Environmental Kuznets Curve: evidence from Ukrainian data on pollution concentrations

Размещено на http://www.allbest.ru/

Размещено на http://www.allbest.ru/

Environmental Kuznets Curve: evidence from Ukrainian data on pollution concentrations

Oleksandr Kubatko, Oleksandra NilovaKubatko Oleksandr, Student at National University “Kyiv-Mohyla Academy”, specialty «Economics»; Nilova Oleksandra, Student at Sumy State University, Specialty «Finance».

© Oleksandr Kubatko, Oleksandra Nilova, 2008

This paper deals with modeling of the relationship between different pollutants (dust, СО2, SO2, NO2) and such influencing factors as income per capita and climate conditions (temperature, rain, wind, smog). For the first time in Ukraine it was quantitatively estimated the influence of per capita income on pollution. In paper the special attention is paid to the forecast of the per capita income level, when concentration of pollutants in the air will decrease.

Introduction

The relationship between economic growth and pollution has been a focus of research by economists for many years. There are two basic competing views with respect to this relationship: the first one states that economic growth is harmful to the environment due to ineffective use of recourses, while the second one states that technological process and economic growth improve environmental quality. The initial debate between the two approaches has been mainly on theoretical grounds because of the lack of indicators to reflect environmental quality. Starting in the 1990s, ambient concentrations of harmful ingredients became the most widely used approximation of environmental quality.

In 1995, Grossman and Krueger (1995) on the basis of cross-country analysis introduced the inverted U-shape relationship between pollution and per capita income. Due to the form of the relationship the curve was named the Environmental Kuznets Curve (EKC), after Simon Kuznets, who in 1955 showed that at the early stages of a country's development the gap between poor and rich increases, while when the country becomes wealthier the inequality gap decreases. Simone Borghesi (1999) argues: “It was probably Panayotou who first coined the term Environmental Kuznets Curve”.

In their seminal work, Grossman and Krueger (1995) tested different pollutants and found that in countries with low GDP per capita concentration of dangerous chemical substances initially increased but then, after some specific level of income (which was different for different pollutants), concentration was decreasing. That is why the per capita income is considered to be one of the decisive factors to determine the level of pollution for a specific territory. The authors estimated break points for per capita income (measure of people well-being when pollution starts to decline), and found that they were at the level of $4,772-$5,965 (in 1990 prices).

The main objective of this study is the estimation of the functional form of the EKC for different air pollutants in Ukraine. We want to see whether or not Ukraine follows developed and developing countries that do exhibit the EKC relationship. As a matter of fact, regions in Ukraine differ in terms of GDP per capita and regional level of pollution, which may show that regions choose appropriate levels of pollution on their own. With the development of research on the EKC, measurement of pollution was subdivided into two categories: ambient concentration of pollutants and emissions per capita. Ambient concentrations are measured in milligram per cubic meter, while emissions are measured in tons (kilogram's) per capita. kuznets ukraine air pollutant

Using concentrations of pollutants as a dependent variable, we are going to include such variables as per capita income, population density, and some new factors that were not tested yet as independent variables. These new factors are: atmospheric precipitation, the number of days with low and strong wind, percentage of days with smog, closeness to the sea, and area of a city. According to the methodology of the Ukrainian Central Geophysical Observatory weather (climate) conditions have strong influence on the concentrations of pollution in particular region in given time. For example, higher winds reduce observed concentrations; fog makes chemical substances to hang in the air. An atmospheric precipitation makes chemical substances fall down to the ground, and that is why concentrations observed are smaller on those days. If we add these new variables to the model by introducing the so-called Vector of Climate-Related Variables (VCV), it will be possible to test whether this vector affects the level of pollution. Evaluation of climate change impacts on the per capita income could be a valuable bi-product of this study in addition to the Ukrainian EKC. It could be done by estimating the per-capita income equation with usual factors such as capital and labour plus some additional like VCV, which shows how climate influences per-capita incomes.

