ECONOMIC GROWTH AND ENVIRONMENTAL DEGRADATION

An Empirical Study on Pollution in Sichuan Province

Yunxi Hu

8586728

Major Paper presented to the

Department of Economics of the University of Ottawa
in partial fulfillment of the requirements of the M.A. Degree
Supervisor: Professor Nguyen Van. Quyen

Ottawa, Ontario
December 2017
ABSTRACT

The paper is an empirical study on the relationship between economic growth and

pollution, using the data from the province of Sichuan, China. Three pollutants –

industrial wastewater, industrial solid wastes, and industrial gas waste – are considered.

The results – obtained from the Johansen co-integration test and the Granger causality

test – indicate that both industrial wastewater and industrial solid waste have a negative

impact on economic growth, but not industrial gas waste.

1
 1. INTRODUCTION

In an economy, the production uses energy, labor, capital, and raw materials to produce

goods and services. The environment plays an important role in the economy. In the

production of goods and services, raw materials, such as metals, minerals, water, tree,

etc. are extracted from the environment. Energy resources – coal, oil, natural gas – are

also extracted from the environment required to transform raw materials into goods and

services – intermediate as well as final consumption goods. In addition, the environment

acts to provide many ecosystem services, such as filtering air pollution, sequestering

carbon, and protect against flood risk. It also provides us with recreational activities and

improves our health. Economic growth is an important driver of the economy. It

stimulates consumption and advances production. It makes improvements in health,

education, and life quality possible. It is also vital for the wellbeing of the economy and

its citizens in both developed and developing world. (Richard Price, 2010, p.12).

According to the first law of thermodynamics, matter can never be destroyed or created:

the materials and energy that do not go into the final product are returned to the

environment as wastes. The wastes can also come from consumption. When people

drive cars, the exhaust from the cars adds to the stock of greenhouse gases in the

atmosphere. The environment thus acts as a supplier of resources and a receptor of

waste products.

With the development of society and economic growth, the attendant environment

pollution has become a serious issue. There are those who argue that Industrialization

2
and urbanization eventually leads to environmental degradation (Peng, 2006), while

others contend that economic growth and environment quality interact with each other

and will be promoting each other mutually (Beckerman, 1992). No clear-cut conclusion

has emerged from the debate, and the issue still needs to be studied further.

There is no denying that economic growth brings many benefits. However, the ever-

increasing scale of production also degrades ecosystems and depletes natural

resources. The debate is whether or not it is possible to achieve sustainable economic

growth without the degradation of the ecosystems and the depletion of the natural

resources. As a result of human activities, the level of carbon dioxide (𝐶𝑂2 ) in the

atmosphere has being rising dramatically, and the world is facing the great challenge of

keeping the rise in the global temperature below two degrees. Researchers have found

that 15 out of 24 ecosystems were used unsustainably, and the consumption of natural

resources such as metals and minerals was rising at an increasing rate (Millennium

Ecosystem Assessment, 2003). Some take the view that the cost of inaction would be

far greater than the cost of action now, and using environmental resources sustainably

is consistent with the economic growth. Others think that the stocks of natural resources

are finite, and this places a limit on the extent to which the economy could expand.

There is a large body of empirical studies on the relationship between income and some

measures of environmental qualities. The regression model used in the empirical

studies typically link GDP per capita – the dependent variable – to three independent

variables – industrial waste water, industrial solid waste, and industrial gas waste –

3
which represent the environmental quality. The World Bank Development Report (1992)

relates the percentage of the population that has no access to safe water or urban

sanitation to the level of income by using internal World Bank data and carries out an

analysis on the relationship between income and municipal solid waste per capita, using

OECD and World Resources Institute estimates for 39 countries in 1985. Seldon (1992)

uses a similar method to investigate the relationship between the estimated rate of

emission of several air pollutants with the country’s national income. These studies tend

to find an inverted “U” relationship between environmental quality and level of income.

This paper is an empirical study on the relationship between pollution and economic

growth for Sichuan, a province of China. The main employs reliable data and a common

methodology to investigate the relationship between the scale of economic activities

and environmental quality for a broad set of environmental indicators. We use the

available data from the “Statistical Yearbook of Sichuan Province”, 1 which tracks

industrial wastes, in order to include dimensions of environmental quality of which actual

measurements have been taken by comparable methods in other analyses. Though

these measures are far from comprehensive, they at least represent the largest part of

the relevant variables as industrial production process cause the majority of

environmental pollution. The main results that come out of our analysis is that indicators

of environmental quality no longer exhibit an inverted “U” relationship with GDP per

capita: industrial water waste and industrial solid waste show an “N” relationship with

1
Source: http://www.sc.stats.gov.cn/tjcbw/tjnj/

4
GDP per capita, and industrial gas waste exhibits no linear or nonlinear relationship with

GDP per capita.

The paper is organized as follows. In Section 2, the most prominent hypothesis linking

environmental degradation an economic growth – the environmental Kuznets curve

(EKC) – is presented. Most empirical studies on the linkages between economic growth

and environment are dedicated to confirming or rejecting the EKC. Section 3 is a short

review of the empirical literature on the relationships between pollutants and economic

growth. A theoretical model of production and pollution is presented in Section 4. The

econometric model is presented in Section 5. The results of the Johansen co-integration

test as well as the Granger causality test are presented in Section 6. In this section, the

main results of the paper are discussed. Section 7 contains some concluding remarks.

The data and the results of some tests are relegated to several appendices.

2. THE ENVIRONMENTAL KUZNETS CURVE

In the 1950s, the American economist Kuznets (1955) examined the relationship

between the economic growth (level of income per capita) and income inequality, and

drew the conclusion that as the economy embarks on its development path, income

inequality would firstly widen, and then gradually narrows down after the economy

reaches a certain stage. That is, income inequality exhibits an inverted U shape, when

plotted against income.

