Balancing Climate Change Policy and Industrialization in the Short-Run: the Case for Transition Economies

(1)

Department of Economics Master Thesis (D level) Supervisor: Runar Brännlund Spring 2010

Balancing Climate Change Policy and Industrialization in the Short-Run: the Case for Transition Economies

Abubeker Shafi

(2)

1

Acknowledgement

The author would like to thank Runar Brännlund, for giving me his kind guidance, corrections and comments on this paper.

(3)

2 Abstract

This study attempts to look into a GHG abatement mechanism for developing transitional economies, that make environmental and economic goals complementary even in the short run (prove wrong EKC hypothesis) or limit emissions in the short run (shorten the course to the turning point of EKC). By assuming transfer of technology from the developed countries as a short-run effective solution to the GHG abatement problem, I tried to make empirical estimates of the effects of technology transfer on GHG emissions of Eastern European transition economies.

The results are in contrary to my research assumption and questions. Basically there is an indication of negative effect of technology transfer on GHG emissions but it is a small effect compared to the roles played by other variables. There also exists some support for the theory of an inverted U-shape relationship between income and pollution; and turning point per capita GDP values are higher for countries that are considered more open compared countries that are less open. Therefore I concluded, passive technology transfer is not the right approach in order to be taken as the way forward to achieve short-run abatement goals.

(4)

3

Table

of

Contents

Abstract 2

1. Introduction 4

1.1. Background 4

1.2. Statement of the Problem 5

1.3. Objectives 6

1.4. Importance of the Study 6

1.5. Scope and Limitations 6

2. Literature Review 7

2.1. Basic Theory and Argument of EKC 7

2.2. Empirical Validity and Critique of EKC 13

3. Research Method 20

3.1. Data 20

3.2. Models and Variables 24

4. Result and Analysis 27

5. Summary and Conclusion 32

References 34

Appendices 37

Tables

1. Summary of Empirical Literatures 18

2. Data Summary Statistics (1990-2007) 22

3. Estimates of Model Coefficients for Individual GHG 28 4. Turning point per capita GDP by Country (US$) 30 Figures

1. Theoretical EKC (Relationship between Pollution and Per Capita Income) 7

(5)

4

1. Introduction

1.1. Background

In recent years, economists have become increasingly aware of the significant implications of economic growth on the environment, hence many literatures have been written on this subject. This issue has gained interest among the developing countries, especially those that are neither part of the least developed countries, nor of the industrialized countries which are located largely in Eastern Europe, Latin and Central America and some countries in Central and South East Asia. In the past many countries in this category were centrally planned economies; nowadays this group of developing countries are being referred by a vague term as transition economies, which basically means countries with an economy that is changing from a centrally planned economy to a free market.

The early stages of industrialization are generally accompanied by rising incomes and worsening environmental conditions. Cross-sectional analyses of numerous countries at different levels of income suggest that pollution tends first to rise with national income and then to fall, this effect has been dubbed the ‘Environmental Kuznets Curve’ (EKC) (Todaro & Smith 2006). If the EKC hypothesis held generally, it implies that instead of being a threat to the environment, economic growth is the means to environmental improvement (Perman et al.

1999).

Green House Gases (GHG) are principal pollutants, their impact is independent of the location of the emission source. The green house effect is an example of what is called a reciprocal spillover; emissions in each country have adverse effects upon citizens in the other (Perman et al. 1999). Such problems require stronger international cooperation towards the shared goal of cutting back GHG emissions. As a result the issue of ‘climate change’ is now on the political agendas of governments, at least at the level of rhetoric, around the world.

Addressing climate change in transitional developing countries poses a fundamentally different challenge. For most, emission reduction is not a viable option in the near term. With income levels far below those of developed countries and per capita emissions on average just one- sixth those of the industrialized world, without any practical alternative for the future, transitional developing countries will continue to increase their emissions as they strive for economic growth and a better quality of life (Chandler et al. 2002).

(6)

5

So should the transitional developing countries simply accept emission limits at the cost of economic growth? Or continue on the traditional path of industrialization while putting their own efforts that serve to reduce or avoid greenhouse gas emissions? Or is there a much better unconventional option which will allow them to meet emission cuts while maintaining and accelerating industrialization and economic prosperity?

1.2. Statement of the problem

Based on the background it is clear to see the dilemma facing transitional developing countries.

One of the most contentious issues in the debate over global climate change is the perceived divide between the interests and obligations of developed and developing countries. It is demanded that developed countries, the source of most past and current emissions of GHGs, they must act first to reduce emissions. That principle is embedded in the 1992 United Nations Framework Convention on Climate Change and in the 1997 Kyoto Protocol, which sets binding emission targets for developed countries only (Chandler et al. 2002) .

But in recent years the focus is turning increasingly to the question of developing countries’

emission, particularly of the transition economies. Their resistance to the idea of limiting their emissions has led to claims that developing countries are not doing their fair share. In addition in developing countries, rising populations, income levels, energy use, deforestation and some agricultural practices are leading to rapid increases in GHG emissions (Chandler et al. 2002).

On the other hand developing countries are considered to be the most exposed to the adverse effects of climate change and will endure a heavy cost from it, thus posing a serious threat to development and poverty reduction in the poorest and most vulnerable regions of the world.

Hence it is clear that neither it is affordable and justifiable nor the rest of the world is patient enough to wait all the developing countries to go through the EKC. So in order to solve this problem or complexity associated with the decision on binding emission cuts, this research seeks to answer the following two critical questions:

Primary: Is there a path to industrialization with regard to GHG emission alternative to the one presented by EKC and make environmental and economic goals complimentary even in the short run?

Secondary (depending on the outcome of the primary question): How to limit GHG emissions in the short run and shorten the course to the turning point of EKC hence modify it in order to meet emission cuts while sustaining industrialization and economic development?

