Microsoft Word - CET--006.docx
CHEMICAL ENGINEERING TRANSACTIONS
VOL. 59, 2017
A publication of
The Italian Association
of Chemical Engineering
Online at www.aidic.it/cet
Guest Editors: Zhuo Yang, Junjie Ba, Jing Pan
Copyright © 2017, AIDIC Servizi S.r.l.
ISBN 978-88-95608- 49-5; ISSN 2283-9216
The Impact of Trade on Environmental Pollution of China -An
Empirical Analysis Based on Double Difference Model
Shurong Han, Xiaowei Zhou
Shaanxi Normal University, Xi’ an 710062, China
srhan2009@21.cn
China has joined the WTO for almost fifteen years so far. It is no doubt that WTO does great help in boosting
China’s economic growth. However, with the rapid development of our trade and economy, the environmental
problem in China has drawn increasing attention. This paper investigates the impact of joining the WTO on
environmental pollution and analyzes whether coastal cities are more affected than inland cities. We use
difference-in-differences approach to analyze data from approximately 300 cities from 1997 to 2006 in China,
taking the emission of industrial SO2 and the discharge standard-meeting rate of industrial wastewaters as an
example, and then examine that the environmental pollution has been relieved since China joined the WTO.
Also, the improvement effect on the environment of coastal cities is greater than that of inland cities.
1. Introduction
Since 2001, China's accession to the WTO, China has quickly integrated into the global liberalization of the
world market (Bischoff et al., 2015). The trade growth has become the most important macroeconomic goals
of governments, and WTO is also seen as promoting the majority of China's cities in the economic prosperity
and progress of the catalyst (Nishigaki et al., 2015). However, behind the prosperity, China's serious
environmental degradation has also aroused our thinking. In China, the pollution-intensive industries account
for a large proportion of the import and export trade volume (Galati et al., 2015). As the State Environmental
Protection Administration has pointed out, the contradiction between environment and development has
become more and more prominent (Wang et al., 2013). There are two major international environmental
economics. According to the environmental impact of trade: environmental protection factions believe that
foreign trade is bound to exacerbate the consumption of resources and worsen the natural environment (Guo,
et al., 2016). On the other hand, the trade protectionists believe that the globalization of trade liberalization is
not the root cause of environmental pollution intensified (Caimo and Lomi, 2015). In order to further answer
this question, in this paper, we have established a double difference model, using 1997 to 2006 China 251
cities.
The ecological environment of the earth is becoming worse and worse. The greenhouse effect, ozone hole,
acid rain and soil erosion have brought serious threat to the survival and development of mankind, and this
challenge is worldwide, occurring in every corner of the world (Borghesi et al., 2015). According to a 2012
report by the United Nations Environment Programme, in twenty-first Century, there were many extreme
events of environmental pollution, such as the European heat wave in 2003 and the flood disaster in Pakistan
in 2010 (Bi et al., 2014). The report says annual emissions of greenhouse gases worldwide will reach 580
tonnes if no action is taken in 8 years. The problem of global environmental pollution has been involved in air,
ecology, energy, water, waste and many other aspects, and there is a growing trend. In many countries,
especially in developing countries, environmental pollution is becoming more and more serious with the
continuous development of economy. As the largest developing country, China has also paid a huge price for
environmental pollution while the economy is developing at a high speed (Al-Mulali, and Ozturk, 2015).
In recent years, the ecological environment in our country is deteriorating day by day. The emission of major
pollutants exceeds the carrying capacity of the environment, and the natural resources are destroyed in
varying degrees. This can also be seen from an endless stream of reports about China's growing
environmental pollution. Fog and haze clearly show that China's pollution problem is increasing. The
DOI: 10.3303/CET1759191
Please cite this article as: Shurong Han, Xiaowei Zhou, 2017, The impact of trade on environmental pollution of china -an empirical analysis
based on double difference model, Chemical Engineering Transactions, 59, 1141-1146 DOI:10.3303/CET1759191
1141
Environmental Status Bulletin issued by the Ministry of environmental protection of China also shows that the
area of acid rain in the country accounts for 10.6% of the total land area (Katircioglu, 2014). Among the 74
cities monitored under the new air quality standards, air quality standards in 3 cities, namely Haikou,
Zhoushan and Lhasa, were up to standard. The proportion of cities exceeding the standard is 95.9%. In 2014,
the central economic work conference also mentioned that China's current environmental carrying capacity
has reached or near the upper limit. Therefore, the situation of environmental pollution in China is no longer
optimistic.
