Abstract: I investigate the theoretically ambiguous relationship between social inequalities and pollution using a panel dataset, with a substantially larger number of observations than most existing literature, including 161 developed, emerging and developing economies for the 1990-2013 period. I examine specifically the effect of not only income but also other forms of inequality on various pollution measures, namely carbon dioxide emissions, access to safe water, water pollution, and deforestation, by integrating inequality into the so-called environmental Kuznets curve. Results from an empirical analysis indicate that while inequality in both income and power are important predictors of pollution levels, the direction of the effects may vary substantially between local and global pollutant indicators. I find that in most cases there is a trade-off between reduction of income inequality and reduction of environmental degradation. An improvement in power inequality, measured by political rights and civil liberties and literacy rates is found to be associated with less local environmental degradation but more carbon dioxide emissions. I show that the relationship between inequality and pollution differs between emerging economies and other countries.
Keywords: pollution; inequality; environmental Kuznets curve; emerging economies
Introduction
The concept of sustainable development (WCED, 1987) suggests that economic, social and ecological factors should interact in such a way that long-term growth brings benefits for both people and the planet. Exploring the connections between the three pillars of sustainability has been an issue of interest for decades.
However, economic growth and improved wellbeing in last decades have come at the cost of high environmental pressure and led to a human-induced climate change (Schandl et al., 2016). Since the 1950s, average GDP per capita has increased over 20 times (World Bank, World Development Indicators Dataset) and this economic growth has been crucial for fighting poverty, improving living standards and promoting shared prosperity (Dollar et al., 2016). On the other hand, for the same period, the greenhouse gases have raised substantially: the global carbon emissions from fossil fuels alone have increased by over 400 percent (Boden et al., 2010). With the climate change accelerating (Karl et al., 2015), it is vital for economic growth to be sustainable and, thus, for people to explore the mechanisms that assure this. The question of how and who should govern the relationship between economic systems and the environment is yet to be answered. However, a line of thought supported by some empirical evidence has introduced the idea that “economic growth and respect of ecological constraints are compatible in the long run” (Clement & Meunie, 2010).
The World Bank’s World Development Report 1992, which was focused on environment, claimed that some environmental problems such as “most forms of air and water pollution… initially worsen but then improve as incomes rise”. This relationship is called the environmental Kuznets curve (EKC) and hypotheses that the environmental degradation is an inverted U-shaped function of income per capita (Stern, 2004). This implies that eventually, growth reduces the negative impacts of economic activities on the environment.
Examining this relationship, a following wave of scholars argues that there might be some major omitted socio-economic variables in addition to economic growth that explain the level of environmental degradation. For instance, income inequality might have a key role in determining the evolution of environmental pollution (Magnani, 2000). Since the 1980s inequality has been rising in most countries worldwide and has reached levels comparable to the 1920s (World Wealth and Income Database). This is not only an issue of great concern in OECD countries but also one of the central problems for developing countries, and particularly for rapidly growing ones (Klasen, 2009). Therefore, the mechanisms underlying the relationships between income, its distribution, and the environment may differentiate between countries at different stages of development (Drabo, 2010; Torras & Boyce, 1998). Older studies are examining those relationships for a limited number of countries and years due to data availability, while more recent ones tend to cover more observations and to differentiate between developed (high-income) and developing (low-income) countries.
Therefore, the purpose of this study is to re-examine the growth-inequality-environment nexus for the recent period of 1990-2013, by using an expanded and improved data on inequality (Solt, 2016), that allows the use of a larger number of countries and observations than the existing literature. I will also contribute to the literature by introducing emerging economies as a separate group of countries, rather than distinguishing between developed and developing countries only. While emerging economies (EEs) are playing a growing role in the world economy, since their share in the world GDP has increased more than two-fold for the period of study (World Bank, World Development Indicators Dataset), there is little knowledge of the effects of inequality on environmental degradation in those countries. The EEs are characterized by rapid economic growth, which is generally expected to put higher pressure on the environment (Pao & Tsai, 2010). Therefore, the strength and even the direction of the relationships between per capita GDP, inequality, and environmental degradation in emerging economies may differ from those in developed and developing countries. While this study focuses mainly on CO2 emissions as an indicator of environmental degradation, other variables (water pollution, access to safe water and deforestation) are also considered. Furthermore, following Torras and Boyce (1998), I use two proxies for inequality of power that contribute to capturing other dimensions of social inequality, rather than income distribution only. To sum up, this study will (i) empirically test the existence of EKC for various environmental quality measures for the recent period of 1990-2013; (ii) introduce inequality into the relationship using an expanded dataset on income inequality as well as additional social inequality measures; (iii) investigate if the results for emerging economies differ from those for developed and developing countries.