The bell-shaped relationship between pollution and income can be explained by several assumptions. Later on using these assumptions some theoretical models were constructed. The theoretical background of EKC we start from the considering the EKC assumptions. Thus, starting the bulk of EKC assumptions the Lopes (1994), wrote that EKC can be observed only due to nonhomothetic preferences of economic agents. Under the homothetic individual preferences, an increase in income leads to higher consumption, which causes higher pollution. Individuals with nonhomothetic preferences along with rising income may desire less consumption and pollution, depending upon relative risk aversion between consumption and environmental safety. Continuing the series of EKC assumptions, Dasgupta and Laplante (2002) proposed to consider following assumptions to explain the “bell-shaped” relationship between income and pollution:

1) with rising income, marginal propensity to consume should decline or at least be constant;

2) marginal disutility of polluted environment should increase;

These assumptions are quite reasonable and little critique followed.

Having above mentioned assumptions in a mind, several theories appeared to explain the income-pollution relationship. First, the study done by de Bruyan and Ecins (1997) suggested to subdivide pollution into two effects: technical and composition. The technical effect is associated with the use of more productive technology, less harmful inputs, and more environmentally friendly equipment. All of these are possible only along with an increasing per capita income. However, at early stages of a country's development, the technical effect brings negative impact on environmental quality due to intensive exploitation of the recourses. The composition effect explains the EKC hypothesis from a structural standpoint. In the process of development, when nations become richer share of industrial sector diminishes relative to the service sector. New industrial sectors appear within an economy, which are less environmentally damaging.

American scholars Millet and List (2002) used the U.S. state level data on sulfur dioxide and nitrogen oxides in terms of concentrations to test the inverted U-shape relationship. They found that the data followed the EKC assumptions at the country level, and their tests suggested a semi-parametric specification of the EKC. Millimet D. and List J. (2002) made a through research for the US on “test the appropriateness of the traditional parametric regression specification of EKC against semi parametric partly linear regression model (PLR model).” The authors claimed that they had a proper data set for NO_x and SO₂ at state level starting from 1924 up to 1994. The advantage of such data is that the data covers long period, and it is more precise then cross-country analysis data. As a result, there was a greater possibility that such data would capture the whole Kuznets Curve. Parametric regression results of different model specification (cubic, squared) supported the hypothesis of inverted U-shape relationship between income and pollution for both NO_x and SO_2. Coefficients of income terms were significant at 99% confidence interval; moreover they were jointly significant at 99% confidence interval too. The EKC was also estimated for the shorter period of 1985-1994. The results showed that coefficients of income were significant for SO_2, but not for NO_x. The study done by List and Millimet showed that short term model did not capture the whole EKC. Parametric estimation showed the EKC to be monotonic function with a break at the level of $13000- $20000. Semi parametric estimation results were much more optimistic, and emissions declined when income per capita was about $7000-$9000. The problem with PLR approach is that researchers “failed to reject the assumption of no serial correlation in the model” (Millimet D and List J, 2002)”. As a separate problem, Millimet and List estimated the EKC for separate states - nine states for SO₂ and nine for NO_x. Semiparametric approach showed the inverted U-shape relationship. Parametric curves for NO_x in Alabama, Georgia, Iowa, Ohio and South Carolina area were almost horizontal, while semiparametric approach suggested the U-shape structure.

Studies devoted to the estimation of the pollution-income relationship at the individual country level were usually performed for the developed countries, and the developing countries were not tested. Main reasons for that were the data suitability, and availability of economic institutions willing to perform such analyses. Below we present some works devoted to the estimation of EKC for developing countries.

A study by De Groot et al. (2002) shed some light on developing countries, particularly China. China is a developing country, and also could be considered as transition economy. The data set for China consisted of several parts: waste, water pollutants, solid pollutants, SO₂ and industrial gas, waste gas. Data was taken from the Chinese statistical yearbooks for the period of 16 years (1982-1997). One peculiarity of the data is that Chinese statistics does not provide data on CO₂ and NO₂. Serious analysis of regional disparities was performed before econometric modeling. It appears that the Coastal areas and South were developing at much faster rates than the inland and North of China (12% vs. 6%). Northern regions in China are mainly agricultural, and they contribute only small share to GDP (de Groot et al 2002). Econometrically the model was specified in such a way that allowed intercepts to change from region to region, but slope coefficients of the GDP per capita were the same. That could be done under the assumption of similar development of pollution trends in the regions when income increases. The main finding of this study is that China failed to support the EKC hypothesis, but still the correlation between income and per capita income is negative. Waste water analysis showed a downward sloping monotonic pattern. Possible explanation by de Groot et al (2002) is that initially water was already heavily polluted, endangering the human lives, and there was no other way out as to only improve the quality of water. For the solid waste emissions, situation was the same: No inverted U-shape relationship was observed. Instead only linear coefficients were significant, showing negative correlations. Emissions per gross regional product decreased as people became richer. The model for the waste gas in levels showed the inverted U-shape relationship but per capita terms as well as relative to GRP terms were monotonically increasing.

