Introduction

A number of recent studies on environment and development have reported a systematic relationship between change in per capita income and environment quality. They argue that as economic development takes place environmental quality declines initially only to improve later on. This systematic relationship between per capita income and environment quality has come to be known as the Environment Kuznets Curve - EKC (Grossman and Krueger, 1991). Before the advent of the idea of EKC, it was believed that with economic development environmental quality declines. It was thought that richer economies damage and destroy their natural resources as faster rates that the poorer economies (Yandle, Vijayaraghavan, and Bhattarai, 2002). Many believe(d) that environmental quality can only be maintained by escaping the clutches of industrialisation and the desire for a higher income. The new idea, inherent in EKC, explains that the best way to maintain environmental quality and quality of life is to become rich and technologically efficient. This paper attempts to address the following question/issues: How has EKC got its name? How does EKC relationship arise? What is the empirical evidence on EKC? What may be possible EKC shapes under various policy regimes? And, what are possible relationships between per capita income and environmental quality in the long-run? The paper is divided into five sections. Section I deals with the naming of EKC, while Section II examines the causes which give rise to EKC and also reviews empirical literature on it. Sections III and IV deal with possible shapes of EKC under various policy regimes and on the relationship between per capita income and environmental quality with a long-run perspective, respectively. The last Section presents conclusions of the study.

Naming of EKC

While delivering the presidential address on ‘Economic Growth and Income Inequality’ at the 67th annual meeting of the American Economic Association in December 1954, Simon Kuznets suggested that initially with increase in income inequality increases. However, the inequality reaches to a point (income) after which with a further increase in income, it starts declining (Kuznets 1955). The point after which the inequality starts decreasing is called ‘turning point income’. In other words, Kuznets believed that at the early stages of economic growth, the income distribution is more polarised or unequal. But the income distribution eventually moves towards greater equality as economic growth continues. This changing relationship between income and inequality can be presented by a inverted U-shaped/bell-shaped curve, popularly known as Kuznets curve, as shown in Fig. 1 (Yandle, Vijayaraghavan, and Bhattarai, 2002). Figure 1Kuznets Curve

Researchers through empirical evidence in the early 1990s found that there exists a similar kind of relationship between increase in per capita income and environmental degradation as between increase in per capita income and inequality in distribution of income. They found that initially with an increase in per capita income environmental degradation increases; however, after a point (turning point income) the environment degradation starts declining with any further increase in per capita income. Thus, in the 1990s the Kuznets curve took on a new existence. Now, EKC has become a vehicle for describing the relationship between existing level of per capita income and environmental degradation like carbon dioxide emission, sulphur dioxide emission, forest cover, water quality, etc (Shaban, 2004). Without changing the shape of the Kuznets curve of inequality, one can convert it to an Environmental Kuznets Curve (EKC) as shown in Fig. 2 (Yandle, Vijayaraghavan, and Bhattarai, 2002). Figure 2Environment Kuznets Curve

The logic of EKC is very simple and appealing. In pre-industrial and simple agrarian economies there is a low level of per capita income. The environment is almost unaffected by economic activities — at least with those pollutants which are associated with industrial activities. As industrialisation and agricultural development progress, more and more natural resources are utilised to increase per capita income and improve quality of life. Owing to the lower level of technological development or inefficient technologies (which degrade environment at faster rates) and also unconcern and ignorance of looming environmental problems, the environmental degradation continues. It can also be seen that many developing and underdeveloped countries even being aware of the environmental problems certain economic activities can cause, continue with those activities for faster rates of economic growth. However, as the environment becomes progressively degraded with increases in income, the demand for and concern for cleaner environment, like cleaner water, improved air quality, cleaner habitat become more important. The inefficient technologies used in the initial stages of industrialisation start giving place to environmentally efficientally efficient and cleaner technologies. High polluting activities are slowly substituted by low polluting activities. Structural change in the economy become more prominent as tertiary activities or information and service-based activities start dominating the economy (One may note here that transport activities which fall under the service sector are generally highly polluting activities. However, technological improvement in the sector is leading to an increase in environmental efficiency). Hence, the environmental degradation with a further increase in per capita income starts declining, (see the next section on causes for environmental improvement with development) forming the EKC. One should note here that if we take a long-period or entire history of a country/globe and expected future development into consideration, the curve of the relationship between per capita income and environmental quality will have tails at both ends. However, the right hand side tail will be higher than the left hand side tail because there is no possibility that we shall ever be able to get the same level of environmental quality with which we (humans) started our journey. In fact, the EKC shows only a partial fact, vide sections III and IV of Fig. 5. This mainly because the empiricists have built up their model from the past data available, and on that they have forecasted the shape of the curve.

