Impact and Spatial Spillover Effect of Transport Infrastructure on Urban Environment

Available online at www.sciencedirect.com

ScienceDirect Energy Procedia 104 (2016) 227 – 232

CUE2016-Applied Energy Symposium and Forum 2016: Low carbon cities & urban energy systems

Impact and spatial spillover effect of transport infrastructure on urban environment Rui Xiea,* Jiayu Fanga,Cenjie Liub a b

School of Economics and Trade, Hunan University, Changsha, Hunan, 410079,China Business School, Hunan University, Changsha, Hunan, 410082,China

Abstract This paper expounds the influence of transport infrastructure on environment as spatial agglomerative, economic growth, innovation and technology diffusion effects. Within the STIRPAT model, we use a spatial Durbin model to estimate the impact of transport infrastructure on the environment in 281 Chinese cities during 20032013. The results show that transport infrastructure, technical progress, and energy intensity have negative direct effects on urban environment. Additionally, we find an inverted U-shaped curve relationship between GDP per capita and urban environment. As spatial effect, transport infrastructure has negative impacts, while technical progress has a positive effect. © 2016 Published by Elsevier Ltd. This 2016The TheAuthors. Authors. Published by Elsevier Ltd.is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Selection and/or peer-review under responsibility of CUE Peer-review under responsibility of the scientific committee of the Applied Energy Symposium and Forum, CUE2016: Low carbon cities and urban energy systems. Keywords: transport infrastructure; STIRPAT model; urban environment; spatial spillover effect

1. Introduction Since its reform and opening up policy, China has made great achievements in the construction of transport infrastructure. However, with the advance of urbanization and industrialization in China, environmental pollution problems cannot be ignored. As amassing spaces for both population and industry, cities are also facing environmental pollution in this process of construction and development. Although transport infrastructure construction reduces transportation costs, promotes the centralization of population and economy, spatial agglomeration of population and economy, and urbanization also has impacts on the urban environment. Therefore, the effect of transport infrastructure and its channels of influence on urban environment need to be determined. Currently, scholars mainly study the influence of transport infrastructure on economic growth, total factor productivity, and agglomeration. For example, Zhang [1] finds that transport infrastructure

*Corresponding

author. Tel: +86 18684675789; fax: +86 731 8684825 E-mail address: [email protected].

1876-6102 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the scientific committee of the Applied Energy Symposium and Forum, CUE2016: Low carbon cities and urban energy systems. doi:10.1016/j.egypro.2016.12.039

228

Rui Xie et al. / Energy Procedia 104 (2016) 227 – 232

investment has positive impacts on economic growth. Bronzini and Piselli [2] discover that the improvement of transport infrastructure has a positive effect on total factor productivity of the region. Liu [3] argues that transport infrastructure would cause agglomeration effects in cities. Current research on energy and environment is primarily focused on the impact of energy consumption and energy structure. For instance, Xu et al. [4] decompose energy consumption into the effects of energy structure, energy intensity, and industrial structure. Wang et al. [5] find that there is a bidirectional causality between energy consumption and CO2 emission. Zheng et al. [6] identify that the increase of energy caused by road transport would lead to greenhouse gas emission. Recently, the STIRPAT model has been widely used in energy environment research as follows. Shi [7] finds that population change was the main factor affecting global carbon emissions, and Fan et al. [8] find that the impact of different income levels on carbon emissions was significantly different. Moreover, Wang et al. [9] study the urbanization, economic development, and industrialization levels with positive effects on carbon emissions in Beijing. Liu et al. [10] study the driving factors of China's industrial pollution emissions, and the results support the environmental Kuznets curve (EKC) hypothesis. Although many scholars at home and abroad have conducted in-depth research on the economic effects of transport infrastructure, there are limited studies on the environmental effects of transport infrastructure. As such, this paper is based on the STIRPAT model, using data on 281 Chinese cities during 2003-2013. First, we elaborate the mechanism of transport infrastructure influence on environment. Second, we analyze the spatial correlation of the urban environment in China. Subsequently, we use a spatial Durbin model to estimate transport infrastructure on urban environment and its spatial spillover effects. 2. Mechanism analysis This paper reviews the economic effect of transport infrastructure and highlights that transport infrastructure affect urban environment as follows (see Figure 1). First, regarding the spatial agglomerative effect, Krugman [11] believes transportation cost reduction would concentrate the industry in a region, forming a "center periphery" model and generating spatial agglomeration. However, the impact of agglomeration on environmental pollution is still controversial. The agglomeration of economic activity will lead to crowding, and increase environmental pollution [12]. Agglomeration is conducive to reducing the level of pollution in the unit [13]. Second, the economic growth effect shows that the new economic geography theory [14] points out that improvement of transport infrastructure would generate economies of scale through trade expansion within regions, and economic growth effects. Furthermore, transport infrastructure investment also has positive impact on economic growth [1]. The relationship between economic growth and environmental pollution take the shape of an “inverted U.” Third, innovation and technology diffusion effect. Transport infrastructure construction has a significant positive impact on total factor productivity [3], which enhances the technical level as well. Meanwhile, technical progress exhibits strong “path dependence” [15]. 3. Model and data 3.1. Model specification Dietz and Rosa [16] reformulate the IPAT model in a stochastic form. The specification of the STIRPAT model is: (1) I i aPi b AicTi d ei ,

