The impact of energy consumption on environmental quality: empirical evidence from the MINT countries

Abstract

Rapid increases in energy consumption and economic growth over the past three decades are considered the driving force behind rising environmental degradation, which remain a threat to people and healthy environment. This study investigates the impact of energy consumption on environmental quality in the MINT countries using a panel PMG/ARDL modelling technique, and the Granger causality test spanning from 1971 to 2017. The empirical results confirm the existence of long-run nexus among the variables employed. The results also reveal that economic growth, energy consumption and bio-capacity have a positive and statistically significant effect on environmental degradation during the long run period. We find that a 1% increase in primary energy consumption leads to 0.4172% increase in environmental deterioration in the long-run period, but it is insignificant in the short run. This implies that energy consumption deteriorates environmental quality through a negative effect of ecological footprint. The result also suggests that as MINT countries increase the use of energy to accelerate pace of economic growth, environmental quality would deteriorate through increased ecological footprints. The coefficient of the error correction term (ect) is negative and significant (− 0.2306), suggesting that ecological footprint, a measure of environmental degradation would converge to its long-run equilibrium in the MINT region by 23.06% speed of adjustment every year due to contribution of economic growth, energy consumption, urbanization and biocapacity. The Granger non-causality test results reveal a unidirectional causal relationship from economic growth, energy consumption, and urbanization to ecological footprint and from economic growth to biocapacity. The results further show bi-directional causality between biocapacity and ecological footprint as well as between biocapacity and economic growth. Moreover, urbanization causes economic growth and biocapacity Granger-causes urbanization. Based on these findings, policy implications are adequately discussed.

Introduction

The concept of environmental degradation has drawn the attention and concern of many global leaders and researchers recently. Based on the European Commission (2006), human beings continuously and increasingly utilize natural resources at a faster rate than the environment could replenish. Environmental sustainability, human productivity and access to clean technology can be promoted if the environment is used economically, e.g. through dams and water reservoirs and nations can adopt alternative measures that lessen greenhouse gases and carbon dioxide to prevent a rise in the average global climate above 2 °C (UNFCCC 2017). Despite the global demand requiring countries to lower greenhouse gas and carbon emissions, there have been growing changes in economic growth and energy consumption which are the leading reasons for the rising environmental pollution. For example, primary energy use account for more than 85% of CO₂, causing harm to mankind including animal species, and claims the lives of about seven million people every year (WHO 2018). To enable protection of the environment, countries must fight against the intensive exploitation of natural resources, particularly energy consumption and other economic activities that require much application energy resources. In this regard, the World Bank action plan on climate change for 2016–2020 target at improving renewable energy consumption by 30 GW, develop early warning systems for 100 million individuals and to develop some climate-smart agricultural investment policies for more than 40 countries. In order to realize the global sustainable development, the Kyoto convention of 1997 was a watershed in the annals of sustainable environmental that reiterate global efforts made at tackling the problems related to Greenhouse gases and global warming by lowering CO₂ globally. The convention also spelt out the need for all countries across the globe to pursue energy and environmental policies that can reduce CO₂ and health hazards and also meet the global sustainable development goals.

The nexus between energy consumption and environmental quality with the consolidation of the concepts of key variables like economic growth, urbanization, biocapacity and ecological footprints became increasingly considered by researchers, beginning from the review of wealth in a limited world of resources (Wackernagel and Rees 1996). It has been recognized, for many decades, that human needs, including food, clothes, shelter and energy resources, are provided by nature. Theoretically, environmental scientists have admitted that energy consumption and economic growth pursuit have negative environmental effects and this has been proven in the recent empirical studies of Ali et al. (2021), Alola et al. (2019b, 2020), Bekun et al. (2019), Caglar et al. (2021), Danish et al. (2019), Güngör et al. (2021), Ike et al. (2020a, b, c), Onifade et al. (2021), Phong (2019), Usman et al. (2020a) and Zhang (2021), among others. Some of these studies have particularly exposed that energy consumption and ecological footprint have a long run positive nexus, but Alola et al. (2019b, 2020), Caglar et al. (2021), Ike et al. (2020a, b, c), Onifade et al. (2021), Sarkodie and Strezov (2018) and Usman et al. (2020b) have also established the existence of a negative relationship between renewable energy consumption and environmental deterioration. These studies alien with the environmental Kuznets curve hypothesis propounded in Grossman and Krueger (1991); also recently extended in (Destek and Sarkodie 2019; Güngör et al. 2021; Ike et al. 2020a, b, c; Phong 2019; Ssali et al. 2019; Usman et al. 2019) that assumes CO₂ emission rises with increases in energy use and income at early stage of economic development, until it reaches an income peak beyond which as an economy keeps growing, CO₂ begin to reduce. Globally, ecological footprint has risen progressively over the last five decades in the same way energy consumption has been steadily increasing. As a result, these two variables have expanded globally. If the global hectare per person is high, it implies that the environment or ecosystem can longer adequately reproduce the natural resources needed to support human life and consequently, environmental quality will be threatened. In other words, it means reduction of the available natural resources and people will suffer lack and other natural hazards. The World Energy Council (2019) ranks nations based on their energy synthesis policies to deal with energy security and environmental sustainability issues. Accordingly, WEC (2019) listed three primary policy areas that focus on productivity and efficiency of energy generation, transmission and distribution, decarbonisation and the quality of air.

The impact of energy consumption in industrial production and economic development is often significant but wastage in energy production and consumption, dependence on primary energy consumption, over exploitation of resources have prevented the mitigation of carbon CO2, economic growth and environmental sustainability. Countries are becoming increasingly aware of the impacts and prospects of alternative energy consumption making the ratio of renewable energy in global energy consumption to rise. Wastage often occurs in extracting, transforming, and utilizing energy resources. This wastage reflects in environmental pollution, high costs of energy resources and rapid draining of energy resources which consequently, can threaten the environment, health and energy sustainability. Since the early 1980s, the ecological footprint is said to exceed the biocapacity and the gap is widening at the global level. This reflects the capacity of the essential natural resources to support human beings and animal species has reduced to some large extent. With regard to policy measures for improving energy efficiency and environmental quality, recent works (Akram et al. 2020; Ali et al. 2021; Alola et al. 2019a; Bekun et al. 2019; Caglar et al. 2021; Ike et al. 2020a, b, c; Köksal et al. 2020; Onifade et al. 2021; Rafindadi and Usman 2019, 2020; Sethi et al. 2020; Usman et al. 2020b; Zhang 2021) have suggested the implementation of energy conservation policies, improving research and development, promotion of renewable energy consumption as a substitute for fossil oils, sustainable land recuperation and focus on effective policy measures for different components of total ecological footprint, for instance, crop-land and development of fishing dams instead of concentrating on reducing carbon dioxide (CO2) alone. In 2017, Ewing has also submitted that to tackle environmental catastrophes with promptness and determination, policy makers need to adopt policies that can create the right path to ecological recuperation.

Figures 1 and 2 show graphical representation of ecological footprint and biocapacity in global hectares per person for MINT countries from 1980 to 2016. It can also be viewed that while the biocapacity has been continuously decreasing for all countries, the ecological footprint on the other hand has been on the increase although with little falls in certain years.

