The Energy Kuznets Curve: Evidence from Developed and Developing Economies.

• Author: Moosa, Imad A.

INTRODUCTION

The environmental Kuznets curve (EKC) is a graphical representation of the relation between measures of environmental degradation and income per capita. The underlying idea is that at low income levels, growth causes (or is associated with) environmental degradation (measured, for example, in terms of per capita emissions). Beyond a certain point, however, growth leads to (or is associated with) environmental improvement. The implication of this changing pattern is that the EKC takes the form of an inverted U-shaped curve. The inversion is explained, inter alia, in terms of the proposition that higher levels of income per capita are associated with a gradual shift towards information-intensive industries and services, increased environmental awareness, enforcement of environmental regulation, better technology, and a higher level of environmental expenditure. These factors lead to a levelling off followed by a gradual decline of environmental degradation (that is, improvement in environmental quality).

A possible explanation that has not been dealt with adequately in the literature is the role played by energy consumption. What is called the "energy environmental Kuznets curve" is effectively an augmented EKC obtained by including some energy-related variables in addition to income and income squared as well as other variables that affect environmental quality. What we suggest here is that the energy Kuznets curve can be derived and estimated separately as a quadratic functional relation between energy consumption and income on a per capita basis. We also suggest that the environmental Kuznets curve implies and is implied by the presence of a corresponding energy Kuznets curve.

The objective of this paper is to demonstrate that if environmental degradation rises then falls (beyond a certain threshold) as income per capita rises, then a similar behaviour is observed in energy consumption per capita--this is not to say that the two thresholds are identical. If no turning point is observed, this does not necessarily mean that it will not appear, in the sense that the threshold has not been reached yet. For the purpose of this exercise, environmental degradation is measured in terms of CO2 emissions per capita, but this does not mean that the results should be valid for each and every measure of environmental degradation such as emissions of different kinds.

The contribution of this paper is the presentation of a different concept of the energy Kuznets curve from what is found in the literature, which is typically a conventional EKC augmented with some energy-related variables. We derive the energy Kuznets curve mathematically and estimate it empirically. By estimating the model in a TVP framework with unobserved components, we achieve two objectives. The first is that we overcome the problem associated with the impossible task of introducing a full list of "control variables." The second is that the estimated equations allow the responses of emissions and energy consumption to vary over time, a task that cannot be executed adequately (or at all) by using rolling regressions, a point that is correctly made by Alptekin et al. (2018). The time-varying property in the energy Kuznets curve may be triggered by the fact that different fuel mixes produce different levels of emissions, which means that the response of emissions to energy consumption may vary over time as the fuel mix changes. (1) Relatedly, changes in income generate changes in the fuel mix and thus changes in emissions over time (even if energy consumption remains constant). This is why the energy EKC can be derived from, and is implied by, the traditional EKC.

As a preview, the results show that both the environmental and energy Kuznets curves are observed for high-income (developed) countries but not for low-income (developing) countries. For the latter, the result may be interpreted to mean that the threshold for a turning point has not been reached yet. One has to bear in mind that the appearance of a turning point does not happen automatically because it is contingent upon political choices (for example, regulation with respect to the environment and energy efficiency).

LITERATURE REVIEW

A number of theoretical and intuitive explanations can be put forward as to why the impact of growth on the environment becomes less severe (even positive) as income rises. The first explanation is that environmental quality can be considered to be a normal good, which means that as income rises, the demand for environmental quality rises as well (Beckerman, 1992; World Bank, 1992; Barrett and Graddy, 2000). The second possible explanation is that as the level of income rises, a tendency appears to use less pollution-intensive technology (Grossman and Krueger, 1995). Another explanation is that the share of pollution-intensive sectors in total output goes down as the share of service sectors goes up, which is associated with rising income (for example, Janicke et al., 1997). Yet another explanation is that the tendency of people to breed declines as they become richer, thus alleviating pressure on the environment (see also Moosa, 2017a).