Gallanger (2005) conducted a comprehensive analysis of income-environmental quality relationship for Mexico. Cross-country analysis suggests that break points for pollution occur at the level of per capital income of $5,000-$15,000. It appears to be that the per capita income of $5,000 was associated with the year of 1995 when Mexico started to liberate its economy. So, as argued in the paper, “Mexico is a pure laboratory for estimation of EKC relationship”. The econometric modeling of income-pollution relationship did not find any bell-shaped relationship. Instead the environmental quality was deteriorating with respect to income. Gallanger (2005) tested the assumption of Mexico being a “pollution heaven”. The main hypothesis was associated with NAFTA: Due to trade liberalization the US transferred its “dirty” production to Mexico. However, the “pollution heaven” hypothesis appeared to be wrong. The share of industrial output in Mexico's manufacturing was decreasing from 1988 to 1998. It was found that at the beginning of the period the share of dirty industry accounted for 30.1%, while at the end - 26.5% of production. The fact that not only income influences pollution but the opposite is also true can be found in some studies. De Bryan (2002) claims that “Environmental degradation not only reduces the productivity of workers and man made capital, it also reduces natural resources as inputs”. Barbier (1994) also claimed that decrease in environmental quality has negative influence on production and well-being of individuals.

Melnik (2006) described the influence of environmental degradation on efficiency of economic system. The environmental degradation causes loses in agricultural and forest industries; causes corrosion of industrial equipment; stipulates loses related to the worsening of workers health status, and higher mortality rates. Overall bad environmental quality stipulates such expenditures as:

- Additional expenditures on conditioners, filters in order to protect people from dangerous chemical substances.

- Additional expenses to protect equipment, (the use of anticorrosion metals); selection of more resistible agricultural plants. The last factor includes costs on R&D due to the fact that more “stable” agricultural plants are associated with genetic engineering.

- Additional cost to compensate for the reduction in productivity (costs of labor flow, medical insurance, the use of mineral fertilizers, etc.).

It is also necessary to mention that opportunity costs are rarely taken into consideration. Due to degradation of the environment some sensitive production should be reduced (usually agricultural products and some manufacturing products). In fact, the highest opportunity costs arise due to closing of such industries as recreation and tourism. In general, pollution as a negative externality reveals itself through such channels as

1) Underproduction of goods and services (health problems, additional machine servicing)

2) Reduction in productivity and quality of goods and services. Acid rains destroy agricultural plants, kill fish and generally negatively influence on farming Melnik (2006)

Ming-Feng Hung and Daigee Show (2004) applied new methodology to examine the EKC hypothesis for Taiwan. Both authors suggest that it is more appropriate from a theoretical point of view to use simultaneous equation method (SEM) to estimate income-pollution relationship. The main critique of the reduced form specification of the EKC is that there was no feedback from pollution to economic growth, and what is more important, pollution was considered as “outcome of economic growth”. The Hausman test was used to clarify the hypothesis of simultaneous equation, and the results supported the SEM. Ming-Feng Hung and Daigee Show (2004) claim that previous estimation was biased due to the omission of important factors. The basic idea of the paper is that income influences pollution, and, in turn, pollution influences income.

Data description

The data set used in this study consists of three blocks: (i) income block, (ii) pollution block, and (ii) meteorological block.

The Income block includes- data on a city level for 50 big Ukrainian cities. Basic variables in the income block are average annual wages in regions and per capita income. Data for per-capita income and wages is taken from the Ukrainian Statistical Year Books. On average, each region is represented by two cities. However, there are some exceptions: data on pollution is better represented in eastern part of Ukraine, where population density is higher and mining industries are better developed. The per capita income in each region is attributed to 1-3 big cities in this region, and the same is done with respect to wages. There are 25 annual observations on per capita income, which are subscripted accordingly for all 50 cities in each region respectively. On average, the regional per capita income is attributed to two cities. We have done such approximation because the per-capita income at a city level is not published in the Statistical Year Books.

