URBANIZATION, LAND DEVELOPMENT, AND LAND FINANCING ... · urbanization with panel data covering 1999–2009. Empirical findings and policy discussions will follow before conclusions

URBANIZATION, LAND DEVELOPMENT, AND LANDFINANCING: EVIDENCE FROM CHINESE CITIES

LIN YESun Yat-sen University

ALFRED M. WUThe Hong Kong Institute of Education

ABSTRACT: China’s urbanization is significant worldwide. This process is characterized by under-urbanization of population and fast urban land expansion. The driving forces behind this expansion andtheir rationale are not fully understood and empirically tested. This study fills this gap by analyzing paneldata from 1999–2009 for all 286 prefecture-level cities in China. The findings reveal that land financing,using different measures, significantly contributed to land urbanization in China. Economically strongercities with higher real estate investment more aggressively pushed for land urbanization. The true pur-pose of urbanization should be improving the living standard, not to generate revenue. It is suggestedthat urbanization can serve its justified goals only if fiscal and political relations between central andlocal governments can be adjusted. As more data become available, future studies are encouraged tofurther explore the subject by investigating additional factors and the latest trend of urbanization inChina.

Nobel laureate in economics Joseph E. Stiglitz has cited urbanization in China, along with high-tech development in the United States, as the two most important issues that will shape the world’sdevelopment during the 21st century. In 2011 China’s urban population surpassed that of its ruralareas for the first time in the country’s history, following three decades of astonishing economicdevelopment. China’s urban population proportion increased from 18% in 1978 to 52% in 2012. It isestimated that China’s urban population will exceed 60% in 2020 and 70% in 2050. It can be seen inFigure 1 that between 1980 and 1995 China’s urban and rural populations both increased, with the

urban population growing at a much faster rate. Between 1995 and 2010, China’s rural populationexperienced a decline from 860 million to 671 million, recording a decrease of 189 million. In thesame time period, the urban population grew from 352 million to 670 million, recording an increaseof 318 million.

In fact the percentage of urban population has been a widely used measurement of China’surbanization in many media reports and academic studies. However, this measurement requires morecareful examination. China’s dramatic urban and rural demographic changes can be attributed totwo reasons. First, millions of rural residents in China have migrated to cities since the country’sopening-up and reform in the late 1970s. It is estimated that in 2011 there were at least 250 millionmigrant workers in China, a 12% growth from 2008 (National Bureau of Statistics, 2012). It isextremely difficult for most migrants to obtain urban household registration status (hukou), so theyare not entitled to urban welfare coverage, such as pensions, health care, and favorable educational

Direct correspondence to: Lin Ye, Center for Chinese Public Administration Research, School of Government, Sun Yat-senUniversity, 135 Xiangangxi Road, Guangzhou, China 510275. E-mail: [email protected]; or Alfred M. Wu, Department ofAsian and Policy Studies, The Hong Kong Institute of Education, 10 Lo Ping Road, Tai Po, New Territories, Hong Kong.E-mail: [email protected].

JOURNAL OF URBAN AFFAIRS, Volume 36, Number S1, pages 1–15.Copyright C© 2014 Urban Affairs AssociationAll rights of reproduction in any form reserved.ISSN: 0735-2166. DOI: 10.1111/juaf.12105

2 II JOURNAL OF URBAN AFFAIRS II Vol. 36/No. S1/2014

FIGURE 1

Demographic Change and Urbanization in China, 1950–2010

Note: Population figures are in millions.Source: China Statistical Yearbook, various years.

opportunities. Second, tens of millions of other rural residents had their farm land taken by localgovernments and their resident status administratively changed from rural to urban. Although theyare granted urban resident status, their levels of welfare coverage are far below that of originalurban residents. These two types of rural to urban demographic transition contribute significantly toChina’s rapid “urbanization.” However, the physical presence in urbanized areas does not portendmuch meaning of urbanization for these “urban populations” since they have fundamentally differentwelfare coverage, employment opportunities, and daily lifestyle, as compared to China’s official urbanresidents (Ye, 2011a). As described by Chan (2010), China’s incomplete urbanization is achievedmainly by “allowing ‘temporary’ migration (of a ‘floating population’) to cities but denying themigrants access to urban welfare and many other benefits” (p. 66).

Therefore, the demographic transition can hardly be a meaningful measure of China’s urbanization.This article postulates that urban land expansion is a more valid measurement of urbanization inChina. More importantly, this article tests the research hypothesis that land financing has drivenland-centered urbanization in China and addresses its policy implications. The remainder of thisarticle will first review the literature on land urbanization in China. A quantitative database of all286 prefecture-level Chinese cities,1 mainly based on city demographics, socioeconomic statistics,and local government fiscal figures, will be employed to test the link between land financing andurbanization with panel data covering 1999–2009. Empirical findings and policy discussions willfollow before conclusions are drawn.

LAND-CENTERED DEVELOPMENT AND LAND FINANCING IN CHINA

Urbanization in China is defined by Gu and Wu (2010) as “a complex and multifaceted processinvolving population migration from rural to urban areas, rural and urban land conversions, spatialreconfiguration of settlements, and changing governance” (pp. 1–2). Land-centered development hasbeen identified as the most significant feature of China’s urbanization (Heikkila, 2007; Lin, 2007;Xu, Yeh, & Wu, 2009). Lin (2007) argues that rapid outward development of Chinese cities wasdriven by a city-centered urbanization model as a strategy of “place-making” and “place-promotion”(p. 1832; see also He & Wu, 2005; Shen & Wu, 2012). Urban land was exploited as a source ofcapital formation. Li, Xu, and Li (2010) further explain the economic and fiscal nature of such aland development process. In China, land ownership is divided. Urban land is fully owned by thestate while rural land is collectively owned by villages. Heikkila (2007) depicts different approacheson how land can be urbanized in Chinese cities. Municipal governments acted as the representatives

II Urbanization, Land Development, and Land Financing in China II 3

FIGURE 2

Built-up Areas in Chinese Cities (1000 Sq. Km.), 1996–2010

Source: China Statistical Yearbook, various years.