The rest of the paper is organized as follows. Section 2 provides an overview of the theoretical framework on the growth-pollution and inequality-pollution links. Section 3 presents a literature review of the main empirical findings on the environmental Kuznets curve (EKC) and the environmental impact of inequality. The relevance of examining those relationships in the context of the emerging economies is discussed in Section 4. Section 5 describes the data used. The econometric analysis and the results are presented in Section 6, while the sensitivity analysis in made in Section 7. In Section 8 the policy implications of the results are discussed before concluding remarks are made in Section 9.
Theoretical framework
An introduction to the Environmental Kuznets curve (EKC)
The environmental Kuznets curve (EKC) derives its name from the similarity with the inverted U-shaped relationship between inequality and economic growth observed by Kuznets (1955). It hypothesizes that environmental degradation will increase at early stages of economic development but will slow down at some turning point of higher income levels. The effect of growth on the environment can be explained through three main channels: scale effects, composition effects, and technological effects (Grossman & Krueger, 1991; Dinda, 2004). Firstly, when industrialization begins, people in a particular country will give more importance to increasing output and incomes rather than having clean air and water, and that will result in a higher environmental pressure due to a greater use of common resources (Dasgupta et al., 2002). Thus, economic growth will negatively impact the environment through the scale effects. However, a further increase in income levels activates the positive impact of the composition effects on environmental quality: initially, the structural change from agricultural to industrial society causes degradation but the following transition towards services will have the opposite effect. Also, as their wealth grows, people start having a higher preference for environmental quality and regulations against pollution become more effective. Therefore, more R&D activities and technological progress also spur from economic growth and ‘dirty’ technologies and processes are being replaced by cleaner ones (the so-called technological effects). Thus, the EKC reveals that economic growth could be compatible with environmental improvement if appropriate policies are taken.
It is important to note that since the environmental Kuznets curve (EKC) depicts the level of pollution as a function of per capita income only, its presence alone does not provide much clarity on the links between the two variables. Thus, criticism has been raised by scholars arguing that economic growth should not be considered a safeguard for environmental quality (Arrow et al.,1995; Copeland and Taylor, 2004). They stress that recent environmental improvement in developed countries is not an automatic process but rather comes as a result of the action of authorities, induced policy response, and partially innovation and technology. So, rising incomes alone shall not be misinterpreted as if there are ‘no environmental limits to growth’ (Stern et al. 1996). Major policies need to be implemented to put the global economy on a sustainable track, and EKC alone does not give much of indications for policy making.
Both Arrow et al. (1995) and Stern et al. (1996) argued that, even if the EKC hypothesis holds, this might be due to the effects of trade on the distribution of polluting industries. As explained by Stern (1996, 2004), the Heckscher–Ohlin trade theory proposes that as a consequence of free trade, each country would specialize in the production of goods that require an abundance of its own factor endowments. For instance, developed countries would specialize in capital intensive activities, such as manufacturing, requiring high-skilled labor, whereas developing countries would specialize in labor intensive activities, such as natural resource extraction, requiring low-skilled labor. Then, it could be that due to specialization the so-called ‘race to the bottom’ occurs: decreasing environmental degradation in developed countries occurs only at the expense of increasing environmental degradation in developing ones. Also, increasing environmental regulations in developed countries might further stimulate the outsourcing of pollution industries towards poorer countries (Lucas et al., 1992). However, when the poorer countries become wealthy and start implementing similar regulations, they will face a difficult task as there might be nowhere else to outsource them (Stern et al., 1996).
Dasgupta et al. (2002) provide a critical review of EKC discussing two additional scenarios to the conventional EKC and the ‘race to the bottom’ scenario. Firstly, the ‘new toxic’ scenario claims that while some pollutants are getting reduced with income, industrial societies create new ones so the there is no overall improvement in emission levels. Secondly, they also present an optimistic review of the conventional curve arguing that pollution may start falling at lower income levels as industrialization is no longer in its early stages. So, developing countries’ environmental degradation peaks will be lower than the ones reached earlier in developed countries.