(7)

6 1.3. Objectives

The objective of this study is primarily to identify an alternative path of industrialization for developing countries which can make environmental and economic goals complimentary in the short run, consequently avoiding the EKC. Secondly to look for an approach that can reduce GHG emissions in the short run and shorten the path to the turning point of EKC so as to meet emission cuts whilst sustaining industrialization and economic development.

1.4. Importance of the Study

This study will be of an importance to:

• Understand the complex issues facing transitional developing economies in the decision to agree on targeted GHG emissions reduction.

• Indicate what transition economies can do by themselves to make emission cuts possible in the near future.

• Highlight the important role that developed countries and various international organizations can play in assisting developing countries and achieve the shared goal of GHG emission reduction through global cooperation.

1.5. Scope and Limitations

In carrying out this project, certain constraints will inhibit effective study, firstly due to short deadline period to finish this write-up, time factor will render certain aspects not to be examined in details. Secondly, the study assumes that transition economies can effectively and efficiently abate GHGs in the short run through transfer and adoption of technologies from the developed world. Thirdly, the data for this study focuses only on three primary GHGs i.e., carbon dioxide, methane and nitrous oxide; and five eastern European countries i.e., Belarus, Bulgaria, Romania, Turkey and Ukraine. I select these countries, because there is hardly any empirical study of EKC in relation to the Eastern European region. Finally, since I have chosen econometrics as my research method, I have to admit that my knowledge of econometric analysis is at the basic level.

(8)

7

2. Literature Review

2.1. Basic Theory and Argument of EKC

The question “what impact do economic activities have on the natural environment?” is of a very significant one. By the mid of the twenty century, in trying to answer whether personal income inequality will increase or decrease in the course of a country's economic growth, Nobel Laureate Simon Kuznets (1955) illustrated on his seminal paper that the shape of the relationship between income per capita and income inequality is inverted-U, i.e., income inequality first rises and then falls as economic development proceeds.

A similar relationship to Kuznets’ hypothesis has often been found (e.g. Grossman and Krueger 1995, e.g. The World Bank 1992) between economic development and pollution (environmental degradation), i.e., pollution initially increases as income increases but eventually declines once income has crossed some threshold (Grossman and Krueger 1995).

Consequently, the relationship between income per capita and the rate of environmental degradation takes its name from Kuznets’ paper and has been labeled an ‘Environmental Kuznets Curve’ (EKC). Suggesting that environmentalists’ concern about the consequences of economic development could be wrong and after sufficient economic growth, income and environmental quality improve together; and it is the world’s poorest and richest countries that have relatively clean environments, while middle-income countries are the most polluted.

Figure 1: Theoretical EKC (Relationship between Pollution and Per Capita Income)

(9)

8

The EKC concept emerged in the early 1990s and Grossman and Krueger (1995) and the World Bank (1992) first popularized this idea, using a simple empirical approach. In the years since these original observations were made, researchers have examined a wide variety of pollutants for evidence of the EKC pattern. policymakers, theoretical and empirical researchers alike has focused on developing models that replicate the inverted U shaped link (Brock and Taylor 2004) and many has come in support of this hypothesis. “The view that greater economic activity inevitably hurts the environment is based on static assumptions about technology, tastes and environmental investments” (Stern 2003).

How does the economy improve environmental quality?

Several recent papers propose competing theoretical explanations for the empirical observation of an inverted U-shape relationship between environmental degradation and per- capita income, focusing on different explanatory factors, such as innovation, structural change, institutions and others. The most explicit, compelling and comprehensive explanation I found is in Perman et al. (1999) which was given by Panayotou (1993) and he puts it in the following way:

At low levels of development both the quantity and intensity of environmental degradation is limited to the impacts of subsistence economic activity on the resource base and to limited quantities of biodegradable wastes. As economic development accelerates with the intensification of agriculture and other resource extraction and the take off of industrialization, the rates of resource depletion begin to exceed the rates of resource regeneration, and waste generation increases in quantity and toxicity. At higher levels of development, structural change towards information-intensive industries and services, coupled with increased environmental awareness, enforcement of environmental regulations, better technology and higher environmental expenditures, result in leveling off and gradual decline of environmental degradation.

Numerous theoretical models for the EKC have been presented. Prominent justifications in these models are “threshold effects in abatement that delay the onset of policy, income driven policy change that get stronger with income growth, structural change towards a service based economy, and increasing returns to abatement that drive down costs of pollution control”

(Brock and Taylor 2004). In the same line Stern (2005) also have suggested it is important to look at the relations between the efficiency component and the level of adopted technology and a variety of variables, such as GDP per capita, size of the economy, and population density.

(10)

9

Grossman and Krueger (1995) argued that economic growth affects environmental quality both negatively (through scale effects) and positively (through composition effects and technological effects). When an economy is in the early stage of development, increasing output requires more natural resources and thus puts a heavier burden on the environment, the so called scale effect. As the level of income grows, the economy shifts from being energy intensive to knowledge and technology intensive in production, known as the composition effect. With economic growth also come cleaner production technology and more effective abatement procedures that improve the environment while maintaining high output, known as the technological effect. A turning point eventually occurs when the scale effect is outweighed by the composition and technological effects.

The inverse U-shape pattern of income per capita to the rate of environmental degradation, as illustrated in Figure 1, can be thought of a three stages of development which are dominated by different production technologies where : first output is initially dominated by agriculture and light assembly, which has a relatively low level of pollution; second production progresses toward heavy industry, which creates a relatively high level of pollution; and third the output of high-tech industry subsequently dominates, which generates a relatively low level of pollution (Webber and Allen 2004).