Fig. 1

Graphical representation of ecological footprints in g/ha for MINT Countries. Source: WDI (2019)

Fig. 2

Graphical representation of biocapacity in g/ha for MINT Countries. Source: WDI, (2019)

From OECD (2019), World Bank Group (2019) and World Data Atlas (2019), the economic profile of MINT countries is quite interesting. World Bank (2019) predicts that MINT shall rise to be among the first 20 largest economies with a total GDP value of approximately $22.4 trillion by 2030. Mexico is classified among the upper middle-income, ranks 2nd largest economy in Latin America with GDP growth of 2.14%, the 11th largest global economy with $2.6 trillion and GDP per-capita of $20,602 in 2018. Exports and imports of commodities formed 83.3% of GDP with FDI of about $32 billion. The country had economic freedom score of 66, the 67th freest economy in the 2020 global index and 11th among 32 nations in America. Indonesia, however, is classified among lower middle-income, a major producer of oil and natural gas which forms 10% of the country’s GDP; and is the largest economy in South Eastern Asia. The country had a growth rate of about 5.2%, with real GDP of $13,230, the 4th global producer and top exporter of coal, 10th global gas producer and 7th gas exporter in 2018. Indonesia had an economic freedom score is 67.2, the 54th freest in the 2020 global index and the 12th among 42 countries in the Asia–Pacific. Similarly, Nigeria is classified among lower middle-income, ranks the largest producer and exporter of oil with highest natural gas reserves in Africa. The country had 1.9% growth rate, $1.2 trillion GDP, with real GDP of $2033 in 2018 and an economic freedom score of 57.2%, and the116th economy in the 2020 global index. Lastly, similar to Mexico, Turkey is being classified among upper middle-income, ranks 36th economy among 45 nations in European region, the 13th largest global economy with GDP of $2.6trillion and real GDP of $27,956, with a growth rate of 2.6% in 2018. Total exports and imports equal about 60% of GDP with $12.9 billion FDI inflow. The country had an economic freedom scores of 64.4, and became 71st freest economy in the 2020 global index. Furthermore, considering the economic growth, large oil and agricultural production and other extractive activities in MINT, the region is significant not just for global economic development of the oil and gas markets, but also for the international commodity market. Exceptionally, MINT countries have abundant natural resources such as rubber, cocoa, palm trees, petroleum, coal, natural gas as well as hydropower, geothermal and solar potentials (World Bank Group 2019; OECD 2019).

With respect to challenges facing the MINT countries, the region faces hazards from energy inefficiency and increasing ecological footprint (Durotoye et al. 2019; Scherer et al. 2019). The wastage in energy production and consumption in MINT reflects in high CO₂ emissions, high costs of energy products and rapid draining of energy resources. Heavy reliance on primary energy consumption. For example, MINT countries accounted for about 9036 Twh, of about 7% of global fossil fuel consumption in 2014. There is also over-extraction of the natural resources, growing deforestation, wildfire, natural disasters and wrong disposal of urban wastes. All these factors culminate into high environmental pollution, growth volatility, energy waste, among others, resulting in severe harm to environmental quality in MINT. Intensive energy use and ecological footprint can constitute danger to environmental quality. Moreover, for countries to maintain sustainable energy and clean environment with fewer pressures on natural resources, individuals and nations must utilize smaller than 1.8 ha per capita of the available ecological footprints with 0.8 human development index (Cellier 2007). However, economists believe ecological footprint needs to be low under effective production and consumption of goods and services for sustainable growth and development. Therefore, it is imperative to investigate the effect of energy consumption on ecological footprint in MINT to expose the level of environmental degradation which to the best of our knowledge is the first of its kind.

More so, the percentage of annual CO₂ in MINT has been increasing based on its oil resources and intensified economic activities. As the natural resources like grazing land, forest reserve and game animals a country has are consumed constantly, the environment is overstretched and suffers from danger of extinction. Activities as intensive mining, cultivation and logging can destroy the environment including the resources involved over time. For example, Indonesia’s air pollution recently accelerated while forestry declined at an increasing rate in the last 10 years mainly due to bush fires in an attempt to clear land, particularly for oil palm and timber plantations giving rise to environmental pollution with its related health hazards (OECD 2018). Besides, fossil fuels consumption is generally high in MINT as is the case of other emerging countries, and being heavily depended on, although it is exhaustive in nature. For example, some experts believe oil can run out in the next 35 years, and gas in the next 55 years time, if global consumption continues at the current rate. Recently, rising economic growth has resulted in increased energy consumption causing trade-off between economic growth and environmental quality, a difficult task (Dogan et al. 2019). Additionally, statistics reveal that energy consumption in MINT is still low even though the bloc has enormous potential energy sources. Therefore, as they increase energy consumption in order to achieve economic growth, more environmental degradation could occur.

This study investigates the impact of energy consumption on environmental quality with focus on the MINT countries from 1971 to 2017. This study contributes to the extant literature, first, by investigating the impact of energy consumption on environmental quality in MINT, a region that has been neglected in the literature. From the extant literature, it is obvious the MINT region has been neglected in the energy-environment research as only a few researchers have focused on the region. Second, majority of the extant literature have also failed to extend the period of study beyond the year 2014, particularly for MINT, and only few have incorporated key variables employed in this study. For instance, most scholars employed CO₂ emissions as the DV, but we employ ecological footprint as DV which is regarded as a more comprehensive indicator of environmental degradation, and include biocapacity among the IV, a measure of environmental sustainability. Third, although there are numerous recent literature on the energy-environment nexus, the findings of these studies still differ considerably as shown in the reviewed literatures. Hence, we perceive that there is no compromise in the MINT region, so this study is imperative. Fourth, the study will be significant based on the policy implications of the study findings that can be applied to synchronize the energy and environmental policies to improve economic growth, energy efficiency and environmental quality in MINT countries. The significance of using ecological footprint as a measure for environmental quality centers on its calculation for not only the biological yield but also the technology and institutions available to make gains; this identification will update the data on MINT.

Review of related literature

Energy consumption and environmental quality

The role of energy consumption in energy-environmental quality by recent studies (like Alola et al. 2019a, b, 2020; Asongu et al. 2020; Bekun et al. 2019; Charfeddine 2017; Danish et al. 2019; Güngör et al. 2020, 2021; Ike et al. 2020a; Onifade et al. 2021; Paramati 2017; Rafindadi and Usman 2020; Sarkodie and Strezov 2018; Usman et al. 2020a) has been conducted which exposed the nexus between energy consumption and environmental degradation in the extent literature. In the extant literature, ecological footprint is accepted as a more comprehensive indicator of environmental degradation (Asongu et al. 2020; Cellier 2007; Ulucak and Bilgili 2018; Sarkodie and Strezov 2018; Usman et al. 2020a), among others. Empirical works have also shown that energy consumption plays a significant role towards environmental quality. A pattern change from some energy concentrated industrial firms to energy low firms can cut down the demand for energy, hence it decreases energy utilization, and consequently leads to less effects on environmental quality (Sarkodie and Strezov 2018). Although the relationship between energy consumption and environmental quality abounds in the literature, it has been observed that such relationship cannot be established without considering ecological footprint. Therefore, environmental variables like energy consumption, economic growth, biocapacity and urbanization can affect the productivity of the environment through ecological footprint; however, this fact does not exist for MINT. For example, economic growth and urbanization can present grave danger to environmental quality since energy consumption must match the increasing economic activities well as the urban population (Wang et al. 2016; Bayer and Gavriletea 2019; Güngör et al. 2021). The studies of Sarkodie and Strezov (2018) among others have also shown that energy consumption plays a significant role in environmental sustainability. Their PMG coefficients indicate that a rise in electricity consumption enhances energy intensity by 0.35% in the panel of 4 countries (Australia, China, Ghana and the USA) studied. The study by Alola et al. (2019b) validated the feedback hypothesis for France, Germany and the UK using the FMOLS from 1990 to 2016. Similarly, Odugbesan and Rjoub (2020) in their study on MINT have confirmed the energy–growth hypothesis for Nigeria and Indonesia, and the feedback hypothesis for Mexico and Turkey. Their findings also confirm a feedback causality between renewable energy consumption and CO₂. The FMOLS result indicated negative elasticities of CO₂ to renewable energy with −0.13, and renewable to CO₂ with −0.14. Phong (2019) has also focused on the ASEAN-5 countries and established that energy consumption reduces environmental quality and found a positively significant impact on greenhouse gas and CO₂. Rafindadi and Usman (2020) have assessed the role of economic growth, globalization and ecological footprint towards sustainable electricity consumption in Brazil, and found a negative and significant effect of ecological footprint on energy consumption at a 1% level. The PMG estimation result indicates that a 0.13% increase in electricity consumption is caused by a 1% positive reaction of ecological footprint. The study carried out by Ike et al. (2020a) on fiscal policy and CO₂ from heterogeneous fuel sources confirms that energy consumption reduces environmental quality through increased CO₂. Similarly, Paramati (2017) examined the significance of renewable energy use for economic output and environmental protection in the next 11 developing countries from 1990−2012. His FMOLS estimation result suggests that a 1% increase in fossil fuels consumption reduces environmental quality through CO₂ emission by 1.08% but a 1% rise in renewables improves the environment by −0.05%.