The complicated and sometimes contradictory theoretical foundations of the EKC have resulted in concerns about the omission of major key variables that could modify the growth-pollution link. This line of thought provoked interest in the third pillar of sustainability- social factors.
Mechanisms linking inequality and environmental degradation
Higher income inequalities are generally considered undesirable and are found to have harmful secondary impacts on various factors, such as health and economic growth (Wilkinson & Pickett, 2010). The increasing attention on environmental worsening have brought about the question if environmental degradation could be yet another phenomena impacted negatively by greater income differences within countries. As today both inequality and environment are experiencing crises, it might be that those effects are reinforcing each other (Berthe & Elie, 2015). Figure 1 summarizes schematically the main theories linking different social inequalities to environmental degradation.
Figure 1. Channels linking inequality and environmental degradation
The backbone of these theories is a paper published by Boyce (1994) that develops a hypothesis called the ‘power-weighted social decision rule’ (i, Figure 1). Boyce explains that choices concerning the environment are based on a balance of power between those benefiting from a particular environmental destruction action (‘winners’) and those who bear the net costs (‘losers’). Therefore, Boyce suggests that more inequality would be tied to more environmental degradation, as those who benefit from it (generally the wealthy people) would prevail in the long run, and the overall environmental effects would diverge from the social optimum. The main reason for this relationship is that, since greater wealth grants more power, economics and politics within a country would favor the preferences of the wealthy people. That suggests an indirect effect of income inequality on the environment through power inequality. Further, greater inequalities of power and wealth would also impact the valuations of the costs and benefits of activities that lead to environmental degradation. Given that the ‘winners’ have a higher purchasing power, they also have a higher ability to alter market values, shape others preferences and influence the future path of technological advancements. The last dimension that Boyce (1994) describes is environmental time preference. He hypothesizes that greater inequality leads to both rich and poor generating short-run benefits and being less concern about long-term environmental costs. According to him, in countries with high degree of economic inequality, poor people are more careless about the future because they focus their lives on survival, while rich people have a high savings rate, and tend to extract more resources in a short term since they fear reallocation of power and wealth.
Opposing this view, Scruggs (1998) says that the hypothesis offered by Boyce (1994) implies that rich people prefer a degraded environment. He argues that the environment is a superior good and as people get richer, they would demand more of it and will have a greater willingness to pay for more environmental protection (iv, Figure 1). Thus, a more unequal society should generate less environmental degradation. This debate underlines the complexity of mechanisms behind the relationship between inequality and the environment. Martini & Tiezzi (2014) seek to test whether the environment is a luxury good and therefore, estimate the income elasticity of willingness to pay for air quality improvement for five income groups in Italy from 1999 to 2006. According to the results of their methodological approach, air quality improvements are a normal good, thus income neutral.
Another channel through which inequality may affect environmental quality is social cooperation (ii, Figure 1). As suggested by Borghesi (2000), it is more difficult to achieve public policy solutions to environmental problems in unequal societies because of less trust and cooperation. Even though a country experiences economic growth, and on theory the capacity to pay for environmental degradation increases, if the inequality is high in this country, the effect might be reduced. The median voter preferences would not be switched towards ‘pro-environment public expenditure’ (Magnani, 2000). Studies suggest that social trust between individual stimulates collective action towards protecting the environment such as recycling (Sonderskov, 2011) and using public transportation (Van Lange et al., 1998), and that income inequality within a country may hamper the development and diffusion of new of new environmental technologies (Vona & Patriarca, 2011).
Further, the consumption patterns in a country can also be linked to the level of inequality. Veblen (1899) argues that in more unequal societies individuals consume more as they tend to compare themselves with the wealthier social classes (iii, Figure 1). Poorer members of the society will try to meet the living standards they desire by earning and spending more, and overall the consumption and pollution levels will be higher than those in an egalitarian society (Cushing et al., 2015). This line of reasoning fits, for example, with the recent developments in the United States: simultaneous increase in income inequality, household debts, and working hours (Frank, 2012).
