Public Land Leasing and the Changing Roles of Local Government in Urban China

February 12, 2003

F. Frederic Deng*

(forthcoming: Annals of Regional Science) ABSTRACT:

Since late 1980s China’s urban land use began to change from administrative allocation to public land leasing. Based on the assumptions of fiscal decentralization and growth of private economy, this paper provides another perspective on public land leasing in addition to the well-known efficiency of market allocation of land resource. My focus is on the endogenous evolution of local public finance due to rent capitalization. Especially, I model a two-sector urban economy and a multi-task local government that allocates its effort to maximize fiscal revenue. The findings indicate that complete wage capitalization under administrative allocation of land contributes to the pressure on local governments to look for alternative revenue source. Public land leasing helps to include private firms into local public finance and re-orient local governments’ interest from SOEs to public goods provision. A simple test on the institutional change shows the efficiency of public land leasing relative to the old system. KEY WORDS: public land leasing, local public finance, rent capitalization, land reform, China *Department of Geography & Planning, AS218, University at Albany, SUNY, Albany, NY 12222. Email: fdeng@albany.edu I thank Rena Sivitanidou and Yongheng Deng for their helpful comments on an early version of the paper.

1

I. INTRODUCTION After almost 10 years of economic reform, China began in late 1980s to abandon the

traditional administrative allocation of urban land and adopt public land leasing as the foundation for its urban land use. Most existing studies focus on efficiency issues of the emerging real estate market. No doubt, market allocation of land greatly increases the efficiency of urban land use. But, it cannot explain why local governments are so enthusiastic about land reform. In some sense, local government is the driving force behind China’s urban land reform (Deng 2003). They have been strong advocates for public land leasing and are often leading the central government in the reform process. A narrow focus on land market itself also cannot explain why this particular form of property rights arrangement is adopted. A simplistic way of thinking may rather point directly to fee simple ownership, which is more common around the world. Hence, understanding public land leasing has to go beyond land market. It is the purpose of this paper to provide another perspective by looking at how the capitalization of local public goods into land rent interacts with land reform and consequent re-orientation of local government functions.

The most influential theory of rent capitalization is probably Henry George’s (1879) proposal for single tax. Because local public goods are capitalized into land rent, he argued that it is efficient for local government to collect land rent that, in turn, should be enough to finance local public goods. In the same vein, Fischel (2001) used rent capitalization to explain the persistence of property tax in American local public finance and why land use control is the most important function of local government. Glaeser (1996) showed how property tax creates incentives for local government to invest for the future. Rent capitalization is also an important issue in the literature on local public finance and Tiebout model (see, for example, Oates 1969; Hamilton 1976; Starrett 1981). Many scholars have realized that some sort of entrepreneurial behavior is necessary to achieve Tieboutian efficiency (Sonstelie and Portney 1978; Brueckner 1983). Their studies imply important links between local institutions and rent capitalization. However, most existing studies related to rent capitalization focus on Western countries where market institutions have been functioning for quite a long time, and they often take for granted existing institutional arrangements (such as property tax and private property rights). Given the nature of China’s transitional economy, can the theory of rent capitalization, or more generally, its interaction with local public finance, provide more insight into its ongoing economic reform and institutional transition?

By constructing a simple general equilibrium framework, this paper studies the relationship between urban land reform and the changing role of local governments in an urban economy that is characterized by fiscal decentralization and labor market segregation of SOEs (State-Owned Enterprises) and private firms. I am especially interested in explaining the adoption of public land leasing from the perspective of rent capitalization and endogenous transition of local government functions. A simple Pareto efficiency test is designed to see whether or not the new public land leasing system is an improvement.

The ground lease system is not a new invention, but quite limited in today’s world. China’s public land leasing (or the transfer of so-called “land use right”) was introduced in the late 1980s and early 1990s. Local officials in Shenzhen City first initiated land leasing in 1987 even though the constitution of PRC explicitly banned any land transaction at that time. Only in 1988 was the Constitution amended and in 1990 was the ground lease system formally approved by the central government. Nation-wide adoption of public land leasing started in 1992. It was first applied only to foreign and private firms, while SOEs and collectively-owned-enterprises

2

(COEs) operate on land allocated by the government in the past. Even up to now most land used by SOEs and COEs are still free to them.1 This process is typical of China’s general economic reform strategy—a gradual and partial approach (Naughton 1995).