EKC Literature: A Review

A number of studies on relationship between environmental degradation and per capita income have been published since 1991. Several studies along with explaining causes behind EKC type relationship between income and environmental degradation have supported EKC for various pollutants. However, there are studies, which negate EKC and show either no relationship between income and environmental pollution, or monotonically increasing, or u-shaped curve, or cubic relationship, etc.

The EKC relationship is believed to arise due to the combined effect of the following factors (Selden and Song, 1994; McConnell, 1997; Barbier, 1997, Shaban, 2004). First, economic structure or the composition effect: as the economy shifts from a subsistence to an agricultural and then to an industrial economy, pollution levels increase due to a fundamental change in the nature of the production processes. However, as the economy matures and becomes service oriented, there is a downturn in pollution levels. Secondly, regional shift: relatively dirty industries from developed regions/economies shift to relatively underdeveloped regions/economies. Thirdly, trade policy: the more open an economy, the greater the possibilities for importing (exporting) pollution intensive commodities, the higher (lower) the domestic pollution levels. Fourthly, technology effect: over time technological change lowers the pollution intensity of production. The could be due to greater efficiency in production (i.e., lower input use per unit of output produced) and/or improved environmental management (i.e., lower emissions per unit of input used). Fifthly, income elasticity of environmental quality: it is hypothesized that environmental quality (pollution) eventually has a positive (negative) income elasticity. This leads to the conceptualisation that environmental quality is a luxury good (Ruttan, 1971). As income increases, all else held constant, the demand for improved environmental quality increases (Baumol, 1967; Echevarria, 1997; Rowthorn and Ramaswami, 1997; Saeger, 1997). This stimulates public pressure for stricter environmental policies, as well as the use of cleaner technologies and management practices (Khanna, 2002). Sixthly, polluting activities increase with income, pollution intensity decreases with income and the product of these two happens to follow an inverted U-shaped curve or EKC. Lastly, action/reaction by citizens, businesses, policymakers, regulators, non-governmental organizations, (NGOs), and market agents act to minimize environmental degradation and to improving its quality (Shaban, 2004).

It is also a fact that the EKC model has generated/evoked conflicting responses from researchers and policy makers. Applied statisticians/econometricians who have generally focused on measuring its parameters have accepted the basic tenets of the model. Typically, using cross-sectional or panel data in their regression models they have shown that air and water pollution increase with economic growth until per capita income reaches a range of $5000 to $8000. However, when income rises beyond that range, pollution starts declining, as shown by the conventional EKC in Fig. 3. The model gives a general impression that one should ‘grow first and then clean up’. However, critics argue that evidence of EKC from cross-sectional data is ‘a snap-shot of dynamic process’, and ‘over the time the curve will rise to a horizontal line at maximum existing pollution level’ as globalisation promotes a ‘race to the bottom’ in environmental standards (see Fig. 3). Some critics even opine that industrial society continuously creates new, unregulated, and potential toxic pollutants. Therefore, even if some pollutants decline with an increase in income, the overall environmental risk from new pollutants may continue to grow, as shown by the ‘new toxic line’ in Fig. 3. Empiricists with their new results argue that the EKC is actually dropping and shifting to the left (due to technological improvements, transfer of environmentally efficient technologies from one country to another-particularly to developing countries, public pressure, stricter environment regulations, etc.) as growth generates less pollution in early stages of industrialisation and pollution starts declining at lower income levels, as depicted by ‘revised EKC’ in Fig. 3 (Dasgupta, et. al., 2002). Figure 3EKC under different Scenarios

Dasgupta et al. (2002) argue that EKC may get flatter and shift leftward because of the following reasons. First, it is due to environmental regulations (Grossman and Krueger, 1993). The available evidence suggests that this is the dominant factor for the leftward shift of the conventional EKC. Analysing data for thirty developing and developed countries, Panayotou (1997) finds that effective policies and institutions can reduce environmental pollution at low-income levels. De Bruyn (1997) finds a greater impact of environmental regulation on sulphur dioxide pollution, while Mani, Hettige and Wheeler (2000) find stricter environmental regulations as the main cause for improvement in water quality, besides impacts of sectorial composition and technology effects. Some suggest that the reasons why developed countries are more strict regulators of pollutants are, (i) after ensuring basic social securities and services they invest in environmental damage control and improvement, (ii) plenty of budget and technological efficiency to monitor and enforce activities, and (iii) environmental consciousness with higher income empowers local communities to enforce higher environmental standards (Dasgupta & Wheeler, 1997; Pargal and Wheeler, 1996; Dean, 1999; Dasgupta et. al., 2002). Dasgupta, et. al. (2001) using data for thirty-one countries show a closer relationship between pollution and pollution regulation.