2. Literature review
In trade policy, the impact of trade on the environment this topic is very research value and research
significance. In the 1990s, Krueger and Grossman studied the impact of NAFTA on the environment and
explained the per capita GDP in terms of GDP per capita. With SO2, dust and suspended particles as the
explanatory variables, supplemented by population density, time trend, geographical position and other control
variables, the analysis shows that environmental pollution and economic growth in the long-term show
inverted U-shaped curve (EKC), inspired by this literature, we add per capita GDP and its square term to the
model as the control variable, which is called the environmental Kuznets Curve (EKC). In 2003, Copeland and
Taylor proposed the "Pollution Shelter Hypothesis" (PHH) by comparing the two sets of data between the
developed countries of the North and the developing countries of the South. There are many literatures
supporting this hypothesis abroad. For example, Rock (1996), after studying the relationship between trade
policy and pollution intensity, concluded that open trade policy is more pollution-intensive than inward trade
policy. The same view was supported in some literature using the input-output model of the empirical analysis
model: some scholars have concluded that the carbon content of non-energy products in Brazil is less than
Export content. The "slope effect" hypothesis proposed by Revesz in 1992 is also a dynamic extension of the
"pollution shelter hypothesis". The developing countries will gradually "degenerate" in the face of the loosening
of environmental policy, and will gradually "degenerate" and raise the standard of environmental protection,
and for the developed countries, they are likely to gradually missing the competitiveness of strategic industries,
and structural unemployment and other social problems, forced to reduce the environmental standards. In
general, increased competitiveness resulting from trade liberalization will force countries to relax their
environmental protection policies and move towards the lowest levels.
However, in the domestic and foreign literature research on the relationship between domestic trade and
environment, the conclusions are very different from the pollution shelter hypothesis (Xiao, et al., 2015). The
first is the use of Grossman as a regression model to analyse it. They use three years of industry panel data
and OECD developed countries data show that exports from OECD countries make China's carbon emissions
increased (Yao et al., 2015), while the OECD Imports but reduce China's carbon emissions. In 2007,
Mukhopadhyay and Dietzenbacher found that India is a net importer of carbon under the I-O model, which
means that trade mitigates India's pollution emissions also used in the input-output model is carbon emissions
as an explanatory variable. Moreover, some scholars have creatively used four types of pollutants through the
input-output model. The study found that China's export products contain less pollution than imported products,
thus proving more forcefully (Sánchez-Franco and Roldán, 2015).
Based on the literature, we can find that the econometric model and the input-output model are the main
models for studying trade and environmental problems at present. However, the dual-difference model is very
rare in related research. In recent years, some foreign scholars have used the DID model to analyze the
impact of regional trade agreements (RTA) with or without environmental policies on carbon emissions in 182
countries from 1980 to 2008 (Tang et al., 2012), only the RTA with environmental provisions can improve the
environment. Inspired by this article, we argue that the impact of China's trade on the environment under the
double-differencing model is particularly innovative and provides a more comprehensive understanding of the
relationship between the two.
3. The empirical analysis
3.1 Model settings
We will be 251 prefecture-level cities in China are divided into coastal cities and inland cities, respectively. In
contrast to the Baghdadi study, the variables selected in this paper are all exogenous variables and do not
involve endogeneity. Therefore, we can test that the rapid trade development of the coastal and inland cities
after entering WTO is exacerbating the environmental pollution or to improve their environmental conditions.
We will begin with a simple list of determinants of environmental pollution, such as per capita GDP, urban land
area, secondary industry as a percentage of GDP, total urban population, and distance from the city to the
port as a control variable. Taking into account that these control variables are not necessarily linearly related
1142
to emissions, we have also made some logarithms or squares of variability for some indicators, such as
incorporating square values of GDP per capita and population logarithms into the control variables to better
explain pollution emission. In order to measure the impact of trade on China's urban environmental pollution
emissions, we estimated the following simple linear regression equation: the industrial SO2 emissions per unit
GDP, the industrial wastewater discharge compliance rate as the explanatory variable,
E𝑚𝑗𝑐 = 𝛽0 + 𝛽1𝐶𝑖𝑡𝑦𝑗𝑐 + 𝛽2𝑊𝑇𝑂𝑗𝑐 + 𝛽3𝐼𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑗𝑐 + 𝜀𝑗𝑐 (1)
𝐼𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑗𝑐 = 𝐶𝑖𝑡𝑦𝑗𝑐 · 𝑊𝑇𝑂𝑗𝑐 is the interactive term, j means before and after the accession to the WTO, c is the
coastal city, Emjc is the explanatory variable, and εjc is the residual term.