The main rationale for EKC is that the demand for environmental improvements is income elastic (Stern 2007). This of course raises a very important the question on whether environmental services are luxuries or necessities. Consumers who have achieved a high standard of living will increase their demand for environmental amenities, not only are richer consumers more willing to pay for green products, they also ask for better institutions to protect the environment (Selden and Song 1994). For the poor, income may be a higher priority than pollution control and Positive income elasticity for environmental quality and hence clean environment is a luxury.

According to Stokey (1998) much of the hump-shaped relationship between income and pollution levels comes from static models. He articulated, the cost of higher pollution is a direct utility cost, so it is independent of income and the benefit, however, is the higher output (consumption) it allows, which depends on the level of conventional inputs. As economies become wealthier and their environments dirtier, eventually the marginal utility of income falls due to the rise in marginal disutility from pollution, to the point where people choose costly abatement mechanisms. After that point, the economies are at interior solutions, marginal

(11)

10

abatement costs equal marginal rates of substitution between environmental quality and income, and pollution declines with income. In frameworks of this type, there is typically zero pollution abatement until some threshold income level is crossed, after which abatement begins and pollution starts declining with income.

However the argument that people demand for a clean environment when they achieve a high income level is not generally valid since it cannot stand the climate change debate. Today we are talking about global pollutants (GHGs) and their impact is independent of the location of their emission source. Also it is the poor countries which are considered to be the most exposed and will endure a heavy cost from climate change risk. So here we can even argue the reverse, “the lower their income the more people demand GHG abatement” since they do not have a robust coping mechanism.

Andreoni and Levinson (2001) presented a simple model where they argue that the EKC might arise due to basic microeconomic forces. According to them EKC “does not depend on the dynamics of growth, political institutions, or even externalities, and can be consistent with market failure or efficiency” and “it can be derived directly from the technological link between consumption of a desired good and abatement of its undesirable byproduct”. Without preferences for environmental quality or externalities and institutions to internalize them, increasing returns to scale in abatement technology can explain the appearance of an EKC. In their standard microeconomic approach, pollution is a by-product of consumption that causes disutility and consumers may devote their monetary resources either to consumption or to pollution abatement. If the pollution abatement technology exhibits increasing returns to scale (and they argue that it does) then the EKC arises endogenously, also implying abatement will be greater in a larger economy.

On the contrary Plassmann and Khanna (2006) argued increasing returns to scale in abatement by themselves are not sufficient for pollution to fall with income unless the returns to scale of abatement exceed the returns to the production of gross pollution. Hence they concluded the existence of an EKC depends on the relative magnitudes of the returns to scale in abatement and in gross pollution, not on their absolute values. They further refuted Andreoni and Levinson (2001) by saying “there is nothing special about increasing returns to scale in abatement, and that even under decreasing returns to scale in abatement, pollution will eventually decline to zero as income increases as long as the returns to scale in gross pollution are less than the returns to scale in abatement”.

(12)

11

Brock and Taylor (2004) in their “Green Solow Model” have presented a very simple model linking growth rates, income levels and environmental quality. They said their model produces an EKC relation between both the flow of pollution emissions and income per capita, and the stock of environmental quality and income per capita. In their model they argued that the force of diminishing returns and technological progress identified by Solow as fundamental to growth process, may also be fundamental to the EKC finding, indicating the same ongoing dynamic processes responsible for income growth and convergence are also at play in determining the EKC finding and emission convergence, with this they tried to show, the EKC and the most influential theory in the macro literature – the Solow model are intimately related. Brock and Taylor (2005) also have supported and based their explanation on the scale, composition and technique argument which is presented earlier by Grossman and Krueger (1995).

Some economists point that trade and investment across regions or nations are two of the most important forces in explaining EKC, where residents of poor countries are willing to trade environmental quality for income. Underdeveloped economies usually have a comparative advantage in producing labor intensive goods that are often more pollution intensive. They also tend to have relatively loose environmental legislation and regulation, which may attract dirty industries from developed economies through trade and direct investment, moving pollution from developed to the underdeveloped economies, resulting in an EKC characterized by the underdeveloped countries (regions) on the rising segment of the curve and the developed ones on the declining segment (Cole 2004).

A hypothesis put forward by Stern (2005) is that high levels of pollution will result in action to reduce pollution, so that countries with high levels of emissions per unit area are aggressive in adopting pollution abating technology. However he further said “the problem with this is that obviously pollution emissions are an endogenous variable partly determined by the state of technology”. Some economists have suggested explanation for the inverse-U notes that pollution involves externalities and that appropriately internalizing those externalities requires relatively advanced institutions for collective decision-making that may only be implementable in developed economies (Andreoni and Levinson 2001). Stern (2005) also has expressed the idea that with higher population densities, more people will be affected by a given amount of pollution and hence more abatement would optimally be undertaken.

Torras and Boyce (1998) agree with Grossman and Krueger that in the developing countries the growth of per capita income can be accompanied by improvements in at least some

(13)

12

important dimensions of environmental quality, however they hypothesize that a more equitable distribution of power can largely contributes to these outcomes, by enhancing the influence on policy of those who bear the costs of pollution, relative to the influence of those who benefit from pollution-generating activities. Promoting more equitable power distributions in the developing world, for example, via more equitable income distribution, wider literacy, and greater political liberties and civil rights, can positively affect environmental quality. From an environmental standpoint, then, the distribution of power is not a peripheral concern.

Other theoretical contributions to include Selden and Song (1994), who described a variety of possible pollution-income paths. According to them as a result of both market forces and changes in government regulation it is rational to expect that economies would pass through

“stages of development” in which at least some aspects of environmental quality first deteriorate and then improve.

Proponents of ‘going-for-growth’ perspective invite the emphasis on achieving faster economic growth rather than on forming environmentally friendly policies because economic growth is perceived to be able to achieve both economic and environmental goals, whereas implementing environmental policies may impede economic growth (Webber and Allen 2004).

