Utilizing panel data collected in 27 countries from 1960 to 2010, we demonstrate that the relationship between per capita urban water consumption and per capita gross domestic product exhibits an N-shaped pattern. Following the environmental Kuznets curve literature, this relationship can be named the cubic water Kuznets curve. We also demonstrate that water policies significantly influence per capita urban water consumption, which implies that appropriate policy interventions might allow developing countries to achieve economic development with less per capita water consumption.

Introduction

Industrialization, population growth and improved quality of life over the past two centuries have dramatically changed water consumption patterns around the world. Between 1900 and 1995, water consumption throughout the world increased by a factor of six, more than double the rate of population growth (World Meteorological Organization, 1998). Increases in water consumption in urban areas have been even more dramatic due to urbanization.

Urban populations survive on water that is either brought in from hundreds, even thousands, of miles away through an elaborate water infrastructure or that has been locally extracted from ground water. Traditionally, water needs have been met through the design and construction of water supply systems, ranging from dams, aqueducts and wells to water treatment and distribution facilities. Until the 1970s, water was considered abundant and inexpensive (Duarte et al., 2011), and no efforts were devoted to its conservation. Water was underpriced, and no externalities were considered. However, over the past three decades, this traditional water management approach has been challenged both by the continuous increase in water demand and the serious deterioration of the water environment. Water is no longer cheap and plentiful but, rather, has become a costly and scare resource; therefore, innovations in water management are urgently required to respond to this challenge.

Studies of water consumption indicate that economic activity is the primary determinant of global water consumption (Vorosmarty et al., 2000). A long-term, positive correlation between water use and economic growth exists in some regions (Hoehn & Adanu, 2008). However, in the early 1980s, water consumption in both Europe and North America peaked and then decreased despite continued economic expansion (International Hydrological Programme, 1999). This phenomenon is even more apparent for per capita water use (Gleick, 2003). Thus, it would be interesting to examine the relationship between per capita water consumption and economic growth, especially in the context of urbanization in the developing world, which has rarely been investigated.

To fill this gap in the existing research, we examine the relationship between per capita urban water consumption and per capita gross domestic product (GDP) utilizing panel data for 26 member countries of the Organisation for Economic Co-operation and Development (OECD) and China for the period 1960–2010. We observe an N-shaped relationship between per capita measures of GDP and water consumption and identify three stages of this relationship. During the first stage, per capita urban water consumption increases with economic development and reaches a peak consumption volume. During the second stage, water consumption per capita decreases despite continuously increasing GDP per capita. During the third stage, as economic development reaches a very high level, water consumption per capita stabilizes or begins to increase slightly. This may reflect changes in household sizes or economic structure (such as the distribution of the agriculture, manufacturing and services sectors). Because this N-shaped relationship is similar to the cubic form of the environmental Kuznets curve (EKC) proposed by Grossman & Krueger (1991, 1995), we propose a cubic water Kuznets curve (CWKC) to describe the urban water consumption pattern.

Following the EKC literature, we attribute urban water consumption per capita along the CWKC to economic scale and structure, technological change and policy factors (Grossman & Krueger, 1991; Shafik, 1994). Lieb (2003, 2004) argued that the economic determinants of the Kuznets curve can be also influenced by the policy measures. As a result, environmental policies are responsible for the downward trajectory of the EKC (Ayres, 1997). Therefore, we are interested in the effects of policy on the evolution of the CWKC and we investigated it based on an econometric analysis. We distinguish between indirect and direct water policies. Indirect water policies consist of policies with socioeconomic goals other than water management but that nonetheless influence water consumption, such as policies to promote the development of the manufacturing or service sectors. Direct water policies are designed and implemented to promote efficient and effective water management. The results of econometric analyses indicate that policy variables and per capita urban water consumption exhibit an inverted U-shaped relationship, which implies that governments should engage in continuous and long-term efforts to encourage water conservation in urban areas. The peak value of the CWKC and corresponding GDP per capita value are strongly affected by both direct and indirect policies.

China was selected for this study due to its rapid economic development over the past three decades. The Chinese data alone illustrate the change in urban water consumption with development; China was a low-income country and has become a middle-income nation. Important water management policy interventions over the last few decades make China suited to a policy study, but our methodology is applicable to any country for which water consumption data are available. Moreover, the CWKCs for China, Mexico and Turkey reflected in our data imply that proper, long-term and consistent policy interventions might allow developing countries to achieve economic growth with lower per capita water consumption in urban areas than the developed countries were able to along their growth trajectories.

Our study provides several contributions to this body of literature. First, this study broadens the research on EKCs by studying water as a natural resource and urban water consumption patterns. Second, this study investigates the effects of policy on the shape of the CWKC in the context of development. Third, this study considers a longer period than the extant research, which often utilizes cross-sectional or shorter time series data. Different countries might have different water Kuznets curves due to their various economic scales and structures, technological characteristics and policies. However, cross-sectional data can produce only one curve based on the observations from various countries during the same period. Therefore, it is problematic to use only cross-sectional data to examine the shape of the water Kuznets curve. In contrast, panel data allow us to construct an accurate curve for each country to illustrate how water use varies with development. Fourth, most of the existing research focuses on either domestic or industrial water use or addresses these sectors separately. However, domestic and industrial water use are closely linked. Indeed, these two sectors are interdependent components of urban water use. The quantity of water available in a region is fixed at a point in time such that a one-unit increase in consumption in one sector reduces the water available to the remaining sectors. In developing countries, this phenomenon is more pronounced because industries are concentrated in urban and semi-urban areas. Given the relationships among urbanization, economic growth and water use, it is logical and necessary to analyze all urban water use rather than examining domestic and industrial water use separately.

The remainder of the paper is organized as follows: in Section 2, we review the previous research on EKCs, the water Kuznets curve and their determinants; in Section 3, we demonstrate the existence of a CWKC and then investigate the factors contributing to the formation of the CWKC; Section 4 concludes.

Literature review

The EKC

The EKC describes the relationship between an indicator of environmental quality and income per capita, which usually exhibits an inverted U-shape. At earlier stages of development, economic growth leads to environmental degradation and the depletion of natural resources; however, beyond a certain level of per capita income, the trend reverses and higher income levels lead to environmental improvement. The concept of an EKC was introduced by Grossman & Krueger (1991) in their work on the potential environmental impacts of the North American Free Trade Agreement and popularized by World Bank reports (Shafik & Bandyopadhyay, 1992; World Bank, 1992). The most common explanation for the inverted U-shaped EKC is that the demand for environmental quality begins to increase when a country achieves a sufficiently high standard of living; the willingness to pay for a clean environment rises by a greater proportion than income (Selden & Song, 1994; Baldwin, 1995; Roca, 2003).