Secondly, many countries have liberalised their economies which has resulted in adjustment towards economic activities that reflect a comparative advantage. As a result, there has not only been a change in sectorial compositions but also in emissions of pollutants. Many developing countries have become specialised in labour-intensive activities like garment production which is hardly pollution intensive. However, activities like electronic assembly and fabric production use and generate certain toxic pollutants (Hettige, Martin, Singh, and Wheeler, 1995). Basic metals, petrochemicals and other chemical industries are the heaviest polluters and they also received subsidies in several developing countries as they operate in crucial sectors. Privatisation and reduction of subsidies not only reduce the scale of these activities but also encourage production in assembly and service sectors that produce pollutants (Dasgupta, Wang, and Wheeler, 1997; Lucas, Hettige, and Wheeler, 1992; Jha, Markandya, and Vossenaar, 1999; Birdsall, and Wheeler, 1993). Goklany (2001), and, Grossman and Krueger (1991) also not that open economies improve their environments. More open economies have had higher growth rates of labour-intensive assembly activities that have low toxic intensity. It is observed that highly protected economies have had more rapid growth of capital-intensive smokestack sectors (Yandle, Vijayaraghavan, and Bhattarai, 2002).

Thirdly, action/reaction against polluters by citizens and civil society also act as a deterrent against pollution. There are many instances within India (like Chipko Movement, Narmada Bacchao Andolan, Silent Valley Movement, etc.), and in other countries where agitations by such organisations have restricted the environmental damage. A cement factory in Jakarta compensates local people with an ex-gratia payment of Rp. 5000 and a tin of evaporated milk every month for the dust it generates (Cribb, 1990) In Rio de Janeiro a neighbourhood association protest led to the relocation of a polluting tannery to the city's outskirts (Stotz, 1991). Community action prevented opening up of a chemical complex in Korea until pollution control equipment was installed (Clifford, 1990). Studies report that local farmers near Jakarta burned down a government-owned chemical factory that was polluting their local irrigation channel (Cribb, 1990).

Fourthly, market agents like credit institutions can also play a vital role in preventing environmental pollution by refusing to extend credit to heavy polluters or technologically inefficient industries. These agents may fear that the consumer may avoid the product of heavily polluting firms. Even firms are found to minimise their pollution emissions fearing consumer's fury. US and Canadian stock markets react significantly to environmental news, and generate loss and gains in the range of 1–2 percent from bad and good environmental news, respectively (Moughalu, Robinson, and Glascock, 1990; Lanoie, and laplante, 1994; Klassen and Mclaughlin, 1996; Hamilton, 1995; Lanoie, Laplante, and Roy, 1998; Dasgupta, et al., 2002). In Argentina, Chile, Mexico and Philippines stock price gain averages 20 percent in response to good news and lose 4–15 percent in wake of bad news (Dasgupta, Laplante, and Mamingi, 2001). It has been reported that multinational firms operating in several developing countries are often environmentally friendlier than domestically owned firms because of their efficient technologies (Afsah and Vincent, 1997; Dasgupta, et. al., 2002).

Fifthly replacement of traditional command-and-control policies by market-oriented forms of regulation has also played a notable role in pollution abatement. Pollution charge have been successfully levied in developing countries like China (Wang, and Wheeler, 1996), Colombia, Malaysia, and Philippines (World Bank, 1999). A pollution charge programme in Laguna Bay region in Philippines reduced pollution by 88 percent within two years (World Bank, 1999). Similar findings have been reported from Malaysia (Jha, Murkandya, and Vossenaar, 1999; Khalid and Braden, 1993).