However from the above theoretical discussion we can learn that the explanation for EKC is not a simple black and white thing but it is a simple phenomenon that can be influenced and explained by various complex variables. So at least on a theoretical basis EKC may arise both spontaneously as technology, consumption and production patterns change and deliberately by implementing active policy on environment and the creation of institutions to internalize externalities. It is also very important to note that each of these explanations yield different policy implication, i.e., do nothing and act.

(14)

13 2.2. Empirical Validity and Critique of EKC

The EKC has been the subject of much empirical examination, a number of empirical papers have found evidence of an inverted U-shape relationship between environmental degradation and economic development. On the other hand it also has been extensively criticized on theoretical, empirical and econometric (Harbaugh, Levinson, & Wilson, 2002; Koop&Tole, 1999; Millimet, List,&Stengos, 2003; Perman&Stern, 2003; Stern & Common, 2001) grounds (Stern 2005). Numerous critics have challenged the conventional EKC, both as a representation of what actually happens in the development process and as a policy prescription (Dasgupta et al 2002).

He (2007) surveyed the existing EKC studies and discussed to what extent they may be valid and applicable for developing countries. He proclaimed “given the shortcomings in both the theoretical and empirical aspects of the analyses applied to this hypothesis, no one-fit-for-all inverted-U-shaped curve can describe adequately the relationship between growth and pollution”.

Shafik and Bandyopadhyay (1992) estimated the coefficients of relationships between environmental degradation and per capita income for ten different environmental indicators , the indicators are: lack of clean water, lack of urban sanitation, ambient levels of suspended particulate matter in urban areas, urban concentrations of sulfur dioxide, change in forest area between 1961 and 1986, the annual rate of deforestation between 1961 and 1986, dissolved oxygen in rivers, fecal coliforms in rivers, municipal waste per capita, and carbon dioxide emissions per capita. Lack of clean water and lack of urban sanitation were found to decline uniformly with increasing income. Both measures of deforestation were found not to depend on income. River quality tends to worsen with increasing income, two of the air pollutants were found to conform to the EKC hypothesis. Note, however, that CO₂ emissions do not fit the EKC hypothesis, but rising unambiguously with income, as do municipal wastes. From their result they stated, it is possible to overcome some environmental problems, but it is not automatic rather actions are required (Perman et al 1999).

The famous study by Grossman and Krueger (1995) covered four types of pollutants from around the globe: urban air pollution, the state of the oxygen regime in river basins, fecal contamination of river basins, and contamination of river basins by heavy metals. They found no evidence that environmental quality deteriorates steadily with economic growth. “Rather, for

(15)

14

most indicators, economic growth brings an initial phase of deterioration followed by a subsequent phase of improvement. The turning points for the different pollutants vary, but in most cases they come before a country reaches a per capita income of $8000”. Another recent paper by Bartz and Kelly (2008) supported Grossman and Krueger’s findings. Using U.S.

national emissions data from 1947-1998, they affirm all (carbon monoxide, nitrous oxide, sulfur dioxide, particulates, volatile organic compounds and lead) pollutants show some evidence of an inverted U-shape over time as GDP grows.

Selden and Song (1994) investigated EKC relationships between pollution and economic development using a cross-national panel of data on emissions of four important air pollutants:

suspended particulate matter, sulfur dioxide, oxides of nitrogen, and carbon monoxide. They found that per capita emissions of all four pollutants exhibit EKC relationships with per capita GDP. Despite their findings, they forecast rising global emissions over the foreseeable future.

Using Chinese provincial data from 1993 to 2002 Shen (2006) examined the existence of EKC.

He first constructed a simultaneous equations model and then employed a Hausman test to check whether or not a simultaneous relationship between income and pollution exists in the dataset of China and confirms that the simultaneity between income and pollutant emission exists in all the pollutants (SO2, Dust Fall, COD, Arsenic and Cadmium). EKC relationship is found in COD, Arsenic and Cadmium emissions in China with high significance in the estimated coefficients and consistent with his expectation. Meanwhile, SO₂ shows a U-shaped curve and Dust Fall indicates no relationship with income level. His result suggests government pollution abatement expense has a significant and negative effect on pollution, at least for the air and water pollutants' emissions he examined. As a result, he concluded “environmental improvement does not depend exclusively on income growth. Poor provinces need not therefore wait passively to become wealthy before doing something else to improve their environment.”

Harbaugh, Levinson & Wilson (2001) reexamined the empirical evidence of an EKC relationship using the air pollution data studied by the World Bank (1992) and by Grossman and Krueger (1995), with the benefits of a retrospective data cleaning and ten additional years of data. They analyzed three common air pollutants: sulfur dioxide (SO2), smoke, and total suspended particulates (TSP), according to them “pollutants for which the most complete data are available”. Two of the three, SO2 and smoke, exhibit the most dramatic inverse-U-shaped patterns in the World Bank's report (1992) and in Grossman and Krueger (1995). They also tested the sensitivity of the pollution-income relationship to the functional forms and

(16)

15

econometric specifications used, to the inclusion of additional covariates besides income, and to the nations, cities, and years sampled. Their conclusion is that the evidence for an inverted- U is much less robust than previously thought. They found that the locations of the turning points, as well as their very existence, are sensitive both to slight variations in the data and to reasonable permutations of the econometric specification. They argue merely cleaning up the data, or including newly available observations makes the inverse-U shape disappear.

Furthermore, econometric specifications that extend the lag structure of GDP per capita as a dependent variable, include additional country specific covariates, or include country level fixed effects, generate predicted pollution-income relationships with very different shapes, and conclude that there is little empirical support for an EKC relationship between several important air pollutants and national income in these data.