However, this explanation has received various criticisms. The traditional inverted U-shaped Kuznets curve is challenged by recent research in the field (Stern, 2004). Some scholars argued that the downward trajectory of the EKC in the North might be caused by the relocation of heavily polluted industry to the South. Another criticism of the Kuznets curve is that pollution affects more the lower income class than the higher income one. Richer people could easily move away from a polluted area, whereas poor people hardly have a choice. Therefore, environmental policy is crucial to environmental improvement, which could affect the formation of the EKC.

Economic growth affects environmental quality through three different channels: scale, composition and technique (Grossman & Krueger, 1991). Scale effects occur as an economy expands; resource use increases with additional economic growth. Composition effects arise from the relative sizes of the economic sectors within an economy, which vary in their resource intensity. Composition is related to national income because low-income economies are likely to produce different goods than high-income economies. Technique effects refer to the impacts of different technologies on the environment. Technological improvements respond to factor prices and, more generally, to regulation. At low income levels, economic development is mainly achieved through the expansion of the economy, that is, the scale of the economy. At this level of development, the economic structure is dominated by agriculture but gradually shifts toward manufacturing. The production technologies are more rudimentary and polluting. The combination of these factors increases pollution. However, as the economy further develops, technological change effects can overtake scale and economic composition effects and reduce pollution (Heerink et al., 2001).

The majority of the EKC literature examines environmental quality as a function of income. Pollution levels are the most common indicator of environmental quality used in these studies. Empirical studies suggest that an EKC is more likely to occur with short-term, local impact and human health environmental indicators, such as sulfur dioxide emissions and biochemical oxygen demand, than with long-term global indicators, such as carbon dioxide and other greenhouse gas emissions (Selden & Song, 1994; Arrow et al., 1995; Barbier, 1997; Cole et al., 1997; Harbaugh et al., 2002). Some studies, including the original paper by Grossman & Krueger (1991), observe an N-shaped relationship, which they call a cubic EKC (e.g., List & Gallet, 1999; Stern, 2004). The positive correlation between environmental indicators and cubic GDP implies that environmental indicators will stabilize or fluctuate within a narrow range when income reaches a very high level.

The majority of the EKC literature focuses on pollution indicators, which has led to criticism that the entire field tends to ignore the component of environmental quality represented by natural resources (Arrow et al., 1995). Of the studies that examine natural resource use, the majority focus on deforestation (Shafik & Bandyopadhyay, 1992; Panayotou, 1993; Cropper & Griffiths, 1994; Koop & Tole, 1999; Ehrhardt-Martinez et al., 2002; Culas, 2007) and a few address other types of resource use, such as energy, pesticides and water (e.g., Suri & Chapman, 1998; Cole, 2004; Managi, 2006).

The water Kuznets curve

Only a handful studies in the EKC literature have examined how water use is correlated with national income. Utilizing cross-sectional data on US national water withdrawals and panel data on US state-level water consumption, Rock (1998) observes that per capita water withdrawal and water consumption follow an inverted U-shaped path, which is consistent with an EKC. In a subsequent study, Rock (2000) argues that water use intensity in an economy is closely related to its natural water endowment, economic structure and government policies. Gleick (2003) observes a similar inverted U-shaped relationship between per capita water consumption and income in the USA and Finland. He argues that efficient and effective water management maintains or reduces water consumption. Cole (2004) utilizes estimated water consumption data for five continents in 1900, 1940, 1950, 1960, 1970, 1980, 1990 and 1995 to test for an EKC. His results are consistent with the existence of an inverted U-shaped relationship between water consumption and income. Based on industrial water data collected by the World Resources Institute for 1995, 1998, 1999, 2000, 2001 and 2002, Jia et al. (2006) demonstrate that industrial water use in most OECD countries follows an EKC. Utilizing United Nations Educational Scientific and Cultural Organization (UNESCO) data for 65 countries for 1960, 1970, 1980, 1990, 1995, 2000 and 2008, Duarte et al. (2013) claim that water use is described by an EKC and that countries facing water scarcity tend to use water more efficiently. They also argue that sound institutions might contribute to environmental improvement, thereby reducing per capita water consumption.

Although the above-mentioned studies support the existence of a water Kuznets curve, their conclusions are limited by the use of cross-sectional data rather than panel data and inaccurate water use indicators, such as withdrawal and industrial water use. For example, Rock (1998, 2000) analyzes a cross-section of 68 countries in 1990 and time-series data for the USA during the period 1960–1995. However, as Vincent (1997) argues, using only cross-sectional data to investigate the existence of a water Kuznets curve is problematic. In such cross-sections, all the low-income observations come from developing countries and high-income observations come from developed countries. Nonetheless, the observations from developing countries cannot represent the situation of the currently developed countries for the period during which they were industrializing. Nor can the data of the currently developed countries predict the future of the currently developing ones. In addition, a cross-section produces only one curve based on the observations from various countries at the same time. In contrast, panel data allow us to construct a curve for each country to provide accurate water Kuznets curves.

Rock (1998, 2000) and Duarte et al. (2013) use water withdrawal data as indicators of water use. As Rock (1998) himself admits, some of the water consumed by a country has been recycled. Therefore, water consumption is a more accurate indicator of water use than water withdrawal. Water consumption is divided into agricultural, industrial and domestic water uses, and industrial water use alone which Jia et al. (2006) analyze provides an incomplete measure of water consumption. Gleick (2003) demonstrates the existence of an inverted U-shaped relationship between water consumption and income through scatter plots only, which is less rigorous than regression analysis. Finally, Cole (2004) analyzes water consumption data estimated by Shiklomanov (2000) rather than actual water use.

The lack of high-quality data explains why the water consumption-income relationship is rarely tested for developing countries. In the limited studies of developing countries, the existence of an EKC is difficult to prove (Gleick, 2003; Cole, 2004). This may be because developing countries are still in the early stages of development, that is, an EKC exists but is not observed because developing countries are located on the left side of the curve. Alternatively, many developing countries have already implemented environmental protection policies, such as moving from command-and-control policies to market-oriented regulation (Dinda, 2004). Therefore, it is questionable whether developing countries will follow the water use trajectory followed by developed countries. Currently developing countries can improve or have already improved their water use based on an improved understanding of the relationship between environmental quality and income.

The determinants of the water Kuznets curve

Scholars have argued that factors other than income contribute to decreased pollution and resource consumption (Ng & Wang, 1993; Horvath, 1994; Antle & Heidebrink, 1995; Baldwin, 1995; Hettige et al., 1996). Economic structure, i.e., the distribution of economic activity among the agricultural, manufacturing and services sectors, is one such factor. Changes in economic structure affect water consumption by changing the portfolios of water-using activities. In Germany, Japan and the USA, as the economy developed, the sector accounting for the largest share of the GDP shifted from agriculture to manufacturing and finally to services. As the structure changes, water consumption per capita increases with the growth of national income but eventually stabilizes or decreases when national income reaches a high level (Infrastructure Development Institute, 1997; Solley et al., 1998).