Forty years later, another American economist, Grossman and Krueger (1991)

suggested an inverted U-shape relationship between GDP per capita and many

5
pollutants by analyzing panel data from 42 countries in the world. The intuitive logic

behind the EKC can be explained as follows. At low levels of income, the limited

income is devoted to meeting the basic needs of the population. As income rises, and

basic needs have been fulfilled, the quality of the environment – a particular type of

good – might yield more utility than an additional unit of consumption, and resources will

be deemed to be better used in pollution abatement than in consumption, with the

ensuing consequence that environmental quality now rise with economic growth.

The environmental Kuznets curve (EKC) – named after Kuznets, but not developed by

this researcher – hypothesizes an inverted U-shaped relationship between economic

development and environment quality. The hypothesis asserts that pollution rises in the

early stages of economic development, but the trend is reversed after GDP per capital

surpasses a certain level. Figure 1 depicts the EKC, when the level of environmental

degradation is plotted against GDP per capita.

6
 Figure 1.-Environmental Kuznets Curve (inverted “U” shape)

3. LITERATURE REVIEW

Most empirical studies on the linkages between environment and economic growth

attempt to test the EKC hypothesis.

Shafik and Bandyopadhyay (1992) discovered that emissions of sulfur dioxide (𝑆𝑂2 )

originally increased and then reduced as income per capital rises, confirming the EKC

hypothesis. Panayotou (1993), who conducted an analysis on forest deforestation and

rising income, found an inverted U-shape relationship between the two variables.

Seldon and Song (1994) confirmed the existence of the EKC by analyzing the

relationship between 𝐶𝑂2, 𝑆𝑂2, 𝑁𝑂2 , 𝐶𝑂, and suspended particular matter. Guo and Li

7
(2010) in conducted an empirical study on the linkages between pollutants and

economic growth, using panel data on 29 provinces of China, and concluded that

industrial solid waste (ISW) and GDP per capita were in line with the EKC.

Not all the empirical studies on the linkages between economic growth and environment

support the EKC hypothesis. Zhou, Yuan, and Xue (2009) carried out a study on Shanxi

province, using the industrial wastewater and GDP per capita, and the results they

obtained showed that the two variables exhibit a stable long-run N-shape relationship.

The same results were found by Ding (2012), who used data in on Guangxi Province.

Wang (2013), in a study on a costal area economy and marine environmental pollution,

found an N-shaped curve. Figure 2 depicts the N-shaped relationship between per

capital income and environmental degradation.

Figure 2.-Environmental Kuznets curve (“N” shape)

8
Gene and Alan (1995) used environmental indicators, such as urban air pollution, the

oxygen regime state in river, fecal contamination of river, and contamination of river by

heavy metals, to represent environmental quality in studying the linkages between

economic growth and the environment. They found that environmental quality does not

steadily deteriorate with economic growth. Furthermore, economic growth first brings

deterioration to most indicators, and then improvement after that. The turning points of

different indicators vary, but in general, they are around the GDP per capita of $8000;

see Grossman and Krueger (1995).

Shaik Feroz (2014) formulated a VAR model and used the variance decomposition

method to investigate the long-term relationships between environmental pollution

indicators and GDP per capita in western China from 1992 to 2010. This researcher

concludes that economic growth is a major cause of environmental pollution and that

there is also a reverse effect. However, the reverse effect is not immediate, but has a

lag. In addition, the variance decomposition shows that economic growth has little

explanatory power for the different kinds of the environmental pollution indicators. The

paper shows that the original EKC hypothesis of EKC cannot be generalized to the

global level since the empirical evidence only shows the results for developed countries.

What is even more striking is that there is no correlation between higher levels of

income and lower gas emissions (including six atmospheric pollutants in Spain), except

for 𝑆𝑂2 . The author argues that the relationship between income level and types of

emissions may depend on many factors.

9
In a paper entitled “Economic Growth and Environmental Pollution in Myanmar: An

Analysis of Environmental Kuznets Curve,” Aung et. Al (2007) tested the validity of the

EKC, and investigated the short-run as well as the long-run relationships between

environmental pollution indicators and economic indicators. The environmental pollution

indicators are 𝐶𝑂2 , 𝐶𝐻4 , 𝑁2 𝑂, and the economic growth indicators are GDP. The authors

also considered other proxies, such as trade intensity and financial openness,

urbanization and structural breaks in the country. The authors used annual data in

Myanmar over the period 1970-2014, and carried out the analysis using an

autoregressive distributed lag (ARDL) model. The empirical results show an absence of

the EKC for 𝐶𝑂2 and GDP, while evidence of the EKC is found for 𝐶𝐻4 and 𝑁2 𝑂. The

authors finally concluded that financial openness and trade liberation will contribute to

the environment improvement in the long-run.

Omisakin (2009) argued that environment pollution is inevitable in developing countries.

According to this author, the developing countries are at an early stage on the EKC

because developing countries desire economic growth and tend to consume cheaper

sources of energy. Massive infrastructures, such as roads and rail tracks, are in heavy

need to promote industrialization. These activities at the take-off stage of economic

development are substantially energy intensive. And the concluded that compelling

developing countries to pursue environment goals requires substantial financial and

technological support from developed countries to compensate for the losses

associated with pollution reduction.

10
George E. Halkos (2003), in a paper entitled “Growth and Environmental Pollution:

Empirical Evidence from China,” investigated the relationship between 𝐶𝑂2 and GDP

per capita by using time series data from 1960 to 2006, with additional variables in trade

and agriculture, industry and services sectors. The results indicate an inverted “U”

shape relationship between 𝐶𝑂2 and GDP per capita and that trade is also a

determinant of pollution.

Luo et al. (2014), in a research paper entitled ”Relationship Between Air Pollutants and

Economic Development of the Provincial Capital Cities in China during the Past

Decade,” chose GRP per capita (Gross Regional Product per capita) as economic

indicators, and 𝑃𝑀10 , 𝑆𝑂2 , 𝑁𝑂2 and the air pollution index (API) as air pollutants

indicators to find out if there is an EKC between the economic indicator and the

pollutant indicators. The analysis was conducted by using the 31 provinces’ data over

the time period 2003-2012. The results of the analysis indicated that three main

industries have a quadratic relationship with 𝑁𝑂2 , but a negative relationship with 𝑃𝑀10

and 𝑆𝑂2 . The API showed an inverted U-shape curve with the GRP per capita which

was the typical type of the EKC.

