Quantifying urban land expansion dynamics through improved land management institution model: Application in Ningxia-Inner Mongolia, China

Land Use Policy 78 (2018) 386–396

Contents lists available at ScienceDirect

Land Use Policy journal homepage: www.elsevier.com/locate/landusepol

Quantifying urban land expansion dynamics through improved land management institution model: Application in Ningxia-Inner Mongolia, China

T

⁎

Yongjiao Wua,b, Suocheng Dongb, , Haosheng Huangc, Jun Zhaid, Yu Lib, Dingxuan Huanga a

Research Center of Modern Enterprise Management of Guilin University of Technology, Guilin 541004, China Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China c Department of Oceanography & Coastal Sciences, School of the Coast & Environment, Louisiana State University, Baton Rouge, LA 70803, USA d Satellite Environmental Center, Ministry of Environmental Protection, Beijing 100094, China b

A R T I C LE I N FO

A B S T R A C T

Keywords: Urban land expansion Land management institution model (LMIM) Path dependence Ningxia-Inner Mongolia region along the Yellow River China

It is of great importance to investigate mechanism of urban expansion as to policy-making and implementation of urban planning in China. The purpose of this study is to carry out a thorough investigation on quantifying dynamic process of urban land expansion through an improved Land Management Institution model (LMI), which was further applied in Ningxia-Inner Mongolia region, China. Therefore, an improved LMI model was developed by introducing the varying panel approach which can well determine the major contributors of spatial differences of urban land expansion and quantify the evolution of land management institution. Through this investigation, following findings were achieved. Firstly, the improved LMI model could efficiently quantify and determine spatial differences of endogenous factors of urban land expansion, and the land management institution evolution. The land management institution evolution presented a path dependence with a stable trajectory, which meant urban land expansion were well coincided with the evolution of land management institution. Secondly, the improved LMI model application in Ningxia-Inner Mongolia region disclosed that urban land expansion had an obvious negative effects of lock-in phase characters in land management institution evolution. Marginal effects of major influencing factors of urban land expansion contributed greatly to the differences of urban land expansion in terms of different sectors and regions. Finally, dynamics analysis of urban land expansion in this study area suggested that its urbanization had similar urbanization patterns with cities in elsewhere of China, and meanwhile it presented distinct characters. The massive conversion of land use has stimulated high-speed economic growth in the short term at the expense of eco-environment in the long term. It is suggested that governments at all levels should pay more attention to land management institution reform, industrial adjustment, technological innovation as to realize sustainable urbanization.

1. Introduction As a crucial symbol of modernization, urbanization is a natural and historical process that population and non-agricultural industries shift from rural to urban area (Lu, 2013; Lu and Chen, 2015). Over the past decades, mass urbanization has occurred in China along with its fast industrialization process. However, the urbanization in China was primarily driven by the government, which has generated many problems (Bai et al., 2014). Firstly, compared to the industrialization process, the urbanization in China was seriously lagged behind. Secondly, the growth of the urban built-up land area was faster than the urban

⁎

population growth. Thirdly, the urban construction land has been utilized in a relatively inefficient manner. Fourthly, the spatial distribution of urban areas are usually unparalleled with the carrying capacity of local environments, and the environment problems in many urban areas of China keep growing (Liu, 2008; Kuang et al., 2009; Pickett et al., 2011). Thus, aiming to address these issues and negative consequences accompanied with rapid urbanization, the central government of China has issued a number of policies and laws to regulate the land-use change, especially to constrain urban land expansion. Urban land expansion can greatly promote the economic growth. Meanwhile, it may also bring contradictions on social and

Corresponding author. E-mail addresses: [email protected] (Y. Wu), [email protected] (S. Dong), [email protected] (H. Huang), [email protected] (J. Zhai), [email protected] (Y. Li), [email protected] (D. Huang). https://doi.org/10.1016/j.landusepol.2018.06.018 Received 25 January 2018; Accepted 14 June 2018 0264-8377/ © 2018 Published by Elsevier Ltd.

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

Fig. 1. Location of Ningxia-Inner Mongolia region along the Yellow River and the artificial surface expansion from different land type.(1990–2010).

(Alonso, 1964; Kuang et al., 2009, 2014; Ding, 2003; Wang et al., 2012; Verburg et al., 1999; Kumar et al., 2011; Yue et al., 2013; Zheng et al., 2012; Dorning et al., 2015; Wu et al., 2010; Zhang et al., 2011a; Taylor, 1998; Huang et al., 2015; Huynh, 2015; Christensen, 2014). At present, some existing studies show that LMI plays a key role in driving urban land expansion (Taylor, 1998; Christensen, 2014; Huang et al., 2015; Wu et al., 2016). However, the key processes and major factors that underpin urban land expansion remains unclear. Specifically, there are some questions are yet to be answered: (1) How does the LMI work in different patterns, i.e. how to use LMI model to quantify its effects and its dynamic influences on urban land expansion? (2) What are the spatial differences of major factors on urban land expansion with regard to the evolution of LMI? In this study, our primary focus will be on the second one, i.e. the spatial differences and the process of LMI evolution. When referring to the LMI and related effects as the major factors on urban land expansion, only describing the process of the LMI evolution is obviously not enough. This is because that qualitative description without firm empirical evidences may lead to misrepresentations of the urban land expansion and misleading consequences. Thus, based on the data collected from our study area, the objectives of this study are: (1) to test whether the marginal effects of major factors with the quantified LMI on urban land expansion are varying across different cities; (2) to explore the process of LMI evolution; (3) to figure out how the local governments make their LMI. The empirical results of this study are used to explore the nature of the process of the LMI evolution, and to compare the spatial differences of major factors on urban land expansion and quantifications of the LMI evolution across our study area. Meanwhile, this study will also provide empirical implication and theoretical support for resolving the existing problems that caused by urban land expansion in China, especially from the perspective of urban planning regulations and land management policy.