Finally, available of improved information about pollutants and their impact has also created pressure from citizens, government, NGOs, and market agents for stricter environmental regulations. Studies have shown how large cities in developing countries suffer very high costs from pollution, even when damage is evaluated at conservative estimates of local opportunity costs (Dasgupta, Wang, and Wheeler, 1997; Von Amsberg, 1997; Calkins, 1994). It is found that a public disclousre programme in Indonesia and Philippines have reduced emissions from hundreds of water polluters by 40–50 percent within a two years period (Afsah and Vincent, 1997; World Bank, 1999). Table 1

Water Pollutants and Income

Pollutants	EKC Turning Points

	1985 to US $	2001 US $

1. Arsenic	4900	8000
2. Biological Oxygen Demand	7600	12500
3. Biological Oxygen Demand	7600	12500
4. Cadmium	5000	8200
5. Chemical Oxygen Demand	7900	13000
6. Dissolved Oxygen	2700	4400
7. Faecal Coliform	8000	13100
8. Nitrates	2000	3300
9. Led	10500	17200
10. Smoke	6200	10200
11. Sulphur dioxide	4100	6700
12. Total Coliform	3000	4900

Note: The value for 2001 US $ are approximates; data cover the period from 1979 to 1990.
Source: Grossman and Krueger (1995).

Grossman and Krueger (1995) using cross-country panel data for 58 countries reported the following turning points (Yandle, Vijayaraghavan, and Bhattarai, 2002).

Cole, Rayner, and Bates (1997) estimated the relationship between per capita income and a wide range of pollutants using cross-country panel data for OECD countries. The range of turning points reported by them is shown in Table 2 (Yandle, Vijayaraghvan, and Bhattarai, 2002). Seldon and Song (1994) using cross-country data for 34 countries focused on emissions of common local pollutants. Their quadratic in level of GDP peaks somewhere below $10,000 of GDP per capita for particulates and sulphur emissions, and above that for nitrogen emission. Importantly, they note that peaks are high enough above the per capita incomes of most of the countries. Therefore, global emission of these pollutants will continue to increase for the foreseeable future. Holtz-Eakin and Seldon (1995), using cross-sectional panel data for 130 countries for the year 1951–86 with quadratic in levels and logs of GDP specification, found that carbon emissions increase monotonically, only peaking far out of the sample at GDP per capita above $8 Million. They feel that as the carbon emission effects the entire planet and the emission reduction has the nature of a global public good, countries are unlikely to impose unilateral carbon regulations. And this is why the turning point is above $8 million per capita. An over view of other studies is given in the Table 3 (Khanna, 2002; Stern, Auld, Common, and Sanyal, 1998). Note that the majority of the studies report inverted u-shaped curve for most of the pollutants. For a more complete review, see Stern and Common (2001) and Agras and Chapman (1999). Table 2

Selected Pollutants and Income

Pollutants	EKC Turning Points

	1985 US $	2001 US $

1. Carbon Dioxide	22500–34700	37000–57000
2. Carbon Monoxide	9900–10100	16300–1600
3. Nitrates	15600–2500	25600–4100
4. Nitrogen Oxide (Industrial)	14700–15100	24100–24800
5. Nitrogen Oxide(Transport)	15100–17600	24800–28900
6. Sulphur Dioxide	5700–6900	9400–11300
7. Sulphur dioxide (Transport)	9400–9800	15400–16100
8. Suspended Particulates (Non-transport)	7300–8100	12000–13000
9. Suspended Particulates (Transport)	15000–1800	24600–29600

Note: The value for 2001 US $ are approximates.
Source: Cole, Rayner, and Bates (1997).

