2009;Halicioglu, 2009)  where: LCO2 it -is the logarithm of the CO 2 emissions of each country (i) in the period (t), LEC it -is the logarithm of the Energy consumption of each country (i) in the period (t), LGDPP it -is the logarithm of the GDP of each country (i) in period (t

Unit root tests for the panel
The order of integration of the series of CO 2 emissions, energy consumption and GDP, expressed in per capita terms, is considered first. The tests for panel data developed by (Im, Pesaran, Shin, 2003) known as the IPS test, (Levin, Lin, Chu, 2002;Breitung, 2000;Maddala, Wu, 1999) (Fisher type ADF) and Choi are used (2001) (Fisher type PP). The tests of (Maddala, Wu, 1999;Im et al., 2003) allow for heterogeneity between individuals in the panel data.
The panel unit root tests are based on the tests developed for the time series, but have an advantage over the latter, and that is that when combining time series and cross-section data, more degrees of freedom are obtained, which improves the properties of the estimators, and correct for unobserved heterogeneity.

Cointegration test for the panel
After verifying that the series are integrated of order one, that is, that they contain a unit root in the panel, the cointegration test is continued, in order to find evidence of the existence of a relationship between the variables in the long term, and thus avoiding the presence of spurious relationships. The foregoing is proven through the well-known heterogeneous panel cointegration test of Pedroni (2000Pedroni ( , 2004. This test is based on seven statistics divided into two groups; the first is based on the (Phillips, Ouliaris, 1990) statistic, which is defined as: where: ˆi t ε is estimated from equation (1) (Pedroni, 2000).

Data
Below we can see the variables that were used in this paper. GDP per capita measured (in dollars at constant prices 2010), energy consumption (kWh per capita), CO 2 emissions (metric tons per capita).
All series were obtained from the World Bank. Figures 1 to 3 show the behaviour of each of the series of the selected MENA countries.  Figures 1 and 2 show that Libya is the country with the most CO 2 emissions and energy consumption per capita, compared to the other MENA countries. This is due primarily to the high level of primary energy consumption per capita (more than 3 tons/capita). In the case of Algeria, can be observed that its behaviour is within the range of the other countries of the group, and that Egypt and Morocco are the most homogeneous countries with a low share of CO 2 emissions per person.
The relationship between Economic Growth, Energy Consumption and CO 2 Emission...

139
The energy consumption of the MENA is reflected in Figure 2 where the energy consumption of Turkey stands out in contrast to the other countries of the group, which present homogeneous behaviour, which may be due to consumption habits, such as the minimal use of fossil fuels or the high rate of economic growth during the last decade.  Figure 3 allows us to infer that Turkey is the country that has presented the highest economic growth per capita throughout the series, with an average increase of 2.40%. The economy of Libya depends primarily on revenues from the petroleum sector, which represents over 95% Prof. Mohammed Touitou 140 of export earnings and 60% of GDP. These oil revenues and a small population have given Libya one of the highest per capita GDP in Africa. Libya after 2000, recorded favourable growth rates with an estimated 10.6% growth of GDP in 2010. This development was interrupted by the Libyan Civil War, which resulted in the contraction of the economy by 62.1% in 2011. After the war, the economy rebounded by 104.5% in 2012, but then crashed again following the Second Libyan Civil War as of 2017, Libya's per capita GDP stands at 60% of its pre-war levels.