Stern (2005) reformulated the EKC as the best practice technology frontier; countries’

distances from the frontier reflect the degree to which they have adopted the best practice technology in emissions abatement. Then he assembled a panel dataset for sulfur emissions and the explanatory variables for the period from 1971 to 2000. He used the Kalman filter to model the state of sulfur emissions abatement technology in a panel of 15 mainly developed countries. The results show that with the exception of Australia, countries are converging toward the frontier but have settled into low pollution abatement and high pollution abatement groups. Pre-abatement levels of pollution, income per capita, population density, and perhaps cultural factors might partly explain the level of abatement adopted. The results show that technology does not evolve independently in each country, but neither is it a question of adopting different amounts of the same technology in all countries.

Using panel data technique Cole (2004) and Dinda (2006) examined the pollution haven hypothesis (PHH), i.e., due to differences in environmental standard, comparative advantages and globalization dirty industries move from rich to poor nations. Dinda (2006) looked at impacts of globalization and international trade on pollution level, pollution intensity and relative change of pollution for the OECD and Non-OECD country groups and the world as a whole. In the study Dinda (2006) uses CO₂ emission, and observed that the impact of globalization on environment heavily depends on the basic characteristics of a country and comparative advantage. His empirical results suggest that globalization is helping developed countries to reduce CO2 emission while developing countries CO2 emission rises, net impact increases CO2

emission. Evidence of PHH effect is found, although such effects do not appear to be

(17)

16

widespread and appear to be relatively small compared to the roles played by other explanatory variables (Cole 2004).

Jiang, Lin & Zhuang (2008) examined the EKC hypothesis in the People’s Republic of China by empirically estimating models using provincial level panel data from 1985 to 2005. The results show that there exists an inverted-U shaped relationship between per capita income and per capita emissions in the cases of waste gas from fuel burning and waste water, with a turning point at per capita gross domestic product of $12,903 and $3,226 respectively. This relationship does not hold in the case of waste gas from production or solid waste. The estimation results from the model allowing region specific slope coefficients show that the EKCs of the more developed coastal region have a flatter rising portion with turning points occurring at a higher income level than those of the less developed central and western regions. They argued that this may reflect technology diffusion and leapfrogging and institution imitation across regions at different stages of development.

Hökby and Söderqvist (2003) provided estimates of income and price elasticity of demand for reduced marine eutrophication effects in the case of the Baltic Sea, using data from five Swedish contingent valuation studies. Point estimates indicate that reduced marine eutrophication effects can be classified as a luxury and an ordinary and price elastic service.

Confidence intervals show however that the classification as a luxury is not statistically significant. A basic finding is that income tends to influence willingness to pay positively and significantly. The elasticity estimates are in most cases greater than zero, but less than unity.

Environmental improvements thus tend to be relatively more beneficial to low-income groups.

With reference to the discussion on the shape of the environmental Kuznets curve, this result does not give any room for concluding we are dealing with a general finding for environmental services.

Assuming gross pollution is generated by income and abatement cannot exceed gross pollution as physical constraints, Rodrigues and Domingos (2003) showed that EKC is no longer valid. They argue “If the environmental damage of pollution is internalized, even if an EKC curve appears for a low level of income, the level of pollution will eventually rise again for higher income (the rebound effect)”. In their model, from a certain level of income onward economic growth will actually lead to a decrease in welfare. They concluded EKC is the result of a static and fully myopic economic analysis of the income-environment relationship and

“EKC might emerge but it will be followed by a rebound effect”.

(18)

17

Torras and Boyce (1998) tested seven distinct pollution variables; sulfur dioxide, smoke, heavy particles, dissolved oxygen and fecal coliform which include observations from more than 1000 locations worldwide and other two variables are national level data on the percentages of the population with access to safe water and sanitation facilities. Their regression results generally shows literacy and rights appear to be particularly strong predictors of pollution levels in the low-income countries. The estimated effects of per capita income on pollution are generally weaken once they account for inequality effects, but they do not disappear altogether. They further commented that it cannot be assumed that environmental enhancements will continue to accompany further growth of per capita income in those countries which have already achieved high average incomes; there is support that increasing average income is associated with renewed deterioration in some dimensions of environmental quality. “The extent to which this trade-off can be relaxed through social, political and technological changes remains an open question.”

In keeping with the original EKC He (2007) stated, the inverted-U-shaped curve concerning the pollution-income correlation estimated from cross-country data is suspected to be a static and descriptive estimate. Hence, some economists believe it cannot predict a dynamic trajectory, not to mention the optimal trajectory that an individual developing country must follow for its future environmental situation.

Pessimistic critics argue that cross-sectional evidence for the EKC is nothing more than a snapshot of a dynamic process. Over time, they claim, the curve will rise to a horizontal line at maximum existing pollution levels, as globalization promotes a "race to the bottom" in environmental standards. Other pessimists hold that, even if certain pollutants are reduced as income increases, industrial society continuously creates new, unregulated and potentially toxic pollutants and the overall environmental risks may continue to grow. “Although both pessimistic schools make plausible claims, neither has bolstered them with much empirical research. In contrast, recent empirical work has fostered an optimistic critique of the conventional EKC. The new results suggest that the level of the curve is actually dropping and shifting to the left, as growth generates less pollution in the early stages of industrialization and pollution begins falling at lower income levels” (Dasgupta et al 2002).

According to Stern (2003) most EKC literatures are econometrically weak, “it is very easy to do bad econometrics and the history of the EKC exemplifies what can go wrong”. The EKC idea

(19)

18

rose to prominence because few paid sufficient attention to econometric diagnostic statistics and little or no attention has been paid to the statistical properties of the data used such as serial dependence or stochastic trends in time series and few tests of model adequacy have been carried out or presented. He further pointed out four issues on the econometric criticisms of the EKC: heteroskedasticity, simultaneity, omitted variables bias, and co-integration issues.