In a paper entitled “Research on the Relationship of Environmental Pollution and

Economic growth in China,” Zhang Jiansheng and Li Qing (2011) investigated the

relationship of environmental pollution and economic growth, using the data during the

time period 1980-2009 in China. The final results are: industrial wastewater (IWW) and

11
GDP per capita have an “N” shape relationship with GDP per capita, while industrial gas

waste (IGW) and industrial solid waste (ISW) display a “U” shape relationship with GDP

per capita. The authors argued that the economic development has been achieved at

the cost of environmental degradation since the time of China’s reform and opening-up

policy. The authors suggested that China’s government should take measures to deal

with the conflict between environment and economic growth to achieve sustainable

development.

4. A THEORETICAL MODELOF PRODUCTION AND POLLUTION:

THE ENVIRONMENTAL KUZNETS CURVE

It would be common in developed countries, which have already passed the turning

point, to support economic growth and avoid environmental regulations. However, early

implementation of tight environmental regulation in developing world would be more

likely to harm the economy. Here are several reasons whether EKC hypothesis is

effective in policy making. First, EKC analysis is typically based on limited set of

pollutants so the conclusions made by these analyses cannot be applied to all types of

environmental damage. For instance, no clear evidence shows an EKC relationship in

the Ecological Footprint, which is a measure of the pressure that man places on the

environment, unless energy use is removed from the measure (Caviglia-Harris et al,

2009). Furthermore, EKC relationship is strong when impacts are local. When impacts

are global, when income rises, carbon and greenhouse gas emissions increase, even in

the developed countries. Second, the supporting evidence has been found less reliable

than previously thought (Stern, 2004): EKC’s predictions of the relationship between

income and pollution are heavily influenced by the mathematical model adopted

12
(Millimet et al, 2003). Third, the cost of improving environmental quality and repairing

damage after the turning point may possibly be higher than the cost of preventing the

damage or undertaking the mitigation earlier. For example, the cost of avoiding the

pollution of a waterway is much less than the cost of cleanup after the damage has

been done. Fourth, there is evidence showing that similar level of wealth may perform

differently under EKC hypothesis (Magnani, 2000).

An economy produces a consumption good from four inputs: capital (K), labor (L),

energy (E), and materials (M). Following Stokey (1998), we assume that the output of

the consumption good at any time t is given by the following Cobb-Douglas production

function:

(1) 𝑌(𝑡) = 𝐴𝐾(𝑡)𝛼 𝐿(𝑡)1−𝛼 𝑧(𝑡),

where 𝑌(𝑡) is the output at time 𝑡; 𝐴 is the total factor productivity a; 𝐾(𝑡) is the capital

input at time 𝑡; 𝛼, 0 < 𝛼 < 1, is a parameter; 𝐿(𝑡) is the labor input at time 𝑡; and 𝑧(𝑡) is

an index of environmental degradation at time 𝑡.

The index of environmental degradation 𝑧(𝑡) is assumed to lie inside the unit interval

[0,1], with the interpretation that the higher the value of 𝑧(𝑡), the dirtier the production

technology and the higher the output of the consumption good. The production function,

as specified by (1), embodies the assumption that dirtier technology yields higher output.

When 𝑧(𝑡) = 1 , the production technology is the one with the maximum pollutant

emissions. When 𝑧(𝑡) falls below 1, the environment improves with the introduction of

13
cleaner and more advanced technologies, but the output decreases. When 𝑧(𝑡) = 0,

there are no emissions, but output also falls to 0.

We are not interested in population growth, and thus, shall assume that the population

is constant and normalize it to be equal to 1. Under the assumption of constant

population, (1) simplifies to

The production function (1) can be expressed under the per capita form as follows:

(2) 𝑌(𝑡) = 𝐴𝐾(𝑡)𝛼 𝑧(𝑡),

Although consumption also generates pollution, it will be ignored. The model only

consider pollution generated in the production process. More specifically, emissions

from production activities are given by

(3) 𝑥(𝑡) = 𝐴𝐾(𝑡)𝛼 𝑧(𝑡)𝛽 ,

where 𝑥(𝑡) represents the emissions at time 𝑡; and 𝛽 > 1 is a parameter.

The instantaneous utility function of the representative consumer at time t is

𝐶(𝑡)1−𝜎 −1 𝐵
(4) 𝑢(𝑡) = − 𝛾 𝑋(𝑡)𝛾 ,
1−𝜎

where 𝑢(𝑡) denotes the utility of the representative agent; 𝐶(𝑡) is the consumption of the

representative agent; 𝜎 > 0 denotes the parameter of the Arrow-Pratt measure of

relative risk aversion. Also, 𝑋(𝑡) is the stock of pollution stock in the economy at time 𝑡,

and 𝐵 > 0 and 𝛾 > 1 are two parameters, which characterize the disutility from the

pollution stock.

The motion of the stock of pollution is governed by the following differential equation:
𝑑𝑋
(5) = 𝑥(𝑡) − 𝜂𝑋(𝑡)
𝑑𝑡

= 𝐴𝐾(𝑡)𝛼 𝑧(𝑡)𝛽 − 𝜂𝑋(𝑡),

14
where 𝜂 > 0 is a parameter that represents the degree of purification capacity of the

environment.

The motion of the stock is governed by the following differential equation:

𝑑𝐾
(6) = 𝐴𝐾(𝑡)𝛼 𝑧(𝑡) − 𝛿𝐾(𝑡) − 𝐶(𝑡),
𝑑𝑡

where 0 < 𝛿 < 1 is the capital depreciation rate.