Unlike structural change, technological change improves the efficiency of water use, which in turn decreases water consumption per capita. Carr et al. (1990) demonstrate that technological changes were the likely cause of reduced water consumption intensity by an average 2.2% per annum in the US manufacturing sector between the 1950s and 1980s.

Because scholars have long argued that policy interventions are the dominant factors driving the decline in pollution after national GDP per capita reaches the middle-income level, efforts have been made to incorporate policy factors in tests of the EKC (Andreoni & Levinson, 2001; Dasgupta et al., 2002; Chavas, 2004). Panayotou (1997) uses indexes of contract enforcement and bureaucratic efficiency to test the EKC hypothesis and observes that the effectiveness of policies and institutions can significantly reduce environmental degradation at low income levels and accelerate improvements at higher income levels. De Bruyn (1997) also observes that environmental policy plays an important role in the shape of the EKC. Hettige et al. (2000) reveal that stricter environmental regulations significantly improve water quality. Similarly, Magnani (2001) question the automatic formation of an EKC with GDP growth and argue that an EKC is more likely to emerge where public support for environmental protection, such as regulations and policies, exists. Echoing these arguments, Yandle et al. (2004) contend that improved policies will alter the EKC and yield environmental improvements at lower income levels.

The demographic characteristics of individual countries also affect the environmental quality-income relationship (Shafik, 1994; Yandle et al., 2002). Population growth and urbanization are the main driving forces of increased domestic water use in cities. Moreover, population growth augments the demand for consumer goods that indirectly increases industrial water use. Household size also influences water demand. Some water consumption among households, such as water use for cleaning or gardening, is constant regardless of the number of residents in a house. Therefore, in an aging society, as the average household size decreases, the average per capita water use within a household may increase (Birrell et al., 2005). Note that although population factors strongly affect domestic water use, the income effect remains the dominant factor, which is estimated to be five to eleven times stronger than the population effect on water withdrawal (Alcamo et al., 2007).

Physiographic factors, which include hydrological differences among regions and climate change, are also important determinants of water consumption (Adeloye et al., 1999). Scientific evidence has indicated that climate change will alter rainfall patterns, so that there will be heavier but less frequent rainfall. Therefore, climate change might decrease the availability of water on average and increase the variance. Moreover, climate change might also affect urban water use. First, higher temperatures and less precipitation will increase water use (Gato et al., 2007). Second, industrial water use is expected to increase with temperature because cooling processes and heat-sensitive industries will require additional water.

Empirical analysis

We propose two models of urban water consumption to test the CWKC hypothesis. The first model describes a conventional Kuznets curve and uses quadratic and cubic functions of GDP to capture the effect of economic development on urban water consumption. The second model investigates the relationship between per capita urban water consumption and socioeconomic, demographic, climate and policy factors. We explore the effects of policies on water consumption in urban areas.

The CWKC

The following equation specifies the first panel data model: 
formula
1
where is the per capita urban water consumption of the country i in the year t, is the GDP per capita, is the square of GDP per capita and is the cube of GDP per capita. We test whether an N-shaped relationship between per capita urban water consumption and GDP per capita exists, i.e., a CWKC. Like the EKC, the CWKC is influenced by economic scale and structure, technological improvement, and policy intervention. These factors are aggregated in model (1) as a function of GDP per capita. All factors that cannot be explained by GDP per capita are captured by the year dummy.

Policy effects in the CWKC regression

The following model investigates the relationship between per capita urban water consumption and socioeconomic, demographic, physiographic and policy factors: 
formula
2
where is the per capita urban water consumption of country i in year t, refers to per capita fixed capital formation, which represents the economic scale of the country. represents the share of total GDP generated by the secondary and tertiary sectors, which captures the structure of the economy. measures inward foreign direct investment per capita, which represents the openness of the economy and learning from other countries. denotes the technological development level by the number of the triadic patent families (Perman & Stern, 2003; Popp, 2005, 2010; Lantz & Feng, 2006). A triadic patent family is a group of patents applied for in Europe, Japan and the USA to protect the same invention. Triadic patents are of substantial innovative value because international patent applications are typically more costly than domestic applications (van Pottelsberghe & Mejer, 2010). Second, patents applied for in these three regions span the most advanced and technologically sophisticated countries in the world. This variable is thus used by the governmental agencies, such as the National Science Foundation in the USA, and intergovernmental organizations, such as Eurostat and the OECD, as an indicator of national or regional innovativeness. measures the average household size of country i over time. is the quadratic term of household size of country i over time. We include this variable in the regression because household size may have a curvilinear relationship with urban water consumption per capita. and control for two climate factors. Year dummy variables are included to capture other time-variant factors.

Because the central research question is whether water policies implemented by governments impact per capita urban water consumption, i.e., affect the shape of the CWKC, the variables represented by are the key variables in this model. We classify water policies affecting the CWKC into indirect and direct water policies. Before the 1990s, the formation of a CWKC in most developed countries was a self-motivated process without purposive government intervention. However, per capita urban water consumption might have been influenced by government policies primarily aiming at stimulating technology diffusion and promoting the development of secondary and tertiary industries, but it may have also affected water consumption. These are indirect water policies.

We are more interested in direct water policies, which governments in developed countries implemented beginning in the late 1980s and early 1990s after policy makers recognized the benefits of policy intervention in water conservation. These water policies can be further classified into two categories: efficiency and effectiveness policies. Efficiency policies aim to improve water use efficiency and achieve higher output while consuming less water without emphasizing management of the water cycle system or wastewater treatment. In contrast, effectiveness policies seek to improve the management of the urban water system to minimize the impact of human activities on the natural water system without highlighting the goal of saving water.

We introduce four variables to represent these types of policies. is defined as the total GDP generated by the secondary and tertiary sectors divided by urban water consumption volume, which proxies for water efficiency policy. The following proxies represent water effectiveness policy: , , and Public Pollution Control Expenditure on. All three variables are proxies to measure the policy effects of water management, which might affect water consumption patterns in urban areas. In theory, as the wastewater treatment rate and sludge recycling rate approach one, the urban water system approaches zero-emissions and becomes more environmentally friendly. Similarly, higher public expenditures on water and wastewater management should be associated with lower per capita water consumption in urban areas. All variables and statistics are described in Tables 1 and 2, respectively. Table 3 indicates modest correlations between these pairs of variables.

Table 1.

Definitions of variables.