More recently, Alola et al. (2020) invested domestic material consumption and greenhouse gas emissions in the EU-28 countries: Implications for environmental sustainability targets and the PMG results reveal that a 1% rise in domestic material consumption exacerbated GHG emissions with 0.14% and 0.10% in both short run and long run periods, but a rise in renewable energy consumption reduced GHG. Using PMG Autoregressive Distributed Lag method, Asongu et al. (2020) have investigated the effects of economic growth, urbanization, electricity consumption, fossil fuels and total natural resources rent on CO₂ for Africa from 1980 to 2014 and affirmed bi-lateral causality nexus between electricity consumption and CO2. They also found a 1% rise in non-renewable energy leads to 1.34% increase in CO₂; this result is similar to Wang and Dong (2019) who found that 1% rise in non-renewable energy leads to 0.23% rise in ecological footprint in the long run. However, applying innovation accounting, Usman et al. (2020a) have investigated the role of renewable energy consumption and trade policy on environmental degradation in the USA, and have found that a 1% rise in renewable energy consumption reduces ecological footprint by 0.35% during the long run. Güngör et al. (2020) have also tested the EKC by incorporating the role of energy consumption and democratic accountability using the PMG in nine countries from 1990 to 2014 and confirmed that energy consumption significantly increased CO₂ in both the short run and long run periods, such that 1% rise in energy consumption would increase CO₂ by 0.79% and 0.81% in the short run and long run respectively. Further, mostly recently, the study of Zhang (2021) found 1% increase in energy consumption increases CO₂ by 0.9% in China in the long run period. Onifade et al. (2021) have employed the pooled mean group estimation technique to examine environmental aspect of energy transition and urbanization in eleven OPEC nations and find that 1% increase in fossil fuels consumption impacts positively on CO₂ by 0.25% is consistent with Bekun et al. (2019) and Güngör et al. (2020) but renewable had a negative effect on CO₂. Similarly, Güngör et al. (2020) tested the environmental Kuznets curve in the role of globalization, energy use and regulatory quality using the pooled mean group in South Africa from 1996 to 2016 and found that energy consumption has a long run positively significant relationship with environmental quality through ecological footprint. The long run PMG result indicates that a 1% rise in energy consumption will cause a 0.50% rise in ecological footprint. Finally, Caglar et al. (2021) have tested the role of ICT and renewable energy consumption in ecological footprint for global ten most pollutant footprint nations, and again, the positive link between energy consumption and environmental degradation is confirmed. Their pooled mean group result shows that a 1% rise in fossil fuels use will reduce environmental quality through 0.46% rise in ecological footprint; however, a unit increase in renewables decreases ecological footprint by 0.05%.

Economic growth, urbanization, biocapacity and environmental quality

The relationship energy consumption and environmental degradation has been conducted with the inclusion of key variables like economic growth, urbanization and biocapacity, among other factors by many recent literatures (see Alola et al. 2019a, b, 2020; Asongu et al. 2020; Bekun et al. 2019; Danish et al. 2019; Caglar et al. 2021; Güngör et al. 2020; Güngör et al. 2021; Ike et al. 2020a; Onifade et al. 2021; Rafindadi and Usman 2019, 2020; Sarkodie and Strezov 2018; Scherer et al. 2019; Sethi et al. 2020; Usman et al. 2020a; Zhang 2021), among others. Economic growth, urbanization and biocapacity have been proven empirically to influence ecological footprint besides energy consumption, but this fact does not exist for MINT. For example, population and urbanization have been reported to reduce environmental quality since energy consumption must match the increasing population size as well as the urban population (Charfeddine 2017; Wang et al. 2016; Bayer and Gavriletea 2019; Güngör et al. 2021). Twerefou et al. (2017) examined the environmental Kuznets curve in 36 Sub-Saharan African countries and analysed the effects of economic growth and globalization on environmental quality by applying the GMM. They proved the negative impact of economic growth on environmental quality, and the EKC theory was confirmed. Using a nonlinear ARDL method, Rafindadi and Usman (2020) conducted a study for Brazil and the results indicate that 1% increase in economic growth causes electricity consumption to rise approximately by 2.11% while a fall in economic growth will lower electricity consumption by 0.35%. This result is consistent with Baz et al. (2019) and they suggested environmental measures to be focused on sustainable economic growth as a positive or negative change in economic growth affects energy use unevenly. In their study on China, He and Yao (2017) tested the validity of the environmental Kuznets curve and pollution haven hypothesis; they found a negative correlation between economic growth and CO₂ but found a positive relationship between GDP and FDI, thereby affirming the pollution haven hypothesis. Danish et al. (2019) has examined the nexus between economic growth and the ecological footprint as it relates to biocapacity and human capital using the ARDL econometric approach and have reported that both economic growth and biocapacity impacted on ecological footprint positively. However, their causality estimates show no causality between economic growth and ecological footprint. Contrary to Danish et al. (2019), Ozatac et al. (2017) have validated existence of long run correlation between economic growth and ecological footprint as they recorded ecological footprint impacts on income during the long run for both middle- and high-income countries. The study of Sarkodie and Strezov (2018) shows a one-way causality running from economic growth to biocapacity and a bidirectional nexus between ecological footprint and biocapacity. And they concluded that economic growth is being propelled by intensive polluting energy resources which aggravate GHG and damages the environment. Caglar et al. (2021) also found positive impact of fossil fuels on economic growth and ecological footprint, and a bidirectional interaction between GDP and ecological footprint, thereby validating the feedback hypothesis, although the EKC hypothesis was not validated for any the countries studied.