Many studies on China’s ground lease system focus more on its pricing mechanism and effects on land price (World Bank 1993; Li 1997). On the other hand, some studies point to the efficiency properties of ground lease-based land use system, albeit, under private land ownership (Deng 2002).2 Beyond the apparent benefits of market allocation of land, there is relatively little discussion, especially formal analysis, about the relationship between public land leasing, local public finance and the roles of local government.

The approach in this paper is related to three strands of literature. The first is studies on rent capitalization and its implications to land use and urban institutions. The most recent study is Fischel’s (2001) argument on how rent capitalization shapes local government functions and makes property tax a persistent and primary source for local public finance in America. Glaeser (1996) analyzed how property tax provides incentives for adequate amenity provision when local governments maximize their revenues. This line of thinking more or less regards urban institutions as endogenous. In contrast, some other studies on wage and rent capitalization (Roback 1982; Sivitanidou and Wheaton 1992) usually treat public goods as exogenous variables. Second, this paper models local government as a multi-task entity that is simultaneously the landowner, the public goods provider and the owner of SOEs.3 A multi-role local government may not be common in the developed countries, but the problems facing China’s local governments arise exactly from their multiple roles. Many theoretical studies (Sonstelie and Portney 1978; Brueckner 1983) also suggest the importance of looking at the different functions of local government. For example, Sonstelie and Portney’s profit-maximizing community is essentially a property owner who profits from renting houses. Finally, in order to construct a simple test on the efficiency of urban land reform, I follow the approach of Greenwald and Stiglitz (1986) in their analysis of externalities with incomplete information and incomplete markets.

By relying on these three strands of literature, this paper addresses the following research questions. First, what are the impact of urban land reform on Chinese cities’ local public finance and the changing role of local government? The second question is how to evaluate urban land reform, especially from the perspective of local public finance. There is no doubt about the efficiency gains from market allocation of urban land. But, does it also facilitate change in local public finance and result in government shifting to more efficient roles? Third, why does China’s urban land reform adopt this particular form of property rights arrangement? Although a comprehensive answer to the last question is beyond the scope of this paper, I hope it can shed some light on the relationship between urban property rights and local public finance.

This paper constructs an urban economic model with two segregated sectors—private firms and SOEs. Fiscal decentralization encourages inter-city competition among local governments, which are modeled as multi-task entities who are simultaneously the owners of SOEs, the providers of local public good, and the owners of urban land. Competitions among firms and cities represent the two driving forces behind China’s economic reform—foreign &

1 Another important category of enterprises in China is township and village enterprise (TVE). Since most of them are located in rural areas or small towns, I will not analyze them in this paper. 2 Gordon and Richardson (2001) also called for attention to market institutions in the urban setting. 3 This treatment of local government is similar to a multi-task agent in the principal-agent analysis (Holmstrom and Milgrom 1991).

3

private investment and fiscal decentralization. Without public land leasing, local public goods are completely capitalized in wage. With SOEs dwindling over time, wage capitalization increases wage cost for private firms, draining the sources of local public finance and local economic development. This vicious cycle pressures local governments to look for other revenue sources and triggers institutional reform. From this perspective, the adoption of public land leasing is no longer an accidental choice in urban land reform. I argue that public land leasing is Pareto improving because it eliminates free riding on the consumption of local public goods and establishes the link of rent capitalization. It also helps to shift local government’s role from a production manager to public goods provider. Hence, public land leasing not only establishes land market but also helps the transition of Chinese local governments.

The rest of the paper is organized into five parts. The following section is a brief introduction to China’s urban economy. The third section presents the basic models with and without land leasing. A Pareto efficiency test on the institutional change is analyzed in the fourth section. The fifth section contains discussion of some related issues. The final section is the conclusion.