environmental sustainability (Angel et al., 2011; Bai et al., 2014). Specifically, China has developed its own mode of urban economic development depending on urban land since 1990s (Liu et al., 2012b). To investigate the mechanism of mass urban land expansion in China and its effects on the content, process and results, a large number of studies have been conducted. Among those studies, there are two major progresses in understanding the effect and evolutionary mechanism of the urban land-use change (Ding, 2003; Wang et al., 2012; Wu et al., 2016). Firstly, the driving forces of land-use change has been advanced from the single economic factor to multi-factors, i.e. social, institutional, and technical factors, etc. Secondly, the dynamic mechanism of the land-use evolution has been discovered from the surface morphological evolution to inner mechanism. Specifically, some representative efforts in such fields are listed as follow: (1) The coupling-mechanism among and within urban economic development, industry structural evolution, and environmental change has been discovered. This finding has provided the theoretical foundation to optimize urban industry development and environment pollution management (Grossman and Krueger, 1995; Wan and Dong, 2012; Huang, 2010; Liu and Deng, 2009; Li, 2011). (2) The study on the effect of urban land change on environment and urban ecosystem sustainability has been furthered. Specifically, the effect from climate and environment factors have been used in urban land change research by employing the GIS and remote sensing techniques. From the similar perspective, the understanding on the interaction between urbanization, urban ecosystem, and global climate change have also been furthered (Liu and Deng, 2009; Bowler et al., 2010; Huang et al., 2011; Dai et al., 2015; Yang et al., 2010; Liu et al., 2012a; Guo et al., 2011; Nicole et al., 2014; Guo et al., 2011; Liu et al., 2012a; Yu et al., 2015). (3) The pattern of urban land expansion, the major driving factors, and the their driving mechanism has been disclosed by employing the GIS spatial analysis techniques, along with the combining the factor analyzes in terms of social-economics, institution, technology, and etc. 387

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

2. Study Area

paper (Wu et al., 2016). Therefore, some datasets of Baotou may have not been described in the followings.

This study focus on several cities locating along the upper and middle reaches of the Yellow River in Ningxia and Inner Mongolia Autonomous Regions, China. As shown on Fig. 1, the study area of this project includes eight cities, i.e. Wuzhong (WZ), Yinchuan (YC), Shizuishan (SS), Wuhai (WH), Bayan Nur (BE), Baotou (BT), Hohhot (HT) and Ordos (OS). Specifically, YC is the capital city of Ningxia Autonomous Region and HT is the capital city of Inner Mongolia Autonomous Region. The WZ and SS are the prefecture-level cities of the Ningxia Autonomous Region, and the WH, BE, BT and OS are the prefecturelevel cities of the Inner Mongolia Autonomous Region. As an important agglomeration area of minority population in the Northwestern China, such area plays a key role in China Great Western Development Movement. Those cities in the study region are facing not only a fragile ecosystem but also the less developed economy and imbalanced social and economic development status. Under the scenario of the Great West Development Movement, the central and local governments have implemented a number of measures and policies to stimulate economic development and urbanization in this area. As a result, the portion of heavy industries in local economy has shown a significant increase in this region (Fan et al., 2014;Li et al., 2015b). In fact, the economy of such region depends heavily on Mining and energy industries. Along with this heavy-industry-driven economic development, the urban lands in such area have also been expanded remarkably over the past decades which has been largely converted from the farmland, grassland and wetland. However, this heavy-industry-driven development have also generated some negative consequences, such as serious environment problems, such as air and water pollution, and ecological degradation. The fast economic growth in such area is continuously increasing the pressures on the local ecosystem and threatening the sustainability of economic development in the Yellow River basin. Hence, these challenges highlight an urgent need to deeply investigate the underlying causes of urban land expansion in the Ningxia-Inner Mongolia along the Yellow River region.

3.2. Theoretical framework 3.2.1. Path dependence theory Path dependence refers to the mechanism of increasing returns and self-strengthening in institutional evolution (North, 1994). Theories of path dependence has been widely applied in different fields (Hämäläinen and Lahtinen, 2016). It emphasizes the significance of historical background and has embodied the characteristics of institutional evolution, which means that peoples' previous behavioral decision would affect their current decision. This theory has been employed to explain why changes always happen following a particular direction (Lin et al., 2015). The classic path dependence has been divided into three phases, which are the contingent phase, the self-reinforcement phase, and the lock-in phase (Vergne and Durand, 2010). The first phase is a contingent process that the initial situation is based on unconstrained options. At this phase, individuals take rational decisions, but which may still result in unintended and irrational consequences on a systematic level. The second phase is a process of self-reinforcement phase, where a dominant action pattern is likely to emerge, rendering the irreversibility of the whole process (Miller, 1992). At the third phase, i.e. the lock-in phase, the dominant pattern becomes fixed and gains a deterministic character which will eventually lead to a path. In this study, we employed a varying-coefficient panel model to determine and quantify the process of LMI evolution. 3.2.2. The varying-coefficient panel model The empirical parts of this study are carried out based on a varyingcoefficient panel model. A longitudinal, or panel, data model is one that modelling the panel dataset, which means that there are multiple measurements on the same observations over several time periods. The panel data model enables researchers efficiently analyze the relationship within and among economic variables. Meanwhile, it can be used to build and test behavior model that is more realistic than the crosssectional model or time series model (Li and Ye, 2000). When coefficients are varying over cross-sections, the function form of such varying-coefficient panel model is (Cheng, 2005),

3. Data and methods 3.1. Data processing

Yi = Xi βi x x ... xki1 ⎡ βi1 ⎤ ⎛ 1i1 2i1 ⎞ y x1i2 x 2i2 x ki2 ⎢ ⎥ ⎡ i1 ⎤ ⎟, β = ⎢ βi2 ⎥, ⋮ y Xi = ⎜ i Yi = ⎢ i2 ⎥ , ⋮ ⎜ ⎟ ⎢⋮ ⎥ ⎢⋮ ⎥ ⎜ ⎟ ⎢ βik ⎥ ⎢ yit ⎥ ⎝ xki1 xki2... xkiT ⎠ ⎣ ⎦ ⎣ ⎦t × 1

The datasets used in this study were obtained from the City Statistical Yearbooks (1993–2014) and the City Construction Statistical Yearbooks (1993–2014) of those cities in the study areas. Socioeconomic datasets, including manufacture and services industries, population, Gross Domestic Products (GDP), per capita disposal income, policy implementation, and technological progress, were collected and trimmed to satisfy model simulation. The price-relevant data has been adjusted by consumer price index (CPI, P1992 = 100) to rule out the effect of inflation. In this study, the urban built-up areas from Yearbooks are used to represent the urban land areas, and the population of permanent residents reported by Yearbooks are considered as the urban population in corresponding regions. To compare the spatial differences across cities, Landsat ETM/TM images were collected for LUCC classification from 1990Landsat ETM/ ET (126/31–131/31) and 2010Landsat ETM/ET (126/31–131/31) datasets, where the spatial resolution is 30 m. Based on 36 scene images of 1990Landsat ETM/ET and 2010Landsat ETM/TM, land use data was derived by artificial digital interpretation (Tables 1 and 2 and Figs. 1 and 2). In Table 1, the numbers are the changed areas of different land type during the period 1990–2010. For example, the number -413.46 (row2, column2) means that the farmland has been reduced by 413.46km2 in BZ. Likewise, the number 326.59 (row4, column2) means that the farmland has been increased by 326.56 km2 in HT during the given time period. The datasets of Baotou are derived from the previously published