Furthermore, a more recent study in Güngör et al. (2020) records a positive and significant impact of GDP on environmental degradation at 1% level for 9 countries. Their PMG estimates reveal a 1% rise in economic growth increases CO₂ by 2.28% while real GDP square reduces CO2 by −0.25%. Similarly, Ike et al. (2020a, b) in their study on oil-producing countries have proven a 1% rise in economic growth promotes environmental degradation by a 1.24% while GDP square improves environmental quality by −0.03% in the long run. Bekun et al. (2019) also have affirmed positive impact of economic growth on CO₂ emissions at a 1% level of significance. However, Alola et al. (2020) have investigated domestic material consumption and GHG in the EU-28 countries with implications for environmental sustainability targets and obtained a contrasting result with a negative sign of GDP on greenhouse emissions; a 1% rise in GDP per capita reduces GHG by−0.06% in the long run but a positive impact in the short run increases GHG by 0.23%. Most recently, Zhang (2021) also finds that 1% increase in GDP per capita increases CO₂ by 6.62% but real GDP square would reduce CO₂ by −0.82% but a 1% rise in urbanization reduces CO₂ by −0.26% in the long run. Contrary to the negative effect of urbanization on CO₂ reported in Sarkodie and Strezov (2018), Ike et al. (2020a, b, c), Zhang (2021) and other researchers, Phong (2019) in his study on ASEAN-5 countries finds that urbanization positively impacts CO₂, indicating that an increase in urbanization will also worsen CO₂. Dogan et al. (2019) in their study on MINT have revealed that urbanization has positive effect on CO₂ in Mexico, Indonesia and Turkey but established a negative impact for Nigeria in the long run period. Similarly, Asongu et al. (2020) affirms a significant and positive impact of urbanization on CO₂ for Africa. Accordingly, a 1% rise in urbanization worsens environmental quality through CO₂ by 0.39%.

Additionally, in a completely new light, the PMG result of Onifade et al. (2021) shows a negative nexus between both economic growth and urbanization with CO₂ at 1% level for OPEC member countries in the long run. They find a 1% increase in GDP, and urbanization reduces CO₂ by −0.11% and − 2.71%, respectively while a unit rise in GDP square impacts positively on CO₂ by 0.01% in the long run. These results contradict that of Alola et al. (2019a), Asongu et al. (2020), Bekun et al. (2019), Güngör et al. (2020), Ike et al. (2020a) and Zhang (2021), among others. Sarkodie and Strezov (2018) also found that 1% increase in urbanization will reduce energy intensity by 0.10% in Australia, 0.02% in the USA in the short run, but urbanization increases energy intensity in China and Ghana by 0.17% and 0.07% respectively. Their findings equally reveal that a 1% rise in biocapacity will reduce environmental impact in by 0.23% and 0.22% in Australia and China respectively during the short run. Employing the dynamic ARDL simulations, Ali et al. (2021) modelled the effects of income, agricultural innovation, energy consumption and biocapacity on CO₂ in an Agrarian based-economy from 1981 to 2014 and their ARDL result reveals that biocapacity has a negative impact on CO₂; a 1% rise in biocapacity decreases CO₂ by about −1.12% and − 0.19% in the short run and long run in respectively. Their causality result shows that changes in CO₂ can be predicted by biocapacity. Finally, Sarkodie and Strezov (2018) have established a one-way causality relationship from GDP growth to biocapacity. And they concluded that economic growth is driven by intensive polluting energy resources which aggravate greenhouse gases and environmental pollution because of too much dependence on primary energy consumption (Sarkodie and Strezov 2018).

Given the trend of the extant literature (see Alam et al. 2016; Ali et al. 2021; Alola et al. 2019a, b, 2020; Asongu et al. 2020; Bekun et al. 2019; Danish et al. 2019; Caglar et al. 2021; Dogan et al. 2019; Güngör et al. 2020, 2021; Ike et al. 2020a, b, c; Onifade et al. 2021; Phong 2019; Rafindadi and Usman 2020; Riti et al. 2018; Shao et al. 2019; Sarkodie and Strezov 2018; Scherer et al. 2019; Sethi et al. 2020; Ssali et al. 2019; Tariq 2018; Ulucak and Bilgili 2018; Usman et al. 2020b; Zaman and Moemen 2017; Zhang 2021) among others on the energy consumption-environment nexus, there appears a wide gap in MINT countries. Using the FMOLS method, Alola et al. (2019a) examined the role of renewable energy, immigration and real income in environmental sustainability target giving evidence from Europe largest states. They focused on economic growth-CO₂; however, taking another part, we examine energy consumption-ecological footprint nexus using a panel data from 1971 to 2017 for MINT. We also used the PMG technique instead of FMOLS and our variables are entirely different from each other except GDP. Alola et al. (2019b) also investigated the dynamic impact of trade policy, economic growth, fertility rate, renewable and non-renewable energy consumption on ecological footprint in 16 Europe from 1997 to 2014. Our variables differ considerably except for RGDP and Alola et al. (2019b) used a short panel data, but we employed data from 1971 to 2017 for MINT. Asongu et al. (2020) assessed the impact of economic growth, urbanization, electricity and fossil fuel consumption on the environment for 13 SSA nations from 1980 to 2014. Their study is on income-CO₂ nexus; contrary, this study is focused on energy consumption-ecological footprint nexus on the MINT countries (Mexico, Indonesia, Nigeria and Turkey) with extension of data up to 2019 and inclusion of biocapacity. Bekun et al. (2019) invested economic growth-CO₂ nexus in 16 EU-countries employing balanced data from 1996 to 2014. Our study applies unbalanced data from 1971 to 2017 with inclusion of two new variables; biocapacity and ecological footprint to examine energy consumption-ecological footprint nexus for the MINT countries. Onifade et al. (2021) studied the environmental aspect of energy transition and urbanization in eleven OPEC member states from 1990 to 2014 using PMG/ARDL. Their work focuses on fossil fuels-CO₂ nexus with a short panel, departing; this study focuses on energy consumption-ecological footprint nexus in MINT with the expansion of study period from 1971 to 2017 using an unbalanced data with the inclusion of new variables; ecological footprint and biocapacity. Alam et al. (2016) examined the relationship among CO₂, economic growth, energy consumption and population growth in Brazil, China, India and Indonesia using the ARDL and VECM methods from 1970 to 2012. They employed time series data, but our study uses a panel data analysis to investigate the impact of energy consumption on environmental quality for the MINT countries (Mexico, Indonesia, Nigeria and Turkey, extending data over more current years (1971–2017)).

Furthermore, Ali et al. (2021) modelled the effects of income, agricultural innovation, energy consumption and biocapacity on CO₂ in Nigeria from 1981 to 2014. While their study applies time series data, our study uses a panel data from 1971 to 2017 and also uses ecological footprint as indicator for environmental quality instead of CO₂. Usman et al. (2020b) study the role of energy consumption on environmental degradation in the USA with quarterly data from 1985 to 2014. They used PMG and Toda-Yamamoto causality test while we applied PMG and H-D Granger no-causality test, and also focused for renewable-ecological footprint nexus while this study focuses on energy consumption-CO₂ nexus for MINT with yearly data extended beyond 2014. Güngör et al. (2021) investigated the impact of energy use on ecological footprint for South Africa from1996 to 2016 using VECM model. They used time series data but this study uses PMG for MINT using panel data from 1971 to 2017. Rafindadi and Usman (2020) have studied the role of economic growth, globalization and ecological footprint on energy consumption in Brazil (1971–2014) using non-linear ARDL model. Our methodological techniques, variables and study period differ from each other. Shahbaz et al. (2018) examined the link among globalization, economic growth and energy consumption in BRICS region using ARDL from 1970 to 2015. Study applied time series analysis, but our study takes a different path by applying PMG technique, a panel data model to investigate the relationship among energy consumption, real GDP, urbanization, biocapacity and ecological footprint for MINT. Shao et al. (2019) examined the pollution haven hypothesis in BRICS and MINT between 1980 and 2014, using the VECM model. The study was limited by inadequate indicators for environmental performance. Danish et al. (2019) studied Pakistan from 1971 to 2014, and applied the ARDL method. Study employed a time series analysis at country level, results cannot be generalized. On the contrary, our study is based at a regional level with findings that can be comfortably generalized across groups, using a heterogeneous data, the PMG and MG methods. Dogan et al. (2019) were the first to examine the determinants of ecological footprint in MINT and validated the EKC for each country using time series from 2071 to 2013. They used four independent variables; energy use, exports, financial development and urbanization. This study extends the span period beyond 2013, and includes a new variable, biocapacity. Besides, we applied the PMG, a panel data analysis to cover more recent years in MINT. Sarkodie and Strezov (2018) tested the environmental sustainability curve hypothesis in Australia, China, Ghana and USA from 1971 to 2013 using the PMG, and Dumitrescu and Hurlin (2012) non-causality test. They suggested extension of the ESuC to determine the main contributing factors in other regions.