And he asserted “when we do take such statistics into account and use appropriate techniques we find that the EKC does not exist.”

A review of the empirical literatures on the EKC shows the empirical evidence is “neither consistently supportive of its traditional inverted-U shape nor uniform across pollutants”. Even if EKC exist, several decades may pass before turning points are reached, and extensive environmental degradation may occur in the mean time. The turning point on the EKC is probably associated with the dynamics of individual environmental elements that change with income (Webber and Allen 2004).

Table 1. Summary of Empirical Literatures

Authors Methods Findings

Shafik and

Bandyopadhyay 1992 Time-series and cross-country OLS estimations

Evidences exist for EKC. There is nothing automatic, but policies and investments are necessary.

Selden and Song 1994

Pooled cross-section, fixed-effects and random- effects estimations

Generally confirms EKC findings for urban air quality, however substantially higher turning points.

Grossman and

Krueger 1995 GLS random effects estimations

Most indicators support EKC and a turning point of $8000 in most cases.

Torras and Boyce 1998 Cross-country OLS estimations.

Literacy, political rights, and civil liberties are found to have particularly strong effects on environmental quality in low- income countries.

Harbaugh, Levinson &

Wilson 2001

Panel data estimation for worldwide urban pollution.

Little empirical support for EKC. Authors argue cleaning up the data, econometric specifications and including new observations make EKC disappear.

Hökby and Söderqvist 2003

Followed a survey-based approach for modeling the demand for public goods to estimate income and price elasticities of demand

Income tends to influence willingness to pay positively and significantly for reduced marine eutrophication effects, and it is also price elastic service.

Webber and

Allen 2004 Survey of existing empirical literatures

Empirical evidence is neither consistently supportive of EKC nor uniform across pollutants.

Cole 2004

Estimations of panel data on North–South trade flows for pollution intensive products

Evidence of pollution haven effects is found, but appear to be relatively small compared to the roles played by other explanatory variables.

Stern 2005

EKC is reformulated as the best practice technology frontier. The Kalman filter is used to model the state of sulfur emissions abatement technology in 15 mainly developed countries.

With the exception of Australia, countries are converging toward the frontier. Pre-abatement levels of pollution, income per capita, population density, and perhaps cultural factors might partly explain the level of abatement adopted.

Dinda 2006

Fixed-effects and random-effects panel data estimation.

Globalization transfers CO2 emissions from developed to developing countries, net effect, an increase in CO₂emissions.

Shen 2006

Simultaneous equation model estimations and Hausman test to check simultaneity between income & pollution.

Confirms EKC and simultaneity between income & pollution.

Government abatement expenses has significant and negative effect.

He 2007

Author surveyed the existing empirical EKC

studies There was no one-fit-for-all growth-pollution relationship.

(20)

19

Jiang, Lin and

Zhuang 2008 Chinese provincial level panel data estimation

EKC relatioship does not hold for all pollutants, and turning point at a higher income level for the more developed regions than the less developed ones.

Bartz and Kelly 2008

Calibration of econometric estimates, using

panel data only from developed countries The result is mixed, some support exists for EKC.

The outcome of the empirical studies is mixed; even those who support the EKC hypothesis admit there is no generality in applying the EKC effect for all pollutants. It is more of a specific phenomenon which can be influenced by multiple specific variables, not forgetting the various academic concerns and debates in estimating the EKC. But the fact still remains that there are new ways and technologies which allow abating significant amount of pollution for some types of pollutants, we may also discover new effective abatement mechanisms for those pollutants which we are not enabled to efficiently abate at this moment. However the question to what extent they will be applied and effectively abate emissions to the required level while maintaining productivity and sustainable economic development is still unknown.

From the empirical literature review, it is observable that most of the studies are largely focused on the developed countries and few studies on the developing countries, which are mainly on China. There is hardly any empirical study of Environmental Kuznets Curve in relation to the Eastern European region which mostly consists of countries in economic transition, both in terms economic growth and economic liberalization. Thus one of the most important aspects of my study is that it makes a new contribution to Environmental Kuznets Curve (EKC) empirical studies, since it is entirely concerned about the eastern European region. This will provide an additional approach when viewing the generality of EKC, which is an issue becoming more and more important in recent studies that are critical of the EKC hypothesis.

(21)

20

3. Research Method

The intention of this study is to empirically identify the short-run effect of technology transfer on transitional developing countries’ GHG emissions and understand how this effect implicates on the EKC hypothesis. The better way to approach this is to examine how this works in a specific context and my study is only suitable in such contexts. Here the specific context is to examine for the eastern European transitional economies. The method through which this research will be done is discussed below.

In general, there are two basic types of research associated with the scientific method. (1) Quantitative research-this is based on collecting facts and figures, which can be sub-divided further into descriptive & inferential; and (2) Qualitative research which is based on collecting opinion and attitudes. For my research, quantitative method is the appropriate method to probe into the subject, as the aim is to analyze an ex-post inference concerning some unknown causal relationship between GHG emissions and technology transfer. Thus, I consider econometric technique to be the best approach and will employ it to examine how technology transfer from the developed to the eastern European transitional economies can help GHG emission abatement in the short-run.

3.1. Data

A panel data on three primary GHGs i.e., carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O) and a set for the explanatory variables i.e., GDP, population, foreign direct investment net inflows, imports of goods and services, percentage share of fossil fuel energy consumption, percentage share of industry to GDP and percentage share of services to GDP for five eastern European countries (Belarus, Bulgaria, Romania, Turkey and Ukraine) for the years between 1990 to 2007 is used. These countries are selected because they have a close relationship and proximity with the developed countries mainly in Europe which is a major source of new technologies for renewable energy sources.