FIGURE 3

Central Government’s Finance as a Proportion of the National Total, 1980–2010

Source: China Statistical Yearbook, various years.

of the state to control urban land and its commodification process, extracting large profit from theprocess (Xu et al., 2009). Conveyance of land user rights from local governments to private developersprovided a significant amount of local revenues. In some cases, local governments earn over 50% oftheir revenues on land transfer fees (Ho & Lin, 2003; Lin, 2007). Under local governments’ strongdrive to expand, urban built-up areas in Chinese cities almost doubled from 1996 to 2000 (Figure 2).

It has to be said that local governments in China have little option but to explore land financing. Thefiscal relationship between the central and local governments dramatically altered after an importanttax reform in 1994. Previously, local governments collected taxes and shared some part of them withthe central government. Nevertheless, the central government altered the rules of the game in 1994by setting up its own tax collection bureaus at the local level. The central government has claimedthe stable, easily collected taxes for its own treasury (Feng, Ljungwall, Guo, & Wu, 2013; Gong&Wu, 2012; World Bank, 2002). As a consequence, the extractive capacity of the central governmenthas grown steadily over the past two decades. As shown in Figure 3, the central government usually


controls over 51% of the national fiscal revenues while covering less than 20% of the expenditures.In other words, local governments have to be responsible for over 80% of the fiscal expenditures withless than 50% of the revenue, despite some intergovernmental transfers from the central government.The gap has continued to widen in recent years. Therefore, in order to keep up with the budgetshortfalls, local governments in China have to look for additional revenue sources that are notcontrolled by the central government. What they found is the “land financing,” which appears inmany developing countries in recent years (Peterson & Kaganova, 2010, p. 1).

As a consequence, municipal governments have to capitalize on the land over which they havepowers of expropriation, allocation, and conveyance to finance urban services and developmentprojects. This is hardly surprising, because land, unlike capital and labor, is the only immobile assetfirmly in the control of municipal government (Lin, 2007). The current land market in China is acombination of free administrative allocation and paid conveyance of land use right, which has madeland transfer a lucrative fiscal source for municipal governments to exploit (Ho & Lin, 2003; Lin,2007; Yeh & Wu, 1996).

Existing literature has tried to build the conceptual framework of land-centered urbanization inChina (see, e.g., Chan, 2010; Heikkila, 2007; Xu et al., 2009) and provide theoretical analysis (see,e.g., Li et al., 2010; Yeh & Wu, 1996; Zhu, 2005). The few empirical studies on this subject tend tofocus on a single city (Ding, 2004; Liu, Yue, & Fan, 2011; Wei & Zhang, 2012; Xu & Yeh, 2005; Yu& Ng, 2007), such as Beijing, Guangzhou, and Hangzhou. Lin (2007) compares land developmentpatterns in the eastern, central and western regions in China and finds that rapid urban expansion ofcities is driven by the conversion of farm land, extension of ring roads, and new zone development. Arecent study by Tao, Su, Liu, & Cao (2010), using prefecture-level city data, analyzes the evolution oflocal governments’ roles in different periods of China’s growth and explores local fiscal incentives touse subsidized land and infrastructure as key instruments in regional competition for manufacturinginvestment. Very limited empirical studies have been able to assess the impact of land financing(leasing) on China’s urbanization at the city level with a nationwide sample. This article fills this gapby investigating prefecture-level cities in China to provide statistical evidence of the patterns anddeterminants of China’s land-centered urbanization. Particular attention is paid to land financing asthe important factor driving such development.

DATA AND METHODOLOGY

This article measures urbanization in China by the expansion of built-up areas within city bound-aries. As discussed above, land financing or land-leasing revenue has been argued to be one of themost important factors that stimulated urban expansion in Chinese cities and thus offers a distinctiveexplanation of the country’s urbanization (Lichtenberg & Ding, 2009; Ho & Lin, 2003; Lin, 2007;Yeh & Wu, 1996). Fixed-effects, random-effects, and two-stage least squares estimations are run totest the driving forces of urbanization in China by analyzing panel data from 1999 to 2009 for all 286prefecture-level cities in China (except those in Tibet). Fixed-effects and random-effects estimationsare traditional approaches for panel data analysis; one of the purposes of two-stage least squaresestimation is to mitigate the problem of reverse causality (for example, urbanization may have animpact on economic development).

Dependent Variable

We use the share of built-up area as a percentage of total city land as one of the proxies for ourdependent variable of urbanization to reflect land development in Chinese cities. Many Chinesecities today typically encompass an extensive area which contains an urbanized core (a high-densitybuilt-up area with nonagricultural land), surrounded by numerous scattered towns and large stretchesof rural territory, usually with a dense farming population. A principal criterion to define the bound-ary of a Chinese city is the presence of a minimum average population density of at least 1,500persons per sq. km., or contiguity of built-up areas (Chan, 2007). Built-up areas are defined as any


officially designated and constructed urbanized land for nonagricultural use in a city, including allthe continuous urban land and surrounding areas with supportive urban infrastructures (i.e., airport,waste treatment plant, radio station, etc.). It has to be cautioned that the boundary of a Chinesecity includes the central city districts (shi xia qu) and outer suburban and rural areas. The landarea of central city districts is usually a proportion of the total area of the city. The reason totalcity land instead of the land of the central city district is used as the denominator is because theboundary of a central city district in Chinese cities often changes. According to a Ministry of CivilAffairs report, there were 717 central city districts in Chinese cities in 1996. The number grew to852 by 2005 (Ministry of Civil Affairs, 2008). Therefore, using land area of total city land pre-vents us from getting inconsistent measures of the percentage of built-up areas at different points intime.