There are two dummy variables in the model. The variable "Cityjc" indicates whether the sample city is a
coastal city. If the value is 1, that is, the city is coastal, if it is 0, the city is inland city; "WTOjc" that the process
of accession to the WTO, if taken 1, then said that accession to the WTO, if taken 0, then that before joining
the WTO, in order to test the effect of accession to the WTO, the establishment of "WTO jc". The interactive
item "WTO" refers to the interaction of the coastal cities after the accession to the WTO. In other cases, the
interactions between the two variables is 0.
On the basis of (1), we also need to add related control variables. Considering the different levels of
development of each city, it is necessary to include per capita GDP (Incomejc) as a control variable. The
environmental Kuznets curve (EKC) also tells us that income and the environment Pollution is a curve, rather
than a simple linear relationship, so we need to join the quadratic GDP per capita GDP to better explain the
variables; followed by the total urban population, taking into account the natural population growth rate, where
we select the logarithm The proportion of secondary In(Popjc) industry to GDP is also an important control
variable, because the greater the proportion of urban industry, the greater the amount of pollution emissions;
Finally, according to the trade gravity The principle of the model, the logarithm of the distance from city to port,
In(Disjc) can also be counted as a control variable. In summary, the final model is set as follows:
E𝑚𝑗𝑐 = 𝛽0 + 𝛽1𝐶𝑖𝑡𝑦𝑗𝑐 + 𝛽2𝑊𝑇𝑂𝑗𝑐 + 𝛽3𝐼𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑗𝑐 + 𝛽4𝐼𝑛𝑐𝑜𝑚𝑒𝑗𝑐 + 𝛽5𝐼𝑛𝑐𝑜𝑚𝑒𝑗𝑐
2 + 𝛽5𝐼𝑛(𝑃𝑜𝑝𝑗𝑐) + 𝛽6𝑅𝑎𝑡𝑖𝑜𝑗𝑐 +
𝛽7𝐼𝑛(𝐷𝑖𝑠𝑗𝑐) + 𝜀𝑗𝑐 (2)
Based on the model (2), this paper will test the related data.
3.2 Data analysis and typical facts
In this model application, since China's accession to the WTO in December 2001, we use data of 251
prefecture-level cities in China between 1997 and 2006, which is derived from the years of China's urban
statistical yearbook. In the model (2) the variable data sources are: GDP (million), industrial output value
(million), the end of the total population (million), land area (square kilometers), industrial sulfur dioxide
emissions Ton), industrial wastewater discharge compliance rate (%) and the distance to the coastal ports
(km). In this paper, we choose two explanatory variables, one is SO2 emission per unit GDP (SO2 EMjc), and
the other is industrial wastewater discharge compliance rate (WW EMjc). Table 1 is the descriptive statistics of
the relevant explanatory variables in this paper.
In this paper, we use the twice-calculated data and the double-difference model to regress the Eq(2). Through
the inspection, we observe the coastal cities and inland cities after the accession to the WTO after the
environmental pollution emissions are not systematic differences, if any, then the WTO does have a significant
impact on environmental pollution. Specific test results are in Table 2 and Table 3:
Table 1: describes the statistical results
Variable Name Explanation Mean Standard Deviation
SO2 EMjc Emissions per Unit of GDP 0.0346078 0.1611538
WW EMjc industrial wastewater discharge compliance rate 74.76253 29.57979
Cityjc coastal dummy variable 0.149682 0.356816
WTOjc time dummy variable 0.5 0.50008
Interactjc Interaction Item 0.074841 0.263176
Incomejc GDP per capita 9.398173 0.906007
Incomejc
2 GDP per capita2 89.1462 16.79063
Ln (Popjc) The population logarithm 4.747621 1.050338
Ratiojc Proportion of the secondary industry 0.494814 0.126863
Ln (Disjc) Distance to harbor 6.200698 0.904956
1143
Table 2: Regression analysis of SO2 emissions per unit of GDP under mixed effects, random effects and fixed
effects (2)
POOLED POOLED Random Random Fix Fix
SO2 EMjc SO2 EMjc SO2 EMjc SO2 EMjc SO2 EMjc SO2 EMjc
Cityjc -0.00000245 0.0118 -0.0174 0.0113 0 0
(-0.00) (1.25) (-0.79) (0.91) (.) (.)