Compared to other transitional economies around the world such as Russia, central Asian countries, China, and few others in different parts of the world, the Eastern European nations are in a much better position to attract FDI from Central and Western Europe. The more developed nature of their economies potentially provide western transnational corporations which have control over the bulk of the world's technology with more developed markets in which they supply products of the transferred technology.

(22)

21

However if we look at and compare these five countries to other countries in east Europe such as Poland, Czech Republic and Latvia in terms of political and economic ties to the more advanced regions of central and west Europe, these five countries still have a considerable gap. Turkey, Ukraine and Belarus are not European Union (EU) members, while Romania and Bulgaria joined the EU only recently in 2007. Economically, most European nations are categorized as high income developed countries and upper middle income developing countries but the five countries in my data set are categorized as lower middle income developing countries (Todaro & Smith 2006). This nature of their economy mimics a number of characteristics of many developing countries around the world to some extent. In this way it is possible to somehow relate the implications of this study to other developing countries to some degree.

The data on GHGs are taken from United Nations Framework Convention on Climate Change (UNFCCC) and it is adjusted for Land Use, Land-Use Change and Forestry (LULUCF) which has both an increasing and decreasing effects on emission levels. All of the country level data on the explanatory variables except for percentage share of industry and services sectors to GDP have been collected from The World Bank Statistical database. Percentage shares of industry and services sectors to GDP are collected from various sources, mainly from CIA’s The World Factbook, country specific statistical bodies and other different websites.

In terms of economic aspects there are no major differences among the first three countries i.e., Belarus, Bulgaria and Romania in table 2 next page. All these three countries are categorized among lower middle income countries (Todaro & Smith 2006). While Ukraine has many similarities with the three countries also being considered as a lower middle income country, it considerably differs from the three in terms of the sheer magnitude of GDP and size of its economy. All these four countries has undergone a severe economic crisis during 1990s due to an extensive political and economic shift in this region, this facts can be seen in the full data set in appendix C.

Turkey is unique in many facets from the other eastern European countries in the data set.

During this period (1990-2007) the country has went through an economic transformation which has resulted in a continuous and momentous growth in the economy and GHG emissions over the entire period. Though categorized as a lower middle income country, it has now become the world's 17th and Europe’s 6th largest economy and is in the top 10 emerging

(23)

22

markets after BRIC countries (Brazil, Russia, India and China) (Todaro & Smith 2006; Turkey Homes 2009).

Table 2. Data Summary Statistics (1990-2007)

Min Max Mean Std. Dev.

All Countries

Per capita CO2 emissions(Kg CO2 eq.) 1,580.81 12,380.56 4,505.55 2,177.28

Per capita CH4 emissions (Kg CO2 eq.) 520.76 2,917.24 1,373.54 489.03

Per capita N2O emissions (Kg CO2 eq.) 22.42 1,280.77 609.47 314.94

Per capita GDP (US$) 635.71 8,874.15 2,374.49 1,643.76

FDI net inflows (Millions US$) 0.01 22,046.00 2,055.04 3,972.67

Imports of goods and services (Millions US$) 4,290.32 178,053.25 27,245.42 29,939.71

Share of fossil fuel energy consumption (%) 70.65 96.15 85.77 6.28

Share of industry to GDP (%) 26.00 49.00 36.10 6.04

Share of services to GDP (%) 28.00 63.00 49.14 9.01

Belarus

Per capita CO2 emissions(Kg CO2 eq.) 2,386.78 7,856.95 3,769.74 1,641.94 Per capita CH4 emissions (Kg CO2 eq.) 1,138.53 1,512.97 1,271.81 130.12

Per capita GDP (US$) 1,209.61 4,666.64 1,933.77 959.47

FDI net inflows (Millions US$) 5.00 1,785.20 244.70 408.86

Bulgaria

Per capita CO₂ emissions(Kg CO₂ eq.) 5,158.09 9,198.72 6,377.33 953.36

Per capita CH4 emissions (Kg CO2 eq.) 1,514.89 2,321.40 1,785.76 241.34

Per capita N₂O emissions (Kg CO₂ eq.) 564.61 1,280.77 705.72 172.56

Per capita GDP (US$) 1,150.55 5,163.25 2,120.87 1,157.64

Romania

Per capita CO₂ emissions(Kg CO₂ eq.) 2,335.13 5,883.83 3,577.31 787.52

Per capita CH4 emissions (Kg CO2 eq.) 1,092.25 1,748.17 1,277.38 165.79

Per capita N₂O emissions (Kg CO₂ eq.) 651.55 1,216.16 785.37 137.11

Per capita GDP (US$) 1,100.98 7,856.45 2,473.39 1,827.27

Turkey

Per capita CO2 emissions(Kg CO2 eq.) 1,580.81 3,125.88 2,120.38 430.80

Per capita CH4 emissions (Kg CO2 eq.) 520.76 746.10 684.12 60.49

Per capita N2O emissions (Kg CO2 eq.) 22.42 132.21 73.40 25.94

Per capita GDP (US$) 2,158.62 8,874.15 4,070.49 1,882.44

(24)

23

Ukraine

Per capita CO2 emissions(Kg CO2 eq.) 4,829.77 12,380.56 6,682.99 2,041.68 Per capita CH4 emissions (Kg CO2 eq.) 1,487.86 2,917.24 1,848.60 450.34

Per capita GDP (US$) 635.71 3,068.61 1,273.92 624.58

Kg CO2 eq. = Kilograms of Carbon dioxide equivalent

Source: The World Bank, UNFCCC, CIA’s The World Factbook, Romania Central, Various Websites and Author’s Calculations

Table 2 above reports summary statistics of the variables used in estimation. With an 18-year period of observations for each individual country, there are ninety observations overall. The first summary is for the overall data set and then for the concerned variables of the five countries on individual basis. Some missing observations particularly on percentage share of industry and services to GDP are replaced by mean of nearby points and by looking the linear trend at point when computing this summary statistics and estimations. Refer appendices for the full data set and some more detailed descriptive graphical analysis.

