CC BY
35
YflK 338.49:311.33
EMPIRICAL RESEARCH ON ECONOMIC GROWTH AND ENVIRONMENTAL POLLUTION BASED ON VAR MODEL
LUO JUAN1, A.N. ILCHENKO2
Wuhan Textile University,
Wuhan City, China 2Ivanovo State University of Chemical Technology,
Ivanovo City, Russian Federation
Introduction. The relationship between environmental pollution and economic development has been a hot issue to environmental economics and ecological economics research. In recent years, as the growth of the economy, many provinces and cities in China are facing increasingly serious environmental pollution problems. In this article, we mainly make a discussion about the problems of environmental pollution and economic of Hubei Province in China during 1999-2009.
Hubei Province is located in center of China, the total area is 186000 square kilometers, and the total population is about 57 million. The transportation and communication are very convenient here. It is rich in natural resources, Changjiang River which is the world's third largest river is through here. Hubei Province is one of China's important industrial and agricultural base.
In recent years, It has a very good development opportunity. The province's GDP was 3229.29 billion yuan in 1999 and 12961.10 billion yuan in 2009. The per capita GDP was 3857.99 yuan in 1999 and 22677 yuan in 2009. During this period, the industrial structure changed a lot. Proportion of three industrial structure was 20.3: 40.7: 39.0 in 1999, and it became to 13.8: 46.6: 39.6 in 2009. Due to historical reasons, the industrial structure of Hubei was major heavy industry at long time. Since the reform and opening up, structure of light and heavy industrial has adjusted, but the heavy industrial characteristics are still evident [1].
At the beginning of industrialization, the industrial model of economic development were basically based on the resource-based, high material consumption, high energy consumption, heavy pollution and heavy chemical industry in Hubei Province. The rapid economic growth of Hubei Province in China has led to increase pressure on resources and environment. Industrial waste water emissions were respectively 1159.85 million tons in 1999 and 913.24 million tons in 2009. The industrial waste gas emissions were 556.60 billion cubic meters in 1999 and 1125.30 billion cubic meters in 2009. The amount of industrial solid waste production achieved 25.11 million tons in 1999 and 55.61 million tons in 2009. Look at the
situation from 1999 to 2009, the amount of industrial waste gas emissions and solid waste were in rapid
growth to the emissions of industrial waste water. Meanwhile, as a water-rich province, Hubei is also inevitably facing a serious problem of water pollution [1].
It is necessary to research the relationship between environmental pollution and economic growth. In this paper, we mainly use the VAR model, and apply the method of impulse response function and variance decomposition to make up long-term dynamic characteristics between per capita GDP and three kinds of pollution indicator of Hubei Province.
Research method.
Vector Autoregressive (VAR) Model.
The VAR model is an unstructured modeling method to study the relationship among each variable which was first use by Sims. The expression of the model is:
yt = Ay-i +■ ■ ■+Apyt-p + Bxt +£t f = 12 ■T
There yt is a dimension variable dimension vector, xt is a exogenous variable vector, p is the number of lags, T is the number of samples. A1 •••AP are k x k dimensional matrix and B is kx d dimensional matrix which will be estimated. St is a k-dimensional matrix of error terms.
We analyze dynamic impact to VAR model when an error term changes, this method is called Impulse Response Function (IRF) method [2]. We analyze the contribution of each structure shock to endogenous
variables, and further evaluate the importance of different structural shocks, this method is known as Variance Decomposition [2].
Index selection and data source. In order to build the model, we select the data of industrial waste water, waste gas and solid waste of Hubei Province as environmental pollution indicators. Real GDP per capita is selected as the economic indicator. All the data is from Hubei Statistical Yearbook during 1999 to 2009 [1]. In this paper, x :GDP per captia; y1 :the emissions of waste water; y2 :the emissions of waste gas; y3 :the emissions of solid waste. In order to removal the heteroscedasticity of the data, we use lnx, ln y ,ln y2 ,ln y3 to model and Eviews 5.0 as a mathematical tool.
Modeling and analysis.
Unit Root Test and Co-Integration Test.
Many real time series of economic variables are non-stationary. In order to prevent spurious regression, we conduct Unit Root Test by ADF (Augmented Dickey-Fuller) method to see if the time series are stationary. The test results show lnx, Iny, lny2, lny3 are non-stationary time sequence ,but all the firstorder differential series are stationary at the 5% significance level. We can see the results in table 1.
Table 1 - The unit root test of all the variables
Variables ADF test statistic Test critical values: Inspection form (C,T,K) Results
Alnx -9.059207 -3.320969** (C,0,2) stationary
Any -4.566553 -3.320969** (C,0,2) stationary
Alny2 -7.977313 -3.320969** (C,0,2) stationary
Alny3 -6.189583 -3.320969** (C,0,2) stationary
(C: containing a constant term in ADF test statistic; T: containing a trend term;K: the number of lag;
A: first-order differential series. **:5% significance level).
So the original sequences satisfy the integrated of 1, and there may be a co-integration relationship. Thus our use Johansen co-integration to test, and the results in table 2 show that: lnx and lnyl, lnx and lny2, lnx and lny3 are through the 5% critical test. So there exists long-term stable relationship (cointegration relationship) between the variables in each group.