In addition to that, for the cities, the data on total population and capital assets (measured in billions of hryvnas) was collected. For the regions, we consider total employment and also total capital assets in each region. Unfortunately, we did not find data on employment in each city, and only the data on population was available. Nonetheless, we consider both employment and population to be appropriate representatives for labor.

The pollution block consists of concentrations. Concentrations are measured in mg/m³ The data set includes concentrations of such pollutants as CO₂, NO₂, SO₂, dust and IAP (index of air pollution). Construction of IAP is discussed later in methodology. The concentration data is presented at the city level as annual concentration of pollution in the air. The city level concentration data on 50 Ukrainian cities is based on observations from 162 meteorological stations of the Central Geophysical Observatory with annual data from 1994-2006 and 1997-2006 depending on a pollutant. We use aggregate data prepared by the Central Observatory. The data on emissions at both levels - regions and cities - is taken from the statistical yearbooks “Environment of Ukraine”.

The meteorological block is presented by such indicators as the number of days in a year with smog, precipitations, winds, and annual average temperature. Based on these indicators, a vector of climate variables was constructed which includes: percentage of days with smog, winds, precipitation during a year; average temperature. All these indicators are given at the city level. However, we also use these indicators as representation of weather condition for the region as a whole. The approach is as follows: we take 2-3 big cities in a region and calculate average with respect to each climate variable. It appears to be that regional average is not much different from any city in that region.

The city level data and regional data are presented by their own concentrations (CO₂, NO₂, SO₂, and dust), emissions, climate variables, and income variables. The entire data set is constructed for the period of nine years 1998-2006, which gives us 450 observations on each indicator at the city level and 225 at the regional level.

Methodology

It is necessary to start with some definitions and descriptive statistics of our dependent variable - pollution.

Maximum Permissible Dose (MPD) is such concentration of a substance in any medium (water, air, ground, meals) that during long period of time does not cause health problems for human beings.

The most frequently used method for the MPD is day-average, which produces concentration in mg/m³. All air pollutants are subdivided into 4 classes according to their influence:

Class 1 - extremely dangerous (benzaperin, lead)

Class 2 - highly dangerous (nitric oxide, phenol)

Class 3 - relatively dangerous (dust, sulphur dioxide)

Class 4 - not very dangerous (carbon dioxide, ammonia)

The quality of air is acceptable if the following inequality holds

, (1a)

where is the existing concentration of a pollutant, mg/m³, is its maximum permissible dose.

Atmospheric pollution is a special case because of the so-called additive affect. Additive effect is associated with a situation when several pollutants together are much more dangerous than the sum of these pollutants based on their individual MPDs. This effect is reflected by the following formula:

(1b)

For comparison of air quality in different cities (territories), integral indicator of pollution is used - the Index of Air Pollution (IAP)

, (2)

where K_i is coefficient defined with respect to a pollutant's class. Values of K_i are given below:

The following table presents day-average MPDs for different countries

Basic model that we are going to test is taken from Egli (2004), who tested the EKC hypothesis for Germany using pooled data. Egli (2004) found a reduced form model with only squared terms for income that underlies the inverted U-shape relationship. He used the following specification:

, (3)

where Y stands for per capita income, S - industry's share in GDP, I - sum of imports and exports from pollution intensive production relative to GDP, D is the reunification dummy for Germany. Egli states that because of time series data two econometric problems may arise: serial correlation and non-stationarity. Therefore, he proposed the use of GLS estimates to control for serial correlation in time series analysis. If two or more time series are non-stationary, they can be regressed on each other only if the series are integrated of the same order. The process is known as cointegration.

In our model, the pollution-income relationship is based on theory using available data. Due to the fact that we have a panel data for 50 big Ukrainian cities, the model (3) will be changed slightly and expanded.