In 2009 the city with the highest share of built-up area as a percentage of total city land in Chinawas Shenzhen, which exceeded 40%. Shenzhen is one of the Special Economic Zones (SEZs) andone of the most rapidly developing cities in China. The city has quickly become one of the topfive largest cities in China, with an estimated population of over 14 million (including permanent,temporary, and migrant populations). Beijing and Shanghai respectively had 8.23% and 14.1% oftotal city land as built-up areas in 2009. Hulunbeir City, a traditional nomadic city in the InnerMongolia Autonomous Region, had the lowest share of built-up area as a percentage of total cityland, around 0.01%. In 1999, the average share of built-up area as a percentage of total city land forthe 286 cities was 0.79%. The figure grew to 1.37% in 2009, which indicated a 7.3% rate of urbanexpansion annually.

Independent Variables

Drawing on the existing literature, this section includes several explanatory factors for urbanization,in particular the economic and fiscal factors. These are (1) dependence on land financing, (2) economicdevelopment, (3) average annual wage of staff and workers, (4) total population, (5) the share ofnonagricultural population, (6) the share of manufacturing employment, and (7) real estate investmentas a share of GDP. Tables 1 and 2 present the data sources and descriptive statistics.

Dependence on Land Financing

This is measured by two factors—the share of land lease fee as a percentage of budgetary revenueor budgetary expenditure.2 Urban expansion, Yew (2012) notes, is engineered by fiscally starvingstates to improve local fiscal capacity. In China, land-centered development and the pursuit of landrevenue is argued to be the strongest drive for urbanization. Therefore, measuring dependence onland financing will help empirically test this relationship.

Tian and Ma (2009) use land lease fee as a share of local revenue to evaluate a local government’sdependence on leasing lands to developers (see also Zhang & Li, 2010). We argue that two measures—contribution of land revenue to local revenue or the share of land revenue in local expenditure—areequally important. Figure 4 depicts the increasing reliance on land lease fees in both local revenueand local expenditure from 1999 to 2009. The average share of land lease fee as a share of budgetaryrevenue for the 286 cities jumped from 5.6% in 1999 to 56.4% in 2009, although it fluctuated insome years. For the share of land lease fees as a share of budgetary expenditure, the pattern wassimilar. It increased from 3.9% in 1999 to 28.1% for all cities in 2009. This suggests that land leasefees became increasingly important to both local revenue and local expenditure while local revenuewas more reliant on land lease fees.

Economic Development

We use GDP per capita (local GDP divided by de facto residents, who by census definitioneither have urban hukou or have resided in the city for more than 6 months during the previous


TABLE 1

Variable Measurement and Data Source

Variable Measurement Data source

Land urbanization The share of built-up area as apercentage of total city land under thecontrol of city governments and theirsubordinate county governments

China data online and ChinaStatistical Yearbooks

Dependence on landfinancing

Land lease fee as a percentage ofbudgetary revenue or budgetaryexpenditure.

Land lease fee was retrieved fromChina Land and ResourceAlmanac (zhongguo guotuziyuan nianjian); revenue andexpenditure from China dataonline

Economic development Real GDP per capita China data online and China CityStatistics

Average wage of staff andworkers

The average wage of staff and workersin real terms

China data online

Total population Population in the urban core city (undercity districts (chengqu))

China data online

Share of nonagriculturalpopulation

Share of nonagricultural population inthe urban core city (under city districts(chengqu))

China data online

Real estate investment asa share of GDP

Real estate investment as a share ofGDP in the city

China data online

Share of manufacturingemployment

The share of manufacturing employmentin the city total labor force

China data online

TABLE 2

Descriptive Statistics, 1999–2009

Variable Obs. Mean Std. Dev. Min. Max.

Land urbanization 3,029 −0.683 1.227 −4.505 3.705Land lease fee as % of budgetary revenue 2,490 3.136 1.364 −3.840 5.886Land lease fee as % of budgetary expenditure 2,461 2.374 1.433 −4.521 5.385Economic development 2,702 8.906 0.656 7.189 10.891Average wage of staff and workers 3,065 9.083 0.295 8.232 11.064Total population 3,146 5.786 0.723 2.678 8.094Share of nonagricultural population 3,146 3.373 0.539 1.995 5.727Real estate investment as a share of GDP 3,039 1.388 0.867 −4.668 4.034Share of manufacturing employment 3,041 42.163 13.488 7.399 81.648

Note: All variables excluding the share of manufacturing employment are in logarithms.

year) as a proxy of economic development. Deng, Huang, Rozelle, & Uchida (2010) find thateconomic development contributes significantly to urban land expansion in China. We hypothesizethat economic development spurs urban expansion.

The Average Wage of Staff and Workers

This variable refers to the average wage of employees in the formal sector in the Chinese cities.This often is misinterpreted as household income in urban areas in China. Household income iscommonly used as an independent variable affecting urbanization in the Western context (Wassmer,2008). Nevertheless, the National Bureau of Statistics of China does not report comparable data.The average wage may represent only a fraction of household income, because investment gains andinformal incomes are excluded, but this is the closet proxy that can be identified.


FIGURE 4

Land Lease Fee as a Percentage of Revenue and Expenditure, 1999–2009

Source: Land lease fee was retrieved from China Land and Resource Almanac (zhongguo guotu ziyuan nianjian). Revenue andexpenditure data from China data online.

Total Population

This variable is the city population at the end of 1 year. Population generates the pressure for urbanexpansion. Torrens (2006) reports that population growth is “one of the most important engines ofchange in any urban system” (p. 251). Classic urban economic models proposed by Brueckner andFansler (1983) also emphasize the effect of total population size on urban land development. Thus weexpect population size to be positively related to urban land development in China. It should be notedthat we use the urban core population as a proxy in this study. According to the Chinese administrativesystem, the urban core population means those living in the urban area where municipal governmentsare located.

