Interdisciplinary Public Finance, Business and Economics Studies—Volume V

1 Introduction

In the recent period, economies in the world focused more on the issues of economy, energy, and environment, which are briefly expressed as “EEE” (Rauf et al., 2018: 32066). In this process, while implementing the policies aiming to encourage economic development and growth, on the one hand, the policymakers also make an effort in order to minimize the negative effects of global warming and environmental pollution in parallel with the sustainable development objectives on the other hand (Dogan and Lotz, 2020: 12717). However, it can also be seen that these efforts do not always yield the desired outcomes because, even though industrial advancements, increases in energy consumption, and technological innovations bring economic growth and an increase in welfare, they also degrade the environmental quality and cause environmental problems, which are irreversible or very costly to reverse (Intergovernmental Panel on Climate Change–IPCC, 2014; United Nations Environment Programme–UNEP, 2019). The issues of global warming, climate change, and environmental pollution, which reached a level that worries societies, are closely followed by researchers, international organizations, and policymakers today.

Studies carried out on the environmental Kuznets curve (EKC) hypothesis in the literature are based on the hypothesis of Kuznets (1955) that there was an inverted U–shaped relationship between economic growth and income inequality. According to this hypothesis named the Kuznets curve, an increase in the income inequality would continue until a certain threshold income level and the increase in income after the peak point would decrease the income inequality (Kuznets, 1955). Upon the increases in problems arising from environmental pollution, global warming, and uncontrolled use of natural resources, the Kuznets curve has started to be called the EKC hypothesis addressing the relationship between income and environmental pollution or quality since the 1990s.

Making use of Kuznets curve, Grossman and Krueger (1991, 1995), Shafik and Bandyopadhyay (1992), Shafik (1994), Panayotou (1993, 1997), and Selden and Song (1994) ←13 | 14→concluded in their studies that there was an inverted U–shaped relationship between income level and various environmental pollution indicators. This inverted U–shaped relationship was defined as the EKC first time by Panayotou (1993). According to the EKC hypothesis, the production structure based on secondary technologies causing environmental damage during the first phase of economic growth would increase environmental pollution. At this phase, environmental pollution would increase as the income level increases until a turning point (Grossman and Krueger, 1991). In the second phase after the turning point, the pressure on the environment decreases as a result of the increase in environmental awareness, use of environment–friendly production technologies, and structural changes and, consequently, environmental pollution decreases as the level of income increases (Shafik, 1994; Vincent, 1997).

One of the indicators most frequently used in representing environmental pollution in literature is carbon dioxide (CO₂) emission (see Shahbaz and Sinha, 2019; Wang et al., 2017). However, CO₂ emission measuring the carbon concentration in the air ignores the pollution in the water and soil. Representing a limited dimension of environmental pollution, CO₂ emission is criticized from this aspect (Solarin, 2019: 6167). For this reason, in recent studies, instead of CO₂ emission, ecological footprint (EFP) which is a more comprehensive indicator has been used as an indicator of environmental pollution (Nathaniel, 2021; Peng et al., 2019; Rehman et al., 2021; Sarkodie, 2021). Introduced first by Rees (1992) and then improved by Wachernagel and Rees (1996), EFP measures the demand of human beings on nature and consists of six components, namely, croplands, grazing lands, forest lands, fishing grounds, built–up lands, and carbon footprint. (Global Footprint Network, 2021).

The relationship between environmental pollution and economic growth (or development) is investigated for different countries and country groups within the context of the EKC hypothesis. In these studies, the validity of the EKC hypothesis varies depending on the sample countries, analysis methods used, and variables included in the model. In recent studies testing the EKC hypothesis, some researchers use the economic complexity index (ECI) instead of GDP per capita (see Neagu, 2019; Yilanci and Pata, 2020; Pata, 2021; Chu and Le, 2022). ECI, from a general perspective, reflects all aspects of a country’s production, productivity, knowledge, and capability (Hidalgo, 2021). Higher ECI levels indicate the export of technologically more advanced or more complex goods and a better position in terms of economic growth rates and development level (Hidalgo and Hausmann, 2009). A lower ECI level indicates less diversification of exports and exports of qualified goods, and a lower level of economic development (Hausmann et al., 2014; Mealy et al., 2019: 1–2).

←14 | 15→The aim of this study is to test the EKC hypothesis in five countries, which have the highest average ECI value in the period of 1980–2017. The countries having the highest average ECI value in this period were Japan [2.338], Germany [2.119], Switzerland [2.098], Sweden [1.982], and Singapore [1.156]. Fig. 1 shows the time–paths of the ECI in these countries.

The reasons for environmental pollution, which has turned into a global threat, especially after the late 1970s, have been investigated in literature for a long time. Within this context, energy consumption, trade openness, economic growth, financial development, shadow economy, and globalization can be listed as the variables, effects of which on pollution have been examined. Besides them, the ←15 | 16→studies examining the effects of ECI on environmental pollution have drawn attention in recent years. The study carried out by Can and Gozgor (2017) for France, Lapatinas et al. (2019) for 88 developed and developing countries, Neagu and Teodoru (2019) for EU member countries, Yılancı and Pata (2020) for China, Romero and Gramkow (2021) for 67 countries, Shahzad et al. (2021) for the USA, Ikram et al. (2021) for Japan, Alvarado et al. (2021) for the Latin America countries, Leitão et al. (2021) for the BRICS countries, and Dogan et al. (2021) for 28 OECD member countries can be given as examples. However, while these studies focused on the linear effect of ECI on environmental pollution, they ignored its nonlinear effects. On the other hand, even though the number is limited, there also are studies investigating the nonlinear effects of ECI on environmental pollution within the context of the EKC hypothesis in the literature. In a panel data analysis on 25 EU member countries, Neagu (2019) used ECI instead of GDP per capita and tested the EKC hypothesis. The results showed that there was an inverted U-shaped relationship between ECI and CO₂ emissions (indicating the validity EKC hypothesis) in only six EU member countries. Swart and Brinkmann (2020) examined the relationship between ECI and environmental variables (deforestation, solid waste generation, forest fires, and air pollution) for Brazil. For this purpose, the researchers using Fixed Effect (FE) panel data estimator couldn’t achieve any result supporting the EKC hypothesis. Emphasizing that there were few studies on the effect of ECI on the environment, Chu (2021) carried out a system – GMM analysis on 118 countries and determined that there was an inverted U-shaped relationship between ECI and CO₂ emissions proving the EKC hypothesis. Such that, the increases in ECI initially increased the CO₂ emissions and, after reaching a turning point, the increases in ECI decrease the emissions and, consequently, the environmental degradation. Finally, an inverted U-shaped relationship between ECI and CO₂ emissions and between ECI and EFP, which shows the validity of the EKC hypothesis, was reported by Pata (2021) for the USA for the period of 1980–2016 and by Chu and Le (2022) for the Group of Seven (G7) countries for the period of 1997–2015.

3 Data, model, and method

4. Empirical results

The findings of CD_LM and LM_adj. used for determining CSD and delta (Δ˜) and adjusted delta (Δ˜adj.) tests employed for determining the slope homogeneity are shown in Tab. 1.