(1)

where yit is the numerical value of the explained variables in crosssection i and time t , xkit is the numerical value of the kth explanatory variable in cross-section i and time t ; βik , the parameter of the kth explanatory variable of the ith cross-section. 3.3. The improved LMI urbanization model This study is intended to further the research in (Wu et al., 2016), which has quantified the influencing effects of the LMI on urban land expansion and discovered the endogenous factors of urban land expansion. Specifically, This study focuses on quantifying the process of the LMI evolution, and exploring the spatial differences of the influencing effects of these endogenous factors on urban land expansion by employing the path dependence theory and the varying-coefficient panel model. 3.3.1. The varying-coefficient panel model of LMI urban land expansion The function form of model used in the study is similar to the one that was proposed in Wu et al. (2016). Meanwhile, the same and similar techniques are utilized to process and quantify the technology progress 388

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

Table 1 Changed areas of different land type from 1990 to 2010 unit:km2. Land Type City

Farmland

Forestry

Grassland

Wetland

Artificial land

Wasteland

Others

YC SS WZ BE OS HT WH

315.31 110.17 592.38 −413.46 −133.45 326.59 4.18

34.18 9.79 208.48 110.64 368.09 70.08 0.70

−344.52 −163.55 −894.47 80.95 −1141.92 −520.12 −39.32

51.22 34.66 −22.43 71.23 −30.33 17.64 −6.00

163.19 106.97 81.54 31.75 147.24 68.59 28.72

−236.42 −100.36 −68.37 116.66 882.43 37.44 14.01

17.09 2.12 109.40 7.76 −81.71 3.18 −2.93

(Note: data is obtained from TM image interpretation.).

the urban land usage increases by 10%. Likewise, the value of LMI will be decreased by one when the urban land usage decreases by 10%. The LMI index is calculated following the Eq. (2):

Table 2 Artificial land converted from farmland and grassland during the period 1990–2010 unit:km2. Land Type City

Farmland

Grassland

YC SS WZ BE OS HT WH

157.91 65.73 97.77 457.37 139.88 235.37 9.92

36.51 46.85 33.51 170.08 243.78 161.52 41.65

⎧ LMI1992 = 1 (Landt − Landt − 1) / Landt ⎨ LMIt = LMIt − 1 + 1 × Integer ⎡ ⎤ 10 % ⎣ ⎦ ⎩

(2)

where LMIt and Landt represents the LMI and urban land area index in time t , respectively. (Landt − Landt − 1) / Landt The term 1 × Integer ⎡ is used to quantify the ⎤ 10 % ⎣ ⎦ change of local governments' LMI in terms of the policy intensity. According to our pilot research, the absolute change rate of urban land area are usually less than 10% when the policy intensity holds constant. Therefore, we set 10% as the threshold of urban land area change caused by the change of local governments' LMI. Based on those calculations, we have formed an improved LMI urbanization model in this study, which can be expressed as the Eq. (3) as follow:

(Note: data is obtained from TM image interpretation.).

and LMI. To measure the urban technological progress objectively, the entropy method has been used to calculate each index weight (Table 3). As we have discovered in previous study, the implementation of a certain policy usually causes a sudden change in the corresponding variables (Wu and Dong, 2010). Thus, like our previous research, the LMI is quantified by using LMI index while assuming the value of LMI in 1992 to be one. Then, the value of LMI will be increased by one when

Lnlandi = βi1 LnInd2/ LnTec + βi2 LnInd3/LnTec + βi3 LnInc (−1) + βi 4 LnPop (−1) + βi5 LnLMI

Fig. 2. land conversion in the Ningxia-Inner Mongolia region along the Yellow River.(1990–2010). 389

(3)

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

Table 3 Indices weight of technological progress, Ningxia-Inner Mongolia region along the Yellow river. City Index

YC

SS

WZ

BE

OS

HT

WH

BT

Education investment (USD) Road areas per capita (km2/104 people) Urban transportation and communication cost (USD) R&D investment (USD) College students (per thousand of the population) Scientific research/technology service and geological prospecting personnel Foreign direct investment (USD) Proportion of foreign trade imports to exports (%) Public finance expenditure (USD) Finance income (USD) Per capita urban population disposable income (USD) Proportion of non-agriculture added value to GDP (%)

0.141 0.029 0.059 0.270 0.029 0.002 0.102 0.066 0.146 0.123 0.032 0.000

0.184 0.017 0.066 0.216 0.089 0.035 0.101 0.037 0.130 0.099 0.028 0.000

0.098 0.050 0.092 0.177 0.041 0.029 0.141 0.139 0.121 0.091 0.020 0.002

0.081 0.083 0.164 0.101 0.104 0.074 0.077 0.054 0.126 0.097 0.038 0.002

0.144 0.062 0.097 0.094 0.051 0.019 0.107 0.085 0.136 0.156 0.047 0.003

0.063 0.039 0.103 0.243 0.067 0.017 0.135 0.123 0.072 0.081 0.057 0.000

0.117 0.005 0.110 0.213 0.039 0.038 0.149 0.061 0.107 0.117 0.043 0.000

0.117 0.061 0.118 0.115 0.076 0.089 0.074 0.035 0.101 0.106 0.076 0.031

(Note: weight calculated by using the entropy method). Table 4 Parameter estimation results of the improved LMI model. Dependent variable: LnLand parameter

YC

SS

β1 (LnInd2/ LnTec )

−0.15

β2 (LnInd3/ LnTec )

0.22**

**

***

β3 (LnInc (−1))

0.22

β4 (LnPop (−1))

−0.21***

β5 (LnLMI )

0.31

***

WZ

−0.26

***

0.14*** 0.46

***

0.27*** 0.33

***

1.26

BE

−1.07*** 0.75

**

−0.05**

***

0.16

−0.11*** 0.30

OS

−0.002

**

***

−0.32***

***

0.27

***

HT

0.11

***

0.24

−0.10*** 0.21

0.07

−0.28*

**

0.05

−0.47* 0.49

WH *

0.16

0.01*

−0.18*

***

0.32

−0.05*

***

BT *

−0.02*

***

0.05*** −0.05***

0.28*

***

0.16

***

0.17***

Note: N = 168, R2 = 0.99; *,**, ***, denote significance at the 1%, 5%, and 10% level, respectively. Most of the parameter estimates are statistically significant.).