The Share of Nonagricultural Population

This variable refers to the share of nonagricultural population in the total core urban population. Inline with the planned economy, Chinese population was separated into agricultural and nonagricul-tural populations. The system is being reformed but remains largely stable in reform-era China. Themajority of urban residents have obtained urban residential status (chengzhen hukou). But not all ofthem have it. The status of nonagricultural residents is related to access to social welfare coveragesuch as education and other basic services in urban areas. We expect there is a positive relationshipbetween the share of nonagricultural population in the total city population and urban land expansion.

Real Estate Investment as a Share of GDP

As a newly urbanized society, real estate is of fundamental importance to China’s urbanization.Since the late 1990s, real estate development has characterized urban development across the country.Cities in China, large or small, coastal or inland, are enthusiastic at promoting the property market.


As noted by Cartier (2001) and Deng and Huang (2004), real estate investment in urban areas hasbeen associated with many fundamental issues, such as urbanization and arable land losses. SinceChinese local governments do not traditionally collect a significant amount of property taxes, realestate tax revenues are important budget items in addition to revenues from land leasing. Propertymarket development is expected to encourage urban land development as local governments striveto generate more revenues.

The Share of Manufacturing Employment

We include the share of the labor force in the manufacturing sector in the total labor force asa proxy of industrial activities. Urbanization is thought to be associated with industrial activities(Antrop, 2004). We postulate that the increase in the share of manufacturing employment has apositive impact on urbanization.

MODEL SPECIFICATIONS

We adopt the standard panel data models in our study. According to Hsiao and Tahmiscioglu (2008,p. 2698), “panel data, by blending inter-individual differences and intra-individual dynamics, havegreater capacity for capturing the complexity of human behavior than data sets with only a temporalor a cross-sectional dimension.” Social scientists are very much interested in dynamic relationships;nonetheless, they cannot be detected in a single cross-section analysis (Wooldridge, 2010, p. 191).Panel data analysis is strong at investigating dynamics in social sciences. The fixed-effects model andthe random-effects model are two common approaches to panel data analysis. The random-effectsmodel can generate more efficient results of the slope coefficients than those produced by the fixed-effects model. In addition, the transformation employed in the random-effects model does not ruleout time-invariant variables (Kennedy, 2003, p. 305). Therefore, as long as a random-effects model isstatistically justified, it should be chosen (Wooldridge, 2010). Nevertheless, the fixed effects modelalways generates consistent results. The Hausman specification test commonly used to comparethe fixed- and random-effects models suggests that fixed-effects models should be preferred in thisstudy. The results of the random-effects models are reported as a reference. We use the standardfixed-effects model to analyze land urbanization in the Chinese cities:

Yjt = αo + βXjt + γCjt + αj + γt + εjt ,

where j indicates cities, t denotes year, Y refers to land urbanization as measured by the share ofcity built-up area as a percentage of the total land under the control of city governments and theirsubordinate county governments, X denotes some important variables such as local governments’reliance on revenues from land transfer fees with C referring to a battery of socioeconomic variables,αj and γt refer to the city fixed effects and the year fixed effects, respectively, while εjt is the randomerror.

Baseline Results

We introduce 12 models in this study. In the baseline models, we use fixed- and random-effectsmodeling to explain urbanization. In Models 5 and 6, we use the two-stage least squares (2SLS)estimations to handle the endogeneity issue. As four centrally controlled cities are different fromthe rest of the samples, we employ fixed-effects and random-effects modeling among Chinese citiesexcluding four centrally controlled cities to evaluate the robustness of the study (Models 7–10).We further use the two-stage least squares (2SLS) estimations among Chinese cities excluding fourcentrally controlled cities to test the results (Models 11 and 12). Table 3 reports the estimation resultsusing the panel data analysis. According to variance inflation factor (VIF) tests, no multicollinearityproblem was detected in any of our models. With the land lease fee as a share of budgetary revenue


TABLE 3

Fixed-Effects (FE) and Random-Effects (RE) Models, 1999–2009

FE (1) RE (2) FE (3) RE (4)

Land lease fee as % of budgetary revenue 0.0157+ 0.0290∗∗∗(0.00839) (0.00485)

Land lease fee as % of budgetary expenditure 0.0179∗ 0.0294∗∗∗(0.00864) (0.00487)

Economic development 0.263∗ 0.506∗∗∗ 0.201+ 0.465∗∗∗(0.123) (0.0338) (0.118) (0.0334)

Average wage of staff and workers −0.325∗ −0.164∗∗∗ −0.283∗ −0.119∗∗(0.127) (0.0471) (0.117) (0.0458)

Total population 0.247+ 0.528∗∗∗ 0.271∗ 0.543∗∗∗(0.130) (0.0314) (0.130) (0.0310)

Share of nonagricultural population 0.112 0.375∗∗∗ 0.0858 0.360∗∗∗(0.122) (0.0397) (0.119) (0.0384)

Real estate investment as a share of GDP 0.0234 0.0616∗∗∗ 0.0246+ 0.0676∗∗∗(0.0148) (0.0110) (0.0149) (0.0106)

Share of manufacturing employment 0.00122 0.00370∗∗∗ 0.000732 0.00317∗∗(0.00222) (0.00108) (0.00220) (0.00105)

Constant −1.995 −7.919∗∗∗ −1.809 −7.932∗∗∗(1.491) (0.371) (1.412) (0.359)

N 2,374 2,374 2,396 2,396R2 0.4289 0.4456

Note: All variables excluding the share of manufacturing employment are in logarithms. Standard errors in parentheses.+p < 0.10; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

being a key independent variable, 42.89% of the variations in land urbanization can be explainedby Model 1. Using the share of land lease fee as a percentage of budgetary expenditure as a keyindependent variable, Model 3 roughly explains 44.56% of the variations of the Y values.