Having identified the gaps in extant literatures above, this study therefore fills these gaps through the above highlighted ways. Following Sarkodie and Strezov (2018), this study, therefore, contributes to the literature, first, by investigating the impact of energy consumption on environmental quality in MINT, a region that has been neglected in the literature. From the extant literature, it is obvious that MINT region has been neglected in the energy-environment research as only a few researchers have focused on the region. Second, majority of the extant literature have also failed to extend the period of study beyond the year 2014, particularly for MINT, and only few have incorporated key variables employed in this study. For instance, most scholars used CO₂ as the DV, but we have employed ecological footprint as DV which has been viewed as a more comprehensive measure of environmental quality, and include biocapacity among the IV, a measure of environmental sustainability. Third, although there are numerous recent literature on the energy-environment nexus, the findings of these studies still differ considerably as shown in the extant literature reviewed. Hence, we perceive that there is no compromise in the MINT region, making this study imperative.

Data and methodology

Theoretical framework

The ecological modernization theory rests on the long run model in environmental economics that posits that economic growth causes environmental hurt, but contends that the progress in development could lessen environmental issues (Grossman and Krueger 1995). The force of modernization is capable of moving human society from its past of environmental degradation towards sustainability in modern societies, which includes markets, know-how and industrialisation (Mol 2002, 2016). The process of modernization has the potential to progressively reduce negative ecological impacts on countries, and although it is possible to achieve an inverse relationship economic growth and environmental degradation at early stage of development in developed nations, as emerging economies continue to grow, the degree of environmental impacts will ultimately reduce (Jorgenson 2016). The ecological modernization theory has also been defined as the restructuring of the society towards effective production, and efficient energy use to achieve environmental sustainability (Mol and Spaargaren 2007). In this regard, the theory is a model of progressive transition from pre-modernization to a modern society, implying that traditional societies develop when they adopt some additional modern technologies (Mol 2002). As the society develops, the environs and natural resources are incorporated into commercial and states accounts with the result that modern civilization experiences development of the economy and conservation of the environmental resources. Therefore, the re-rationalization of society for effectiveness and efficiency of energy consumption to achieve high environmental quality and sustainability is the primary contention of the ecological modernization theory (Mol and Spaargaren 2007). Hence, energy conservative indicates that improvements in growth and development have the potentials to raise energy consumption. However, researchers have noted that an increase in energy consumption might not necessarily contribute to GDP growth. By implication, the conservation hypothesis assumes that energy conservation policies that are directed at curbing environmental degradation, energy efficiency and waste control may not automatically reduce economic growth rate.

The growth hypothesis suggests a unidirectional causal relationship between economic growth and energy use in which energy use has a positive impact on GDP growth, and this indicates that energy use can stimulate economic growth and sustainable development (Nugraha and Osman 2017). The feedback hypothesis indicates a bidirectional causality between GDP and energy consumption with policy suggestions that are akin to the growth hypothesis on the energy-growth related researches (Rafindadi and Ozturk 2015; Shahbaz 2016). These suppositions had directed policymakers and researchers in the energy sector to make updated energy-growth related macroeconomic strategies that could have fewer consequences upon the environment, growth process and economic development of countries (Usman et al. 2020c). The ecological footprint measures the number of land resources that an individual requires to facilitate the production of the items desired for human consumption like foodstuff, clothes, energy needs, housing and the implements required to produce other goods and services (Cellier 2007).

The ecological footprint indicates the level of wasteful utilization of land resources in some specific regions and stipulates vast influence zones that require more responses (Castellani and Sala 2012). The estimation of the ecological footprint is essential for creating awareness of human influence on the environment and the knowledge of minimum hectares of land resources per-capita for individuals and countries is important Moffatt (2000) cited in Cellier (2007). Theoretically, energy consumption and ecological footprint have a direct relationship. But empirically, there remain lots of differences in result estimates. Individuals have ecological footprints since everybody requires a portion of the earth’s natural resources to stay alive, and this fact explains why we should ensure we do not overuse the resources in our environment. The ecological footprint is also said to have advantage because it can be interpreted in the context of environmental quality. On the other hand, energy consumption has the advantage of accessible and accurate computation; hence, this nexus is important for policy making. The ecological modernization theory has continued to be used and the conservation hypothesis debated over many years in the energy–environmental nexus. Since ecological footprint has been empirically proven to be a relevant and more comprehensive indicator of environmental degradation, we used ecological footprint as the dependent variable while incorporating some chief variables as regressors include: energy consumption, economic growth, urbanization and biocapacity that are seen to have a vital influence on the environment. Hence, following Sarkodie and Strezov (2018), we arrive at our model specification as shown in Eqs. (1), (2) and (3).

Model specification

This study diligently follows the model as developed of Sarkodie and Strezov (2018), which takes the following specification form:

$$ EFPT=f\left( RGDP, PEC, URN, BIOC\right) $$

(1)

where EFPT denotes ecological footprint, a measure of environmental quality. RGDP denotes real gross domestic products, used as proxy for economic growth. PEC is primary energy consumption per-capita comprising commercially traded fuels, including modern renewables used to generate electricity, while URN represents urbanization and measures the total number of people living in urban areas. Lastly, BIOC represents biocapacity, an indicator to gauge environmental sustainability. Based on the theory, the expectation signs of the selected explanatory variables can be explained as follows: for energy consumption, it is expected to have a positive sign. This implies that as an economy increases energy consumption to accelerate the pace of economic growth, environmental quality would be depreciating through increases in ecological footprint. Hence, energy consumption increases (causing a boost in economic activities), there is the tendency for quality of the environment to deteriorate due to an increased carbon dioxide emissions from the economic activities (Banday and Aneja 2019; Ike et al. 2020a, 2020b, 2020c; Ssali et al. 2019), among others. For the economic growth (RGDP), the expected sign will be a positive impact on ecological footprint such that as the economic growth rises, environmental pollution rises, reducing the ecological footprint available. In other words, as economies pursue economic growth, more land and resources are exploited tending towards ecological footprint decreases.

Consequently, there is the tendency for the quality of the environment to deteriorate due to an increased carbon dioxide emissions as GDP pursuit is the primary reason for CO₂ since it has been empirically proven that economic growth has a positive relationship with environmental pollution (Ike et al. 2020a, 2020b, 2020c; Özcan and Özkan 2018; Shahbaz 2016; Uddin et al. 2019; Usman et al. 2020b). Then, the urbanization (URN) is expected to be negatively related to ecological footprint. This shows that an increase in urbanization can lead to decrease in ecological footprint which means an increase in urbanization could improve the quality of environment. The implication of this is that as people living in the urban areas in a nation increases, there is tendency for the people to be conscious of their environment, which in turn leads to efficient utilization of energy in the country. Another likely explanation to the negative effect of urbanization is that environmental laws are more stringent in the urban areas compared to the rural areas (Asongu et al. 2020; Bekun et al. 2019; Güngör et al. 2020; Ike et al. 2020a; Liu et al. 2020; Ozatac et al. 2017; Pata 2018; Zhang 2021). This implies that a rise in environmental deterioration is attributed to increase in biocapacity (like Danish et al. 2019; Sarkodie and Strezov 2018; Usman et al. 2020a).