Variable name Definition and notes 
CWKC regression (unit of analysis: 27 countries) 
Per capita urban water consumption Logarithm of the quantity of water consumed divided by the urban population 
GDP per capita Logarithm of GDP per capita 
Quadratic GDP per capita Logarithm of GDP per capita squared 
Cubic GDP per capita Logarithm of GDP per capita cubed 
CWKC policy effects regression (unit of analysis: 27 countries) 
Fixed capital formation Logarithm of gross fixed capital formation per capita 
Secondary and tertiary industries Logarithm of the share of total GDP from secondary and tertiary industries 
Inward FDI Logarithm of the inward foreign direct investment per capita 
Triadic patents Logarithm of number of triadic patent families 
Household size Logarithm of the average number of people in a household 
Quadratic household size Logarithm of the average number of people in a household squared 
Mean temperature Logarithm of annual mean temperature 
Total precipitation Logarithm of total annual precipitation 
Water productivity Logarithm of GDP from the secondary and tertiary industries divided by urban water consumption 
Wastewater treatment rate Logarithm of the wastewater treatment rate in urban areas 
Quadratic wastewater treatment rate Logarithm of the wastewater treatment rate in urban areas squared 
Sludge recycling rate Logarithm of the sludge recycling rate of waste water treatment plants 
Quadratic sludge recycling rate Logarithm of the sludge recycling rate of waste water treatment plants squared 
Public pollution control expenditure on water and wastewater Logarithm of public pollution control expenditure on water and wastewater per capita 
Quadratic public pollution control expenditure on water and wastewater Logarithm of public pollution control expenditure on water and wastewater per capita squared 
Year dummy Base year is 1960 for model (1), 1977 for model (2), 1978 for model (3), 1985 for model (4), and 1990 for model (5) in Table 4 
Variable name Definition and notes 
CWKC regression (unit of analysis: 27 countries) 
Per capita urban water consumption Logarithm of the quantity of water consumed divided by the urban population 
GDP per capita Logarithm of GDP per capita 
Quadratic GDP per capita Logarithm of GDP per capita squared 
Cubic GDP per capita Logarithm of GDP per capita cubed 
CWKC policy effects regression (unit of analysis: 27 countries) 
Fixed capital formation Logarithm of gross fixed capital formation per capita 
Secondary and tertiary industries Logarithm of the share of total GDP from secondary and tertiary industries 
Inward FDI Logarithm of the inward foreign direct investment per capita 
Triadic patents Logarithm of number of triadic patent families 
Household size Logarithm of the average number of people in a household 
Quadratic household size Logarithm of the average number of people in a household squared 
Mean temperature Logarithm of annual mean temperature 
Total precipitation Logarithm of total annual precipitation 
Water productivity Logarithm of GDP from the secondary and tertiary industries divided by urban water consumption 
Wastewater treatment rate Logarithm of the wastewater treatment rate in urban areas 
Quadratic wastewater treatment rate Logarithm of the wastewater treatment rate in urban areas squared 
Sludge recycling rate Logarithm of the sludge recycling rate of waste water treatment plants 
Quadratic sludge recycling rate Logarithm of the sludge recycling rate of waste water treatment plants squared 
Public pollution control expenditure on water and wastewater Logarithm of public pollution control expenditure on water and wastewater per capita 
Quadratic public pollution control expenditure on water and wastewater Logarithm of public pollution control expenditure on water and wastewater per capita squared 
Year dummy Base year is 1960 for model (1), 1977 for model (2), 1978 for model (3), 1985 for model (4), and 1990 for model (5) in Table 4 
Table 2.

Descriptive statistics.

Variable Number of observations Mean Standard deviation Minimum Maximum 
Per capita urban water consumption 503 5.93 0.40 4.88 6.81 
GDP per capita 1,377 9.73 0.78 5.79 11.05 
Quadratic GDP per capita 1,377 95.34 14.00 33.48 122.07 
Cubic GDP per capita 1,377 938.84 191.56 193.69 1348.64 
Fixed capital formation 1,267 7.55 1.33 2.40 10.11 
Secondary and tertiary industries 1,132 4.52 0.11 3.79 4.60 
Inward FDI 1,012 4.60 2.16 − 9.39 11.13 
Triadic patents 900 5.03 2.38 0.00 9.94 
Household size 651 1.06 0.21 0.67 1.75 
Quadratic household size 651 1.16 0.49 0.45 3.08 
Mean temperature 949 2.20 0.62 − 1.20 3.13 
Total precipitation 975 6.78 0.33 5.80 7.63 
Water productivity 510 4.64 1.23 0.80 6.95 
Wastewater treatment rate 317 3.85 0.90 0.26 4.60 
Quadratic wastewater treatment rate 317 15.65 5.66 0.07 21.12 
Sludge recycling rate 196 4.24 0.50 0.87 4.60 
Quadratic sludge recycling rate 196 18.21 3.52 0.75 21.19 
Public pollution control expenditure on water and wastewater 399 3.69 1.29 − 0.08 6.32 
Quadratic public pollution control expenditure on water and wastewater 399 15.30 9.06 0.01 39.93 
Variable Number of observations Mean Standard deviation Minimum Maximum 
Per capita urban water consumption 503 5.93 0.40 4.88 6.81 
GDP per capita 1,377 9.73 0.78 5.79 11.05 
Quadratic GDP per capita 1,377 95.34 14.00 33.48 122.07 
Cubic GDP per capita 1,377 938.84 191.56 193.69 1348.64 
Fixed capital formation 1,267 7.55 1.33 2.40 10.11 
Secondary and tertiary industries 1,132 4.52 0.11 3.79 4.60 
Inward FDI 1,012 4.60 2.16 − 9.39 11.13 
Triadic patents 900 5.03 2.38 0.00 9.94 
Household size 651 1.06 0.21 0.67 1.75 
Quadratic household size 651 1.16 0.49 0.45 3.08 
Mean temperature 949 2.20 0.62 − 1.20 3.13 
Total precipitation 975 6.78 0.33 5.80 7.63 
Water productivity 510 4.64 1.23 0.80 6.95 
Wastewater treatment rate 317 3.85 0.90 0.26 4.60 
Quadratic wastewater treatment rate 317 15.65 5.66 0.07 21.12 
Sludge recycling rate 196 4.24 0.50 0.87 4.60 
Quadratic sludge recycling rate 196 18.21 3.52 0.75 21.19 
Public pollution control expenditure on water and wastewater 399 3.69 1.29 − 0.08 6.32 
Quadratic public pollution control expenditure on water and wastewater 399 15.30 9.06 0.01 39.93 
Table 3.

Correlation matrix.