Some economists suggest that the environmental Kuznets curve and the Solow growth model are intimately related. Brock and Taylor (2010) modify the Solow model by incorporating technological progress in abatement, reaching the conclusion that the EKC is a necessary by-product of convergence to a sustainable growth path. The amended Solow model, which they call "green Solow model", generates an EKC involving both the flow of pollution emissions and income per capita, as well as the stock of environmental quality and income per capita. They demonstrate that when the EKC is derived the way they suggest, it may be humped-shaped or strictly declining. This implies that current methods for estimating the EKC are likely to fail whenever they do not take into account cross-country heterogeneity in either initial conditions or deep parameters. By constructing a green Solow model, Huang and Quibria (2014) demonstrate that the growth of emission is negatively related to technological progress and the rate of expenditure in abatement. They suggest, ceteris paribus, that the emission curve mirrors exactly the Solow fundamental equation of growth and produces the environmental Kuznets curve.

Typically, the EKC is augmented by adding other explanatory variables (called "control variables") that may affect environmental quality, such as globalisation (Tisdell, 2001), education and inequality (Hill and Magnani, 2002), human capital, technology, industrial structure and urbanisation (Jun et al., 2011), poverty (Liu, 2012), and trade (Ang, 2009; Halicioglu, 2009; Jalil and Mahmud, 2009). Some economists point out that energy has not been given the attention it deserves. For example, Agras and Chapman (1997) argue that "one variable consistently omitted in these relationships is energy prices", putting forward an EKC model that includes energy prices as an explanatory variable. Likewise, Luzzati and Orsini (2009) suggest that "energy has been scarcely explored in the Environmental Kuznets Curve (EKC) debate and is taken as the indicator of overall environmental pressure."

More recently, however, energy prices or consumption, as well as other energy-related variables, have been used by Ang (2007) and by Lean and Smyth (2010), giving rise to what has become to be known as the "energy environmental Kuznets curve." For example, Luzzati and Orsini (2009) introduce an energy-related explanatory variable and obtain evidence against the EKC. Likewise, Ozokcua and Ozdemirb (2017) add per capita energy use as an explanatory variable and obtain results supporting the existence of N-shaped and inverted N-shaped curves, hence concluding that "the results do not support the EKC hypothesis." Pablo-Romero and Sanchez-Braza (2017) estimate the "residential energy environmental Kuznets curve" and conclude that "the EKC hypothesis is confirmed for the residential sector in the EU-28 countries" and that "the turning point has been reached in Denmark, Luxembourg, Finland, The Netherlands, and Sweden." On the other hand, they introduce energy prices as an explanatory variable and find that "the turning point of improved environmental quality is not reached in any instance."

Similar studies have been conducted along these lines. Pablo-Romero and De Jesus (2016) obtain results that do not support the "energy-environmental Kuznets curve." Lorente and Alvarez-Herranz (2016) augment the EKC by adding (as explanatory variables) renewable energy promotion, energy innovation processes, and the suppression effect of income level on the contribution of renewable energy sources to total energy consumption. They conclude that air pollution will not disappear on its own as a result of economic growth. Therefore, it is necessary to promote energy regulation measures to reduce environmental pollution. Saboori et al. (2016) find support for the EKC when energy use is introduced as an explanatory variable. Pablo-Romero and De Jesus (2016) identify over 50 studies of the dynamic relation between economic growth, energy consumption and environmental pressure, which they attribute to the increasing tendency to examine the relation between growth and energy consumption from an environmental perspective (for example, Suri and Chapman, 1998; Richmond and Kaufmann, 2006; Ahmed and Long, 2012; Muhammad et al., 2013).

The environmental Kuznets curve is related to the energy Kuznets curve via the energy-income relation, which has been investigated extensively because, as van Ruijven et al. (2009) argue, "it is increasingly important to understand how energy use evolves across the full income continuum from less developed to highly developed countries. Csereklyei and Stern (2015) show that per capita energy consumption is positively correlated with GDP per capita and note...