The model that we are estimating in our study is:

, (4)

where stands for pollution in a city i in year t, Y stands for per capita income in each particular city, T - is average annual temperature in each city i, W - is the percentage of days in the year with wind in each particular city, R - is the percentage of days in the year with precipitation in the city, S - is the percentage of days in the year with smog in each city. In general, model (4) is restricted in a sense that we have a single intercept for all cities. According to that assumption, within one country the pollution would be the same if all economic and climate factors were equal. That assumption can be overcome by incorporating dummy variables for all but one city, which is a control unit. The EKC hypothesis is confirmed if. This would result in a inverse quadratic relationship between income and pollution.

Climate variables such as precipitation, wind, temperature, and smog have strong influence on concentration of pollution in the air, but not on the emissions. The expected sings are as follows: , which shows marginal impact of wind, should be negative, because stronger winds reduce concentration of chemicals in the air; , which shows marginal impact of precipitation, is also expected to be negative, because more rain and snow only increase the quality of air; , which shows marginal impact of smog, is expected to be positive because in such a case particles of a harmful ingredient stay in the air and do not fall on the ground. As for the, which is marginal impact of temperature, expected sign is unknown. We expect that it could be insignificant in influencing pollution.

As for the dependent variable, we are going to use - concentration here. It is also important to know whether general ecological situation in a city improves or not. For that purpose, the IAP is used. We expect the IAP to be correlated with income and maybe with the Vector of Climate Variables (VCV) since weather conditions have different impacts on different types of pollutants. The influence of the VCV on the IAP is difficult to predict, because the IAP in each city has different structure and weights of pollution classes (see table above), and, as a result, the impact of weather is often ambiguous. For example, precipitation eliminates dust more quickly from air than CO2 or NO2 because dust has larger particles, and rain more effectively purifies atmosphere. On the other hand, wind reduces concentration of CO2 and NO2 in the air much quicker because these chemicals are smaller in both size and mass, and they are easier transferred by the wind out of a city. That is why it is difficult to predict theoretically the impact of the VCV on the IAP since it depends on the structure of the IAP. In fact, our data set includes only the IAP but its structure in each particular city is unknown. The IAP is calculated by the Central Geophysical Observatory using up to 60 pollutants as components of the index.

The Economic theory states that income per capita may be also influenced by weather conditions. For example, a study done by Chimeli (2002) indicates that weather is an important factor in corn production. Suman Jain (2007) finds that climate variables are important determinants of net-farm revenues. Jeffrey Sachs (2003) showed a significant influence of geography on per capita income on the basis of cross-country analysis. Deschenes (2004) estimated the reduced value of agricultural lands due to climate changes. Helmy (2007) found a link between climate changes and efficiency of Egyptian economy, and it was proposed for Egypt to implement better technology and more irrigation. All of the above studies show that there could be a significant correlation between income per capita and weather conditions for some specific agricultural regions. The most serious critique in this approach is that weather conditions are not that important for industrialized or service economies. However, study done by Sorenson (2002) deals with seasonal forecasting of Monthly Hotel Night in Denmark, and one of the influencing factors was weather (climate index), which included indicators of Sun activity, rain, humidity, and temperature. The study done by Sorenson (2002) showed that all indicators except rain were statistically significant and were influencing tourism in Denmark. This result shows that even in developed countries weather conditions are important in determining the per capita income.

In the case when there is strong link between income and climate variables, equation (4) will be affected by multi-collinearity due to correlation of income with other variables. One possible econometric solution is the instrumental variable approach.

The relationship between pollution and income based on instrumental variable approach can be specified as follows

(5)

In which is the predicted value of capital for each particular city, which comes from the following regression:

Once the value of capital is obtained through the instrumented income with capital; it can be substituted into equations (4) and in order to estimate the true influence of per-capita income on pollution.

Descriptive statistics

As we have expected in the literature review there could exist inverted U-shape relationship between pollution and income. Looking at the table it is seen that income and income squared are both significant, and indeed represent the inverted U-shape relationship. The sign near the income squared in negative, which states that we have inverted parabola.

Table 1 - The relationship between pollution and per capita income.