WTOjc 0.0831
*** 0.0786*** 0.0643*** 0.0779*** 0.0609*** 0.0816***
(11.89) (13.38) (12.18) (13.95) (11.52) (9.07)
Interactjc -0.0721
*** -0.0461*** -0.0535*** -0.0448*** -0.0502*** -0.0442***
(-4.24) (-3.40) (-4.27) (-3.87) (-4.02) (-3.77)
Incomejc 0.0372 0.0408
* 0.0286
(1.53) (1.79) (1.08)
Incomejc
2 -0.00227* -0.00246** -0.00187
(-1.80) (-2.08) (-1.37)
Ln(Popjc) -0.00353 -0.00470 -0.00881
(-0.63) (-0.86) (-1.36)
Ratiojc -0.00709 0.00407 0.0621
(-0.31) (0.15) (1.34)
Ln(Disjc) 0.00561
* 0.00588 0
(1.91) (1.35) (.)
_cons 0.00000434 -0.0405 0.0176** -0.0423 0.00903*** 0.108
(0.00) (-0.36) (2.01) (-0.36) (2.76) (0.59)
N 2407 1028 2407 1028 2407 1028
R2 0.062 0.200 0.0922 0.1676 0.060 0.238
Note: The standard deviation estimates for heteroskedasticity and autocorrelation in parentheses, *, **, ***, respectively, are
significant at the 1, 5 and 10% confidence levels.
Table 3: Regression analysis of model (2) for mixed effect, random effect, fixed effect on industrial effluent
discharge compliance rate
POOLED POOLED Random Random Fix Fix
WW EMjc WW EMjc WW EMjc WW EMjc WW EMjc WW EMjc
Cityjc 4.953
** -8.217** 5.826** -8.393* 0 0
(2.53) (-2.06) (2.24) (-1.77) (.) (.)
WTOjc 23.82
*** 12.08*** 24.64*** 10.97*** 25.46*** 15.07***
(21.26) (4.75) (23.73) (4.38) (24.21) (3.59)
Interactjc 1.201 1.327 0.45 1.379 -0.335 2.27
(0.44) (0.23) (0.18) (0.26) (-0.13) (0.44)
Incomejc 21.51** 18.47* 3.073
(2.09) (1.82) (0.25)
Incomejc
2 -0.976* -0.785 0.0412
(-1.83) (-1.49) (0.07)
Ln(Popjc) -10.78*** -11.70*** -15.20***
(-4.48) (-4.79) (-5.06)
Ratiojc -31.29*** -35.81*** -28.11
(-3.23) (-3.25) (-1.30)
Ln(Disjc) -2.260* -2.105 0
(-1.79) (-1.30) (.)
_cons 61.37*** -67.42 60.34*** -49.95 61.46*** 91.83
(75.42) (-1.40) (57.02) (-1.00) (89.96) (1.08)
N 2790 1012 2790 1012 2790 1012
R2 0.169 0.248 0.222 0.3061 0.222 0.311
Note: The standard deviation estimates for heteroskedasticity and autocorrelation in parentheses, *, **, ***, respectively, are
significant at the 1%, 5%, and 10% confidence levels
In the choice of regression model, we used the Hausman test, as shown in Table 4:
1144
Table 4: Hausman test
—— Coefficients ——
(b) (B) (b-B) sqrt(diag(V_b-V_B))
m6 m4 Difference S.E.
WTOjc .0816473 .0779483 .003699 .0070242
Interactjc -.0442493 -.0448478 .0005985 .0015117
Incomejc .0285651 .0408404 -.0122753 .0130408
Incomejc
2 -.001865 -.0024627 .0005977 .0006554
Ln(Popjc) -.0088137 -.0047009 -.0041128 .0034168
Ratiojc .0620859 .0040688 .058017 .0375997
b = consistent under Ho and Ha; obtained from xtreg;
B = inconsistent under Ha, efficient under Ho; obtained from xtreg;
Test: Ho: difference in coefficients not systematic;
Chi2 (7) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 9.54;
Prob>chi2 = 0.2163>0.05.
According to the results of Hausman test, we choose random effects as our regression analysis model. From
the third and fourth columns of Table 2 and Table 3, it can be seen that the trade has a significant mitigation
effect on the industrial SO2 emission of coastal cities in China, but the improvement rate of industrial
wastewater discharge is not significant.
4. Conclusions
This paper examines the impact of China's accession to the WTO on China's environmental pollution. We use
the double difference model to analyze the impact of foreign trade on China's environmental pollution from
1997 to 2006, taking two kinds of pollution as the observed value-GDP per unit of GDP, which is the dummy
variable, with several related control variables. Industrial sulfur dioxide emissions and industrial wastewater
discharge compliance rate, inland cities as a control, focusing on the coastal city trade on its environmental
impact. As the variables involved in the model are exogenous variables, so there is no need to explore the
endogenous problems. From the results of the random effects regression, China's accession to the WTO on
the coastal cities of sulfur dioxide emissions has played a very significant role, but for its industrial wastewater
discharge compliance rate did not significantly improve the role. This shows to a certain extent the trade of the
Chinese environment is played a certain role in improving the trade and the more developed the coastal city
environment benefit more from trade.