Next, the Eq. (1) can be also expressed in a panel regression form as follows:

$$ {EFPT}_{i,t}={\beta}_0+{\beta}_1{EFPT}_{i,t-1}+{\beta}_2{RGDP}_{i,t}+{\beta}_3{PEC}_{i,t}+{\beta}_4{URN}_{i,t}+{\beta}_5{BIOC}_{i,t}+{\mu}_{i,t} $$

(2)

where β₀is a constant; β₁₋β₅ represent the coefficients of the selected independent variables, and they are used to capture country-specific effects. μ_{i, t} is the stochastic disturbance or error terms, i represents number of country i = 1,2,...,N: N = 4; while t denotes number of years t = 1,2,...,T: T = 43 years (1971–2014).

Also, for the stability of variances and to enable the empirical results to be interpreted in form of elasticities, we transform the variables into logarithmic form as follows:

$$ {lnEFPT}_{i,t}={\beta}_0+{\beta}_1\mathit{\ln}{EFPT}_{i,t-1}+{\beta}_2\mathit{\ln}{RGDP}_{i,t}+{\beta}_3\mathit{\ln}{PEC}_{i,t}+{\beta}_4\mathit{\ln}{URN}_{i,t}+{\beta}_5\mathit{\ln}{BIOC}_{i,t}+{\mu}_{i,t} $$

(3)

Figure 3 shows the distribution plots for CDF Empirical Survivor of all logs of the coefficients of the dependent varable (DV) and  independent variables (IV). It can be observed that the variables curves are all sloping upwards, from left to right and this implies that the independent variables have a direct relationship with the dependent variable such that; for example, as energy consumption increases, ecological footprint increases also.

Fig. 3

Distribution Plots for CDF Empirical Survivor of the coefficients of all variables

Panel data

The panel data employed in this study include ecological footprint (EFPT) which serves as the dependent variable, while economic growth (RGDP), primary energy (PEC), urbanization (URN) and bio-capacity (BIOC) form the independent variables that covers from 1971 to 2017. Ecological footprint is defined as the number of land resources that an individual requires to facilitate production of some whole lot of items desired for human consumption like foodstuff, clothes, energy needs, housing and the implements required to produce other goods and services (Cellier 2007). It is measured in Global hectares per person (gha) and retrieved from the Global Ecological Footprint Network database. While the explanatory variables include RGDP denotes real gross domestic products, used as a proxy for economic growth. EC represents energy consumption and consists of the total energy consumption of oils, natural gas, electricity and combustible renewable and waste. URN represents urbanization and measures the total number of people living in urban areas. Lastly, BIOC represents biocapacity, an indicator to gauge environmental sustainability. All data are retrieved from the World Bank development indicator (WDI) database and British Petroleum (BP) statistical reviews except ecological footprint which was taken from the Global Ecological Footprint Network database. The measurement, proxies, expected signs and empirical studies that used these variables have been stated under the empirical model specification. Also, a summary of all these variables description, unit of measurement and their sources are given in Table 1.

Table 1 Summary of data description

Table 2 presents the descriptive statistics and the correlation matrix. The results show that economic growth has the highest average score of 9.61 followed by primary energy with 4.41 while bio-capacity has the lowest mean score of 0.95. The primary mean energy consumption is 3.56 with a maximum log score of 4.41. The bio-capacity mean score is 0.32 with a maximum of 0.95 while maximum ecological footprint is 1.36 and maximum bio-capacity is 0.95 for the MINT countries during the period investigated. When we compare the mean score for the ecological footprint to that of bio-capacity, the ecological footprint (0.57) is greater, giving us a deficit of 0.24.

Table 2 Descriptive statistics and correlation matrix

Also, it can be viewed that the correlation between ecological footprints and all the key variables; energy consumption, economic growth, urbanization and biocapacity is positive and significant at a 1% level except for the case of urbanization and biocapacity which has a negative but insignificant correlation. The correlation between energy consumption, urbanization, biocapacity and economic growth is found to be also positively significant at 1%. These results contradict those from Usman et al. (2020c) who find a negative relationship among ecological footprint and key variables like renewable energy, economic growth and trade policy measure but find a positive correlation between ecological footprint and the biocapacity.

Figure 4 presents line plots of logs of all variables employed in the study. Ecological footprint, economic growth and primary energy consumption show a dynamic and unsteady increases across the years covered. Similarly, urbanization shows a continuous steady increase; however, biocapacity shows a wavy and continuous decrease over the years covered.

Fig. 4

Graphical presentation of the logs of all variables used

Estimation techniques

The study employs the pooled mean group technique to estimate the long-run association among variables across countries by not only making homogeneity noticeable among the short-run estimates based on the panel Autoregressive Distributed Lag (ARDL) system. The PMG method estimates both the short-run and long-run coefficients and is suitable for small sample data set as in the case of MINT with only four countries. In other words, based on Pesaran et al. 1999, the PMG technique is applied in cases where T happens to be larger than N and also, has a significant advantage of accommodating variables with some mixed integrations since the bounds range from lower 1(0) to upper bound 1(1), even in the cases when all variables are mutually integrated, that is, 1(1). The pooled mean group works between the Mean Group (MG) and the Dynamic Fixed Effects (DFE), also known to be more efficient than the DFE since it estimates the dynamic heterogeneous panels by pooling effects, applying homogeneity restrictions on the long-run coefficients and averaging across groups for the short-run parameters of the model as well as the error term coefficients. The main advantage of PMG over the conventional DFE is that the model allows for short-run dynamic specification to vary across countries (Pesaran et al. 1999).

Going by the advantages of PMG, this study employs the PMG/ARDL modelling technique in order to investigate the long-run relationship between the dependent variable and its regressors. The panel ARDL equation in (3) is derived from the short-run ARDL model given below.

For an unrestricted specification of ARDL equations for periods t = 1,2, … T, with number of countries i = 1,2, … N, the explained variable Y is:

$$ {y}_{it}={\beta}_0+\sum \limits_{i=1}^p{\beta}_{1 it}\ {y}_{i,t-i}+\sum \limits_{i-1}^q{\beta}_{2 it}{x}_{i,t-i}+{\mu}_i+{\varepsilon}_{it} $$

(4)

where x_{i, t − i} is the (k × 1) vector explanatory variables of countries, i; β_it, are the (k × 1) coefficients’ vectors and μ_i represents the fixed effect of the countries. Model (4) above can be reparametrized into a VCM system. Equation (4) shall be used to derive the long-run model for Eq. (3).

$$ {y}_{it}={\gamma}_0+{\gamma}_1{x}_{it}+{\gamma}_2{x}_{it}+{\gamma}_n{x}_{it}+{\mu}_i+{\varepsilon}_{it} $$

(5)

${\gamma}_0=\frac{\beta_0}{1-\sum {\beta}_{1 it}}$, ${\gamma}_1=\frac{\beta_1 it}{1-\sum {\beta}_{1 it}}$,${\gamma}_2=\frac{\beta_{2 it}}{1-\sum {\beta}_{1 it}}$,${\gamma}_n=\frac{\delta_{nit}}{1-\sum {\delta}_{1 it}}$ (6).