Where the suffix (-1) means the last year, the subscript i denotes the city i, thus βi1, βi2 , βi3 , βi4 , βi5 denote the coefficients of the different explanatory variables in the city i. The landi denotes the urban land area in the city i. In addition, term Ind2, Ind3, Tec, Inc, Pop, and LMI denote the secondary industry, tertiary industry, technological progress, urban population density, land management institution, respectively. Then, we employ EVIEWS6.1 to model the simulation and get the results as Table 4.

Table 5 Parameter estimation results of the path dependence model. Dependent variable: LnLMI Parameter

γ1 γ2

OS ***

0.23 0.78***

WH ***

0.59 0.80***

BT ***

0.08 0.83***

SS *

0.02 0.95***

WZ ***

0.75 0.31***

0.51* 0.86***

(Note: N = 126, R2 = 0.96; *,**, ***, denote significance at the 1%, 5%, and 10% level, respectively.).

3.3.2. The phase of LMI path dependence China's unique political, economic, and historical background are usually considered to play essential roles in shaping land management system (Lin and Kalantiari et al., 2015). Meanwhile, due to the special land management regulation in China, the local governments must ask for approval from the provincial government before they can implement changes on land use, especially for the urban land change (SCNPC, 2004). Therefore, it is very interesting to explore whether the local governments of the prefecture-level cities make their LMI depending on their previous decisions on LMI, and their provincial capital cities’ decision on LMI. To test this idea, we built a varying-coefficient panel model, as shown in Eq. (4).

⎧ LMIi = γi 1LMIHT + γi 2LMIi (−1) ⎨ LMI j = γj 1LMIYC + γj2 LMI j (−1) ⎩

BE

Fig. 3. The annual average growth rate of urban land across cities in the study area from.1992–2013.

(4)

Where the suffix (-1) means last year; the term LMIi denote the LMI of the city i in the Inner Mongolia Autonomous region including BE, OS, WH and BT. The LMI j denote of the LMI of city j in the Ningxia Autonomous Region including SS and WZ; the LMIHT , LMIYC denote the LMI of the city HT and YC, respectively. This mode, thereby, is used to explore whether the path dependence of the LMI in the prefecture-level cities exists and how the local governments of the prefecture-level cities formulate their LMI. Then, we employ EVIEWS6.1 to model the simulation and get the results as Table 5.

4. Results and discussions 4.1. Spatial pattern of Urban Land expansion Fig. 3 show that the urban land has been rapidly expanded in the study area during the period 1992–2013. The annual average growth rate of land urbanization reaches 4.28%, and the annual average growth rate of population urbanization was 0.96% during the same 390

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

Fig. 6. The annual average growth rate of the secondary industry across cities in study area from.1992–2013.

during 1992 to 2013. In general, this is a great development. However, the actual growth rate varies a lot across cities. Specifically, the maximum mean annual growth rate of the secondary industry happened in the OS (22.15%), which leads to an 57 times increase in the total valueadded from the secondary industry during our 22-year study period. And the minimum mean annual growth rate of the secondary industry in the WZ also reaches 11.70%. Meanwhile, technological progress has also been greatly advanced in the study area, as shown in Fig. 7. Generally, the technological progress has directly improved the resources' allocation of the secondary industry, then affected the inputs of urban land in the secondary industry. The estimation results of the panel model are analyzed as follow. At first, the marginal effects of the secondary industry with technological progress on urban land expansion show that it is a major factor but not a critical one compared to other major factors in most cities in our study area, except in WZ and HT. Secondly, the marginal effects of the secondary industry with the technological progress show big differences across these cities, varying between -0.002 and 1.26. Thirdly, there are three cities (YC, SS, and BE) have the negative marginal effects of the secondary industry with the technological progress on urban land expansion. These results imply that the growth modes of the secondary industry in those cities remain intensive ones, and the technological progress have enhanced the efficiency of land use and allocation of resources. Fourthly, WZ, HT, OS, and WH all have the positive marginal effects of the secondary industry with the technological progress on urban land expansion. This indicates the growth mode of the secondary industry in these four cities are relatively extensive, especially in WZ and HT. The marginal effects are 1.26 and 0.24 in WZ and HT, respectively, implying that the development of the secondary industry largely depends on the increase of urban land supply, rather than improving the efficiency of allocating resources.

Fig. 4. The annual average growth rate of population urbanization across cities in the study area from.1992–2013.

period (Fig. 4). The average ratio of the urban land growth rate to the urban population growth rate reached 5, which is far exceeds a widely accepted reasonable level of 1.12 (Yao et al., 2009; Li et al., 2015a). Particularly in WZ and OS, the areas of urban land have been expanded around by eight and seven times with the annual average growth rate of 12.06% and 11.73% in WZ and OZ, respectively. During the same period, the annual average growth rate for WH and BT are 1.39% and 1.29%, respectively, which are the slowest among all cities in our study area. Moreover, most of urban land has been converted from grassland and farmland, as shown on Table 2 and Fig. 2. Specifically, there were 1163.5km2 of farmland and 733.9km2 of grassland have been converted to artificial surfaces. Tables 4 and 5 also examine the major factors of urban land expansion proposed by Wu et al. (2016). They are related to the secondary and tertiary industry with technology progress, per capita disposable income of urban residents, urban population density, and local LMI. Meanwhile, the marginal effects of major factors on urban land expansion show a varying pattern across these cities in our study area, as shown on Fig. 5. Moreover, Fig. 5 also reveal that the LMI plays a critical role during the process of the urban land expansion because its effects on urban land expansion are all large positive numbers, which indicates large scale marginal effects. Especially in OS, the marginal effects reaches 0.49, which means that the urban land would be expanded 0.49 percent when the intensity of LMI increases by one percent, holding other variables constant. 4.2. Spatial differences of marginal effects of major factors on urban land expansion

4.2.2. The tertiary industry with the technological progress During the period between 1992–2013, the tertiary industry has experienced a great development among these cities (Fig. 8). The city

4.2.1. The secondary industry with the technological progress Fig. 6 shows that the average annual growth rates of the secondary industry of our objective cities vary between 11.70% and 22.15%

Fig. 7. The trend from.1992–2013.

Fig. 5. The marginal effects of major factors on urban land expansion. 391

of

technological

progress

indices

across

cities

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

Fig. 8. The annual average growth rate of the tertiary industry across cities in the study area from.1992–2013.