II. BRIEF INTRODUCTION TO CHINA'S URBAN ECONOMY Around the introduction of public land leasing in the 1980s, China’s urban economy

could generally be regarded as consisting of two sectors. One is the state sector controlled largely by local government,4 while the other includes firms set up by foreign investments and other types of private investments. Since the beginning of the economic reform, SOEs and COEs largely operate on fixed factors, i.e., the same amount of labor on the same parcel of land (Byrd 1987). Foreign investment and the growth of the private sector are an important driving force behind China’s economic growth in the recent two decades, but, in the short run, its direct contribution to local public finance is limited due to two reasons. First, the share of foreign firms and private firms was relatively small, especially in the 1980s. The second reason is related to China’s favorable tax treatment to foreign investment. Most foreign firms enjoy much lower tax rates than SOEs because local governments want to attract foreign investment through tax reduction or even exemption.5 Moreover, private firms are more motivated for and better at evading tax than SOEs. Therefore, tax from foreign and private firms is not the main source of revenue for China’s local governments, at least in the 1980s. Over time, the two sectors' relationship has been changing. While the SOEs have been dwindling, the private sector has been growing very fast. This poses a challenge to China’s local public finance system that traditionally relied heavily on SOEs.

Another important feature of China’s economic reform is fiscal decentralization. World Bank (1990) regarded China as one of the most decentralized economies and many scholars have analyzed the benefits and problems of fiscal decentralization (see, for example, Qian and Weingast 1997; Wang 1997). Fiscal decentralization increases inter-city competition, which in turn pushes local governments to maximize fiscal revenue. There are several reasons why maximizing fiscal revenue becomes an important, if not overwhelming, objective of local

4 Here I intentionally avoid discussing the role of the central government given this paper's focus on local government. Some SOEs are actually controlled by the central government. 5 Gordon and Li (1999) studied why Chinese government treated foreign firms more favorably than domestic firms.

4

government.6 First, because of long-time ignorance of consumer welfare in socialist planned economy, many Chinese cities are lagging far behind in basic urban infrastructure such as road, water, and electricity. Any economic development requires huge investment into urban infrastructure. Zhu (1999:102) argues that “the primary goal for local governments is thus to mobilize as much revenue as possible in order to finance local development”. Second, local officials have personal interest in maximizing fiscal revenue. The bureaucratic hierarchy of China’s political system determines that local officials’ personal interest lies in promotion, power and corruption. Since promotion is now to some extent related to local economic development, this motivation of local officials reinforces the first reason. Power and corruption are both related to fiscal revenue under their control. One piece of clear evidence is that, before economic reform or in its early years, there were numerous small local enterprises that were set up by local governments to raise fiscal revenue even though their production scale was too small to be efficient. Rampant local protectionism is also strong in China. The link between local officials’ personal interest and fiscal revenue is, of course, the unchecked and omnipresent government power in China.

Unlike their counterparts in the West, Chinese local governments have multiple roles in addition to providing local public goods. Local governments have great interests in local SOEs simply because they are really their owners, or at least, they have some form of residual interest in these firms (Granick 1990; Walder 1994; Che and Qian 1998). Their profits are collected by local governments and constitute the main source of local public finance, especially before economic reform.7 Local government officials can also enjoy a lot of personal benefits from local enterprises under their control. Furthermore, local government is also the de facto owner of urban land. According to the Constitution of PRC all urban land belongs to “the public”. But, who represents the public remains unclear.8 With this ambiguity of property rights, local government exerts strong control over urban land use given its direct involvement in land use and the ability of hiding revenue.9 In summary, Chinese local governments are simultaneously the owner and even the manager of local SOEs, the provider of local public goods, and the owner of urban land.