We can use the residuals of the long-run model to have the error correction model (ECM) as given below:

$$ \Delta {Y}_{it}={\varphi}_0+\sum \limits_{i=1}^{n-1}{\varphi}_{1 it}\Delta {Y}_{i,t-1}+\sum \limits_{i=1}^{n-1}{\varphi}_{2 it}\Delta {X}_{i,t-1}+{\theta ECT}_{it-1}+{\mu}_i+{\varepsilon}_{it} $$

(7)

where the error correction term, ECT_it − 1, represents the long-run model residuals in Eq. (3), lagged by one period (t-1). It is defined below as:

$$ {ECT}_{it-1}={\mu}_{it-1}={Y}_{it-1}-\left[{\gamma}_0+{\gamma}_1{X}_{1 it}+{\gamma}_2{X}_{2 it}+{\gamma}_n{X}_{nit}\right]\beta $$

(8)

The error correction parameter δ indicates the speed of adjustment, such that when δ = 0, it is evidence of no long-run association between the variables, expected to carry a negative sign and as well as significant to be able to indicate the evidence of convergence or speed of adjustment of the variables to long-run equilibrium. However, based on the PMG assumptions, the element γ is made the same across countries. These assumptions are ε_it, are independently distributed across i, t and the regressors x_it, having 0 mean, and σ_i^t > 0, and limited fourth-order moments. The most required of these assumptions is the independence between disturbances and the explanatory variables, in order to have consistency in the short-run estimates.

In the case of the long-run homogeneity, the parameters are identical across countries given as: θ_i = − γ_i / φ_i, namely θ_i and θ_n, i = 1,2, … N. Maximum likelihood estimation is used in computing both the country-specific short-run parameters and average long-run coefficients. The PMG parameters of interest are the long-run association and the speed of adjustment, with the parameters consistent and asymptotically normal for both stationary I(0) and non-stationary I(1) explanatory variables (Pesaran et al. 1999).

Prior to the PMG test, the Pesaran cross-sectional dependence test is taken as displayed in Table 3. The results show that the Pesaran C-D test for all the variables is significant at 1% level. This implies the presence of cross-sectional dependence for the panel data across all countries. The implication of this result is that applying a first generation panel data estimation technique would lead to misspecification and spurious results. Therefore, our analysis is based on the second generational panel data technique, the PMG under auto-regressive distributed lag (ARDL) model.

Table 3 Pesaran cross-sectional dependence test

Panel unit root tests

In this research, to test for the integrating or stationarity properties of the time series employed, we applied four kinds of panel unit root tests; the Breitung (2000), the Im, Pesaran and Shin (2003) (IPS), the Maddala and Wu (1999); consisting of the Augmented Dickey-Fuller (ADF) -Fisher Chi-square and the Phillips-Perron (PP) Chi-square panel non-stationarity tests. First, we employed the Breitung (2000) test, which assumes a common autoregressive parameter for all individuals in the panel, which is relaxed in the other panel unit root tests. Second, we tested the IPS (2003), which permits for heterogeneity of the coefficient of the lagged dependent variable. The IPS is said to be built on averaging only one unit root of the augmented Dickey-Fuller test data of countries, where the IPS t-bar statistics is usually likened to the IPS critical value. Third, we also tested the ADF-Fisher type test that combines p values from the unit root test of individual countries in the panel data. Finally, we applied Phillips-Perron (PP), where probabilities for the Fisher tests are computed using an asymptotic chi-square distribution. All other tests assume an asymptotic normality. An assumption of non-stationarity is the null hypothesis for all of these tests. We made use of the automatic Akaike information criterion (AIC) to select the minimum and maximum lags. These results are presented in Table 4 for all panel unit root tests conducted. The results display the nonexistence of unit root in the variables at either level, 1(0) or first difference, 1(1) with deterministic trend. These results therefore indicate that all our variables are mutually co-integrated at both level (i.e. order zero) 1(0) and at first difference (i.e. order one) 1(1).

Table 4 Panel unit root tests with deterministic trend

Panel co-integration test

Having found that all the variables employed in this study are integrated of I(0) and I(1), the next stage is to determine if there is existence of co-integration among these variables. To do this, we employed the Pedroni residual-based co-integration test with no deterministic trend and with deterministic intercept and trend. The results as presented in Table 5, panels A and B, reveal to us that the statistical values of Panel-PP and Group-PP as well as its weighted values are statistically different from zero. The table shows that 5 statistics, the Pane v-Statistics, Panel rho-Statistics, Panel PP-Statistics, the Group rho-Statistics and the Group PP-Statistics (out of seven statistics), are significant. The implication of these results is that there is a valid long-run relationship between the variables under consideration.

Table 5 Panel co-integration tests

Furthermore, to test the robustness of the Pedroni co-integration test, we employed a panel co-integration test developed by Westerlund (2008) which controls for the cross-sectional dependence in the series using four panel co-integration tests. The first two tests used mean group statistics while the last two tests used panel statistics. The results in Table 6 reveal that at a 5% and 10% level of significance, the null hypothesis of no co-integration is rejected in two (Gt and Pt tests) of the four-panel tests. This confirms that there is a valid long-run relationship between the variables employed in this study.

Table 6 Westerlund (2008) panel cointegration test

Pooled mean group (PMG/ARDL) result

Table 7 shows the short run and long run coefficients of the PMG estimation results for MINT countries. During the long run period, the results reveal that economic growth, economic growth, primary energy consumption, and bio-capacity all have a positive and significant impact on environmental degradation (ecological footprint), while urbanization has a negative sign. This implies that economic growth, primary energy consumption and biocapacity can lower environmental quality in MINT while urbanization can improve environmental quality but it is insignificant in the case of this study. Particularly, a 1% increase in economic growth would deteriorate the environment via increase in footprint by 0.5076% in the short run and 0.4288% in the long run period. This result implies that as economic growth increases there is tendency for the quality of the environment to deteriorate due to an increased carbon dioxide emission. Therefore, the finding of this study is consistent with Usman et al. (2020a) who found economic growth to be associated with ecological footprint while controlling for trade policy and bio-capacity for the USA.

Table 7 Pooled mean group (PMG/ARDL) result

With respect to primary energy, we find that a 1% rise in primary energy consumption leads to 0.4172% rise in environmental deterioration in the long-run period, while, in the short run period, primary energy is insignificant. This means that as MINT increases the consumption of energy to accelerate the pace of economic growth, the quality of the environment would be deteriorating and consequently, resulting to a reduction in the ecological footprint. This finding is consistent with the results of Asongu et al. (2020), Destek and Sarkodie (2019), Onifade et al. (2021), Uddin et al. (2019), Usman et al. (2019) and Rafindadi and Usman (2020). On the other hand, the positive effect of energy consumption on ecological footprints contradicts the findings of Köksal et al. (2020), who found a negative impact of energy consumption on ecological footprint for some Asian economies. Similarly, our result provides that 0.1409% and 1.3625% increase in environmental deterioration during the short run and long run periods are attributed to a 1% increase in bio-capacity respectively. This result is consistent with Usman et al. (2020a) who found bio-capacity to reduce ecological footprint in the US in both the long run and short run.

The effect of urbanization on ecological footprint is negative for the MINT countries; however, it is insignificant. This implies that a 1% increase in urbanization could improve environmental quality by −0.2637% if it were significant in the long run. The economic intuition behind this result is that as number of people living in the urban areas of the region increases, there is tendency for them to be conscious of their environment by adhering to all the environmental policies. In other words, environmental laws are more stringent in the urban areas compared to the rural areas. This in turn leads to efficient utilization of energy in the MINT region. Another likely explanation to the negative effect of urbanization on ecological footprint is that the migration from rural areas to urban areas enhances technological advancement particularly from non-renewable energy consumption to renewable (clean) energy consumption. The transition from non-renewable to renewable energy consumption accompanied with the movement of the people from rural areas to urban areas could be due to the engagements of the people with the multi-national corporations cited in the urban areas. These corporations are agents of technological diffusions and technical know-how that aid energy transitions. Therefore, our finding concurs with Shahbaz (2016) who found urbanization to have a negative impact on carbon dioxide emissions in MENA countries. The urbanization-triggered environmental quality channel is consistent with Güngör et al. (2021) who found globalization to have reduced environmental degradation for South Africa through technological diffusions stimulated by economic globalization.