Fig. 10. Comparison of per capita housing area between 1992 and 2012 across cities in study area.

with the fastest growth rate in the tertiary industry is OS, which has an annual growth rate of 29.64%. The lowest one is WZ, whose average annual growth rate is 10.63%. Meanwhile, the results disclose that the spatial differences of the marginal effects of the tertiary industry with technological progress on urban land change are significant. The marginal effects are positive in YC and SS with 0.22 and 0.14, respectively. This reflects that the development of the tertiary industry development in these two cities are extensive. On the contrary, the marginal effects are negative. In detail, the estimated values of marginal effects of the tertiary industry development in WZ, BE, OS, HT, WH and BT are -1.07, -0.05, -0.10, 0.28, 0.18 and -0.02, respectively. These negative numbers reflect that the development of the tertiary industry is relatively an intensive growth. The negative marginal effects also imply that the land inputs in the tertiary industrial were over-supplied, and the development of the tertiary industry is not catching up with the speed of urban land expansion in those cities. This finding indicates that the development mode of the tertiary industry should be further revised in WZ, HT, WH, OS and BE.

land expansion across these cities. Secondly, the per capita disposable income of urban residents has a positive and lag effect on urban land expansion in the whole study area. Generally, the increase of disposable income directly stimulates households to improve their living of quality. One crucial improvements in terms of living quality is a better housing condition, e.g. better appliances and larger rooms. The demand in improving housing condition, thereby, will cause the increased supply of construction land to meet such demand. Thus, the rise in construction land can also be concluded as a directly result of higher per capita disposable income. Thirdly, the residents in the whole region may have strong willingness to improve their housing conditions, which directly caused an over-consumption of housing areas in 2012. In fact, the per capita disposable income for these cities were between USD 2810 and USD 5218 in 2012. With this level of personal income, the housing condition in our study area reached the affluent type in 2012. 4.2.4. Urban population density The estimation results show that the effects of urban population density on urban land expansion present big differences across these cities. Firstly, the marginal effects of urban population density on urban land expansion show an obviously spatial difference across these cities, varying from −0.47 to 0.28. Secondly, the negative marginal effects are happened to YC, WZ, BE, OS, HT, and BT. These results imply that the urban land areas would be reduced as the density of urban population increases. Thirdly, the marginal effects are positive for SS (0.27) and WH (0.28), denoting a one percent increase in urban population density may cause a 0.28 percent increase in urban land expansion in SS, and 0.28 percent in WH, holding other variables constant. These results imply that the urban population density may cause the increases of urban land in these two cities. In theory, the urban population density will continue to rise quickly with urban economic development in the developing countries and regions (Bhatta, 2010). However, as Fig. 11 show, the urban population density in 2012 was less than in 1992 for most of cities in our study area, except in BE and WH. Moreover, in most cities in our study area, the urban population density have not increased along with the economic development during the period 1992–2013 (Fig. 11), even though this study area is less developed region and expected to have a urban population agglomeration. Additionally, the urban population density in our study area are less than China’s national standard of 1.00 × 10 4 people per km2 (Lu and Chen, 2015). Specially in SS, the population density was just about 0.44 × 10 4 people per km2 in 2013 (Fig. 12).

4.2.3. Per capita disposable income of urban residents Fig. 9 shows that the annual average growth rates of the per capita disposable income are between 5.91% and 11.08% among cities in our study area, which are great achievements in general. As shown in Fig. 10, accompanied with the increase in per capita disposable income, the areas of urban per capita housing construction have been gradually increased in the study area. In detail, the per capita disposable income of urban residents have increased by 3 to 8 times across cities. The housing conditions for HT, BE,WH, and BT have been improved from the low rent type to the affluent type. Furthermore, the housing conditions for SS, YC, OS, and WZ have been improved from the economic type to the affluent type. There are several conclusions can be extract from those empirical results aforementioned. Firstly, there shows obvious spatial differences on the marginal effects of urban per capita disposable income on urban

4.2.5. The LMI The estimation results also prove that the LMI plays a critical role during the process of urban land expansion. Meanwhile, the actual effects of LMI on urban land expansion vary a lot among these cities. The estimated marginal effects of the LMI on urban land expansion for YC,

Fig. 9. The annual average growth rate of the dispose income per capita across cities in the study area from.1992–2013. 392

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

Table 6 The land management policies of China central governments. Year

the China Central Government's Land Management Policy

1994 1998 1999 2001

the Basic Farmland Protection Regulation issued the New Land Administration Law issued the New Land Administration Law implemented the Notice on Strengthening the Management of State owned Land Assets issued the Notice on Strengthening the Management of State owned Land Assets implemented the Maintain both Economic Development and Land Resources Conservation issued the "decision on deepening the reform and strengthening strictly land management" issued the Notice on Adjusting the Policy of Compensation for the Use of Land for New Construction Land issued the Notice on Adjusting the Policy of Compensation for the Use of Land for New Construction Land implemented Notice on Accelerating Distribution of Land Registration Certificate in Rural Areas, Notice on Land Change Survey and Remote Sensing and Rules on National Land Change Survey (Tentative) the Emergency Notice on strictly Prohibiting the Illegal Land Acquisition issued and implemented; the Application of Adorable Housing Land approved

2002 2003 2004 2006 2007

Fig. 11. Comparison of urban population density between 1992 and 2013.

2011

2013

administrative performances evaluation results of local officials under the current evaluation system in China. At present, the evaluation system for administrative performance evaluation in China is largely depending on the absolute numbers of local gross domestic product (GDP). In addition, the evaluation is only based on the administration performance of officials during their current tenure, but does not account for any of the post-tenure review. Thus, even though many policies that suppress the short-run economic growth may lead to a better long-term outcome, the local government officials still lack motivation to apply it because those policies is contradicting to their personal interests. In this case, some local governments generally take contrary reactions such as Fig. 13 and Table 6 show. Fig. 12. Urban land areas per capita (1992–2013).