Our main independent variables, land lease fee as a share of budgetary revenue and expenditure,are positively related with land urbanization. In Models 1 and 3, land lease fee as a share ofbudgetary revenue or expenditure shows a significant relationship with the dependent variable in thefixed-effects models. A 1% increase in dependence on land financing is positively associated with a0.0157%—0.0179% increase in land urbanization.

Supporting the literature on economic strength in a region and urbanization, this study finds thateconomic development is positively associated with land urbanization in four models. It shouldbe noted that the significance levels of economic development are much higher in random-effectsmodels.

The average wage of staff and workers in the city exerts a negative and significant impact onland urbanization. A caveat is in order that the average wage in Chinese cities is not equivalent tothe household income used in the Western literature; rather, the Chinese average wage reflects theformal wage of employees working in the formal sector. The role of the informal sector has beenincreasingly important in Chinese cities, but it is extremely difficult to obtain income statistics oninformal employment.

As expected, there is a significantly positive relationship between city population size and landurbanization in Chinese cities.

As a common phenomenon in developing countries, real estate investment in urban areas hasa significantly positive impact on land urbanization in the Chinese cities. It is nonetheless onlyconfirmed in Model 3 at a significance level of 10%.

Our baseline results indicate the following findings: (1) Dependence on land financing has a sig-nificantly positive impact on land urbanization, (2) population size in Chinese cities has a significantand positive relationship with land urbanization, and (3) real estate investment also plays a significantrole affecting land urbanization.


TABLE 4

Two-stage Least Squares (2SLS) Estimations, 1999–2009

2SLS (5) 2SLS (6)

Land lease fee as % of budgetary revenue 0.0180∗(0.00765)

Land lease fee as % of budgetary expenditure 0.0186∗(0.00758)

Economic development 0.418+ 0.405+(0.235) (0.234)

Average wage of staff and workers −0.276∗ −0.271∗(0.123) (0.122)

Total population 0.187 0.192(0.147) (0.146)

Share of nonagricultural population 0.0115 0.0186(0.131) (0.129)

Real estate investment as a share of GDP 0.0397∗ 0.0405∗(0.0172) (0.0175)

Share of manufacturing employment −0.000338 −0.000317(0.00206) (0.00205)

N 2,011 2,029R2 0.3520 0.3583

Note: All variables excluding share of the manufacturing employment are in logarithms. Standard errors in parentheses.+p < 0.10; ∗p < 0.05.

Robustness Checks

We employ two robustness checks on the baseline models. In Models 5 and 6, we attempt toaddress the endogeneity issue, and Models 7–10 exclude the four centrally controlled cities (zhi xiashi) of Beijing, Shanghai, Chongqing, and Tianjin.

The models used in this study raise issues about a potential endogeneity (e.g., reverse causation)problem. Economic development and urbanization may have a two-way relationship. As economicdevelopment affects the level of urbanization, measured by the built-up area as a share of the totalland of the city, this may contribute to GDP growth in a region. As suggested by Wooldridge (2009),the fixed-effects model may, to some extent, address the problem of endogeneity. Nevertheless, itis not enough without introducing additional methods. In this study, we employ the two-stage leastsquares (2SLS) estimations to mitigate the issue of endogeneity and especially reverse causation.Therefore we would like to investigate whether changes in independent variables cause change in thedependent variable, not the reverse, although we acknowledge the limitations of 2SLS estimations.

We enter lag two regressors (economic development) as the instruments to handle the endogeneitythreat. The instruments pass the weak instrument test and overidentification test (Hayashi, 2000;Stock & Yogo, 2005). The result suggests that the impact of our main explanatory factors on landurbanization remains significant (Table 4). Nevertheless, urban population is not significantly relatedto land expansion, which means that it offers no explanatory power to the dependent variable. In the2SLS models, real estate investment in urban areas has a significantly positive impact on urbanization.

We also exclude four centrally controlled Chinese cities in the models with both fixed- and random-effects tests (Table 5). Centrally controlled Chinese cities refer to the cities of Beijing, Shanghai,Tianjin, and Chongqing. Two reasons stand out for the tests. First, the political structure is differentbetween the above four cities and all other Chinese cities. For one thing, in contrast to other cities,they share a very high political status—the administrative rank of the mayor of Beijing, for example,is equivalent to the provincial governor. whereas an ordinary city mayor ranks much lower thanthe provincial governor administratively. Second, perhaps more importantly, demographic factors inthe above four municipalities are distinct from those in other cities. This suggests that, comparedwith the baseline models, the regression excluding the four municipalities remains unaffected. In


TABLE 5

Fixed-Effects (FE) and Random-Effects (RE) Models Excluding Four Centrally Controlled Cities, 1999–2009

FE (7) RE (8) FE (9) RE(10)

Land lease fee as % of budgetary revenue 0.0183∗ 0.0290∗∗∗(0.00829) (0.00488)

Land lease fee as % of budgetary expenditure 0.0195∗ 0.0290∗∗∗(0.00862) (0.00490)

Economic development 0.266∗ 0.510∗∗∗ 0.202+ 0.468∗∗∗(0.125) (0.0339) (0.120) (0.0335)

Average wage of staff and workers −0.341∗∗ −0.202∗∗∗ −0.298∗ −0.154∗∗∗(0.130) (0.0476) (0.119) (0.0463)

Total population 0.245+ 0.515∗∗∗ 0.269∗ 0.532∗∗∗(0.130) (0.0322) (0.130) (0.0317)

Share of nonagricultural population 0.111 0.363∗∗∗ 0.0853 0.350∗∗∗(0.123) (0.0400) (0.119) (0.0386)

Real estate investment as a share of GDP 0.0293∗ 0.0674∗∗∗ 0.0307∗ 0.0736∗∗∗(0.0149) (0.0111) (0.0148) (0.0106)

Share of manufacturing employment 0.00116 0.00391∗∗∗ 0.000607 0.00334∗∗(0.00230) (0.00109) (0.00227) (0.00106)