Under the above setup, a social planner solves the following maximization problem:

Choose a time path for the production technology 𝑡 → 𝑧(𝑡), 0 ≤ 𝑧(𝑡) ≤ 1, 𝑡 ≥ 0, and a

time path for the consumption of the representative agent 𝑡 → 𝐶(𝑡), 𝑡 ≥ 0, to maximize

∞ 𝐶(𝑡)1−𝜎 −1 𝐵
(7) ∫0 𝑒 −𝜌𝑡 ( 1−𝜎
− 𝛾 𝑋(𝑡)𝛾 ) 𝑑𝑡

subject to
𝑑𝐾
(6) = 𝐴𝐾(𝑡)𝛼 𝑧(𝑡) − 𝛿𝐾(𝑡) − 𝐶(𝑡),
𝑑𝑡

𝑑𝑋
(5) = 𝐴𝐾(𝑡)𝛼 𝑧(𝑡)𝛽 − 𝜂𝑋(𝑡),
𝑑𝑡

(8) 𝐾(0) = 𝐾0 , 𝑋(0) = 𝑋0 are given.

In (7), 𝜌 > 0 is the social discount rate.

To solve the above maximization problem, first write the Hamiltonian

𝐶 1−𝜎 −1 𝐵
ℋ(𝐾, 𝑋, 𝑧, 𝐶, 𝜆, 𝜇, 𝑡) = 𝑒 𝜌𝑡 ( − 𝛾 𝑋 𝛾 ) + 𝜆(𝐴𝐾 𝛼 𝑧 − 𝛿𝐾 − 𝐶)
1−𝜎

+𝜇(𝐴𝐾 𝛼 𝑧 𝛽 − 𝜂𝑋).

Let 𝑧(𝑡) denote the optimal production technology at time t and 𝐶(𝑡) the optimal

consumption also at time t. The capital stock and the stock of pollution at time t under

the optimal control 𝑡 → (𝑧(𝑡), 𝐶(𝑡)), 𝑡 ≥ 0, are denoted, respectively, by 𝐾(𝑡) and 𝑋(𝑡).

15
Also, let 𝜆(𝑡) and 𝜇(𝑡) denote, respectively, the shadow price of capital and the shadow

price of pollution both at time t along the optimal trajectory.

The optimal consumption at time t satisfies the following first-order condition

𝜕ℋ(𝐾(𝑡),𝑋(𝑡),𝑧(𝑡),𝐶,𝜆(𝑡),𝜇(𝑡),𝑡)
(9) = 𝑒 −𝜌𝑡 𝐶 −𝜎 − 𝜆(𝑡) = 0,
𝜕𝐶

from which we obtain

1
(10) 𝐶(𝑡) = (𝑒 𝜌𝑡 𝜆(𝑡))−𝜎 .

The following first-order condition characterizes the optimal production technology at

time t.

𝜕ℋ(𝐾(𝑡),𝑋(𝑡),𝑧(𝑡),𝐶(𝑡),𝜆(𝑡),𝜇(𝑡),𝑡)
(11) = 𝜆(𝑡)𝐴𝐾 𝛼 + 𝜇(𝑡)𝛽𝐴𝐾 𝛼 𝑧(𝑡)𝛽−1 ≤ 0,
𝜕𝑧

with equality holding if 𝑧(𝑡) < 1.

It is clear from (11) that

(12) 𝑧(𝑡) = 1 if 𝜆(𝑡) + 𝜇(𝑡)𝛽 ≥ 0.

When 𝜆(𝑡) + 𝜇(𝑡)𝛽 < 0, we have

1
𝜆(𝑡) 𝛽−1
(13) 𝑧(𝑡) = (− 𝜇(𝑡)𝛽) .

The motion of the shadow price of capital is governed by the following adjoint equation:

𝑑𝜆 𝜕ℋ(𝐾(𝑡),𝑋(𝑡),𝑧(𝑡),𝐶(𝑡),𝜆(𝑡),𝜇(𝑡),𝑡)
(14) =− = −𝜆(𝑡)(𝛼𝐴𝐾(𝑡)𝛼−1 𝑧(𝑡) − 𝛿)
𝑑𝑡 𝜕𝐾

−𝜇(𝑡)(𝐴𝛼𝐴𝐾(𝑡)𝛼−1 𝑧(𝑡)𝛽 ).

Letting 𝜆(𝑡) = 𝑒 𝜚𝑡 𝜆(𝑡) denote the current shadow price of capital at time t along the

optimal trajectory, we have

1 𝑑𝜆 𝜇(𝑡)
(15) = 𝜚 − (𝛼𝐴𝐾(𝑡)𝛼−1 𝑧(𝑡) − 𝛿) − 𝜆(𝑡) (𝛼𝐴𝐾(𝑡)𝛼−1 𝑧(𝑡)𝛽 )
𝜆(𝑡) 𝑑𝑡

𝛼𝑌(𝑡) 𝜇(𝑡) 𝛼𝑌(𝑡)

=𝜚+𝛿− − 𝜆(𝑡) ( 𝐾(𝑡) 𝑧(𝑡)𝛽−1 ).
𝐾(𝑡)

16
When 𝑧(𝑡) = 1, (15) is reduced to

1 𝑑𝜆 𝛼𝑌(𝑡) 𝜇(𝑡)
(16) =𝜚+𝛿− (1 + 𝜆(𝑡) ).
𝜆(𝑡) 𝑑𝑡 𝐾(𝑡)

When 𝑧(𝑡) < 1, (15) is reduced to

1 𝑑𝜆 𝛼𝑌(𝑡) 1
(17) =𝜚+𝛿− (1 + 𝛽).
𝜆(𝑡) 𝑑𝑡 𝐾(𝑡)

The motion of the shadow price of pollution is governed by the following differential

equation:

𝑑𝜇 𝜕ℋ(𝐾(𝑡),𝑋(𝑡),𝑧(𝑡),𝐶(𝑡),𝜆(𝑡),𝜇(𝑡),𝑡)
(18) =− = 𝑒 −𝜚𝑡 𝐵𝑋(𝑡)𝛾−1 + 𝜇(𝑡)𝜂.
𝑑𝑡 𝜕𝑋

Letting 𝜇(𝑡) = 𝑒 𝜚𝑡 𝜇(𝑡) denote the current shadow price of pollution at time t along the

optimal trajectory, we have

1 𝑑𝜇 𝐵𝑋(𝑡)𝛾−1
(19) =𝜌+𝜂+ .
𝜇(𝑡) 𝑑𝑡 𝜇(𝑡)

Suppose that the initial value of capital, 𝐾0 , is much smaller than its steady state value.