Because FDI and trade have been increasing all over the world in the past decade, it is not surprising that FDI and trade among these countries has followed a similar trend. However if we compare these countries among each other in terms their integration with the global economy through FDI and trade, Turkey stands out as the best. Turkey has now become one of the world’s biggest markets for both capital and consumer goods and services with its population reaching 73 million in 2007. The country has managed to attract major international firms from different parts of the world to invest in the energy, manufacturing and services sectors of its economy.

Being former centrally planned economies, the other four countries have inherited very weak private sectors from the period before 1990. After a serious of reforms during the late 1990s, the share of private sector contribution to GDP and employment has grown significantly.

Nevertheless compared with Turkey these countries still clearly lag far behind. Subsequent to Turkey; Romania and Bulgaria are other major destinations for FDI in this region since the late 1990s, the countries have attracted increasing amounts of foreign investment, especially Romania becoming the single largest investment destination in Southeastern and Central Europe. The economy of Bulgaria is an open free market economy with a moderately

(25)

24

advanced private sector and enterprises, the service and industry sectors now relies on contract agreements with European firms and attracting foreign investment and trade, but still with a number of strategic state-owned enterprises (Wikipedia 2010). Following their accession to the EU in 2007 the two countries are expected experience a large increase in FDI inflows and external trade in the future.

With regard to openness, Ukraine and Belarus have a long way to go. Despite major reforms towards economic liberalization and encouraging foreign investment and trade in Ukraine during the 1990s, there is still a widespread resistance to reforms within the government and from a significant part of the population, which soon delayed the reform efforts and continues to stymie direct large-scale foreign investment in Ukraine. Due to the structure of its economy and some uncertainties about the country’s political stability, Belarus is still not in the position to attract major foreign direct investments. Most of the Belarusian economy remains state- controlled and its largest trading partner is Russia, accounting for nearly half of total trade.

3.2. Models and Variables

The model specifications I employ follows the main stream in the sense that I transform the GHG emissions, GDP, FDI net inflows and Imports data to per capita quantities and estimate per capita GHG emissions for each of the three gases separately.

To study the impact of technology transfer on GHG abatement, I specify the following panel model for each GHG:

GHGPCikt = αk + β1kGDPPCit + β2kGDPPC²it + β3kFDINIPCit + β4kGDPPCit.FDINIPCit

+ β5kIMPPCit + β6kGDPPCit.IMPPCit + β7kSFFEit + β8kSIDUit + β9kSSERit + µik + εikt (1) where:

i = 1, 2, …,5 (countries); k = 1,2,3 or CO₂, CH₄, N₂0

t = 1, 2, …, 18 or 1990,1991,…, 2007; α = constant term (intercept) µ = country fixed effects; ε = error term

GHGPC =per capita Greenhouse Gas emissions GDPPC =per capital Gross Domestic Product

(26)

25

FDINI PC= per capita Foreign Direct Investment Net Inflows IMPPC = per capita value of Imports of Goods and Services SFFE = percentage share of Fossil Fuel Energy consumption SIDU = percentage share of Industry to GDP

SSER = percentage share of Services to GDP

Equation (1) is a general model and I have further included three more models in my estimate, which are modifications of the general model. The modifications are made by just dropping the non key variables from the general model, without changing the functional form of my specification or adding a new variable. Hence there are four model estimates for each of the GHGs; model one is equation (1) which is the general form, model two is a model where the variable SFFE is dropped, model three is a model where the variables SIDU and SSER are dropped and finally model four is a model where all the non key variables i.e., SFFE, SIDU and SSER are dropped.

Both theoretically and empirically, the non-key variables have direct and very significant influence on GHG emissions and also on the other key explanatory variables i.e., GDPPC, FDINIPC & IMPPC. In the literature review, we have seen that the non-key variables in my model hold a central place in the EKC hypothesis explanations, i.e., structural change in the economy and energy sources (fossil or non-fossil fuel). Thus it is clear to notice that, it is not possible to view the issue of EKC entirely separate from the non-key variables. However, since the aim of this study is specifically on the role of technology transfer on GHG abatement and its implication on the EKC hypothesis, it is ideal to extend the general model in such a way. This approach will allow a detailed analysis of the key variables and how the interaction among the independent variable may potentially cause severe statistical problems in the overall model and coefficient estimates, problems such as coefficients having “wrong” sign or implausible magnitudes, most commonly due correlations among explanatory variables.

Foreign direct investment net inflows and the value of imports of goods and services are the two key variables of the models apart from per capita GDP which my analysis will focus. These two variables are proxies to measure and capture the degree of openness and technology transfer for each country. It is widely acknowledged that most technology transfers take place through investment contracts with transnational corporations which are headquartered in

(27)

26

developed countries. Since these corporations are the sources of most of the world's technology, the principal method of transfer is obviously foreign direct investment and trade transactions especially through the import of capital goods to the developing countries’

economy by these corporations.

Other key aspects of this model which my analysis will focus on are the first two coefficients of GDPPC and GDPPC² which are β1k and β2k respectively and the interaction terms GDPPCit.FDINIit and GDPPCit.IMPit. For the EKC hypothesis to hold we must have that β1k ≥ 0 and β2k < 0. The interaction terms are very essential to examine how FDINIPC and IMPPC are related to the turning points (if there exist evidences for EKC) and extend the findings to compare each country’s turning point. In this way I hope we can learn whether or not technology transfer and openness are likely to serve as short-term solutions for GHG abatement.