Constant −1.880 −7.520∗∗∗ −1.684 −7.566∗∗∗(1.490) (0.377) (1.410) (0.365)

N 2,330 2,330 2,352 2,352R2 0.4377 0.3882

Note: All variables excluding share of the manufacturing employment are in logarithms. Standard errors in parentheses.+p < 0.10; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

TABLE 6

Two-Stage Least Squares (2SLS) Estimations Excluding Four Centrally Controlled Cities, 1999–2009

2SLS (11) 2SLS (12)

Land lease fee as % of budgetary revenue 0.0182∗(0.00777)

Land lease fee as % of budgetary expenditure 0.0183∗(0.00768)

Economic development 0.416+ 0.402+(0.234) (0.233)

Average wage of staff and workers −0.280∗ −0.274∗(0.124) (0.122)

Total population 0.181 0.186(0.148) (0.147)

Share of nonagricultural population 0.00855 0.0162(0.132) (0.129)

Real estate investment as a share of GDP 0.0446∗ 0.0453∗(0.0179) (0.0183)

Share of manufacturing employment −0.000283 −0.000286(0.00219) (0.00218)

N 1,975 1,993R2 0.3443 0.3512

Note: All variables excluding the share of manufacturing employment are in logarithms. Standard errors in parentheses.+p < 0.10; ∗p < 0.05.

fixed-effects models, the significance level increases to 5% when using land lease fee as a share ofbudgetary revenue (Model 7).

As shown in Table 6, we include the 2SLS models using the data set excluding the four centrallycontrolled Chinese cities (Models 11 and 12). The result remains almost unchanged with regard tothe main explanatory factors. It shows that findings in Table 6 are comparable to those in Table 4although the R-squared drops slightly in Table 6.


FINDINGS AND DISCUSSION

The statistical analyses reveal several important findings regarding urbanization in China. First,Chinese cities have experienced a significant increase in urban built-up areas. As described above,the share of built-up area as a percentage of total city land increased 73% in the 286 cities surveyedfrom 1999 to 2009. At the same time, dependence on land financing also increased dramatically.The share of land lease fee as a share of local budgetary expenditure, for example, jumped by 616%between 1999 and 2009.

More importantly, strong positive relations were identified between dependence on land financingand land urbanization in Chinese cities. For those cities that are more dependent on land leasingfees as part of their local budget, the government is under greater pressure to exploit the land and togenerate more revenue to support public spending. Our findings support the arguments that China’surbanization has been land-centered and an outcome of strong government-driven forces. The realestate investment significantly drove land urbanization in Chinese cities. In addition, economicallymore developed cities with a higher GDP per capita significantly recorded more aggressive urbanland expansion.

These important factors together reflect the phenomena of land-centered urbanization in China,driven by local fiscal resource extraction and real estate property development. Chinese cities adopteddifferent approaches to make more land available and generate more land lease revenues. In thisprocess, the empirical results in this research indicated that economically more advanced citieswith heavier real estate investment were more aggressive in pursuing urban land expansion tosupport local growth. In China, major cities with stronger economic output usually are equippedwith more favorable status and flexible policies from higher authorities, which give them stronger“teeth” in fighting with their competitors. For example, in 2000 the City of Guangzhou redrew itsadministrative boundary by adjusting Huadu and Panyu from the status of county-level cities tourban districts. Following this reorganization, the area under the direct administration of the citygovernment increased from 1,400 to 3,719 square kilometers, which significantly enriched the localland bank and alleviated development pressure within the city’s jurisdictions (Xu & Yeh, 2005). Inaddition, major cities like Beijing, Shanghai, and Guangzhou were able to seize opportunities ofhosting high-profile international events and to allocate vast urban land for megaprojects (Lin, 2007;Liu, 2012; Ren, 2008; Sun & Ye, 2010). When megaprojects or land annexation are not viable, manyChinese cities adopt the “redevelopment” model to exploit more land in urban villages for real estatedevelopment (He & Wu, 2005; Ye, 2011b). Indeed, many Chinese cities adopt a combination of theseapproaches to maximize land revenues and push urbanization.

Echoing other studies on China’s urban development, our analysis suggests that economic growthdrove up urbanization. Nonetheless, the average wage of staff and workers in the city was found tohave a negative impact on land urbanization. As previously explained, this measurement representswages of workers in formal sectors without considering households’ informal gains, which mayaccount for a sizable proportion of urban household income in China. Inclusion of such comprehensivedata will help better investigate the relationship between urban wealth and land expansion. However,such data are not readily available for China. The analysis of the impact of local formal wage onurbanization serves as a reference. In addition, our analysis indicates that population size and level ofindustrialization do not necessarily determine the urbanization path for a city in the Chinese context.

CONCLUSIONS AND FUTURE RESEARCH

According to the latest national statistics, land transfer in China recorded a historic high in the firstsix months of 2013, in both total area and government revenue generated from land transfer, despiteseveral years of tightened housing market control policies (Ministry of Land and Resources, 2013).Extracting economic benefits from land development has been a dominant character of China’surbanization. Intergovernment fiscal relations and local governments’ strong impetus to compete anddevelop substantially drive the expansion of urban land in China. Land financing is arguably the“invisible hand” of China’s urbanization. This article provides empirical evidence of this missing


link in the existing literature. Other factors, such as heavier investment in real estate and strongereconomic status, were also found to significantly drive land urbanization in Chinese cities.

It is important to realize that urbanization is a multidimensional phenomenon. China’s urbanizationis unique in terms of its speed, scale, and government-driven nature. Future research is called uponto further advance study on this subject. When more data become available, better measurements canbe developed to analyze variables that have been identified as important in the urbanization process,such as household income, commuting time, and urban–suburb disparity. It will also be importantto identify varying urbanization models in different regions and among different types of cities inChina (Ye, 2009, 2013). Some recent trends, such as urbanization in the peri-urban areas, can beincorporated to better analyze urbanization development in Chinese cities (Gross, Ye, & LeGates,2014; Ye, 2014; Zhao, 2013).