Urban households in Chinese cities are relatively immobile at the inter-city level due to various constraints such as the Household Registration System (Cheng and Selden 1994), although their mobility has been improving over time. Their intra-city mobility is much higher. Au and Henderson’s (2002) study showed how severe the restrictions on mobility limit Chinese cities’ agglomeration and productivity. One consequence is that the change of labor force in SOEs is very small, especially in the 1980s. A large percentage of urban residents work in SOEs and their wages are more or less determined by bureaucrats (planned system) rather than by the 6 Starting from Niskanen (1971) and Brennan and Buchanan (1980), many studies in the public choice literature assumes revenue-maximizing government (Romer and Rosenthal 1982; Oates and Schwab 1988). 7 Nominally there was also a tax rate for SOEs, especially after the tax reform that changed profit-collecting system to taxation. However, since local governments still had strong control over these enterprises, their relationship was still very much like owner vs. factories. That is why one way for local governments to evade tax collection from the central government is to reduce or exempt local firms’ tax (Wang, 1997). 8 In the new Land Administration Law, the owner of all urban land is stipulated to be the State Council. This ownership designation is obviously nominal because land management is actually handled by the local government. 9 In early 1990s, China’s central government once tried to grab a share of the revenue from land leasing. Although the bill was passed, central government could not collect much money because local governments hid their land leasing revenue from the central government. This is why it is also very difficult to obtain data on local government’s land leasing activities. As a result, the new Land Law enacted in early 1999 no longer requires the local government to share land leasing revenue with the central government.

5

market. Comparing to urban residents, temporary rural-city migration of peasants has a much higher volume and constitutes a major source of labor for foreign and private firms (Knight and Song 1999). Officially dubbed “floating population,” they are much more mobile across the country. This is partly due to the segregated labor market (Knight and Song 1995) and segregated society (Chan 1994). Wage in foreign and private firms is determined by price mechanism of the market.

Several problems plagued China's urban land use and local public finance before the introduction of public land leasing. The first is that the incentive of providing local public goods was not aligned with its benefits. The existing revenue source of local government in the 1980s determined that local government was more interested in expanding the revenue base of SOEs rather than providing better local public goods. However, in the on-going economic reform SOEs are doomed to decline while economic growth requires large investment in local public goods such as various kinds of urban infrastructure. Therefore, local governments have to explore other sources of local public finance to meet the demand for public goods and face the competition among local governments. Second, private firms’ more or less free riding on local public goods also resulted in both inefficient and insufficient provision of local public goods. The last and most obvious problem is that land allocation under a non-price rationing system greatly distorted resource allocation. The administrative allocation system is often dubbed the “Three-No” (no price, no transfer, and no time limit) system. Due to unlimited demand for a free good, once the land user obtained the land it was extremely difficult to change hand. As typical of socialist cities, urban land use in pre-reform Chinese cities was relatively static in that there were few incentives to recycle land (Bertaud and Renaud 1994; Tang 1994). All these problems indicate the necessity of reform in urban land use and local public finance.

III. THE MODEL Because China’s economic reform has been proceeding in a gradual and partial way,

modeling its urban economy has to incorporate these characteristics of the transitional economy. The open city model that is often applied in the developing countries has to be changed in order to accommodate the relatively static SOE sector. Chinese local governments’ multiple roles and their behavior of maximizing fiscal revenue in a decentralized fiscal system also require broadening the usual assumption of government being only the provider of local public goods. I set up two models for the urban economy with and without public land leasing, respectively, and analyze wage and rent capitalization in both scenarios.

Basic Model without Land Leasing I assume an economic system of n cities that consists of two segregated sectors: private

(foreign) firms and SOEs. Private firms (or capital) are mobile while SOEs are largely fixed in location. In this sense, this is a partial open city model. For SOEs, local governments are their owners who collect profits from these enterprises within each city. Thus, all SOEs within a city are regarded as one big firm under the management of the local government. These SOEs are assumed to produce a numeraire good Y. Private firms produce a good X and compete with each other in the national market. Good X and Y can be the same good or different good. If X and Y are the same, then a different price for Y and a quota system need to be assumed to be imposed

6

on the economy so that this two-sector system can be maintained (Byrd 1987; Murphy, Shleifer and Vishny 1992).