Rui Xie et al. / Energy Procedia 104 (2016) 227 – 232

where a is a coefficient. I, P, A, T represent environmental quality, population size, affluence and technical progress, respectively. On the basis of equation (1), there are several empirical studies [8]. We include the transport infrastructure into the model. After taking logarithms, the model takes the following form: (2) ln I D0  D1 ln Trans  D 2 ln P  D3 ln A  D 4 ln T  H , In this paper, we study the influence of transport infrastructure on urban environment and its spatial spillover effect by using the spatial Durbin model. Additionally, we incorporate energy intensity and the quadric term of GDP per capita. The model can be rewritten in the following form: ln Iit UW ln I it  D 0  D1 ln Transit  D 2 ln Pit  D 3 ln Ait  D 4 ln Tit  D 5 ln EI it  D 6 ln Ait2 , (3) T1W ln Transit  T 2W ln Pit  T3W ln A  T 4W ln Tit  T5W ln EI it  T6W ln Ait2  H it

where i and t are the indices of city and year, respectively; α0 indicates the constant term and εit is the error term; W represents the n×n order of spatial weight matrix. In this paper, the spatial weights matrix W is specified as a row-normalized binary contiguity weight matrix, with elements wij=1 if two regions share a common border, and zero otherwise. 3.2. Data The explained variable is the urban environment (I). We use industrial SO 2 emissions as a substitute index of the urban environmental pollution level. The core explanatory variable is the transport infrastructure (Trans). This paper uses the per capita area of paved roads in a city to measure transport infrastructure. In this study, we consider four control variables. First, we use total population to measure population size (P). Second, we use GDP per capita to measure affluence (A). Third, we use the capital-labor ratio to measure technical progress (T). Furthermore, total fixed capital stock data is used as a substitute for capital. Due to lack of published data, this paper references literature to estimate urban fixed capital stock [17]. Moreover, we utilize urban employment by the end of the year to evaluate the quantity of labor. Finally, due to lack of city energy data, we use electricity consumption per output unit to measure energy intensity (EI). The data are derived from Chinese City Statistical Yearbook and Chinese Statistical Yearbook during the period 2004 to 2014. Considering data integrity, this paper covers panel data of 281 cities. 4. Empirical results 4.1. Exploratory spatial data analysis of urban environment Table 1. Global Moran’ I index of industrial SO2 emissions Year

Moran’s I

Value Z

Year

Moran’s I

Value Z

2003

0.1840

5.0000

2009

0.1747

4.7499

2004

0.1666

4.5473

2010

0.1854

5.0357

2005

0.2212

6.0198

2011

0.2114

5.7242

2006

0.2015

5.4638

2012

0.2484

6.7097

2007

0.1596

4.3458

2013

0.2750

7.4175

2008

0.1871

5.0783

229

230

Rui Xie et al. / Energy Procedia 104 (2016) 227 – 232

We use global Moran’ I index to certify if there exist agglomeration phenomena of urban environmental pollution. Moran’ I index reflects the correlation coefficient between observed value and spatial lag, which ranges between -1 and 1. In this paper, we perform the spatial correlation test to analyze 281 cities’ industrial SO2 emissions, which affect city environment. The results show that value of Moran's I index and the Z-value are above 0 and significant at the 1% significance level. This indicates that China's urban environmental pollution has significant positive spatial correlation. 4.2. Estimation results of spatial Durbin model Due to the existence of spatial factors in the model, endogeneity problems may appear. Anselin [18] proposes the maximum likelihood method to estimate the model, which can effectively overcome the estimation error caused by the endogeneity. Using the model selection mechanism for spatial panel data proposed by Elhorst [19], we select the model before estimation. The results are shown in Table 2. We can see that the Lagrange multipliers (LM) and robust LM tests reject the ordinary panel model without spatial effects. The joint significance LR test shows that, at 1% significance level, the model without spatial and temporal fixed effects should be rejected. The results of the Wald test show that the spatial Durbin model cannot be transformed into spatial lag and spatial error model. The results of the Hausman test show that the fixed effect model cannot be rejected. Table 2. The test of spatial econometric model Test method