In the table1 we should discuss the influence of climate variables. Thus, smog is positively correlated with pollution (case of SO2), the sign near smog is positive and significant. The more the number of days with smog in the city the more polluted the air. The precipitation is not significant in the model and their influence might be omitted. Interesting situation is with wind - according to our data and specified model the higher number of days with wind the lower the concentration of chemicals in the atmosphere. The later fact can explain the situation that according to Central Geophysical Observatory some big industrial cities are “relatively clean”, but people are suffocating because of pollution. For example Zaporizha has extremely polluted air, but according to indicators the air is clean. We showed that example with Zaporizha can be explained by presence of strong winds in the city, and methodological stations (which are stationary) don't observe severe pollution in the city.

The Economic theory doesn't provide background on how the temperature should influence pollution, in our case the influence of temperature is insignificant.

Figure 1 - The pollution - income relationship (case of SO2)

The key point that we may conclude from the table is that according to our predictions the concentration of SO2 should start to decline in Ukraine, when the per-capita income will be at level of about UAN22000. The obtained results are not in contradiction with the previous works of Western scientist, and we say that Ukraine has its own EKC (at least for SO2).

Analyzing the other pollutants we did not find the clear inverted U-shape relationship, the bulk of models states that pollution is increasing together with rise in income. The analysis of CO2 showed that temperature is significant and positively influences the concentration of CO2 in the atmosphere, while the influence of income is significant and positive. Taking into consideration the overall situation in the cities of Ukraine we found that IZA is decreasing along with rise in income. One possible explanation to the fact may that the concentration of the most dangerous chemicals decreases, while others “not so dangerous” are on their places. The model, where the dust was a dependent variable showed that precipitation is negatively influences concentration of dust in the air. That later proves the fact, that rain and smog are called to purify the atmosphere.

Table 2 - The influense of pollution on per capita income

Conclusions

The main finding of the article is that there is an inverted U-shape relationship between pollution and per capita income. We have analysed the data on the concentration of four pollutants in Ukraine and found that the EKC hypothesis is indeed held in Ukraine at least for the SO2. The rest of the pollutants did not show the inverted U-shape relationship, however we assume that it could be due to the small period of observation, and low volatility in data, more econometric analysis should be done in that sphere. Special attention should be devoted to the VCV (vector of climate variables); we found that taking into consideration the climate variables it is possible to better distribute the manufacture across the Ukraine. The influence of weather conditions on the concentrations of chemicals are really in accordance with the theory.

Special attention should be devoted to the per capita income, which is itself is influenced by the climate variables and pollution. Thus it was estimated that temperature brings a positive increment to income, wind reduces the per capita income in the regions, and smog has no influence at all.

Another finding of the paper is that CO2 and dust are positively correlated with per capita income. The higher pollution in the region the higher is the per capita income. High positive, rather than negative influence of pollution on per capita income can be explained by evidence that largest part of national income is produced in industrial cities.

Аppendix A

Table 4 - The relationship between pollution (dust) and main economic indicators (pooled data instrumental variables approach)

Arrow K, Bolin B., (1996) “Economic growth, carrying capacity and environment” Ecological application, Vol.6, No.1 p.13-15.

1. Barbier, E.B. 1994. “Natural Capital and the Economics of Environment and Development” in Ming - Feng hung and Daigee Show. 2004. “Economic growth and the Environmental Kuznets Curve I Taiwan: A Simultaneity Model Analysis”. Internet link http://www.econ.sinica.edu.tw/dshaw/ download/ekc.pdf

2. Beckerman W. 1992. “Economic growth and the Environment: Whose growth! Who's Environment?” World Development. Vol.20. pages 481-496.

3. Borghesi Simone. 1999. “The Environmental Kuznets Curve: Survey of Literature” European University Institute. http://www.feem.it/NR/rdonlyres/1D089671-FFCF-42F9-BA15-DEB9E2A581F1/138/8599.pdf

4. Chimeli Ariaster, Filho Francisko., (2002) “Climate forecasting and emergency policies evidence from opportunities Ceara, Brasil” link in Internet http://www.anpec.org.br/encontro2004/artigos/ A04A118.pdf

5. Dasgupta Sasmita, Laplante Bnoit (2002) “Confronting the Environmental Kuznets Curve” Journal of Economic Perspective, Vol.16, №1 pp.147-168 access on line http://www.owlnet.rice.edu/ ~econ480/notes/Kuznets_curve.pdf

6. de Bruyan S. 1997 “Explaining the Environmental Kuznets Curve: structural change and international agreements in reducing sulfur emissions” Environment and development economics. Vol 2. pp. 485-504.