A segregated labor market is also assumed. Urban households who work in SOEs are immobile at the inter-city level, but employees of private firms (mostly rural migrants) are assumed to be highly mobile across cities. Since private firms face much less constraints in hiring workers, they are more willing to hire rural migrants and thus their labor supply is more elastic. Wages in these private firms are set in a competitive labor market. In contrast, SOEs’ workers enjoy “iron bowls” and it is very difficult for SOEs to fire or hire workers. Although wages in SOEs are not determined by the market, but they face political pressure to keep the wage somewhat indexed to the market wages so that certain living standard can be maintained. Therefore, wages in the two sectors are related. For simplicity, I assume they are the same within the same city and there is only inter-city variations of wages.

Before the introduction of public land leasing, urban land is allocated by some non-price rationing system. There are still non-pecuniary transaction costs to obtain the land, such as bribes or waiting time, and local governments could not obtain financial benefits from these costs.

Traditionally, the SOEs are responsible for their workers’ housing that is allocated according to some administrative measures (Huang and Clark 2002). Since housing reform lags far behind land leasing, residential land for SOEs workers is assumed to be fixed. Land leasing only applies to private firms’ commercial and industrial land. Therefore, SOE workers’ wages are separate from housing, which is really a type of welfare and is associated with their low mobility. They are relatively passive in terms of utility maximization. But for workers in the private sector, they have to find housing by themselves. Housing is no longer welfare and housing cost becomes an important variable in their utility function.10

A public good, which can be consumed by all people in the same city, is produced by the local government. Private firms are assumed to be free riders of this good before the introduction of ground lease. Of course, there is inter-city variation for this public good.

Private Firm Private firm i located in city j employs a constant-return-to-scale production function f(.)

to produce good X for a competitive national market. The price of good X is p, which is exogenous to the model. The cost to obtain land lij is bj per unit of land, and the number of workers employed at wage wj is nij. The economic problem for the firm becomes:

Max ijijjij blnwpx −−

s.t. ( )jijijij alnfx ;,= Here x is the amount of good X produced by the firm. aj is the amount of public good that is provided by the local government. At equilibrium, unit cost equals product price:

( ) pabwc =;, (1) For simplicity, the subscript j is omitted in the equations. I assume the first order derivatives cw>0 and cb>0, i.e., the higher the wage and land cost, the higher the production cost for the private firms.

10 Giving the crowded housing condition in Chinese cities, the private firms usually need to provide housing for their workers, especially those from the countryside. Of course, these housing facilities have to be paid by the workers and the land used for housing is usually part of the industrial land allocated to the firm. In this sense, the housing is still allocated in the market.

7

Workers Workers are assumed to be identical in tastes and skills and derive utility from consuming

good X and Y with prices equal to 1 and p respectively. Because of the assumption of inter-city immobility, SOE workers in different cities might enjoy different levels of utility. Hence, they don’t face any utility maximization problem. In other words, SOE workers are in a very passive position and they don’t have much choice in the urban economic system.11

But, for workers in the private firms, their utility should be constant at equilibrium given their high mobility and competition in the job market and housing market. Since they are not guaranteed housing by government, housing cost and consequently the cost of land enters their utility function no matter that cost is pecuniary or in other forms (such as waiting time, bribe, etc.). Their economic problem is:

Max );,,( jkjhjhjhj aSyxu

s.t. hjjkjhjhj wbSyxp ≤++⋅ where whj is household income for the household k in city j, which is assumed to be equal to wage. Skj is the household’s housing consumption at the cost of bj. Assuming v(.) is the indirect utility function for workers and u0 is their equilibrium utility level, we have the following equation:

( ) 0;,, uabwpv = (2) It is assumed that vw>0, vb<0, which means that the workers' utility is positively associated with wage and negatively associated with land cost.