4.3. The lock-in phase of LMI path dependence

SS, WZ, BE, OS, HT, WH and BT, are 0.31, 0.33, 0.30, 0.27, 0.49, 0.16, 0.16, and 0.17, respectively. These results imply that a one percentage increase in the intensity of LMI will lead to a 0.31, 0.33, 0.30, 0.27, 0.49, 0.16, 0.16, and 0.17 percent increase in urban land area in YC, SS, WZ, BE, OS, HT, WH and BT, respectively, if the other variables remain unchanged. Municipalities have core roles in the process of urban development and often want to stimulate the urban development by changing land use (Finn, 2014). In China, the urban and rural lands are administrated separately. The primary land markets in China are monopolistic markets because the local governments are the only supplier of land (Liu and Zhou et al., 2012). This fact leads to different land-use patterns and changes associated with the spatial-temporal changes (Wang et al., 2012). Thus, manipulation of supply of urban land on the primary and secondary market has become an important measure to utilize the land policy as a tool of Marco-economic control in China. This urban-rural differentiated, dual-layer LMI system has made a great contribution to the rapid industrialization and urbanization in China since the 1990s. However, in the meantime, this dual-layer LMI has also caused some negative consequences, such as the waste of land resources and inefficient land use (Liu et al., 2012b). To prevent those negative consequences from being continuously worsen, Chinese central government has issued many strict policies and regulations, as summarized in Table 6. However, these policies and regulations are not be able to deploy smoothly in many places because most of these measures would slow down the short-term economic growth, and thereby result in material injures to the personal interest of some local government officials. Specifically, slowed economic growth rate will directly hurt the

Table 5 and Fig. 14 show that the path dependence of LMI has existed and located on lock-in phase during the period 1992–2013. The local governments of prefecture-level cities make their annual LMI depending on their own last year LMI and the LMI of their provincial capital governments with lock-in trajectories. Moreover, the estimation results indicate that the trajectories of LMI path dependence present interesting patterns among these cities. Firstly, in cities within the Ningxia Hui Autonomous Region (SS and WZ), the coefficients of their own last year LMI are 0.31 and 0.86, respectively. The coefficients of the LMI of the provincial capital (YC) are for these two cities are 0.75 and 0.51, respectively. Secondly, within the Inner Mongolia Autonomous Region, the coefficients of the LMI of provincial capital (HT) show

Fig. 13. Land management institution indices (1992–2013). 393

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

economic and political interests of the provincial capital governments officials before issuing or implementing a new land management policy or regulation. This is because the provincial capital governments generally take the reactions, which are commonly opposite to the original intentions of the central government of China. For example, in 2011, the central government has issued two regulations, i.e. the Notice on Accelerating Distribution of Land Registration Certificate in Rural Areas, Notice on Land Change Survey and Remote Sensing and Rules on National Land Change Survey (Tentative), aiming to constrain the growing negative consequences from land use change, especially to constrain the conversion of farmland and ecological conservation zone to artificial surface (Wang, 2013). However, the HT government directly adjusted the administration division of the town of Shaerqing, which has a total area of 200km2, put it under the administration of the Hohhot economic and technological development zone in 2012. As a direct result, the urban land area of HT has been expanded by 36km2 and the urban land growth rate achieved 20.69% in the same year. In addition, the HT government also planned in 2012 that it will convert 100km2 of rural land to industrial land before 2020. Finally, Chinese central government should optimize the performance evolution system of the government officials. The new performance evolution system must be able to evaluate officials on the basis of both their current tenure and the long-term development status, which will balance off some officials’ willingness to boost short-term economy growth rate at the cost of causing many long-term negative consequences. To sum up, we have the following findings from this study: (1) the marginal effects of the secondary and tertiary industry with technological progress on urban land expansion present big differences across cities. Both negative and positive effects are happened in some cities; (2) the marginal effects of urban population density on urban land expansion also vary across cities; (3) over the whole study area, the areas of urban land increases as the urban per capita disposable income increases; (4) the path dependence of the LMI evolution exists and locates on the lock-in phase; (5) the local governments make their own LMIs depended on their own last year LMI and their provincial capital LMI; (6) the intensities of the LMI path dependence present different characters among and between different provincial regions.

Fig. 14. The trajectories of the LMI path dependence.

big differences across these cities. In detail, the parameter estimates are 0.23, 0.59, 0.08, and 0.02 in BE, OS, WH and BT, respectively. In the meantime, the coefficients of their own last year LMI are almost same, i.e. 0.78, 0.80, 0.83, 0.95 in the BE, OS, WH and BT, respectively. The concept of lock-in usually was assigned with negative interpretation as rigidification and inflexibility (Arthur, 1989; Lin et al., 2015; Lahtinen and Hämäläinen, 2016). In this study, the lock-in is also treated as a negative concept. The negative consequences would appear within the process of the lock-in trajectory. Thus, we would naturally want to minimize the possibility and effects of lock-in. In this study, the estimation results show that the lock-in process for the prefecture-level cities was enforced by their own last year LMI and the provincial capital LMI. However, it is difficult to block up the lock-in trajectories when the lock-in process are controlled by the governments, unless a total reform can be introduced (Wu, 2011). High switching cost, high sunk cost, and monopoly are usually considered as the crucial reason resulting in lock-in (Lin et al., 2015). Over the 1990s, local governments in China have converted the traditional land functions (production and living function) to capital function for urban construction purpose. These mass land use conversion has made great contributions to the unprecedented economic growth and urbanization, but also generated numerous negative consequences. Thus, Chinese central government has also issued a number of land management policies as response. With the restriction of those policies, it is a low-cost choice for local governments to follow their own last year LMI when making their current LMI, compared to the high switching cost and risk of changing path trajectories. Meanwhile, the local governments, functioning as the state trustee, hold the property rights of urban land directly. This status of land ownership greatly reduces executive cost of applying the LMI in the local market, which also easily results in lock-in. Moreover, the local governments prefer to follow their provincial capital LMI to make their own LMI, which also works as a reason of lock-in. This is because most of changes on urban land use in the prefecture-level cities must be upon formal approval from the provincial capital governments (SCNPC, 2004). In fact, this finding was also confirmed by Zhao et al. (2013), who found that local governments of prefecture-level cities prefer to follow the behavior of their provincial capital governments in making LMI, while they also tried to make their LMI paralleling with other same level cities. To minimize the negative consequences caused by lock-in, the most important task for Chinese central government is to introduce a total reform to resolve these trajectories and switch their LMI trajectories. In this study, we are providing some suggestions with the central government of China. Firstly, the Chinese central government should pay more attentions on guiding the provincial capital LMI because the prefecture-level cities prefer to follow the LMI of those provincial capital cities. Secondly, Chinese central government should consider the