Table 3

An Overview of the EKC Literature

Study	Pollutants	Estimated Relationship with Income

Carson et al. (1997)	NOx, SOx, PM10 (emissions per capita)	Monotonically declining
Cole et al. (1997)	NO₂, SO₂, SPM, CO (emisisons per capita)	Inverted U-shaped
De Bruyn et al. (1998)	NOx, SO₂ (emisisons per capita)	Log linear
Grossman & Krueger (1991)	SO₂ Concentration	Inverted U-Shaped
Grossman & Krueger (1995)	SO₂, SPM (concentrations)	SO₂: cubic
		SPM: monotonically declining
Kaufmann et al. (1998)	SO₂ (connectrations)	Iverted U-shaped
List & Gallet (1999)	SO₂, NOx (emisisons per capita)	SO₂: inverted U-shaped, cubic
		NOx: U-shaped, cubic
Panayotou (1993)	SO₂, NOx, SPM (emisisons per capita)	Inverted U-shaped
Panayotou (1995)	SO₂ (concentrations)	Inverted U-shaped
Panayotou (1997)	SO₂ (concentrations)	Inverted U-shaped
Selden & Song (1994)	SO₂, NOx, SPM, CO	SO₂, NOx, SPM: Inverted
	(emisisons per capita)	U-shaped; CO: insignificant
Shafik & Bandhopadhyay (1992)	SO₂(emisisons per capita)	Inverted U-shaped
Stern & Common (2001)	SO₂(emisisons per capita)	Inverted U-shaped, increasing
Torras & Boyce (1998)	SO₂, SPM (concentrations)	SO₂: cubic;
		SPM: inverted U, in significant
Vincent (1997)	TSP (concentrations)	Monotonically increasing
Khanna (2002)	NOx, PM10, CO, O3, SO2	NOx, PM10: inverted U-shaped; CO, O3, SO2: insignificant.
Egli (2001)	SO₂, NOx, CO₂, CO, NH3 NOx, CH₄, PM, NMVOC	NOx, NH3 inverted U shaped: SO₂, CO₂, CO, CH₄, PM, NMVOC; ambiguous.
Shaban (2004)	CO2 (carbon) (emisisons per capita)	Mixed results for India - total CO2 (carbon), CO2 from solid fuels, CO2 from gaseous fuels, CO2 from gas flaring: inverted U-shaped; CO2 from liquid fuels, CO2 from cement manufacturing; monotonically increasing; while at global level CO2 from all the sources monotonically increasing.

Note: SPM: Suspended Particulate Matter (Heavy particles); TSP: Total Suspended Particulates, NMV = Non-methane Volatile Organic Compound.

Empirical work in this area, as Dasgupta et. al. (2002) suggest, are taking a number of directions. First, increased number of organisations, local as well as international, have encouraged and sponsored data collection on environmental quality in developing countries. New windows of opportunities for studying the relationship between economic development and environmental quality would be opened up when more data become available. Secondly, pollutants data is scarce everywhere, not just in developing countries. Even, Global Environmental Monitoring System (GEMS) effort has focused on a few ‘criteria pollutants’ (so designated because statutes have required regulators to specify their damaging characteristics), like ozone, carbon mono-oxide, suspended particulates, sulphur dioxide, lead and nitrogen dioxide. Some other pollutants have been analysed but thousands of other toxic pollutants remain unregulated and untested for their impacts. Thirdly, in most cases the implied relationship between increase in income and pollution is sensitive to inclusion of higher order polynomial terms in per capita income whose statistical significance varies widely. Many, like De Bruyn, et. al. (1998), believe that cross-sectional estimations have generated spurious estimates of the EKC because they do not adequately capture the dyamic process involved. It will be useful to compare results from cross-sectional and time-series data, wherever data is available. Availability of data may help in removing ad-hoc interpretation of EKC, formulation of appropriate functional form and choices between cross-sectional and time-series analyses.

There is a general optimistic feeling in policy circles that economic growth will be beneficial to the environment and will automatically improve the environment quality. However, there are several reasons to caution against such optimism. First, the empirical evidence in favour of EKC is generally based on a reduced form equation. It assumes that increasing income automatically and by itself causes a change in environmental quality. The model does not rigorously explain the underlying mechanism for EKC. Grossman and Krueger (1995, p. 372) say that a policy response due to increased public pressure for more strictly enforced environmental standards has provided the strongest link between increasing income and declining pollution. Secondly, as mentioned earlier, most of the studies have used cross-sectional panel data mainly from developed countries along with a few developing countries. It is argued by Stern and Common (2001) that the turning point estimates based on such data may be downwardly biased. Therefore, there should be use of a much larger and more globally representative data, not only for the empirical history of developed countries. Thirdly many studies, including Grossman and Krueger (1992, 1995), Shafik and Bandhopadhyay (1992), Panayotou (1997), and Torras and Boyce (1998), are based on the GEMS data. These data are based on concentration of pollutants measured at specific stations. For example, data on ambient concentrations of various gases in India are obtained from monitors in Bombay, Delhi, and Calcutta. The data for specific locations and their averages cannot be used for generalization of the whole country. Moreover, regressions of these pollutants in specific locations are not based on the per capita incomes of their own specific locations but national level per capita incomes are used. It is possible that these studies have miscalculated the turning points. Finally, Stern and Common (2001) argue that the GDP time series from the Penn World Tables (a commonly used economic database in the EKC literature) are integrated and that a global EKC may be miss-specified. They are in favour of a model estimated in the first different form (Khanna, 2002).