Local Government In the setting of fiscal decentralization, many studies have shown the limitations of

Tiebout model (Stiglitz 1977; Bewley 1981). One important limitation is some entrepreneurial behavior of local government is necessary to achieve efficiency. That is why many models in local public goods provision try to set up some optimizing objectives for local governments. Sonstelie and Portney (1978) assumed communities are profit-maximizing firms that are basically real estate developers who also provide public services.12 In Brueckner’s (1983) model aggregate property value maximization is the objective of local government. Although the behavior assumptions in these models capture important aspects of local public finance, their institutional assumptions may not be correct. For example, in Sonstelie and Portney’s model public goods and housing are both provided by the same government, but the interaction between these two local government functions are essentially through market (like rent capitalization). In other words, the two functions are assumed to be performed by completely independent entities that transact in the market. But, in reality, both functions are performed by the same government that has to allocate resources among these different functions inside the same institutional hierarchy. The interaction between the two roles is not in the market but through the coordination and command system inside the government. The objective of maximizing aggregate property value in Brueckner’s (1983) model also requires some additional links to the sole role of government as the provider of local public goods.13 Hence, how local government 11 This is in the same spirit as Hayek’s argument of “the road to serfdom” (1994). 12 Deng (2002) argues that a ground lease system, which integrates landowner and collective goods provider, effectively combines the core merits of Henry George with Tiebout Model. 13 In a sense, Fischel’s (2001) recent work can be regarded as substantiating these links.

8

allocates its effort among different functions is an important issue to the efficiency of local public finance.

This is especially true for Chinese cities where local governments have at least three roles. The first is the owner of SOEs who collects all profits of SOEs. The second role is to provide local public good. The third role is the owner of urban land, who is responsible for allocating land to firms. I assume the total effort from local government is a constant equal to 1 and the efforts allocated to the three tasks are e1, e2, and e3 respectively.14 Of course,

0,, 321 ≥eee . 1321 =++ eee (3)

For the local government’s first task, the production function of its SOEs is assumed to be a constant-return-to-scale function, h(.), with the inputs of labor Ns, land Ls, as well as the public good a. Here, Ns and Ls are exogenous variables because SOEs are assumed to be producing on fixed factors. As the owner of SOEs, local government’s effort in managing the enterprise directly impacts the production output of good Y, which is measured by y. For simplicity, a linear relationship is assumed between government effort and production output.

( )aLNhey ss ;,1 ⋅= (4) The second task of the local government is to provide local public good a. The provision

of the public good is assumed to be related to fiscal revenue B through function g(.). The effort allocated to this task, e2, is also assumed to be positively related to a. Thus we have the following equation:

( ) 02 aBgea +⋅= (5) g’(.) > 0, meaning that the higher the fiscal revenue the better the public good. The constant a0 represents the basic level of public service that is essential for the city. Assuming the revenue to the local government comes from a tax or (profit collection scheme) on the SOE output y with a flat rate t, which is assumed to be set by the central government and exogenous to the model, the local government’s fiscal revenue is

Byt =⋅ (6) Local government’s third task is to supply urban land. Since market mechanism is absent

before land reform, land supply is assumed to be decided by some local planner and is exogenous to the model. However, the cost to the private firms of obtaining land will be reduced if local government increases its effort in this task. If local government puts no effort into land administration, then the cost b is assumed to be a constant.

( )3esb = (7) ( )00 sb = (8)

I have discussed why Chinese local governments are interested in maximizing fiscal revenue. Now, facing the three tasks, local government’s objective is then to maximize its revenue ty. Because tax rate t is exogenous, the maximization problem to the local government then becomes

Max y s.t. 1321 =++ eee

14 The allocation of government effort is not assumed to be directly related to budget allocation or its marginal changes. The reason is that many other factors, such as the ranking of officials and their relationship to city leaders, may also determine the allocation of effort. For example, land administration inside a city may not cost much financially compared to other departments, but it may be more important given its potential to generate revenue.

9

From equation (4) to (6) it is easy to see that y is not determined by e3 and is positively related to e1 and e2. Therefore, in order to maximize y, e3 must be zero. Put in another way, local government has no incentive to reduce bureaucratic costs in land administration. Then the constraint for local government becomes:

121 =+ ee , 03 =e (9) Rearranging equation (4)-(6) yields

( ) ( )( ) 01 .1 athgea +−= (10)

Denoting ( )a

h∂

∂ . by ha and rearranging equations, the first order condition for the local

government’s maximization problem becomes: ( ) ( ) ( ) ( ) ( ) ( ) 0....'.'. 11 =+−− ghehhthgethgh aaa (11)

Equilibrium Conditions Now we have five equations (1), (2), (9), (10) and (11), for five variables (w, b, a, e1, e2).