 WKC regression              
Urban water consumption per capita 1.00                 
GDP per capita 0.08 1.00                
Quadratic GDP per capita 0.08 1.00 1.00               
Cubic GDP per capita 0.08 0.99 1.00 1.00              
 WKC policy effects regression 10 11 12 13 14 15 16 17 
Per capita urban water consumption 1.00                 
GDP per capita 0.23 1.00                
Fixed capital formation 0.39 0.58 1.00               
Secondary and tertiary industries 0.19 0.42 0.29 1.00              
Inward FDI −0.40 0.28 −0.21 0.02 1.00             
Triadic patents 0.24 0.16 0.27 0.58 −0.44 1.00            
Household size 0.31 −0.47 0.15 −0.47 −0.48 −0.17 1.00           
Quadratic household size 0.30 −0.47 0.15 −0.48 −0.49 −0.17 0.99 1.00          
Mean temperature 0.28 0.20 0.37 0.44 −0.14 0.41 −0.11 −0.13 1.00         
10 Total precipitation 0.64 0.34 0.41 0.59 −0.33 0.72 −0.06 −0.07 0.43 1.00        
11 Water productivity −0.70 0.31 0.27 0.22 0.35 0.02 −0.38 −0.37 −0.13 −0.25 1.00       
12 Wastewater treatment rate −0.14 0.23 −0.18 −0.16 0.43 −0.23 −0.41 −0.41 −0.07 −0.32 0.09 1.00      
13 Quadratic wastewater treatment rate −0.15 0.24 −0.17 −0.16 0.44 −0.24 −0.42 −0.42 −0.06 −0.32 0.10 1.00 1.00     
14 Sludge recycling rate −0.34 0.08 −0.28 0.22 0.66 −0.29 −0.29 −0.30 −0.21 −0.35 0.32 0.39 0.39 1.00    
15 Quadratic sludge recycling rate −0.34 0.09 −0.27 0.23 0.64 −0.29 −0.29 −0.29 −0.21 −0.34 0.32 0.39 0.39 1.00 1.00   
16 Public pollution control expenditure on water and wastewater per capita −0.01 0.44 0.47 −0.20 −0.12 0.15 −0.06 −0.05 −0.13 0.12 0.27 0.11 0.12 −0.43 −0.44 1.00  
17 Quadratic public pollution control expenditure on water and wastewater per capita 0.02 0.48 0.47 −0.19 −0.09 0.13 −0.08 −0.08 −0.09 0.12 0.25 0.15 0.16 −0.42 −0.42 0.99 1.00 
 WKC regression              
Urban water consumption per capita 1.00                 
GDP per capita 0.08 1.00                
Quadratic GDP per capita 0.08 1.00 1.00               
Cubic GDP per capita 0.08 0.99 1.00 1.00              
 WKC policy effects regression 10 11 12 13 14 15 16 17 
Per capita urban water consumption 1.00                 
GDP per capita 0.23 1.00                
Fixed capital formation 0.39 0.58 1.00               
Secondary and tertiary industries 0.19 0.42 0.29 1.00              
Inward FDI −0.40 0.28 −0.21 0.02 1.00             
Triadic patents 0.24 0.16 0.27 0.58 −0.44 1.00            
Household size 0.31 −0.47 0.15 −0.47 −0.48 −0.17 1.00           
Quadratic household size 0.30 −0.47 0.15 −0.48 −0.49 −0.17 0.99 1.00          
Mean temperature 0.28 0.20 0.37 0.44 −0.14 0.41 −0.11 −0.13 1.00         
10 Total precipitation 0.64 0.34 0.41 0.59 −0.33 0.72 −0.06 −0.07 0.43 1.00        
11 Water productivity −0.70 0.31 0.27 0.22 0.35 0.02 −0.38 −0.37 −0.13 −0.25 1.00       
12 Wastewater treatment rate −0.14 0.23 −0.18 −0.16 0.43 −0.23 −0.41 −0.41 −0.07 −0.32 0.09 1.00      
13 Quadratic wastewater treatment rate −0.15 0.24 −0.17 −0.16 0.44 −0.24 −0.42 −0.42 −0.06 −0.32 0.10 1.00 1.00     
14 Sludge recycling rate −0.34 0.08 −0.28 0.22 0.66 −0.29 −0.29 −0.30 −0.21 −0.35 0.32 0.39 0.39 1.00    
15 Quadratic sludge recycling rate −0.34 0.09 −0.27 0.23 0.64 −0.29 −0.29 −0.29 −0.21 −0.34 0.32 0.39 0.39 1.00 1.00   
16 Public pollution control expenditure on water and wastewater per capita −0.01 0.44 0.47 −0.20 −0.12 0.15 −0.06 −0.05 −0.13 0.12 0.27 0.11 0.12 −0.43 −0.44 1.00  
17 Quadratic public pollution control expenditure on water and wastewater per capita 0.02 0.48 0.47 −0.19 −0.09 0.13 −0.08 −0.08 −0.09 0.12 0.25 0.15 0.16 −0.42 −0.42 0.99 1.00 

Data

We analyze panel data for 26 OECD countries and China for the period from 1960 to 2010. These 26 OECD member states were selected mainly due to data availability1. Because of data availability, we are unable to include more developing countries into this analysis either. The GDP, GDP per capita and fixed capital formation data were obtained from the World Bank database. The population, urban population and inward FDI data were attained from the United Nations Conference for Trade and Development database. The secondary and tertiary industries data were extracted from the OECD database. The triadic patent family data were obtained from the PATSTAT database constructed by the European Patent Office. We use public water supply data to measure water consumption in the urban areas of each country because urban water is supplied by public water plants and distribution systems. Most of the data on urban water consumption were obtained from local authorities in each individual country, except those that were extracted from the Eurostat database. Temperature and precipitation data were provided by the meteorological authorities in each country. The household size and policy variables were provided by the local statistical bureaus, ministries of the environment, water companies and Eurostat.

Results

We run models (1) and (2) with both fixed and random effects. The random-effects model assumes that random effects (individual, specific unobservables) are uncorrelated with the explanatory variables (individual, specific observables). This assumption is not likely to hold in our case. For example, the volume of per capita urban water consumption is affected by the environmental awareness of urban citizens, which is not measured by the explanatory variables (we cannot obtain consistent and long-run data on the environmental awareness of the urban citizens) and is therefore captured by the random effects. However, these random effects (e.g., the environmental awareness of the urban citizens) would be correlated with GDP per capita, which is an explanatory variable. If the assumption is violated, a random-effects estimator is not consistent. We perform a Hausman test to verify this assumption. The random-effects model is rejected in all the specifications, and the results are presented in Table A of Appendix A (available with the online version of this paper). The results of the fixed-effects model controlling for heteroskedasticity are provided in Table 4.

Table 4.

Determinants of urban water consumption: fixed-effects model.