The recent 18th National Congress of the Chinese Communist Party adopted urbanization as aprimary development strategy to support China’s economic growth in the next few decades. As landbecomes scarce in Chinese cities, urbanization has to change its objective from its primitive goal ofgenerating revenue to serving residents. Local governments have to adjust the logic of their actionsfrom using urbanization as a way to extract fiscal resources to taking the opportunity to improvethe desirability of urbanization. This task will not be accomplished without modifying the fiscal andpolitical relations between the central and local governments. China’s urbanization will continue tobe the most significant event affecting world development. But only when its drives and goals arejustified can that urbanization be more desirable and sustainable.

ACKNOWLEDGEMENTS: This research was jointly supported by a 985 III Project, National Social Science Fundmajor program 13&ZD041, Ministry of Education key research center major project 13JJD630014, excellent researchfund and internal project of Center for Chinese Public Administration Research, and excellent young faculty trainingproject of Sun Yat-sen University. The authors would like to thank Dajian Zhu, Wen Wang, Fangzhi Ye, Xiulian Ma,Huiping Li, Richard LeGates, Jill Gross, and three anonymous reviewers for their constructive input. An earlier versionof this article was presented at the 43rd Annual Conference of the Urban Affairs Association in San Francisco. Theusual disclaimer applies.

ENDNOTES

1 A prefectural city is an administrative division of Chinese governments, ranking below a province and above acounty (or county-level city) in China’s administrative structure.

2 The Ministry of Land and Resources compiled the data on the land leasing fee. The ministry has built up anationwide data collection system with regard to the revenue. According to the official guideline, the land leasingfee is being saved under the local coffer. Though we don’t rule out the possibility of underestimating the land leasefee, in general the data are reliable after various cross-checks.

REFERENCES

Antrop, M. (2004). Landscape change and the urbanization process in Europe. Landscape and Urban Planning, 67(1),9–26.

Brueckner, J. K., & Fansler, D. (1983). The economics of urban sprawl: Theory and evidence on the spatial size ofcities. Review of Economics and Statistics, 55, 479–482.

Cartier, C. (2001). “Zone fever”, the arable land debate, and real estate speculation: China’s evolving land use regimeand its geographical contradictions. Journal of Contemporary China, 10, 445–469.

Chan, K. W. (2007). Misconceptions and complexities in the study of China’s cities: Definitions, statistics, andimplications. Eurasian Geography and Economics, 48, 383–412.

Chan, K. W. (2010). Fundamentals of China’s urbanization and policy. The China Review, 10(1), 63–94.Deng, F. F., & Huang, Y. (2004). Uneven land reform and urban sprawl in China: The case of Beijing. Progress in

Planning, 61, 211–236.Deng, X., Huang, J., Rozelle, S., & Uchida, E. (2010). Economic growth and the expansion of urban land in China.

Urban Studies, 47, 813–843.


Ding, C. (2004). Urban spatial development in the land policy reform era: Evidence from Beijing. Urban Studies, 41,1889–1907.

Feng, X., Ljungwall, C., Guo, S., & Wu, A. M. (2013). Fiscal federalism: A refined theory and its application in theChinese context. Journal of Contemporary China, 22, 573–593.

Gong, T., & Wu, A. M. (2012). Central mandates in flux: Local noncompliance in China. Publius: The Journal ofFederalism, 42, 313–333.

Gross, J., Ye, L., & LeGates, R. (2014). Asia and the pacific rim: The new peri-urbanization and urban theory. Journalof Urban Affairs, 36(S1).

Gu, C., & Wu, F. (2010). Urbanization in China: Processes and policies. The China Review, 10(1), 1–10.Hayashi, F. (2000). Econometrics. Princeton, NJ: Princeton University Press.He, S., & Wu, F. (2005). Property-led redevelopment in post-reform China: A case study of Xintiandi redevelopment

project in Shanghai. Journal of Urban Affairs, 27, 1–23.Heikkila, E. J. (2007). Three questions regarding urbanization in China. Journal of Planning Education and Research,

27(1), 65–81.Ho, S. P. S., & Lin, G. C. S. (2003). Emerging land markets in rural and urban China: Policies and practices. The China

Quarterly, 175, 681–707.Hsiao, C., & Tahmiscioglu, A. K. (2008). Estimation of dynamic panel data models with both individual and time-

specific effects. Journal of Statistical Planning and Inference, 138, 2698–2721.Kennedy, P. (2003). A guide to econometrics. Cambridge, MA: MIT press.Li, X., Xu, X., & Li, Z. (2010). Land property rights and urbanization in China. The China Review, 10(1), 11–38.Lichtenberg, E., & Ding, C. (2009). Local officials as land developers: Urban spatial expansion in China. Journal of

Urban Economics, 66, 57–64.Lin, G. C. S. (2007). Reproducing spaces of Chinese urbanisation: New city-based and land-centred urban transfor-

mation. Urban Studies, 44, 1827–1855.Liu, C. Y. (2012). From Los Angeles to Shanghai: Testing the applicability of five urban paradigms. International

Journal of Urban and Regional Research, 36, 1127–1145.Liu, Y., Yue, W., & Fan, P. (2011). Spatial determinants of urban land conversion in large Chinese cities: A case of

Hangzhou. Environment and Planning B: Planning and Design, 38, 706–725.Ministry of Civil Affairs (2008). The adjustment of urban administrative boundaries since 1997 in China

and its implication. Retrieved April 22, 2013, from http://www.mca.gov.cn/article/mxht/llyj/200801/20080100009424.shtml (in Chinese).