5. Conclusions The process of urban land expansion in Ningxia-Inner Mongolia region along the Yellow River presents the special and common characters of urbanization in China. This region is facing many challenges, one of which is the urban land use (Dong and Li, 2010). In this study, focusing on urban land expansion from the perspectives of spatial difference and path dependence, we have thoroughly investigated the spatial differences of the major factors affecting urban land expansion, and the process of the LMI evolution. The empirical results of this study show that urban land in the study area has been greatly expanded, and most of the newly-added urban land has been converted from farmland and grassland. Meanwhile, to avoid the violating the “zero net loss of farmland policy” in China, the local governments have converted a large amount of grassland to farmland to offset the farmland area that has been converted to urban land. In the short run, this massive conversion of land use has produced a high-speed economic growth. However, in the long run, these land conversions could negatively affect the local environment and bio-diversity, even the food security. The secondary and tertiary industry with technological progress, per capita disposable income of urban residents, urban population density, and LMI are also the major factors influencing urban land expansion in this study area. Meanwhile, their marginal effects are varying in large magnitude across these cities due to the different social-economic development status. As Li et al., 2015a mentioned, the local governments would adjust the residential, commercial, and industrial zoning according to the local social-economic status. Generally, the local 394

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

governments refer to the local social-economic status to adjust the urban land inputs among different social-economic departments, such as industrial land, housing construction land. Therefore, the local governments should pay more attentions on the local industrial transformation and technology update, and guide the local housing with progressive consumption concepts, to shift away from the extensive economic growth mode, which relies on more and more land inputs. In addition, the LMI in our study area presents path dependence, locating on lock-in phase with a stable trajectory. Local governments prefer to follow their own last year LMI and their provincial capital LMI to formulate their LMI. It is difficult to switch these path dependence trajectories because they are controlled by the local governments and provincial capital governments. Meanwhile, the provincial capital governments form their LMI mainly based on both the policies issued/ implemented by the central government and their own social economic situation. Under the current evaluation system of administration performance in China, officials’ administrative performance is highly depends on the numbers of local GDP. Thus, the local governments are generally making and imposing the LMI to boost short-term GDP growth to meet their own interests, while ignoring the possible negative consequences in the long run. The same story also happened to the provincial capital governments, who generally prefer to make their LMI to protect their own interests, which are sometimes opposite to the central government’s. In fact, officials’ interests on short-term GDP growth won’t change unless there is an ex post investigation targeting on those policy makers who have made the LMI which leaded to negative consequences in the long run. Understanding the relationship of the LMI among the central government of China, provincial capitals, and prefecture-level cities should be regarded as an initial step in controlling urban sprawl, achieving ecosystem balance, and realizing new-type urbanization. The findings of this study may further our understanding on the major factors leading to the urban land use change and the LMI evolution process. They may also be used as a reference for making land use policy and urban land use planning in the process of China's new-type urbanization development. For future research, a possible direction to explore is their interaction process in land use change among three level of governments, i.e. the central government, provincial capital governments, and prefecture-level cities’ governments based on game theory.

cities: a systematic review of the empirical evidence. Landsc. Urban Plan. 97, 147–155. Cheng, H., 2005. Analysis of Panel Data. Peking University Press, pp. 141–185. Christensen, F.K., 2014. Understanding value changes in the urban development process and the impact of municipal planning. Land Use Policy 36, 113–121. Dai, M., Zhang, J., Wang, L., et al., 2015. A review on impacts of land use/land cover change on water resources in karst areas. Ecol. Sci. 34 (3), 189–196. Ding, C., 2003. Land policy reform in China: assessment and prospects. Land Use Policy 20, 109–120. Dong, S.C., Li, X., et al., 2010. Study on the strategy of eco-economic zone in the HanxiGansu-Ningxia-Inner Mongolia region along the Yellow River. Geog. Res. 29 (2), 204–213 (in Chinese). Dorning, M.A., Koch, J., Shoemaker, D.A., et al., 2015. Simulating urbanization scenarios reveals tradeoffs between conservation planning strategies. Landsc. Urban Plan. 136, 28–39. Fan, H., Liu, W., Wu, Z., 2014. Spatio-temporal characteristics of internal coordination of intensive urban land use: a case study of the downtown of Wuhan. Sci. Geog. Sin. 34 (6), 696–704 (in Chinese). Finn, K.C., 2014. Understanding value changes in the urban development process and the impact of municipal planning. Land Use Policy 36, 113–121. Grossman, G.M., Krueger, A.B., 1995. Economic growth and the environment. Q. J. Econ. 110, 353–377. Guo, Z., Wang, S., Cheng, M., et al., 2011. Assess the effect of different degrees of urbanization on land surface temperature using remote sensing images. Procedia Environ. Sci. 8, 962–969. Hämäläinen, R.P., Lahtinen, T.J., 2016. Path dependence in operational research—How the modeling process can influence the results. Oper. Res. Perspect. 3, 14–20. Huang, J., 2010. Empirical study on environmental pollution and urban economics is growth: a simultaneous equations approach. Finance Trade Res. 5, 8–16 (in Chinese). Huang, J., Li, Q., Hong, H., et al., 2011. Preliminary study on linking land use &landscape patter and water quality in the jiulong river watershed. Environ. Sci. 32 (1), 64–72. Huang, Z., Wei, Y.D., He, C., et al., 2015. Urban land expansion under economic transition in China: a multilevel modeling analysis. Habitat Int. 47, 69–82. Huynh, D., 2015. The misuse of urban planning in Ho Chi Minh City. Habitat Int. 48, 11–19. Kuang, W., Liu, J., Shao, Q., et al., 2009. Spatio-temporal patterns and driving forces of urban expansion in Beijing central city since 1932. J. Geo-Inf. Sci. 04, 428–435 (in Chinese). Kumar, A., Pandey, A.C., Hoda, N., et al., 2011. Evaluating the long-term urban expansion of Ranchi urban agglomeration, India using geospatial technology. J. Indian Soc. Remote Sens. 39 (2), 213–224. Lahtinen, T.J., Hämäläinen, R.P., 2016. Path dependence and biases in the even swaps decision analysis method. Eur. J. Operat. Res. 249 (3), 890–898. Li, S., 2011. Urbanization, industrial structure and environment pollution. Res. Financial Econ. Issues 06, 38–43 (in Chinese). Li, Z.N., Ye, A.Z., 2000. Advanced Econometrics. Tsinghua University Press (in Chinese). Li, F., Liang, J., Keith, C., 2015a. Urban land growth in eastern China: a general analytical framework based on the role of urban micro-agents’ adaptive behavior. Regional Environ. Change 15, 695–707. Li, J., Dong, S., Li, Y., et al., 2015b. Driving force analysis and scenario simulation of urban land expansion in Ningxia-Inner Mongolia area along the Yellow River. J. Nat. Resour. 30 (9), 1472–1485 (in Chinese). Lin, Q., Mohsen, K., Abbas, R., et al., 2015. A path dependence perspective on the Chinese cadastral system. Land Use Policy 45, 8–17. Liu, Y., 2008. Dynamic econometric analysis of the relationship between urbanization and ecological environment in Jiangxi Province. Resour. Sci. 30 (6), 829–836 (in Chinese). Liu, J., Deng, X., 2009. Progress of the research methodologies on the temporal and spatial process of LUCC. Chin. Sci. Bull. 54 (21), 3251–3258. Liu, L., Li, J., Zhuang, H., 2012a. An ecosystem service valuation of land use change in Taiyuan City, China. Ecol. Modell. 225, 127–132. Liu, S., Zhou, F., Shao, T., 2012b. Reform of Land System and Transition of Development Pattern. China development press (in Chinese). Lu, D., 2013. Research on the urbanization’s framework and content in geography. Acta Geogr. Sin. 33 (8), 897–901 (in Chinese). Lu, D., Chen, M., 2015. Several viewpoints on the background of compiling the "national New urbanization planning, 2014-2020. Acta Geogr. Sin. 70 (2), 179–185 (in Chinese). Miller, D., 1992. Environmental fit versus internal fit. Org. Sci. 3, 159–178. Nicole, W., Dagmar, H., Ulrich, F., 2014. Zooming into temperature conditions in the city of Leipzig: how do urban built and green structures influence earth surface temperatures in the city? Sci. Total Environ. 496, 289–298. North, D.C., 1994. Institutions, Institutional Change and Economic Performance. Shanghai Sanlian Press, Shanghai. Pickett, S.T.A., Cadenasso, Grove, M.L.J.M., et al., 2011. Urban ecological systems: scientific foundations and a decade of progress. J. Environ. Manage. 92 (3), 331–362. The Standing Committee of the National People’s Congress (SCNPC), 2004. Land Administration Law of the People’S Republic of China. China Legal Publishing House (In Chinese). Taylor, N., 1998. Urban Planning Theory Since 1945. SAGE Publications, London. Verburg, P.H., Veldkamp, A., Fresco, L.O., 1999. Simulation of changes in the spatial pattern of land use in China. Appl. Geogr. 19 (3), 211–233. Vergne, J.P., Durand, R., 2010. The missing link between the theory and empirics of path dependence: conceptual clarification, testability issue, and methodological implications. J. Manage Study 47, 736–759. Wan, Y., Dong, S., 2012. Study on interactive coupling mechanism of industrial structure