Statistical value

Probability

LM test no spatial lag

101.2022

0.0000

robust LM test no spatial lag

17.4473

0.0000

LM test no spatial error

114.6712

0.0000

robust LM test no spatial error

30.9163

0.0000

Wald test spatial lag

34.1815

0.0000

Wald test spatial error

41.3552

0.0000

Hausman test

31.5456

0.0028

Spatial fixed

4320.2349

0.0000

Temporal fixed

116.6314

0.0000

Joint significance LR test

This paper explains spatial and temporal effects of the spatial Durbin model. The results are shown in Table 3. The estimated coefficient of transport infrastructure is positive and significant, which illustrates that construction of transport infrastructure increases urban SO 2 emissions. The linear coefficient of GDP per capita is significantly positive, and the quadratic term coefficient is significantly negative, which shows the inverted U-shaped curve relationship between GDP per capita and urban environment, which supports the EKC. Technical progress has a positive impact on SO 2 emissions, which confirms that technical progress has “path dependence” [24]. As such, the coefficient of energy intensity is significantly positive, which indicates that the increase of energy intensity raises SO2 emissions. Due to the existence of spatial characteristics, LeSage and Pace [20] indicate that the spatial econometric model requires special explanations. In this paper, we use direct, indirect, and total effect decomposition methods to analyze the effects of transport infrastructure on urban environment. The results are shown in Table 4. Direct and indirect effects of transport infrastructure are both positive. This shows a positive direct and indirect impact of transport infrastructure on urban industrial SO 2 emissions.

231

Rui Xie et al. / Energy Procedia 104 (2016) 227 – 232

The impact of GDP per capita on the local environment is represented by an inverted U-shaped relationship. The direct effect of technology progress is positive, while its indirect effect is negative. This shows that the technical progress in the surrounding area has a demonstration effect on local enterprises, which enhances their technical level and reduces local pollution. The direct effect of energy intensity is significantly positive, implying that the increase of local energy intensity contributes the urban environment damage. Table 3. Spatial econometric results of transport infrastructure impact on urban environment Coefficient

t-statistic

p-value

0.1025***

3.3943

0.0006

lnP

0.2619

1.4350

lnA

0.3426***

4.3996

lnT

0.0620*

1.6672

lnEI

0.0712**

2.5407

lnTrans

2

lnA

-0.1320***

R2

0.8411

-7.1492

Coefficient

t-statistic

p-value

WhlnTrans

0.1534**

2.4601

0.0138

0.1512

WhlnP

-0.1559

-0.3971

0.6912

0.0000

WhlnA

-0.2034

-1.6213

0.1049

0.0954

WhlnT

-0.2247***

-3.4609

0.0005

0.0110

WhlnEI

-0.0118

-0.2650

0.7910

0.0000

2

0.0731***

2.8896

0.0038

WhlnA

Table 4. Direct and indirect effects of explanatory variables Direct effect

Indirect effect

Total effect

Coefficient

t-statistic

Coefficient

t-statistic

Coefficient

t-statistic

0.0954***

(3.0272)

0.1091**

(2.0108)

0.2046***

(3.7671)

lnP

0.2773

(1.4804)

-0.1955

(-0.5736)

0.0817

(0.2589)

lnA

0.3570***

(4.4100)

-0.2487**

(-2.1380)

0.1082

(1.2367)

lnT

0.0748

*

(1.9071)

***

lnEI

0.0722**

(2.4678)

lnTrans

2

lnA

***

-0.1379

(-6.8532)

-0.2042

-0.0266 0.0905

***

**

(-3.4200)

-0.1294

(-2.4686)

(-0.6540)

0.0456

(1.0223)

(3.6271)

***

-0.0473

(-2.7741)