  Dependent variable: per capita urban water consumption
 
 CWKC regression
 
CWKC policy effects regression
 
 1960–2010 1977–2010 1978–2010 1985–2010 1990–2010 
Independent variables (1) (2) (3) (4) (5) 
GDP per capita 12.54** (5.023) −0.107 (0.094) −0.259 (0.171) 0.394* (0.185) −0.356 (0.204) 
Quadratic GDP per capita −1.420** (0.602) – – – – 
Cubic GDP per capita 0.054** (0.023) – – – – 
Fixed capital formation – 0.308*** (0.052) −0.108* (0.057) −0.086 (0.055) 0.076 (0.056) 
Secondary and tertiary industries – 0.536* (0.304) 2.598** (0.893) −0.913 (0.678) 0.883 (1.789) 
Inward FDI – −0.002 (0.002) −0.006* (0.003) −0.001 (0.006) −0.001 (0.006) 
Triadic patents – −0.016 (0.016) −0.154*** (0.045) −0.029* (0.014) −0.068** (0.028) 
Household size – 1.775 (1.087) 0.154 (2.163) 1.612 (1.601) 1.358 (1.578) 
Quadratic household size – −0.846 (0.548) −0.118 (1.000) −0.810 (0.807) −0.616 (0.545) 
Mean temperature – −0.049* (0.028) −0.024 (0.037) −0.028 (0.017) 0.037 (0.066) 
Total precipitation – −0.019 (0.014) −0.002 (0.027) −0.005 (0.023) −0.037** (0.016) 
Water productivity – −0.509*** (0.031) – – – 
Wastewater treatment rate – – 0.613*** (0.071) – – 
Quadratic wastewater treatment rate – – −0.061*** (0.011) – – 
Sludge recycling rate – – – 0.304* (0.155) – 
Quadratic sludge recycling rate – – – −0.046** (0.020) – 
Public pollution control expenditure on water and wastewater – – – – 0.046* (0.023) 
Quadratic public pollution control expenditure on water and wastewater – – – – −0.006* (0.003) 
Constant −31.22** (13.37) 3.417*** (0.854) −2.919 (3.517) 5.892** (2.248) 4.555 (6.664) 
Year dummy Included Included Included Included Included 
Number of observations 503 213 151 87 128 
Number of groups 27 20 13 12 18 
Chi-square statistic of the Hausman test 405.89*** 422.17*** 97.15*** 53.46*** 92.94*** 
  Dependent variable: per capita urban water consumption
 
 CWKC regression
 
CWKC policy effects regression
 
 1960–2010 1977–2010 1978–2010 1985–2010 1990–2010 
Independent variables (1) (2) (3) (4) (5) 
GDP per capita 12.54** (5.023) −0.107 (0.094) −0.259 (0.171) 0.394* (0.185) −0.356 (0.204) 
Quadratic GDP per capita −1.420** (0.602) – – – – 
Cubic GDP per capita 0.054** (0.023) – – – – 
Fixed capital formation – 0.308*** (0.052) −0.108* (0.057) −0.086 (0.055) 0.076 (0.056) 
Secondary and tertiary industries – 0.536* (0.304) 2.598** (0.893) −0.913 (0.678) 0.883 (1.789) 
Inward FDI – −0.002 (0.002) −0.006* (0.003) −0.001 (0.006) −0.001 (0.006) 
Triadic patents – −0.016 (0.016) −0.154*** (0.045) −0.029* (0.014) −0.068** (0.028) 
Household size – 1.775 (1.087) 0.154 (2.163) 1.612 (1.601) 1.358 (1.578) 
Quadratic household size – −0.846 (0.548) −0.118 (1.000) −0.810 (0.807) −0.616 (0.545) 
Mean temperature – −0.049* (0.028) −0.024 (0.037) −0.028 (0.017) 0.037 (0.066) 
Total precipitation – −0.019 (0.014) −0.002 (0.027) −0.005 (0.023) −0.037** (0.016) 
Water productivity – −0.509*** (0.031) – – – 
Wastewater treatment rate – – 0.613*** (0.071) – – 
Quadratic wastewater treatment rate – – −0.061*** (0.011) – – 
Sludge recycling rate – – – 0.304* (0.155) – 
Quadratic sludge recycling rate – – – −0.046** (0.020) – 
Public pollution control expenditure on water and wastewater – – – – 0.046* (0.023) 
Quadratic public pollution control expenditure on water and wastewater – – – – −0.006* (0.003) 
Constant −31.22** (13.37) 3.417*** (0.854) −2.919 (3.517) 5.892** (2.248) 4.555 (6.664) 
Year dummy Included Included Included Included Included 
Number of observations 503 213 151 87 128 
Number of groups 27 20 13 12 18 
Chi-square statistic of the Hausman test 405.89*** 422.17*** 97.15*** 53.46*** 92.94*** 

Note: The data between the parentheses are robust standard errors.

***Denotes significance at the 1% level.

**Denotes significance at the 5% level.

*Denotes significance at the 10% level.

In support of the CWKC hypothesis, the coefficient of GDP is positive, the coefficient of the quadratic GDP variable is negative and that of the cubic GDP variable is positive. All three coefficients are statistically significant at the 5% level (column 1, Table 4). This pattern indicates that a CWKC exists, and the relationship between per capita urban water consumption and GDP per capita follows an N-shaped curve (Figure 1). The peak value of per capita water consumption in urban areas appears at different values of GDP per capita in different countries, ranging from several thousand to approximately 45 thousand US dollars. This variation is produced by varying socioeconomic, demographic, climate and policy factors.
Fig. 1.

The CWKC.

Fig. 1.

The CWKC.

In Appendix B (available with the online version of this paper), we present the CWKC for each country during the period 1960–2010. Interestingly, we only observe the part of the CWKC to the right of the inflection point for Australia, Sweden and Switzerland. The missing left part of the CWKC indicates that the peak values of water consumption per capita of these countries occurred before the sample period, which may be due to the specific socioeconomic characteristics of these countries. For example, all three countries are well known for strong environmental awareness among their citizens. This early peak might be due to their earlier adoption of environmental policies. The CWKCs for Denmark, Greece, Luxembourg, New Zealand and Sweden indicate stable or slight increases in per capita water consumption at the end of the sample period. This change might be caused by aging populations and shrinking household sizes in these countries (Birrell et al., 2005). Small households become dominant in these societies, where the amount of water consumed by domestic activities, such as cleaning and gardening, will be shared by fewer people, which increases per capita water consumption in urban areas. However, in the four specifications of the CWKC policy effects regression (column 2–5, Table 4), none of the coefficients of household size and quadratic household size are statistically significant2. This indicates that the relationship among household size, aging population and water consumption per capita is not strong and may only exist in some countries.