Shape of EKC Under Various Policy Regimes

There can be two kinds of EKC: one for the whole globe called Global EKC (GEKC), and the other for local area/country called Local EKC (LEKC). The shape of LEKC may very from country to country depending upon the level of per capita income and environment consciousness and action of people/governments. There can be three situations. First, if resource use is subsidised and environmental externalities are unchecked in the process of industralisation of a country, the environmental degradation will increase rapidly and after the turning point the environmental degradation will decline slowly when the technology becomes environmentally efficient because of reduced subsidies and discouragement of environment polluting activities. The argument is based on the premise that subsidies cannot be sustained in the long-run and environmental quality with respect to income and information will have positive elasticity, while quantity and quality of information and per capita income would be mutually reinforcing each other. In this situation, the environment and per capita income relationship will be like the (a) curve shown in Fig. 4. Figure 4Income-Environment Relationship Under Various Policy Regimes

Secondly, there may be a regime which neither encourages nor punishes environment-polluting activities and allows the situation to be taken care of by the ‘invisible hands of market’. In this situation a symmetric curve of relationship of income and environmental degradation may arise as shown by (b) curve in Fig. 4. Finally, if government and public are concerned about the environment from the very beginning of the industralisation in a country, they tax and penalise environment polluting activities and externalities, and encourage and subsidise environment friendly activities. In this situation, the EKC will be flatter than the other two curves and lie leftward below other two curves as depicted by (c) curve in Fig. 4. Although, there is a probability that there may not be a faster rate of economic growth in this type of regime in a short-run as in the earlier two regimes, the pristinity of the environmental quality would be better maintained.

EKC and course of Economic growth

No doubt, EKC reveals only a partial relationship of economic growth and environmental quality. It does not reveal the income-environment relationship from early stages of human history. In fast, the income-environment relationship curve can logically be traced from early times and can be drawn for the foreseeable future. As mentioned earlier, this curve will have tails at both ends, and the right hand side tail will be higher than the left hand side tail. The environment-income curve can be divided into five stages (from beginning to the end): Stage of Pristine environment; Stage of Low but Increasing Environmental Degradaton; Stage of High and Fast Increasing Environmental Degradation; Stage of High but fast Decreasing Environmental Degradation; Stage of Low, Decreasing and Stabilising Environmental Degradation (Fig. 5). The Hill-like curve in Fig. 5 shows environmental degradation with increase in per capita income (economic development), while the valley shaped-curve (which is made by inverting the hill-shaped curve) shows environmental quality with an increase in income. These two curves together look like levers of a pair of pliers. Rostow's (1960) stages of economic growth can be equated with these stages of environmental degradation. The Stage of Pristine environment is similar to Rostow's age of traditional society. The traditional society is structured within limited production functions, low levels of science and technology, and even fails to reach modest levels of output. However, it is not wholly static. Agriculture and commerce continue to expand and per capita income slowly but steadily increases. At the stage, environment, although slowly being degraded, is largely pristine. Figure 5Stage of Enviornmental Degradation