Then at time 𝑡 = 0 the shadow price of capital 𝜆(0) is sufficiently larger than the shadow

price 𝜇(0) of pollution, and 𝑧(0) = 1. If 𝑡 is small, 𝑧(𝑡) = 1, and 𝑌(𝑡), 𝐾(𝑡), 𝑥(𝑡) all rise as

𝑡 rises above 0. However, as capital accumulation progresses, 𝜆(𝑡) decreases and 𝜇(𝑡)

increases. Eventually, 𝜇(𝑡) exceeds 𝜆(𝑡) and then 𝑧(𝑡) < 1. After that, 𝑧(𝑡) decreases

over time, and the EKC shows a downward sloping curve. For a more detailed analysis,

we refer the reader to Uchiyama (2016).

In summary, although pollution emissions increase with an increase in production in a

country’s early stage of economic development, pro-environmental production

technology will be introduced, and pollution emissions will decrease after economic

development has reached a certain level. Moreover, both the pollution flow and the

17
pollution stock decrease in the convergence to the steady state. Figure 3 depicts the

EKC that comes out of the theoretical model of Stokey.

Figure 3.- The Environmental Kuznets curve according to the Stokey model

5. THE ECONOMETRIC MODEL

In most empirical studies, the concentration and emission of pollution are two basic

variables used in analysis. Because industrial wastes are the main sources of pollution,

we shall choose use industrial wastewater (unit: 109 t), industrial solid waste (unit:104 t),

and industrial waste gas (unit: 104 𝑚3 ) as pollution indicators. In our model, waste gas

includes sulfur dioxide and smoke and dust emissions. Also, we use GDP per capita

(unit: RMB) to reflect economic growth. The data, which span the years from 1995 to

2013, are from the “Statistical Yearbook of Sichuan Province”2.

In order to avoid the problems of multi-collinearity and heteroscedasticity, we use the

log form of each variable. Furthermore, it is much likely that we get stationary series

2
Source: http://www.sc.stats.gov.cn/tjcbw/tjnj/

18
after transformation with the characteristics remain unchanged. We use LNIWW to

denote industrial wastewater, LNISW to represent industrial solid waste, LNIGW to

represent industrial gas waste, and LNGDP to represent GDP per capita, respectively.

The data analysis is carried out using Eviews6.0.

The econometric equation we estimate is

𝐺𝐷𝑃 = 𝛽0 + 𝛽1 𝐼𝑆𝑊 + 𝛽2 𝐼𝐺𝑊 + 𝛽3 𝐼𝑊𝑊 + 𝜀,

where ISW represents industrial solid waste (unit:104 𝑡); IGW represents industrial gas

waste (unit: 104 𝑚3 ); and IWW represents industrial wastewater (unit: 109 t).

6. THE RESULTS

6.1. The Unit Root Test

Prior to the co-integration test for a relationship between environmental pollution and

economic growth, we conduct the unit root test for the stability of the time series

variables. The results are presented in Table 1.

Table 1
Results of unit roots test

Variables ADF test data Test type(c,t,p) 5% critical value Conclusions

LNIWW -2.596334 C,T,1 -3.039621 UNSTABLE
DLNIWW -6.63674 c,t,2 -3.081002 STABLE
LNISW -2.429445 C,T,0 -3.051963 UNSTABLE
DLNISW -5.376162 c,t,1 -3.065585 STABLE
LNIGW -1.429445 C,T,0 -3.051963 UNSTABLE
DLNIGW -4.186637 c,t,1 -3.065585 STABLE
LNGDP -1.959071 C,T,1 -3.039621 UNSTABLE
DLNGDP -6.695637 c,t,2 -3.081002 STABLE

Source: author’s own calculations.

Notes: c means intercept, t means trend term, p means hysteresis orders which is determined by SIC information
principles, D means first order difference.

19
At the plain level, we cannot reject the null hypothesis that the time series for the four

variables are stable, i.e., none of them are stable. As a result, we take the first order

difference, according to the Table 1, all of the variables reject the null hypothesis under

5% significant level, which means they are stable under 5% significant level. After taking

the first order difference and we can then carry out the do co-integration.

The results of the co-integration test are presented in Table 2.

Table 2
Result of the co-integration test

Variables Trace statistics 5% critical Maximum 5% critical

included value eigenvalues value
LNGDP & 16.42986 15.49471 16.13027 14.26460
LNIGW 4.299581 3.841466 4.299581 3.841466
LNGDP & 18.77500 15.49471 18.58667 14.26460
LNISW 4.188329 3.841466 4.188329 3.841466
LNGDP & LN 17.12099 15.4971 17.64794 14.26460
IWW 5.473054 3.841466 5.473054 3.841466

Source: author’s own calculation.

From table 2, it is easy to find that there exists co-integration relationships between

LNIWW, LNISW, LNIGW and LNGDP, which means that the growth of GDP contributes

to the decline of industrial waste water, industrial solid waste and industrial gas waste.

6.2. The Environmental Kuznets Curve Analysis

This paper carries out a regression on the variables industrial waste water, industrial

gas waste, industrial solid waste, and GDP per capita. There are two basic models: one

is time series and the other is panel data. In this paper, we use the time series model,

20
which was proposed by Grossman and Krueger (1995). The equation to be estimated

has the following functional form:

𝑦𝑡 = 𝛽0 + 𝛽1 𝑥𝑡 + 𝛽2 𝑥𝑡2 + 𝛽3 𝑥𝑡3 + 𝜀𝑡 ,

where 𝑦𝑡 represents pollution emission indicators, and 𝑥𝑡 represents GDP per capita.