Funding We would like to acknowledge the funding of the National Natural Science Foundation of China (41,761,112, 41,771,182, 71,662,008), and the Key Research Institute of Philosophies and Social Sciences in Guangxi Universities (16YB005). Acknowledgments We would like to thank all the anonymous reviewers for their helpful comments. In addition, we also thank the financial support from the National Natural Science Foundation of China (41761112; 41771182; 71662008). References Alonso, W., 1964. Location and Land Use: Toward a General Theory of Land Rent. Harvard University Press, Cambridge, Mass. Angel, S., Parent, J., Civco, D.L., et al., 2011. The dimensions of global urban expansion: estimates and projections for all countries, 2000-2050. Prog. Plan. 75 (2), 53–107. Arthur, W.B., 1989. Competing technologies, increasing returns, and lock-in by historical events. Econ. J. 99 (394), 116–131. Bai, X., Shi, P., Liu, Y., 2014. Society: realizing China’s urban dream. Nature 509, 158–160. Bhatta, B., 2010. Causes and Consequences of Urban Growth and Sprawl. In: Analysis of Urban Growth and Sprawl from Remote Sensing Data. Advances in Geographic Information Science. Springer, Berlin, Heidelberg, pp. 17–36. Bowler, D.E., Knight, L.M., Pulling, A.S., et al., 2010. Urban greening to cool towns and

395

Land Use Policy 78 (2018) 386–396

Y. Wu et al.

Yang, Y., Gong, J., Zhou, Q., et al., 2010. Impacts of landscape pattern on urban expansion: a case study of Beijing city. J. Nat. Resour. 25 (2), 320–329. Yao, S., Chen, S., Wu, J., et al., 2009. Spatial expansion patterns of Chinese big cities-the case of Suzhou. Sci. Geogr. Sin. 29 (1), 15–21 (In Chinese). Yu, Z., Guo, Q., Zeng, Y., He, Z., 2015. Research progress on urban land use/cover change and its eco-environmental effects in urbanization process. Ecol. Sci. 34 (6), 193–200. Yue, W., Wang, R., Fan, P., 2013. Spatial patterns analysis of urban expansion in Hangzhou city. J. Zhejiang Univ. 5 (40), 596–605. Zhang, L., Lei, J., Li, X., et al., 2011a. The features and influencing factors of urban expansion in China during 1997-2007. Prog. Geogr. 30 (5), 607–614. Zhao, J., Chen, L., Xue, L., 2013. The prototype role of the local government, the interest choice and the behavior difference—A local government theory based on the research of the policy process. Manage. World 02, 90–106. Zheng, K., Xu, X., Zhuan, X., et al., 2012. Spatial-temporal characteristics and future prediction of urban expansion in Shanghai. J. Geo-Inf. Sci. 14 (4), 490–496.

and environmental quality: a case study of gansu province. Areal Res. Dev. 31 (5), 117–121. Wang, W., 2013. Chinese Land management system: Status quo, problems and reforms. J. Nanjing Agric. Univ. Soc. Sci. Ed. 13 (4), 76–82 (in Chinese). Wang, J., Chen, Y., Shao, X., et al., 2012. Land-use changes and policy dimension driving forces in China: present, trend and future. Land Use Policy 29, 737–749. Wu, J.L., 2011. How to break path dependence. New. Econ. Wkly. 10, 14–15 (In Chinese). Wu, Y.J., Dong, S.C., 2010. Dynamic change patterns of farmland in China from the perspective of resource economics. China Popul. Resour. Environ. 5 (20), 5–8 (in Chinese). Wu, D.Q., Liu, J., Wang, S.J., et al., 2010. Simulating urban expansion by coupling a stochastic cellular automata model and socioeconomic indicators. Stoch. Environ. Res. Risk Assess. 24, 235–245. Wu, Y., Dong, S., Zhai, J., et al., 2016. Land management institution as a key confinement of urbanization in Baotou, China–application of proposed endogenous urbanization model. Land Use Policy 57, 348–355.

396