The system is mathematically solvable. For simplicity, I assume a linear relationship between public good provision and fiscal revenue, 02 aBea +⋅⋅= β , and ignore second-order effects. Then, we can obtain the following result by differentiating the five equations.

( )sLsN dLEhdNEhD

da +=1

(12)

where ( ) ( ) ( ) ( )[ ] aaaa hhehhtheethD −++−−−= 111 .2.11 ββ

( ) ( ) ( )( )[ ] ( ).1..1 1111 hteheehthhetE aa +−−+−−= ββ It’s easy to see that inside this system the provision of the public good and the behavior

of local government are not affected by private firms’ cost function or workers’ utility function. The allocation of government effort and the level of public good are determined only by the relationship between government and SOEs, as part of the legacy from the planned economy.15

Because land cost is a constant at equilibrium, local public good is completely capitalized into wage. In other words, since land cost is the same across cities, private firms would like to move into cities with better public goods, resulting in higher demand for labor, the extent of which depends on the interaction of the location choices of both workers and firms. In the same way as for equation (12), we can solve for the wage gradient that measures the degree of the capitalization of local public good into wage.

w

a

w

a

v

vdadv

c

cdadp

dadw

−=

−=

0

(13)

Since the change of wage capitalization is mostly determined by how exogenous variables p and v0 change, it is interesting to see what is the effect on wage capitalization if there are some changes in the exogenous variables. China’s experience in the past two decades shows that SOEs have been dwindling and their products have become less and less important (Naughton 1995). Meanwhile, the average welfare (or utility) of workers has been rising.

15 As qualification, this result is obviously based on the assumption that private firms can free-ride on public goods and don’t need to pay tax. Given the dominating position of SOEs in the early years of reform, this assumption should be approximately true. The indirect effect from the growth of private firms is not considered in this model.

10

Therefore, it is reasonable to assume that over time Ns and Ls are declining while v0 and p are

rising, i.e., p

N s

∂∂

<0, pLs

∂∂

<0, 0v

N s

∂∂

<0, 0v

Ls

∂∂

<0.

To estimate the corresponding change in wage gradient, we need to make assumptions about the effect of the public good on SOEs. Several scenarios are possible (Table 1). First, if the public good is productive to SOEs and also enjoyable to workers, then ha > 0 and va ≥ 0.16 If ha is small enough, which is a reasonable assumption given the low efficiency of SOEs, then D>0 and E>0. With declining Ns and Ls, equation (12) indicates that public good a is also deteriorating. Therefore ∂a/∂p and ∂a/∂v0 are negative given the changes of the exogenous variables. Since vw > 0, ∂w/∂a will decrease in equation (13). This means that, if local public good doesn’t affect SOE production much but increases local workers’ utility, then better provision of public good will result in lower wages. Put in another way, lower wage is compensated by better local public service. Equation (13) also suggests that a rise in worker’s utility v0 has to be accompanied by a greater increase in private firms’ product price p given that ca is negative and va is positive. The improvement of worker’s welfare depends more on the growth of private firms than SOEs.

The second scenario is that the public good is not productive for SOEs, i.e., ha = 0 or even ha < 0. The maximization problem of local government then have a corner solution of e1 = 1 and e2 = 0. Local government will not put any effort into public good provision, but instead will only be interested in the production of local SOEs. The public good is a constant (a0) now and wage is constant across cities. Although this case may fit into some part of the stark picture of pre-reform planned economy, it is probably more useful in pointing out the inherent problems in this local public finance system than in analyzing economic transition. If public good becomes a constant, local government then loses its main method to attract private investment when the goal of reform is now the growth of private economy. Obviously, economic reform and the need to develop the private sector disrupt this corner-solution equilibrium. To succeed in inter-governmental competition, the mechanism of local public finance has to be changed.