Depending on the signs of the coefficients on the RHS of the above equation, one of the

following situations might arise.

 When 𝛽1 = 𝛽2 = 𝛽3 = 0, the equation exhibits a straight line, i.e. economic

growth has no impact on environment;

 When 𝛽1 > 0, 𝛽2 = 𝛽3 > 0, the equation exhibits monotonic increase, i.e.

environment deteriorates with economic growth;

 When 𝛽1 < 0, 𝛽2 = 𝛽3 > 0, the equation exhibits monotonic increase, i.e.

environment improves with economic growth;

 When 𝛽1 > 0, 𝛽2 > 0, 𝛽3 > 0, the equation exhibits an “N” shape curve, i.e.

environment firstly deteriorates, and gradually improves and then deteriorates

with economic growth;

 When 𝛽1 < 0, 𝛽2 > 0, 𝛽3 < 0, the equation exhibits a inverted “N” shape curve, i.e.

environment firstly improves, and gradually deteriorates and then improves with

economic growth;

 When 𝛽1 > 0, 𝛽2 < 0, 𝛽3 = 0, the equation exhibits a inverted “U” shape curve, i.e.

the typical EKC, environment deteriorates before improves with economic growth;

21
  When 𝛽1 < 0, 𝛽2 < 0, 𝛽3 = 0 , the equation exhibits a “U” shape curve, i.e.

environment improves before deteriorates with economic growth;

 When 𝛽1 ≠ 0, 𝛽2 = 𝛽3 = 0, the equation exhibits linear relationship.

Table 3 presents the estimation results for the above equation.

Table 3
Result of EKC model

Pollution 𝛽0 𝛽1 𝛽2 𝛽3
emission
index
LNIGW - 3.456722 - 2.001231
2.984536 (0.0112) 0.991267 (0.0612)
(0.1267) (0.4120)
LNISW - 6.823411 - 0.761133
3.564394 (0.0034) 1.762322 (0.0137)
(0.0348) (0.0003)
LNIWW - 2.110036 - 1.440015
4.795647 (0.0423) 1.336168 (0.0069)
(0.0955) (0.0046)

Source: author’s own calculation.

As can be seen from Table 3, at 10% level of significance, 𝛽1 , 𝛽2 , 𝛽3 are significantly

different from zero with 𝛽1 > 0, 𝛽2 < 0, 𝛽3 > 0 ; we know that LNGDP and LNIWW

exhibits an “N” relationship. At 5% significant level, 𝛽1 , 𝛽2 , 𝛽3 are significantly different

from zero and 𝛽1 > 0, 𝛽2 < 0, 𝛽3 > 0; we know that LNGDP and LNISW exhibits an “N”

relationship.

22
Also, the p-value is greater than 0.1 for some of the coefficients, and we cannot

conclude that the coefficients are significantly different from zero: there is no linear or

non-linear relationship between LNGDP and LNISW.

To summarize the analysis, LN GDP and LNIWW exhibits an “N” relationship instead of

inverted “U” relationship, LNGDP and LNISW exhibits an “N” relationship instead of

inverted “U” relationship, LNGDP and LNIGW exhibits no linear or non-linear

relationship which is different from the conclusions made by other authors.

6.3. The Granger Causality Test

According to the co-integration test, there exist co-integration relationships between

LNIWW, LNISW, LNIGW and LNGDP, and this means the growth of GDP contributes to

the decline of industrial wastewater, industrial solid waste, and industrial gas waste.

However, we do not know whether income changes lead to pollution emissions changes

or whether pollution emissions changes result in income changes. We conduct a

Granger causality test to solve this problem. The results of the Granger causality test

are presented in Table 4.

23
 Table 4
Result of Granger test

Null Observations F- Probability Conclusion

hypothesis statistics
Granger cause 19 2.657433 0.657455 accept
of LNIWW or
LNGDP
Granger cause 19 0.657866 0.710990 reject
of LNGDP or
LNIWW
Granger cause 19 4.697853 0.343545 accept
of LNIGW or
LNGDP
Granger cause 19 1.562342 0.223370 accept
of LNGDP or
LNIGW
Granger cause 19 0.997112 0.170064 accept
of LNISW or
LNGDP
Granger cause 19 0.545655 0.768674 reject
of LNGDP or
LNISW
Source: author’s own calculation.

As can be seen from Table 4, Granger causality only exists between LNGDP and

LNIWW, LNISW and LNGDP, but no Granger causality between LNIGW and LNGDP.

Economically speaking, GDP per capita makes a large contribution to the industrial

wastewater and industrial solid waste, but not to industrial gas waste. Furthermore,

changes in environmental pollution make no contribution to the economic growth.

7. CONCLUSION

The goal of this paper is to determine whether the industrial pollution – industrial gas

waste, industrial solid waste and industrial wastewater – will affect economic growth.

The results of the paper show that the increase in industrial solid waste and industrial

24
wastewater would have a negative impact on economic development. Gas waste, on

the other hand, has a positive impact on economic development. However, this last

result does not necessarily mean we should produce more gas waste thinking that this

would have a positive effect on economic growth. In reality, the air quality in Sichuan

province is becoming worse and worse, and this problem has attracted government’s

attention to release new policies in order to improve the environment condition. The first

action to undertake is to phase out and shut down the business activities that are

technology-backward and that are causing heavy pollution. In 2001 and 2004, the

government shut down 30 thousand factories and companies that heavily consume

resources and energy. Efforts have been made to deal with the heavily polluting

industries such as: steel, cement, electricity, aluminum, and iron. The government also

suspended some projects that were not in accordance with the related industry policies.

The second action is to encourage recycling. Factories should make the best use of the

resources and control the waste generated in the process of production. The

government should consider an overall planning of the development of industry and

farming, production and consumption, urban and rural.