Lastly, developing a theory specifically on carbon dioxide emissions, Ravallion et al. (2000) suggest that income inequality might even improve environmental quality (v, Figure 1). They argue that every individual has an implicit demand function of carbon emissions, and they call the derivative of this function with respect to income marginal propensity to emit (MPE). According to this model, if poor people have a higher MPE than rich people, for example, due to less efficient energy use, reducing inequality will lead to more aggregate environmental degradation. It could even be that initially MPE rises with income for poor segments of the society in developing countries due to lack of access to modern energy and technologies. Therefore, rising income inequality within such a country would reduce the marginal propensity to emit for both richer (as they will have better access to modern technologies) and poorer segments of the population (as they will be basically excluded from the carbon economy).
Empirical finding of the literature
Empirical studies on the environmental Kuznets curve
Studies on the environmental Kuznets curve (EKC) emerged in the early 1990s with Grossman and Krueger’s (1991) paper on the potential alleviation of pollution problems in Mexico due to NAFTA, and Shafik and Bandayopadhyay’s (1992) background study for the World Development Report 1992. Most studies employ the following reduced form equation to test for the presence of the EKC:
〖POL〗_it= β_0+ β_1 〖GDP〗_it+ β_2 〖GDP〗_it^2+β_3 X_it+ε_it , (1)
where POL= specific environmental indicator; GDP=GDP per capita; and X relates to control variables of influence on environmental degradation. Therefore, the inverted U-shaped relationship will exist if β_1>0 and β_2<0. Some studies also include a cubic term of GDP per capita in order to detect a possible second turning point.
Grossman and Krueger (1991) estimate the EKC for sulfur dioxide (SO2), smoke and suspended particles using the GEMS dataset that monitors air quality in urban areas throughout the world. Their regressions include a cubic functional form of per capita GDP (in PPP), various geographical variables, a time trend, and a trade intensity variable. They find that the levels of sulphur dioxide and smoke reach a turning point at around GDP per capita of US$4,000-5,000 in 1985 dollars (but then perhaps raise again at income levels over US$10,000–15,000) and that the mass of suspended particles monotonically decreased with increasing GDP per capita. Shafik and Bandayopadhyay (1992) estimate the EKC for six pollutants including CO2 emissions, deforestation, and lack of safe water and sanitation using three functional forms: log-linear, quadratic, and cubic. They found that the EKC is confirmed for air pollutants as they start decreasing when the countries approach middle-income levels but the results are mixed for other indicators: the access to clean water and sanitation marks a stable improvement with increasing income, whereas river quality worsens with increasing income.
Most of the subsequent studies also confirmed empirically the existence of the EKC for a number of air and water pollution indicators. Grossman and Krueger (1995) find evidence that air concentration of SO2 and 11 indicators of water quality improve together with economic growth after a critical level of income per capita is reached. They reveal the turning point to be at GDP per capita of less than US$8,000 (1985 dollars). De Bruyn et al. (1998) find that that economic growth has a direct positive effect on the levels of emissions of carbon dioxide (CO2), nitrogen oxides (NOx) and sulfur dioxide (SO2) in the Netherlands, Western Germany, the UK and the USA for the period 1960-1993. However, they stress that the main reason for the decline of the emissions could be technological and structural improvements. A recent city-level study of Japan conducted by Kasuga and Takaya (2016), also supports the EKC hypothesis. They find that environmental indicators on air and river quality improve as incomes increase which is consistent with previous literature as income levels of Japan are above the turning point.
Turning to less common environmental indicators, the EKC is most often not supported. Indicators with direct effect on human health, such as access to urban sanitation and access to safe drinking water, tend to improve monotonously with growing GDP per capita (Dinda, 2004). For other environmental quality measures, for example, deforestation, the empirical evidence is quite controversial (Bhattarai & Hammig, 2001; Koop & Tole, 2001).
Notwithstanding the abundant EKC literature, it has failed to provide conclusive results and economic mechanism reasoning on the relationship between economic growth and environmental degradation (Stern, 2004). Even among the studies supporting the EKC, there is a lack of agreement on the turning points beyond which different pollutant levels start improving. Copeland and Taylor (2004) state that “Our review of both the theoretical and empirical work on the EKC leads us to be sceptical about the existence of a simple and predictable relationship between pollution and per capita income.’’
Empirical studies on the nexus between inequality and environmental degradation
A sizeable empirical literature on the effects of inequality on environmental degradation has emerged in the recent period. Table 1 provides an overview of the main contributions to this topic. Generally, the empirical analysis is based on the following function of pollution indicators:
P= f (Y,G,π,X) , (2)
where P=pollution indicator; G=Gini coefficient; π= political variables/power inequality measures; and X= vector of control variables.