In the CWKC policy effects regression, we include different policy variables separately in the model to avoid multicollinearity. In regression (2), which includes the water productivity variable, the coefficient of water productivity is negative and statistically significant at the 1% level, which indicates that an increase in water efficiency is associated with a decrease in per capita urban water consumption. In regression (3), the wastewater treatment rate is included. The coefficient of the wastewater treatment rate is positive, whereas the coefficient of the quadratic wastewater treatment rate is negative. Both are statistically significant at the 1% level. This pattern implies that at the initial stage of policy implementation, per capita urban water consumption will continue to increase. However, if the policy implementation period is sufficiently long, an inflection point will appear and per capita urban water consumption will begin to decrease. In regression (4), the sludge recycling rate is tested as a policy proxy. The coefficient of the sludge recycling rate is positive and statistically significant at the 10% level, whereas the coefficient of the quadratic sludge recycling rate is negative and statistically significant at the 5% level. In regression (5), the coefficient of the public pollution control expenditure on water and wastewater is positive, but the coefficient of the quadratic public pollution control expenditure on water and wastewater is negative. Both are statistically significant at the 10% level. The similar results for the regressions in columns 3, 4 and 5 suggest that only long-term policy interventions lead to decreases in per capita urban water consumption.

Regarding the control variables, the coefficient of fixed capital formation is positive and statistically significant in regression (2), but negative and statistically significant in regression (3) and not significant in other specifications. This implies that the effect of economic scale of a country on per capita urban water consumption is inconclusive. The similar inconsistent results also apply to the variables of secondary and tertiary industries which control for structure of an economy, inward FDI which represents the openness of an economy and the learning effects, triadic patents and total precipitation. The coefficients of triadic patents are negative and statistically significant in all the regressions except for regression (2). The negative and statistically significant coefficients reveal that technological development decreases per capita urban water consumption.

Seen in Figure 2, the inflection point of the CWKC for the three developing countries in our sample, i.e., China, Mexico and Turkey, occurs much later than those of the developed countries, except for Ireland. However, the GDP per capita at the inflection points of these countries is lower than those of the developed countries in the sample (Figure 3). A hypothesis to account for this fact is that the latecomer countries adopted the water-saving technologies that were not available to the currently developed countries when they were industrializing. Therefore, the inflection points of the developing countries appeared earlier in the course of development than those of the developed countries. However, combining the regression results obtained in Table 4, we would like to also highlight the role that policy interventions can play in addressing water efficiency and water and wastewater management and inducing an earlier inflection point in the CWKC in developing countries compared to developed countries. Developing countries should thus implement water-saving policies as they develop.
Fig. 2.

The inflection points of CWKC: actual maximal per capita urban water consumption versus year.

Fig. 2.

The inflection points of CWKC: actual maximal per capita urban water consumption versus year.

Fig. 3.

The inflection points of CWKC: actual maximal per capita urban water consumption versus GDP per capita.

Fig. 3.

The inflection points of CWKC: actual maximal per capita urban water consumption versus GDP per capita.

Conclusion

In this paper, we propose the existence of a CWKC and test its existence in urban areas utilizing panel data collected from 27 countries from 1960 to 2010. The relationship between urban water consumption per capita and GDP per capita, i.e., the CWKC, exhibits an N-shape. In other words, per capita urban water consumption increases during the initial stages of development but decreases after GDP per capita reaches a threshold. When GDP per capita reaches an even higher level, per capita urban water consumption might rise slightly again, a pattern we observe in several countries in our sample.

Following the EKC literature, we regress per capita urban water consumption on the factors representing national economic scale and structure, technological improvement, water policy intervention and climate conditions. We believe that the urban water consumption pattern (the CWKC) is strongly affected by policy interventions. We distinguish between two groups of water policies: indirect water policies, which are implemented for other socioeconomic goals than saving water but positively affect water efficiency, and direct water policies, which aim to improve water management. An inverted U-shaped relationship exists between direct water policy and per capita urban water consumption, which implies that governments can make continuous and long-term efforts to decrease per capita water consumption in urban areas.

With rapid economic growth, population expansion and accelerated urbanization, many developing countries face water shortage and environmental deterioration in their urban areas, which can severely affect the economic development and quality of life of their citizens. However, we observe that the per capita GDP values at the inflection points of the CWKCs for China, Mexico and Turkey are lower than those of more developed countries. This implies that long-term water policies that are properly designed and rigorously implemented can mitigate the stress of water demand in urban areas. These developing countries can then reach an earlier inflection point in the CWKC, thus improving the balance between economic development and environment preservation.

1

The 26 OECD countries included in this study are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Japan, Korea, Luxembourg, Mexico, the Netherlands, Norway, New Zealand, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the USA. The OECD countries excluded from this study include Czech Republic, Hungary, Poland, Slovakia, Chile, Slovenia, Estonia and Israel.

2

If we remove the variable quadratic term of household size and keep the variable household size only, the coefficient of household size is not statistically significant either.