25
 Appendix A
Data Sources

YEAR IWW ISW IGW GDP

1995 19.2 4011 210.9 2443.21
1996 15.1 4323.8 194.4 2871.65
1997 10.7 4988.09 129.6 3241.47
1998 10.3 5284.9 285.5 3474.09
1999 9.5 5011.5 174.5 3649.12
2000 11.7 5832.44 235.3 3928.2
2001 11.5 6664.83 224 4293.49
2002 11.8 6983.54 234.1 4275.01
2003 12 7290.76 226.1 5333.09
2004 11.92 7811.08 229.86 6379.63
2005 12.3 8099.89 215.9 7385.1
2006 11.5 9122.34 196.6 8690.24
2007 11.5 9651.3 154.5 10562.39
2008 10.9 9236.9 132.79 12601.23
2009 10.7 10672.12 125.57 14151.28
2010 9.34 11239.2 133.87 17185.48
2011 8.04 12684.47 118.48 21026.28
2012 7 13187.3 106.18 23872.8
2013 6.48 14006.62 104.24 26392.07

Notes: Data is from “Statistical Yearbook of Sichuan Province”. (http://www.sc.stats.gov.cn/tjcbw/tjnj/)

26
 Appendix B
OLS test result

Dependent Variable: GDP

Method: Least Squares
Date: 10/02/17 Time: 14:52
Sample: 1995 2013
Included observations: 19

Variable Coefficient Std. Error t-Statistic Prob.

C 1442.774 5230.759 6.275825 0.0067

ISW -2.954359 1.933562 -2.553331 0.0230
IGW 81.52811 104.4573 3.838063 0.0049
IWW -12.19064 22.30734 -3.363557 0.0216

R-squared 0.953979 Mean dependent var 9566.096

Adjusted R-squared 0.940831 S.D. dependent var 7591.177
S.E. of regression 1846.534 Akaike info criterion 18.10094
Sum squared resid 47735637 Schwarz criterion 18.34948
Log likelihood -166.9590 Hannan-Quinn criter. 18.14300
F-statistic 72.55294 Durbin-Watson stat 2.043077
Prob(F-statistic) 0.000000

27
 REFERENCES

Arrow K. Bolin B., Costanza R., Dasgupta P., Folke C., Holling C.S., Jansson B., Levin
S., Mäler K., Perrings C. and D. Pimentel,1996, “Economic growth, carrying capacity,
and the environment.” Environment and Development Economics, 1, 104-110.

Beckerman W., 1992, “Economic growth and the environment: whose growth? Whose
environment”, World Development, (20):481-496.

Berman E. and L. Bui, 1998, “Environmental Regulation and Labor Demand: Evidence
from the
South Coast Air Basin”, National Bureau of Economic Research Working Paper No.
6776.

Bimonte S., 2009, “Growth and environmental quality: Testing the double convergence
hypothesis”, Ecological Economics, 68, 2406–241.

Blazejczak J., Braun F. and D. Edler, 2009, “Global Demand for Environmental Goods
and Services on the Rise: Good Growth Opportunities for German Suppliers”, German
Institute for Economic Research, 5, 20, 138-145.

Boulding K. E., 1966, “The Economics of the Coming Spaceship Earth”, Environmental
Quality in a Growing Economy, ed., H. Jarrett, Baltimore, MD: John Hopkins Press.

Carraro C., De Cian E. and M. Tavoni, 2009, “Human Capital Formation and Global
Warming
Mitigation: Evidence from an Integrated Assessment Model”, CESIFO WP 2874.

Caviglia-Harris J., Chambers L., Dustin K. and R. James, 2009, “Taking the “U” out of
Kuznets: A comprehensive analysis of the EKC and environmental degradation”,
Ecological Economics, Elsevier, 68(4), 1149-1159.

Choudhury K. and L. Jansen, 1997, Terminology for integrated resources planning and
management. Rome, FAO, Soils Resources Management and Conservation, 59.

Cole M. A., 2004, “The pollution haven hypothesis and the environmental Kuznets
Curve: Examining the linkages”, Ecological Economics, 48(3), 71-81.

Coondoo D. and Dinda S., 2002, “Causality between income and emissions: a country
group-specific econometric analysis”, Ecological Economics, 40(3), 351-367.

Davidson C., 2000, “Economic growth and the environment: alternatives to the limits
paradigm”, BioScience 50:433–440.

Davis G. A., 1995, “Learning to love the Dutch disease: Evidence from the mineral

28
economies”, World Development, 23(10), 1765–1779.

Day K. M. and Graften R. Q., 2003, “Growth and the environment in Canada: An
empirical analysis, Canadian Journal of Agricultural Economics”, 51(3), 197-216.

Dietz S., 2000, “Does an environmental Kuznets curve exist for biodiversity?” Institut für
Wirtschaftsforschung, Eidgenössische Technische Hochschule Zürich.

Dutt K., 2009, “Governance, institutions and the environment-income relationship: a

cross-country study”, Environment, Development and Sustainability, 11:4, 705-723.

Easterlin R. A., 1974, “Does economic growth improve the human lot?”, Nations and
Households in Economic Growth, Academic Press.

Eatherley D. and S. Slater, 2009, “Competitiveness Improvements Potentially Available

from Resource Efficiency Savings”, report for Defra.

Ekins P., 2000, “Economic Growth and Environmental Sustainability: The Prospects for
Green Growth”, Routledge, London.

Frontier Economics, 2009, “Alternative policies for promoting low carbon innovation.”
Report for Department of Energy and Climate Change.

George E. Halkos, 2003, “Environmental Kuznets Curve for sulfur: evidence using GMM
estimation and random coefficient panel data models”, Environmental and development
economics 8 (4), 581-601.

Gray W. B. and R. J. Shadbegian, 1998, “Environmental Regulation, Investment Timing,

and Technology Choice”, Journal of Industrial Economics 46:235-256.

Grossman G. and Krueger A., 1992, “Environmental impacts of a North American free
trade agreement”, Garber P., The Mexico-US Free Trade Agreement, Cambridge, MA:
M ⅡPress.