Table 2 - Johansen co-integration test
Variables Eigenvalue Trace Statistic 5% Critical Value Hypothesized No. of CE(s)
lnx 0.889975 21.65312 15.49471 None *
lny1 0.180332 1.789706 3.841466 At most 1
lnx 0.710653 11.49520 15.49471 None
lny 2 0.036436 0.334050 3.841466 At most 1
lnx 0.759816 16.10931 15.49471 None *
lny3 0.304812 3.272163 3.841466 At most 1
VAR Model and Impulse Response Function method.
We respectively build corresponding VAR models for lnx and lnyl, lnx and lny2, lnx and lny3 .The
number of lags of model are choose as AIC(Akaike Information Criterion). After several tests, all the number of lags for three VAR models are 2, these show there are remarkable influence of two variables in each group. Then we see all the reciprocals of root modulus for VAR models are less than 1 by using AR root test, which show that the VAR models are stable. we further get Figure 1 and Figure 2 by impulse
response function analysis[3][4], and try to find the long-term response of single variable pulse disturbance.
—♦— The impact from waste water
The impact from waste gas
The impact from solid waste
Figure 1 - Response of Industrial wastes to GDP Figure 2 - Response of GDP to Industrial wastes
From the Figure 1, we can see the response function curve of industrial waste water caused by the impact of economic growth hovers around zero, indicating that the economic growth had little impact on the increase of industrial waste water. The response function curves of solid waste pollution and industrial emissions caused by the impact of economic growth fluctuate and both the cumulative value are positive, indicating that economic growth has brought an increase in solid waste pollution and industrial waste gas emission, the heavy industry of Hubei Province is on rapid growth and consumption of industry still exists.
From the Figure 2, we can see the response function curves of economic growth caused by the impact of solid waste pollution and industrial emissions hover around zero, indicating that the solid waste pollution and industrial emissions have little affect to economic growth. The response function curves of economic growth caused by the impact of waste water is U-shaped, and the cumulative response values is -
0.540811, showed that an increase of industrial waste water has a negative effect on the economic growth.
Variance Decomposition. Variance decomposition is used to further evaluate the importance of the impact of different structures through analysising each structural shock contribution to inside variable changes (measured by variance ) [3][4], we get the following conclusions (Table 3) by variance decomposition analysis of the VAR models.
Table 3 - Average percent Environmental pollution indicators variance and GDP variance
Environmental pollution indicators Average percent environmental pollution indicators variance due to economical growth indicators Average percent economic growth indicators variance due to environmental pollution indicators
Waste water 0.580036 78.06841
Industrial emissions 0.996646 50.05382
Solid waste pollution 85.49533 4.432238
From the Table 3, we can see the average percent solid waste pollution indicators variance due to economical growth indicators is very high, this is good agreement with pattern of heavy industrial and coal-based energy consumption patterns in Hubei Province. The average percents economic growth indicators variance due to waste water and industrial emissions are high, this shows that the impact of environmental pollution on economic growth relatively more.
Conclusion. According to the economic and environment indicators in Hubei Province from 1999 to 2009, and empirical analysis above ,we can see that the economic growth is an important reason for change of pollution emissions in Hubei Province, the average percent environmental pollution indicators variance due to economical growth indicators is very high, and the increase of polluting emissions hinders economic development. Therefore, we should focus on the proportion of industrial structure of Hubei Province in the future. In order to promote the coordinated development of the economy and the envi-
ronment of Hubei Province, we give the following suggestion: the main way to reduce the provincial pollutants is not the economic growth, but should increase and improve the technology and equipment, pollution treatment facilities to reduce pollution and promote the harmonious development of the economy. At the same time, Government should strengthen the pollution supervision, perfect scientific and effective supervision system, and set up the related legal policy[5].
REFERENCES
1. Hubei Provincial Bureau of Statistics. Hubei Statistical Yearbook[M].China statistics Press, 1999-2009.
2. Gao, Tiemei. Econometric Analysis and Modeling-Eviews Application and Example[M] / Gao Tiemei // Tsinghua University Press, 2006.
3. Koop, G. Impulse Response Analysis in Non-linear Multivariate Models [J] / G. Koop, M. Pesaran, S. Potter // Journal of Econometrics, 1996(74).
4. Pesaran, M. Generalize Impulse Response Analysis in Linear Multivariate Models[J] / M. Pesaran, Y. Shin // Ecomomic Letters, 1998(58).
5. Peng, Wen-bin. Empirical Research on Environmental Pollution and Economic Growth in Hunan Province[J] / Peng Wen-bin, Tian Yin-hua // Journal of Xiangtan University, 2011,35(1).
EMPIRICAL RESEARCH ON ECONOMIC GROWTH AND ENVIRONMENTAL POLLUTION BASED ON VAR MODEL
LUO JUAN, A.N. ILCHENKO Summary
We establish the VAR (Vector Autoregressive )model, and apply the method of impulse response function and variance decomposition to make up long-term dynamic relationship between per capita GDP and three kinds of pollution indicator of Hubei Province during 1999-2009.The results show that: on the one hand, economic growth is an important reason for the impact of environmental pollution changes in Hubei Province, on the other hand, the increase of industrial waste has inverse function to economic growth. So we should pay more attention to environmental problems at the time of economic development.
Keywords: Economic Growth; Environmental Pollution; VAR; Impulse Response Function; Variance Decomposition.
© Ло Джейн, Ильченко А.Н.
Поступила в редакцию 02 апреля 2013г.