7. de Groot Henri, Withagen Cees, Minliang Zhou. 2002. “Dynamics of China's Regional Development and Pollution” Tinbergen Institute Discussion Paper. TI 2001-036-3. acces via Internet http://www.tinbergen.nl/discussionpapers/01036.pdf

8. Deschenes Olivier, Greenstone Michael (2004) “The Economic Impact of Climate Change:Evidence from Agricultural Profits and Random fluctuations in Weather” Working paper 10663. http://www.nber.org/papers/w10663

9. Eid Helmy, Ouda Samiha, El-Marsafawy Samia (2007) “Assessing the economic Impact of Climate Change on Agricultural in Egypt” Policy research working paper 4293. World Bank Development Research Group. Access via Internet http://www-wds.worldbank.org/external/default/ WDSContentServer/IW3P/IB/2007/07/31/000158349_20070731143402/Rendered/PDF/wps4293.pdf

10. Egli Hannes (2004) “The Environmental Kuznets Curve - evidence from Time Series Data for Germany” WIF institute of economic research. Working paper 03/28 http://www.wif.ethz.ch/research/wp_03_28.pdf

11. Giles David, Mosk Carl 2003. “Ruminant Eructation and a Long-Run Environmental Kuznets' Curve for Enteric Methane in New Zealand: Conventional and Fuzzy Regression Analysis” http://web.uvic.ca/econ/ewp0306.pdf

12. Grafton Quentin, Day Kathleen, 2002. “Growth and the Environment in Canada: an empirical analysis”. University of Ottawa. CJAE manuscript # 144. http://een.anu.edu.au/download_files/ een0207.pdf

13. Grossman G., Krueger A (1995) “Economic Growth and the Environment”, Quarterly Journal of Economics, Vol.110 p.353-377.

14. Jaeger William, Kolpin Van.2001.Economic growth and Environmental Resource Allocation. Economics Department of Williams College. Internet Link. http://www.williams.edu/Economics/ wp/Jaeger_growth.pdf

15. Lopez R. (1994). “The environment as a factor of production: The Effect of Economic Growth on the trade Liberalization” Journal of Environmental Economics and Management. Vol. 27. Pp.163-184.

16. Melnik L.G., Shapochka M. K. 2006. Basics of Ecology. Ecological Economics and Environmental Management. The Textbook. Sumy University book. pp. 270-297.

17. Ming - Feng hung and Daigee Show. 2004. “Economic growth and the Environmental Kuznets Curve I Taiwan: A Simultaneity Model Analysis”. Internet link http://www.econ.sinica.edu.tw/ dshaw/download/ekc.pdf

18. Roca J, Padilla E (2003) “Economic growth and atmospheric pollution in Spain: discussing the Environmental Kuznets curve hypothesis” University of Barcelona http://www.ecap.uab.es/ RePEc/doc/wp0101.pdf

19. Sachs Jefrei (2003) “Institutions don't Rule: Direct Effect of Geography on per Capita Income” Working paper 9490. http://www.earthinstitute.columbia.edu/sitefiles/File/about/director/pubs/ nber_w9490.pdf

20. Sheoli Pargal, D. Wheeler (1996). “Informal Regulation of Industrial Polluting in Developing Countries: Evidence from Indonesia”, The journal of political economy. Vol.104. pp1314-1327.

21. Sorensen Karl (2004) “Modelling and Seasonal Forecasting of Monthly Hotel Nights in Denmark” Excess via Internethttp://www.ersa.org/ersaconfs/ersa02/cd-rom/papers/114.pdf

22. Suman Jain (2007) “An Empirical Economic Assessment of Impacts Climate Change on agriculture in Zambia” World Bank Working Paper 4219. Internet access http://www-wds.worldbank.org/external/default/WDSContentServer/IW3P/IB/2007/07/31/000158349_20070731100059/Rendered/PDF/wps4291.pdf

23. IMF ranking of countries based upon purchasing power GDP per capita http://www.imf.org/ external/pubs/ft/weo/2007/01/data/

Размещено на Allbest.ru