References

References
Adeloye
A.
Nawaz
N.
Montaseri
M.
(
1999
).
Climate change water resources planning impacts incorporating reservoir surface net evaporation fluxes: a case study
.
International Journal of Water Resources Development
15
,
561
581
.
Andreoni
J.
Levinson
A.
(
2001
).
The simple analytics of the environmental Kuznets curve
.
Journal of Public Economics
80
,
269
286
.
Antle
J. M.
Heidebrink
G.
(
1995
).
Environment and development: theory and international evidence
.
Economic Development and Cultural Change
43
,
603
625
.
Arrow
K.
Bolin
B.
Costanza
R.
Dasgupta
P.
Folke
C.
Holling
C. S.
Pimentel
D.
(
1995
).
Economic growth, carrying capacity, and the environment
.
Ecological Economics
15
,
91
95
.
Ayres
R. U.
(
1997
).
The Kuznets curve and the life cycle analogy
.
Structure Change and Economic Dynamics
8
,
413
426
.
Baldwin
R.
(
1995
).
Does sustainability require growth?
In:
The Economics of Sustainable Development
.
Winters
L. A.
(ed.).
Cambridge University Press
,
Cambridge
,
UK
, pp.
51
78
.
Barbier
E. B.
(
1997
).
Introduction to the environmental Kuznets curve special issue
.
Environment and Development Economics
2
,
369
381
.
Birrell
B.
Rapson
V.
Smith
T. F.
(
2005
).
Impact of Demographic Change and Urban Consolidation of Domestic Water Use
.
Water Services Association of Australia
,
Melbourne
,
Australia
.
Carr
J. E.
Chase
E. B.
Paulson
R. W.
Moody
D. W.
(
1990
).
National water summary 1987
.
USGS Water Supply Paper, 2350
,
USGS
,
Reston, Virginia
,
USA
.
Cole
M. A.
(
2004
).
Economic growth and water use
.
Applied Economics Letters
11
,
1
4
.
Cole
M. A.
Rayner
A. J.
Bates
J. M.
(
1997
).
The environmental Kuznets curve: an empirical analysis
.
Environment and Development Economics
2
,
401
416
.
Cropper
M.
Griffiths
C.
(
1994
).
The interaction of population growth and environmental quality
.
The American Economic Review
84
,
250
254
.
Dasgupta
S.
Laplante
B.
Wang
H.
Wheeler
D.
(
2002
).
Confronting the environmental Kuznets curve
.
Journal of Economic Perspectives
16
,
147
168
.
Dinda
S.
(
2004
).
Environmental Kuznets curve hypothesis: a survey
.
Ecological Economics
49
,
431
455
.
Ehrhardt-Martinez
K.
Crenshaw
E. M.
Jenkins
J. C.
(
2002
).
Deforestation and the environmental Kuznets curve: a cross-national investigation of intervening mechanisms
.
Social Science Quarterly
83
,
226
243
.
Gato
S.
Jayasuriya
N.
Roberts
P.
(
2007
).
Forecasting residential water demand: case study
.
Journal of Water Resources Planning and Management
133
,
309
319
.
Gleick
P. H.
(
2003
).
Water use
.
Annual Review of Environment and Resources
28
,
275
314
.
Grossman
G. M.
Krueger
A. B.
(
1991
).
Environmental Impacts of a North American Free Trade Agreement
.
NBER working paper no. w3914
.
National Bureau of Economic Research
,
Cambridge, MA
.
Grossman
G. M.
Krueger
A. B.
(
1995
).
Economic growth and the environment
.
Quarterly Journal of Economics
110
,
353
377
.
Harbaugh
W. T.
Levinson
A.
Wilson
D. M.
(
2002
).
Reexamining the empirical evidence for an environmental Kuznets curve
.
Review of Economics and Statistics
84
,
541
551
.
Hettige
H.
Mani
M.
Wheeler
D.
(
2000
).
Industrial pollution in economic development: the environmental Kuznets curve revisited
.
Journal of Development Economics
62
,
445
476
.
Hoehn
J.
Adanu
K.
(
2008
).
Do growth, investment, and trade encourage water use or water conservation?
Letters in Spatial and Resource Sciences
1
,
127
146
.
Horvath
R. J.
(
1994
).
National development paths 1965–1987 – Measuring a metaphor
.
Environment and Planning A
26
,
285
305
.
Infrastructure Development Institute
(
1997
).
Drought Conciliation and Water Rights-Japanese Experience
.
Infrastructure Development Institute
,
Tokyo
,
Japan
.
International Hydrological Programme
(
1999
).
Information about World Water Use and Availability
.
United Nations Educational, Scientific, and Cultural Organization
,
Paris, France
, .
Jia
S. F.
Yang
H.
Zhang
S. F.
Wang
L.
Xia
J.
(
2006
).
Industrial water use Kuznets curve: evidence from industrialized countries and implications for developing countries
.
Journal of Water Resources Planning and Management
132
,
183
191
.
Koop
G.
Tole
L.
(
1999
).
Is there an environmental Kuznets curve for deforestation?
Journal of Development Economics
58
,
231
244
.
Lieb
C. M.
(
2003
).
The Environmental Kuznets Curve – A Survey of the Empirical Evidence and of Possible Causes
.
Discussion Paper Series No. 391
,
Department of Economics, University of Heidelberg
,
Heidelberg, Germany
.
Lieb
C. M.
(
2004
).
Possible Causes of the Environmental Kuznets Curve: A Theoretical Analysis (Vol. 3040)
.
Peter Lang AG, European Academic Publishers
,
Bern
,
Switzerland
.
List
J. A.
Gallet
C. A.
(
1999
).
The environmental Kuznets curve: does one size fit all?
Ecological Economics
31
,
409
423
.
Magnani
E.
(
2001
).
The environmental Kuznets curve: development path or policy result?
Environmental Modelling & Software
16
,
157
165
.
Managi
S.
(
2006
).
Pollution, natural resource and economic growth: an econometric analysis
.
International Journal of Global Environmental Issues
6
,
73
88
.
Panayotou
T.
(
1993
).
Empirical Tests and Policy Analysis of Environmental Degradation at Different Stages of Economic Development
.
Working paper WP238
.
International Labor Office
,
Geneva
,
Switzerland
.
Perman
R.
Stern
D. I.
(
2003
).
Evidence from panel unit root and cointegration tests that the environmental Kuznets curve does not exist
.
Australian Journal of Agricultural and Resource Economics
47
,
325
347
.
Rock
M. T.
(
1998
).
Freshwater use, freshwater scarcity, and socioeconomic development
.
The Journal of Environment & Development
7
,
278
301
.
Rock
M. T.
(
2000
).
The dewatering of economic growth
.
Journal of Industrial Ecology
4
,
57
73
.
Selden
T. M.
Song
D.
(
1994
).
Environmental quality and development: is there a Kuznets curve for air pollution emissions?
Journal of Environmental Economics and Management
27
,
147
162
.
Shafik
N.
(
1994
).
Economic development and environmental quality – an econometric analysis
.
Oxford Economic Papers New Series
46
,
757
773
.
Shafik
N.
Bandyopadhyay
S.
(
1992
).
Economic Growth and Environmental Quality: Time-Series and Cross-Country Evidence
.
World development report working paper WPS 904
.
The World Bank
,
Washington, DC
.
Shiklomanov
I. A.
(
2000
).
Appraisal and assessment of world water resources
.
Water International
25
,
11
32
.
Solley
W. B.
Pierce
R. R.
Perlman
H. A.
(
1998
).
Estimated Use of Water in the United States in 1995
.
USGS
,
Washington, DC
.
Stern
D. I.
(
2004
).
The rise and fall of the environmental Kuznets curve
.
World Development
32
,
1419
1439
.
van Pottelsberghe
B.
Mejer
M.
(
2010
).
The London agreement and the cost of patenting in Europe
.
European Journal of Law and Economics
29
,
211
237
.
Vincent
J. R.
(
1997
).
Testing for environmental Kuznets curves within a developing country
.
Environment and Development Economics
2
,
417
431
.
Vorosmarty
C. J.
Green
P.
Salisbury
J.
Lammers
R. B.
(
2000
).
Global water resources: vulnerability from climate change and population growth
.
Science
289
,
284
288
.
World Bank
(
1992
).
World Development Report 1992: Development and Environment
.
Oxford University Press
,
Oxford
,
UK
.
World Meteorological Organization
(
1998
).
World water day (1997)
.
Retrieved from http://www.wmo.ch
(Accessed March 2 2011)
.
Yandle
B.
Vijayaraghavan
M.
Bhattarai
M.
(
2002
).
The Environmental Kuznets Curve: A Primer
.
Property and Environment Research Center
,
Bozeman, MT
.
Yandle
B.
Bhattarai
M.
Vijayaraghavan
M.
(
2004
).
Environmental Kuznets Curves: A Review of Findings, Methods, and Policy Implications
.
Property and Environment Research Center
,
Bozeman, MT
.

Supplementary data