Results and discussion
This section presents the results of the unit root and cointegration tests in the panel data. Table 1 shows the results of the five-unit root tests presented in the previous section, and is indicated for each of the variables in levels, the statistics and the probability. It can be seen that for none of the tests the null hypothesis of unit root can be rejected, with which it is concluded that the series are not stationary in the levels. The results of the unit root tests applied to the series in the first differences are presented below, in order to rule out the presence of more than one unit root. Table 2 shows these results for each variable, and the statistic and probability corresponding. It can be concluded that the series in differences are stationary, which can be rejected by the null hypothesis that there is a unit root.
The result of the Pedroni (2000Pedroni ( , 2004 cointegration test for the MENA countries is presented in Table 3. It can be seen that five of the seven statistics of this test allow rejecting the null hypothesis of non-cointegration at 5% significance; therefore, there this is strong statistical evidence in favour of a cointegration relationship between the variables CO 2 emissions per capita, energy consumption per capita, and GDP per capita.   Table 4 shows the results of the estimation of the long-term relationship based on the cointegration test of Pedroni (1999Pedroni ( , 2004, for equation (1), and can be observed that all the coefficients are statistically significant at 1%, and also have the expected a-priori sign.  LCO LGDPP LGDPP LGDPP LGDPP This implies that the inflection point is found when GDP reaches a level of 7.77 in logarithms (or 2,368.47 in million dollars of 2010 per capita). These results reinforce the hypothesis of the existence of an EKC, since the level of CO 2 emissions first increases with GDP, and then it stabilizes and finally decreases. The elasticity of GDP over CO 2 emissions will be greater than 1, when GDP per capita is less than (1.71 -1)/0.22 = 3.23 (in logarithms).
In summary, in the long term, CO 2 emissions are inelastic to energy consumption and elastic to GDP if it is less than 3.23, while they are inelastic if GDP is greater than 3.23.
In summary, the results suggest a considerable response of CO 2 emissions to changes in GDP, and a relatively low response of CO 2 emissions to increases in energy consumption. Similar results were found by (Apergis, Payne, 2009;Pao, Tsai, 2011), for a different group of countries.

Conclusions
During the decade 2010-2020, regions and countries in the world have become aware of the negative effect that energy consumption patterns and GDP growth could have, among other things, on the environment. In other words, they have realized that the most important thing, even above economic growth, was the negative impact that global consumption patterns can have on the environment. This paper presents a first approach at studying the determinants of CO 2 emissions for the eight MENA countries for the period 1995-2017. Using the panel data, the relationship of CO 2 emissions, real GDP, and energy consumption is estimated, in order to know the coefficients of this relationship in the long term and to be able to analyze the impact, in addition to contributing to the fulfilment of the hypothesis of the existence of an EKC.
The results presented in this paper indicate that there is a long-term relationship between CO 2 emissions, energy consumption and GDP. In addition, it has been found that the elasticity of CO 2 emissions with respect to energy consumption is less than one (inelastic), and the elasticity of CO 2 emissions with respect to GDP suggests the existence of an Environmental Kuznets Curve, which is U-shaped inverted with a point of inflection at the level of GDP per capita of 7.77 (in logarithms).
Similarly, the results suggest that there is a causal relationship between GDP and CO 2 emissions, and energy consumption with CO 2 emissions, which means that in the long-term economic growth, is one of the determinants of climate change through CO 2 emissions, at least for the studied countries, known as the MENA. Specifically, at equilibrium, a 1% increase in energy consumption generates in the long term an increase of 0.39% in CO 2 emissions from the MENA data panel. On the other hand, CO 2 emissions are elastic to GDP if it is less than 3.23 (in logarithms), while they are inelastic if GDP is greater than 3.23 (in logarithms).
An important finding is that energy consumption has a positive effect on CO 2 emissions, but relatively low, as mentioned above, indicating that policies aimed at promoting efficient use and conservation of energy reduce CO 2 emissions, hence, these could contribute to reducing global warming. On the other hand, the evidence found in favour of the Environmental Kuznets Curve indicates that although economic growth increases the level of CO 2 emissions, this increase stabilizes and then introduces a decrease in emissions. This reduction demonstrates that countries are becoming more efficient at production, with better production processes and less polluting.
To reduce CO 2 emissions, considering that the MENA countries are the most affected and vulnerable in the world because of environmental degradation and climate change, accordingly, the countries of the MENA region are called upon to significantly increase the use of renewable energies and to develop the establishment of a more efficient energy policy. Solar energy is among the most important renewable energy resources available in the MENA region, as it possesses the largest desert in the world, which makes it have sunlight throughout the whole year.
However, several aspects to study remain on the research agenda, such as what is the effect that the financial deepening or capital flows of this region of the MENA have on CO 2 emissions, since as (Frankel, Romer, 1999) argue, that financial deepening and development can attract foreign direct investment and increased investment in research and development, which accelerates the economic growth of economies and affects the dynamics of environmental performance.