Ministry of Land and Resources (2013). China’s 2012 land market statistics. Retrieved July 31, 2013, fromhttp://www.mlr.gov.cn/xwdt/jrxw/201307/t20130730_1246849.htm (in Chinese).

National Bureau of Statistics (2012). Report on the 2011 National Migrant Worker Survey. Retrieved April 22, 2013,from http://www.stats.gov.cn/tjfx/fxbg/t20120427_402801903.htm (in Chinese).

Peterson, G. E., & Kaganova, O. (2010). Integrating land financing into subnational fiscal management. World BankPolicy Research Working Paper No.5409. Washington, DC: World Bank.

Ren, X. (2008). Architecture and national building in the age of globalization: Construction of the national stadium ofBeijing for the 2008 Olympics. Journal of Urban Affairs, 30, 175–190.

Shen, J., & Wu, F. (2012). Restless urban landscapes in China: A case study of three projects in Shanghai. Journal ofUrban Affairs, 34, 255–277.

Stock, J. H., & Yogo, M. (2005). Testing for weak instruments in linear IV regression. In D. W. K. Andrews, T. J.Rothenberg, & J. H. Stock (Eds.), Identification and inference for econometric models: Essays in honor ofThomas Rothenberg (pp. 80–108). New York: Cambridge University Press.

Sun, J., & Ye, L. (2010). Mega-events, local economies, and global status: What happened before the 2008 Olympicsin Beijing and the 2010 World Expo in Shanghai. Journal of Current Chinese Affairs, 39, 133–165.

Tao, R., Su, F., Liu, M., & Cao, G. (2010). Land leasing and local public finance in China’s regional development:Evidence from prefecture-level cities. Urban Studies, 47, 2217–2236.

Tian, L., & Ma, W. (2009). Government intervention in city development of China: A tool of land supply. Land UsePolicy, 26, 599–609.

Torrens, P. M. (2006). Simulating sprawl. Annals of the Association of American Geographers, 96, 248–275.Wassmer, R. W. (2008). Causes of urban sprawl in the United States. Journal of Policy Analysis and Management, 27,

536–555.Wei, Y., & Zhang, Z. (2012). Assessing the fragmentation of construction land in urban areas: An index method and

case study in Shunde, China. Land Use Policy, 29, 417–428.Wooldridge, J. M. (2009). Introductory econometrics: A modern approach (4th ed.). Stamford, CT: Cengage Learning.Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data. Cambridge, MA: MIT Press.


World Bank. (2002). China: National development and subnational finance: A review of provincial expenditures.Washington, DC: World Bank.

Xu, J., & Yeh, A. G. O. (2005). City repositioning and competitiveness building in regional development: Newdevelopment strategies in Guangzhou, China. International Journal of Urban and Regional Research, 29,283–308.

Xu, J., Yeh, A. G. O., & Wu, F. (2009). Land commodification: New land development and politics in China since thelate 1990s. International Journal of Urban and Regional Research, 33, 890–913.

Ye, L. (2009). Regional government and governance in China and the United States. Public Administration Review(Supplement), S116–S121.

Ye, L. (2011a). Demographic transition, developmentalism and social security in China. Social Policy and Adminis-tration, 45, 678–693.

Ye, L. (2011b). Urban regeneration in China: Policy, development, and issues. Local Economy, 26, 337–347.Ye, L. (2013). Urban transformation and institutional policies: A case study of mega-region development in China’s

Pearl River Delta. Journal of Urban Planning and Development, 139, 292–300.Ye, L. (2014). State-led metropolitan governance in China: Making integrated city regions. Cities. Retrieved April 4,

2014, from http://dx.doi.org/10.1016/j.cities.2014.03.001Yeh, A. G. O., & Wu, F. (1996). The new land development process and urban development in Chinese cities.

International Journal of Urban and Regional Research, 20, 330–353.Yew, C. P. (2012). Pseudo-urbanization? Competitive government behavior and urban sprawl in China. Journal of

Contemporary China, 21, 281–298.Yu, X. J., & Ng, C. N. (2007). Spatial and temporal dynamics of urban sprawl along two urban–rural transects: A case

study of Guangzhou, China. Landscape and Urban Planning, 79(1), 96–109.Zhang, S., & Li, D. (2010). Financial base of the second housing reform: An analysis of the relationship between local

land financing and real estate prices. Public Finance Research, 7, 5–11 (in Chinese).Zhao, P. (2013). Too complex to be managed? New trends in peri-urbanisation and its planning in Beijing. Cities, 30,

68–76.Zhu, J. (2005). A transitional institution for the emerging land market in urban China. Urban Studies, 42, 1369–1390.

ABOUT THE AUTHORS

Lin Ye is an Associate Professor in Center of Chinese Public Administration Research, Institute forUrban Governance and Urban Development, School of Government at Sun Yat-sen University inGuangzhou, China. He holds an MPA and PhD in urban and public affairs from the University ofLouisville. His research interests are in urban policy and politics, metropolitan and regional gover-nance, and urbanization in China. His academic work has appeared in both urban studies and publicadministration journals, such as Journal of Urban Affairs, Journal of Urban Planning and Devel-opment, Journal of Planning Literature, Social Policy and Administration, Public AdministrationReview, American Review of Public Administration, and other journals in China.

Alfred M. Wu is a Lecturer in the Department of Asian and Policy Studies at the Hong Kong Instituteof Education. He earned his PhD from City University of Hong Kong. His research interests includecentral–local fiscal relations, public sector reform, corruption and governance, and social protectionin Greater China. He is the author of the forthcoming book Governing Civil Service Pay in China.He is also the author or co-author of articles in such journals as World Development, Public Choice,Publius: The Journal of Federalism, Review of Public Personnel Administration, Social Policy andSociety, and Journal of Contemporary China.

Documents

URBANIZATION, LAND DEVELOPMENT, AND LAND FINANCING ... · urbanization with panel data covering 1999–2009. Empirical findings and policy discussions will follow before conclusions