Grossman G. and Krueger A., 1995, “Economic Growth and the Environment”,
Quarterly Journal of Economic, 110(2), 353-372.

Harris J M. and N. R. Goodwin, 2003, “Reconciling Growth and Environment”, GDAE

Working Papers 03-03, GDAE, Tufts University.

Innovas, 2009. “Low Carbon and Environmental Goods and Services: an industry
analysis.” Report for BERR.

Jaffe A. B., Newell R. G. and R. N. Stavins, 2000, “Technological Change and the
Environment”, WP7970, National Bureau of Economic Research.

29
Jaffe A. B. and K. Palmer, 1997, “Environmental Regulation And Innovation: A Panel
Data Study,” The Review of Economics and Statistics, MIT Press, 79(4), 610-619.

Jaffe A. B., Peterson S. R., Potney, P.R. and R. N. Stavins, 1995, “Environmental
Regulation and the competitiveness of US Manufacturing: What does the evidence tell
us?”, Journal of Economic Literature, 23, 132-163.

Jorgenson D. W. and P. J. Wilcoxen, 1990, “Environmental Regulation and U.S.

Economic Growth”, The RAND Journal of Economics, 21-2, 314-340.

Kuznets S., 1934, "National Income, 1929-1932". 73rd US Congress, 2nd session,
Senate document no. 124, 7.

Kuznets S., 1955, “Economic growth and income inequality,” American Economic
Review, 49.l-28.

Lewis W. A., 1954, “Economic Development with Unlimited Supply of Labour”, The
Manchester School, 22, 2:139-91.

Liang J. and McKitrick R., 2002, “Income growth and air quality in Toronto: 1973-1997”,
Mimeo, University of Guelph, Economics Department.

Liu G., 2006, “A causality analysis on GDP and air emissions in Norway”, Discussion
Papers, Statistics Norway, No.447.

Liu L., 2005, “The Environmental Kuznets Curve of Guangdong Province”, Research of
Environmental Sciences, 18(6), 6-11.

Magnani E., 2000, “The Environmental Kuznets Curve, environmental protection policy
and income distribution”, Ecological Economics 32, 431–443.

Meyer S. M., 1992, “Environmentalism and economic prosperity: testing the

environmental impact hypothesis, MIT Mimeo.

Millennium Ecosystem Assessment, 2005, “Ecosystems and Human Wellbeing:

Synthesis.” Island Press, Washington DC.

Omisakin A. O., 2009, “Economic growth and environmental quality in Nigeria: Does
environmental Kuznets Curve hypothesis hold?”, Environmental Research Journal,
3(1),14-18.

Palmer K., W. E. Oates and P.R. Portney, 1995, “Tightening Environmental

Standards:The Benefit-Cost or the No-Cost Paradigm?” Journal of Economic
Perspectives 9:119-132.

Panayotou T., 2000, “Economic Growth and the Environment.” CID Working Papers 56,

30
Center for International Development at Harvard University.

Peng L., Tong X. and Shen Y., 2008, “Quantitative process of environment and
economic progress in Shanghai”, China Population Resources And Environment,
18(3),186-194.

Peng S. and Bao Q., 2006, “Economic growth and environmental pollution: an empirical
test for the environmental Kuznets Curve hypothesis in china”, Research On Financial
And Economic Issues, (8):3-17.

Porter M. E. and C. van der Linde, 1995, “Towards a New Conception of the
Environmental Competitiveness Relationship”, Journal of Economic Perspectives 9, 97-
118.

Ranjan R. and J. Shortle, 2007, “The environmental Kuznets curve when the
environment exhibits hysteresis.” Ecological Economics, Elsevier, 64(1), 204-215.

Selden T. M. and D. Song, 1994, “Environmental Quality and Development: Is There a

Kuznets Curve for Air Pollution Emissions?” Journal of Environmental Economics and
Management, Elsevier, 27(2), 147-162.

Shen M. and Xv Y., 2000, “A new type of Kuznets Curve: the relationship between
economic growth and environment change”, Zhejiang Social Sciences, (7),53-57.

SQW, 2007, “Exploring the relationship between environmental regulation and

competitiveness – Literature Review. Final Report to the Department for Environment
Food and Rural Affairs”, Defra.

Stern D. I., Michael S. C. and Edward B. B, 1996, “Economic growth and environmental
degradation: The environmental Kuznets Curve and sustainable development”, World
Development, 24(7):1151-1160.

Stokey, N.L, 1998, “Are there limits to growth?”, International Economic Review, 39, 1-
31.

Turner R. K., Morse-Jones S., and B. Fisher, 2007, “Perspectives on the ‘Environmental
Limits’ concept: A report to the Department for the Environment, Food and Rural Affairs”,
CSERGE, Norwich.

Tim Everett, Mallika Ishwaran, Gian Paolo Ansaloni and Alex Rubin, 2010, “Economic
Growth and the Environment”.

Uchiyama, 2016, “Environmental Kuznets curve Hypothesis and Carbon Dioxide

Emissions”, 8(63), p.6.

Wang Y., Cui X., and Chen W., 2006, “Empirical research on the relationship between

31
economic growth and environmental degradation in Nanjing city” [J]. Resources and
Environment in the Yangtze Basin, 15(2), 142-146.

Wang Z., Wen Z.,Yan F., et al., 2007, “Verifying the environmental Kuznets Curve
hypothesis and its conditions in Beijing”, China Population Resources And Environment,
17(2), 40-47.

World Bank, 1992, World Development Report 1992: “Development and the
Environment “, Washington, DC: The World Bank.

Wu Y. and Dong S., 2002, “Evaluating environmental policy of Beijing”, Urban

Environment & Urban Ecology, 15(2), 4-6.

Yang K., Ye M. and XVQ., 2003, “Environmental Kuznets characteristics of municipal

solid waste growth in Shanghai city, Geographical Research”, 22(1), 60-66.

Zhang B., Bi J., Ge J., et al., 2008, “Empirical study of environmental Kuznets curve in
south Jiangsu, Economic Geography”, 28(3), 376-379, 424.

32