5. Conclusions Nowadays, China is engaged in large-scale transport infrastructure construction, while the city environment faces severe challenge. Subsequently, we use a spatial Durbin model to analyze the impact of transport infrastructure on environment in 281 Chinese cities between 2003 and 2013. The conclusions of this paper are as follows. First, transport infrastructure, population size, technical progress, and energy intensity have negative direct impacts on environment. Additionally, we find an inverted U-shaped curve relationship between GDP per capita and environment. Second, transport infrastructure has a negative spatial spillover impact on environment, while the spatial spillover effect of technical progress is positive. The spatial spillover effects of population size and energy intensity are not significant. Therefore, we suggest that the government controls the transport infrastructure investment scale and optimizes the transport layout. Local governments could to enhance coordination and break administrative barriers in ecological environmental governance. Acknowledgements

232

Rui Xie et al. / Energy Procedia 104 (2016) 227 – 232

Our heartfelt thanks go to the Natural Science Foundation of China (NSFC) (71303076 and 71420107027), and the Soft science plan key projects of Hunan (2015zk2002) for funding this research. References [1] Zhang XL. Has Transport Infrastructure Promoted Regional Economic Growth? With an Analysis of the Spatial Spillover Effects of Transport Infrastructure. Soc Sci China 2012;03:60–77+206. (In Chinese) [2] Bronzini R, Piselli P. Determinants of long-run regional productivity with geographical spillovers: the role of R&D, human capital and public infrastructure. Reg Sci Urban Econ 2009;39(2):187–99. [3] Liu YH. The Spatial Spillover Effect and Its Influencing Mechanism Analysis of the Transport Infrastructure in China. Nankai University 2012. (In Chinese) [4] Xu S-C, He Z-X, Long R Y. Factors that influence carbon emissions due to energy consumption in China: Decomposition analysis using LMDI[J]. Appl Energy 2014;127:182–93. [5] Wang S, Fang C, Guan X, et al. Urbanization, energy consumption, and carbon dioxide emissions in China: A panel data analysis of China’s provinces[J]. Appl Energy 2014;136:738–49. [6] Zheng B, Zhang Q, Borken-Kleefeld J, et al. How will greenhouse gas emissions from motor vehicles be constrained in China around 2030? [J]. Appl Energy 2015;156:230–40. [7] Shi A. The impact of population pressure on global carbon dioxide emissions, 1975–1996: evidence from pooled cross1country data. Ecol Econ 2003;44(1):29–42. [8] Fan Y, Liu LC, Wu G, Wei YM. Analyzing Impact Factors of CO2 Emissions Using the STIRPAT model. Environ Impact Assess Rev 2006;26(4):377–95. [9] Wang Z, Yin F, Zhang Y, et al. An empirical research on the influencing factors of regional CO 2 emissions: evidence from Beijing city, China. Appl Energy 2012;100:277–84. [10] Liu Y, Zhou Y, Wu W. Assessing the impact of population, income and technology on energy consumption and industrial pollutant emissions in China. Appl Energy 2015;155:904–17. [11] Krugman P. Increasing Returns and Economic Geography. Journal Polit Econ 1991;99(3):483–99. [12] Zhang K, Wang DF. The Interaction and Spatial Spillover between Agglomeration and Pollution. China Ind Econ 2014;06:70–82. (In Chinese) [13] Lu M, Feng H. Agglomeration and Emission Reduction: An Empirical Study on the Impact of Urban Scale Gap to the Intensity of Industrial Pollution. J World Econ 2014;07:86–114. (In Chinese) [14] Fujita M, Krugman P R, Venables A. The spatial economy: Cities, regions, and international trade. MIT press;2001. [15] Shen M, Li KJ, Qu RX. Technological Progress, Economic Growth and Carbon Dioxide Emissions: A Theoretical and Empirical Study. J World Econ 2012;07:83–100. (In Chinese) [16] Dietz T, Rosa E-A. Rethinking the environmental impacts of population, affluence and technology. Hum Ecol Rev 1994;1:277–300. [17] Ke SZ. Spread-backwash and Market Area Effects of Urban and Regional Growth in China. Econ Res J 2009;08:85–98. (In Chinese) [18] Anselin L. Spatial Econometrics: Methods and models. Springer Science & Business Media;1988. [19] Elhorst J-P. Spatial econometrics: from cross-sectional data to spatial panels. New York: Springer;2014. [20] LeSage J, Pace R-K. Introduction to Spatial Econometrics. CRC Press;2009.