Based on the conclusions drawn in this paper, some pertinent suggestions are put forward. First of all,
technological innovation is one of the important factors to solve environmental problems. Trade promotes
technology transfer. While adapting to globalization and liberalizing trade development, China must seize the
opportunity to learn from it or develop more efficient environmental management technology to improve it. On
the other hand, we also need to enhance the comparative advantage of technology-intensive industries, to
reduce the advantages of pollution-intensive industries in China and to improve the overall trade environment.
Second, China needs to accelerate the upgrading of industrial structure, the waste discharge of high-waste
industries is phased out, and thus to speed up the upgrading of energy structure, to develop the proportion of
services to rise, so that China's domestic industry more green. Moreover, in the international market, the
"green barrier" of the developed countries to developing countries still has certain restrictions on the
development of China's economy.
The results of this study show that trade has a certain mitigation effect on the environmental pollution of
coastal cities in China. It also proves the importance of double difference model in the field of trade and
environment research. Through this study of the relationship between trade and environment in China, the
understanding of the role of trade in the environment has been improved. In this paper, based on relevant
literature, the author sums up the viewpoints of all parties, and test and prove to some extent that the
"pollution shelter hypothesis" cannot be established in China. However, the level of pollution chosen in this
paper is not comprehensive enough. In 2002, SO2 emissions data is missing, thus a more robust conclusion
remains to be further proved.
Reference
Al-Mulali U., Ozturk I., 2015, The effect of energy consumption, urbanization, trade openness, industrial output,
and the political stability on the environmental degradation in the MENA (Middle East and North African)
region. Energy, 84, 382-389.
1145
Bi G.B., Song W., Zhou P., Liang L., 2014, Does environmental regulation affect energy efficiency in China's
thermal power generation? Empirical evidence from a slacks-based DEA model. Energy Policy, 66, 537-
546.
Bischoff F., Emrich E., Pierdzioch C., 2015, Universities of cooperative education between facts and fictions:
an empirical analysis based on an example of the sports and fitness sector. , 14(1), 83-106.
Borghesi S., Cainelli G., Mazzanti M., 2015, Linking emission trading to environmental innovation: evidence
from the Italian manufacturing industry. Research Policy, 44(3), 669-683.
Caimo A., Lomi A., 2015, Knowledge sharing in organizations: a bayesian analysis of the role of reciprocity
and formal structure. Journal of Management: Official Journal of the Southern Management Association,
41(2), 665-691.
Galati A., Crescimanno M., Giacomarra M., Tinervia S., 2015, Organizational models in the sicilian ornamental
plant industry: an empirical analysis based on transaction cost theory. New Medit, 14(4), 58-64.
Guo X., Ren D., Shi J., 2016, Carbon emissions, logistics volume and gdp in china: empirical analysis based
on panel data model. Environmental Science & Pollution Research International, 23(24), 24758.
Katircioglu S.T., 2014, International tourism, energy consumption, and environmental pollution: The case of
Turkey. Renewable and Sustainable Energy Reviews, 36, 180-187.
Nishigaki Y., Maki D., Satake M., 2015, Capital adjustment and limit cycles: an empirical analysis based on
the threshold autoregressive model. Acta Scientiarum Mathematicarum, 42.
Sánchez-Franco M.J., Roldán J.L., 2015, The influence of familiarity, trust and norms of reciprocity on an
experienced sense of community: an empirical analysis based on social online services. Behaviour &
Information Technology, 34(4), 392-412.
Tang K., Yang Z., Yang H., Fan X., 2012, Environmental cost of pond aquiculture in shanghai: an empirical
analysis based on double-bounded dichotomous cvm method. Acta Ecologica Sinica, 32(7), 2212-2222.
Wang C.L., Zhang W., Statistics S.O., 2013, Financial development, capital formation and differential regional
effects: an empirical analysis based on dynamic panel data model. Mathematics in Practice & Theory,
43(18), 17-27.
Xiao Y., Zheng X., Hu L., Hen Q., 2015, Research on the factors of trade growth between china and india —
an empirical analysis based on constant market share model. Journal of Service Science & Management,
08(4), 569-577.
Yao W., Wu H., Kinugasa T., 2015, Financial deepening, asset price inflation, and economic convergence:
empirical analysis based on china’s experience. Emerging Markets Finance & Trade, 51(sup1), S275-
S284.
1146