With the increase in saving and income, the society establishes its first modern industry, which continues to grow. This is the beginning of stage of Low but Increasing Environmental Degradation. This stage is shown as second stage in the Figure 5. In this stage, because of increasing number of industries, which use primitive technologies, environmental degradation starts increasing at faster rate than the earlier stage. This state matches with Rostow's stage of precondition for take-off. The point where environmental degradation and environmental quality curves cut each other, the crucial third stage of environmental degradation starts. This stage can be called as stage of High, and Fast Increasing Environmental degradation. At this stage number of polluting industries and polluting activities increase phenomenally. Technology is still primitive and there is a desire to get rich soon. People want to become rich without much care for environment and therefore natural resources are depleted very fast. This stage is similar to Rostow's stage of take-off. For some countries tentative approximation of the beginning of high and fast increasing environmental degradation years (as presented by Rostow (1971) for take-off stage) can be 1783–1802 for Britain, 1830–60 for France, 1843–60 for United States, 1850–73 for Germany, 1878–1900 for Japan and 1896–1914 for Canada. Development history of Western Developed Countries shows that this stage of environmental degradation may also include Rostow's stage of drive to maturity and some early years of the stage of age of high mass consumption. By maturity of this stage, environment becomes highly polluted and a concern for environmental quality emerges in the pollution. The demand for a cleaner environment becomes prominent. Therefore, by the end of the third stage, further increase in environmental degradation is prevented by the society and government replacing inefficient technologies with relatively environmentally efficient technologies. Thus, a point is reached where there is no further increase in environmental degradation with increase in per capita income. This point is called turning point, and the income at which this point occurs is called turning point income. This point onward, the environmental degradation stops declining or environmental quality starts rising with increase in income. After this point, the environmental degradation enters IV Stage called High but Fast Decreasing Environmental Degradation. Because of high demand of environmental quality, public awareness, punitive action against polluting activities by governments, the inefficient technologies are rapidly replaced by cleaner technologies. Much improvement is made in environmentally efficient technologies at this stage (after this stage declining marginal rate of technological efficiency with respect to environment is assumed). Due to these, environmental degradation may decline at a faster rate. However, it cannot be assumed that average rate of growth of environmental pollution in this stage will be lower than or equal to the average rate of growth of environmental pollution in the third stage. This stage can be compared to the early stage of the age of high mass consumption by Rostow. By the end of this stage much of the economic activities and technologies become environment friendly. Further progress in environmentally efficient technologies proceeds at a much slower rate, as there is not much scope to improve the existing technologies and replace the economic activities. Now the environmental degradation curve starts flattening, and we enter the V Stage of environmental degradation called Low, Decreasing, and Stablising Environmental Degradation. Environmental degradation in this stage declines very slowly with increase in per capita income. Although increasing, the environmental quality will never be able to reach the quality at the Stage I. It is impossible to achieve pristine environment of Stage I because population pressure is still high and the number of economic activities are also very high. Moreover, it is impossible to recover the same environment, which we have lost. That is why the right hand side tail of the environmental degradation curve is higher than the left hand side tail. Socio-economic conditions in fifth stage can be though as similar to the late stage of Rostow's age of high mass consumption.

The above presentation on stages of environmental degradation has been an attempt to prepare a bone-structure of the theoretical model on which future works can be undertaken. The limitation of the above idea, however, must not be ignored. First, the environmental degradation curve and its counterpart is conceptualised and drawn based on the experience of Western Developed Countries. In fact, Developing Countries may learn lessons from the experience of developed countries and may avoid the course of environmental degradation the Developed Countries have gone through by using environmentally efficient technologies developed not only by Developed Countries but also through their own indigenous research. Secondly, the curves may not be as smooth as shown in Fig. 5. They may be wavy because of some short-term fluctuations due to changes in technology, economic activities, and government and public concern towards environment. Thirdly, the change from one stage to another stage may not be very marked; instead a transitional period may be experienced. Fourthly, natural factors bringing out the environmental changes, as have been seen throughout geological history, have been ignored and instead it is assumed that whatsoever change is brought about in the environment is due to human activities.

Conclusions

The relationship between environment quality and economic development has been intensively debated for a while. Nevertheless, the idea that economic development leads to environmental degradation (in the long-run) is now being replaced by the optimistic expectation that economic development in the long-run will lead to environmental improvement. Many empirical studies have supported the EKC type relationship between environment and economic growth. The conception of EKC is very useful for empirical investigation at present. At a theoretical level, however, as the present study has shown the EKC presents only two of various stages of environment and income relationship. Albeit, economic development is the major cause which leads to EKC, but the role of factors like public policy towards environment, availability of information about pollutants and their impact on people/citizens, environmental concern of people/citizens, civil society and market agents, etc., cannot be ignored. The conception of EKC has given a new hope to development thinkers and policy makers. It gives a hope that the trade-off between economic development and environmental quality may finally be overcome.

All these feel-good factors somewhere leave some scepticism/issues unresolved. The realisation of EKC is based on a long-run perspective and are we all not dead in the long-run (as remarked by Keynes in some other context)? If we keep on degrading the environment thinking that we shall overcome it in future when we get rich, may well be too late. Human society is dealing here with a different kind of enemy (environmental pollution) and like individuals it must be careful in choosing enemies. Can we fully rely on human ingenuity in this context, and will human ingenuity finally bail us out if Nature deals us a nasty hand (Nordhaus, 1993)? Or fearing environmental consequences, should we retard our rate of economic growth/progress and continue with human suffering and miseries like poverty, destitution, penury, hardship, impoverishment, famines, homelessness, diseases, (civil) war, etc., which seem to be erupting everywhere?

Environment kuznets curve: Theoretical and empirical issues