Finally, the coefficient of the lagged term of error correction (ect), i.e. − 0.2306, is negative and statistically significant. This suggests that ecological footprint, a measure of environmental degradation, would converge to their long-run equilibrium within the MINT region by 23.06% speed of adjustment every year due to contribution of economic growth, energy consumption, urbanization and biocapacity.

Table 8 shows the short-run cross-section coefficients for individual MINT countries to indicating the country specific effects. The short run coefficients for energy consumption and economic growth are positively significant at 10% and 5%. But urbanization and biocapacity are insignificant. The coefficient of the lagged error correction term (ect) is negatively significant (− 0.2687) and this implies that ecological footprint will converge to its long-run equilibrium for Mexico by 26.87% speed of adjustment yearly due to contribution of GDP growth, and primary energy, urbanization and biocapacity. For Indonesia and Nigeria, energy consumption, economic growth and biocapacity are significant except urbanization. Similarly, the coefficient of the lagged error term for Indonesia and are negatively significant with −0.2945, and − 0.2344, respectively. These results suggest that ecological footprint will converge to its long-run equilibrium for Indonesia and Nigeria by 29.45% and 23.44% speed of adjustment every year due to contribution of the all the explanatory variables. Finally, for Turkey, all variables are significant except urbanization, and the coefficient of lagged error term is also negatively significant with −0.1245. This implies that ecological footprint for Turkey will converge to its long run equilibrium by 12.44% speed of adjustment year by year as a result of contribution of the variables.

Table 8 Short-run cross-section coefficients for MINT countries

Conditional Granger non-causality test

Table 9 presents the results of the Dumitrescu-Hurlin panel causality tests as shown in the table below. We report the Dumitrescu-Hurlin Granger causality tests which are based on the W-stat and Zbar-Stat as well as their probability values with regard to the predictability for the both the dependent and independent variables. From the results reported in Table 9, there is a unidirectional (one-way) causality flowing from economic growth, energy consumption, and urban population to ecological footprint in the region under consideration. This means that the past values of economic growth, energy use and urban population are found to be predictors of ecological footprint. The finding is therefore consistent with the Usman et al. (2020a, 2020b). It is also in agreement with Ike et al. (2020a) who found GDP to Granger-cause environmental pollution in the G7 countries. The results of the Granger non-causality also reveal that biocapacity and ecological footprint have a bi-directional causality while biocapacity and economic growth also established a bi-directional causality in MINT countries.

Table 9 Dumitrescu-Hurlin Granger non-causality test

Moreover, our causality results show that the causal relationship between and ecological footprint is bidirectional, which means that both biocapacity and ecological footprint predict each other. This result corroborates Sarkodie and Strezov (2018) and Ali et al. (2021) who found a bi-directional causality between biocapacity and environmental degradation but contrary to Usman et al. (2020a) who found that ecological footprint not having a causal relationship with biocapacity for the USA. We also find a unidirectional causality running from energy consumption to urbanization and biocapacity. This finding is contrary to Shahbaz et al. (2016) who found urbanization to Granger-cause energy consumption in Malaysia. Furthermore, we found a unidirectional causality running from urbanization to real economic growth and biocapacity to urbanization. However, there is no causality between energy consumption and economic growth. This result is contrary to Rafindadi and Usman et al. (2019), Iorember et al. (2021) and Akadiri et al. (2021) who found a causal relationship between energy consumption and economic growth in South Africa.

Conclusion and policy implications

This study investigated the effects of economic growth, energy consumption, urbanization and biocapacity on ecological footprint in MINT countries—Mexico, Indonesia, Nigeria and Turkey between the periods 1971–2017. We applied the Pedroni residual-based and Westerlund (2008) panel cointegration tests, the pooled mean group (PMG) estimation technique under the auto-regressive distributed lag (ARDL) by Pesaran et al. (1999), and the Dumitrescu-Hurlin (2012) Granger non-causality test. The findings of this study suggest a long-run relationship between the variables as all variables are statistically significant at 1% level except for urbanization. The panel unit roots tests results show that all variables are integrated at order zero or level 1(0), and at first difference 1(1). The PMG results show that both in the long run and short run, energy consumption, economic growth and biocapacity exacerbate the environment through a negative impact of ecological footprint. These results imply that as economic growth, energy consumption and biocapacity rise, there is tendency for environmental quality to deteriorate due to increased carbon dioxide emission, which is associated with economic growth policies. The PMG result also reveals, that urbanization leads to environmental improvement possibly through attaining energy efficiency, the use of clean energy and improvement of environmental policies, although the effect of urbanization is not statistically significant during the short-run period.

In addition, the D-H Granger non-causality test shows that there is a unidirectional causality relationship flowing from economic growth, energy consumption, and urbanization to ecological footprint and from economic growth to biocapacity in the region under consideration. This means that the past values of economic growth, energy consumption and urbanization are found to be predictors of ecological footprint. The results of the Granger non-causality also reveal that biocapacity and ecological footprint have a two-way causality. Two-way causality is also found between biocapacity and economic growth. Moreover, while urbanization causes economic growth, biocapacity Granger causes urbanization. We further found energy consumption to predict both urbanization and biocapacity. However, our sample did not support any causality between energy consumption and economic growth in the MINT countries over the period considered.

Therefore, our study reveals several understanding in regard to policy implications for world leaders, organizations and policymakers generally. For example, governments and private enterprises could utilize energy resources, biocapacity and urbanization as instruments for achieving some level of energy efficiency and reduce global warming and carbon dioxide related problems aimed at abating climate change and improving environmental quality. Based on our findings and challenges facing the MINT countries, here are some possible policy implications. The study findings will enable world leaders to adopt environmental strategies that can enhance energy and environmental sustainability to achieve the global sustainable development goal of preventing a rise in the average global climate of 2 °C. MINT countries in particular could embark on enforcing some more stringent regulations on natural resource exploitation, and environmental laws that reduce pressures on nature and the environment, such as levying taxes on high CO₂ producing industries and firms. MINT governments could also spell out some capital punishments for those who embark on indiscriminate dumping of refuse, setting wild fires, vandalizing oil pipe lines and those that engaged in illegal mining and fishing activities.

To reduce the rate of deforestation and erosion in these countries, afforestation laws may as well be included in environmental policies, and the government can start to practice wildlife conservation and biodiversity initiatives, and must also derive a blue print towards combating natural disasters like earthquake and flood. We also suggest the promotion of renewable energy like development of hydro, solar and wind energies and focusing on the components of total ecological footprint, e.g. developing crops land, fish farms and dams along other environmental measures. These policy suggestions could yield more desirable results to enhance energy and environmental sustainability in MINT, considering the role of primary energy consumption in production and its huge impacts on the environment and risks to mankind and animal species. Policy makers are also encouraged to embark on innovations in financial markets, and natural resources to develop effective energy measures might promote wealth creation and stable economic growth in MINT nations in the long run. For MINT, a sustainable energy framework may facilitate the reduction of CO₂, guarantee accelerated economic development and lower the pressures on the environment based on influence by level of footprint in these countries.