It should be obvious from equation (13) that, once private firms are allowed to operate, the planned system becomes more and more unsustainable unless exogenous changes in ∂p/∂a and ∂v0/∂a (from the growth of private firms) also vary across cities and are big enough to offset the impact from the different signs of ca and va. With shrinking size of SOEs, local fiscal revenue based on SOE profits decreases. This in turn raises wage cost to private firms because the variation of local public good is completely capitalized into wage and they have to compensate workers for worse provision of local public good. Private firms will be less likely to invest and more likely to leave. SOEs are in a similar situation because their wage is indexed to the wage in the private sector and, hence, they also face a rising payroll bill. As a result, local economy will decline since both SOEs and private firms are shrinking. In this sense, any local economic development requires to take private firms into the loop of local public finance. Public land leasing becomes one of such solutions.

16 It is unlikely that the public good will not be enjoyed by workers given Chinese cities’ long-lagging urban infrastructure. In that rare scenario, the sign of wage gradient depends on the relative magnitude of the change of v0 and p versus ca and va.

11

Model with Land Leasing After the introduction of public land leasing and consequent emergence of a land market,

all private firms have to lease land from local government. In the land market, supply and demand has to be equal at equilibrium. SOEs still operate on the land allocated to them under the planned system. Total land leased to private firms in a city is denoted L and land rent is r. Transaction cost b still exists and is assumed to be a constant b0. Denote the total output from all private firms in the city as Q.17 Shepard's Lemma and the constant-return-to-scale production function give the total demand of land from private firms:

( )abrwQCL r ;, 0+= (20) Now the supply of land by the local government is also endogenous to the model and is

determined by both rent r and government effort e3. The land supply function is assumed to be in the following form:

( )relL ,3= (21) Here I assume le > 0 and lr > 0. The more effort local government puts into land supply and the higher is the rent, the more land is supplied by the local government.

The sources of fiscal revenue to the local government now expand from one (SOE profits) to two (SOE profits plus revenue from land leasing). The budget equation (6) is now the following:

BLryt =⋅+⋅ (22) The maximization problem facing the local government becomes: Max Lryt ⋅+⋅

s.t. 1321 =++ eee 0,, 321 ≥eee

The first order condition is then

332211 eL

rey

teL

rey

teL

rey

t∂∂

+∂∂

=∂∂

+∂∂

=∂∂

+∂∂

(23)

This basically means that marginal returns from investing efforts into the three tasks must be equal in order to maximize total revenue to the local government.

So, with land leasing introduced into the model, there are nine endogenous variables (Q, w, r, a, L, B, e1, e2, e3) and nine equations (1, 2, 3, 4, 5, 20, 21, 22, 23). The model is now mathematically solvable. Note that since Q only appears in (20), the system can actually be determined by the other eight equations. Differentiating (2) and (3) we can obtain the following wage and rent gradients.

( )

rwrw

rrrara

cvvcdadv

cdadp

vvccv

aw

−

−+−

=∂∂

0

(29)

( )

wrwr

wwwawa

cvvcdadv

cdadp

vvccv

dar

−

−+−

=∂

0

(30)

17 Because of the assumption of the same constant-return-to-scale production function for all private firms, the total output Q actually measures the spatial distribution of firms given the wage, rent and public good variations across cities.

12

The forms of wage and rent gradients are similar to those in a market economy (Roback 1982) in that they are now largely determined by the parameters in (1) and (2). The change in the relative size of private firms versus SOEs will affect wage and rent capitalization through p and v0. Although exact solution of wage and rent capitalization depends on specification of functional forms and parameters, some general conclusions can be obtained (Table 1).

First, from a cross-sectional perspective, i.e., with no change in p and v0, rent gradient is positive while wage gradient’s sign depends on the magnitudes of relevant parameters. When government shifts its effort to public good provision over time, i.e., ∂p/∂a>0 and ∂v0/∂a>0, it can be generally expected that rent gradient will increase while wage gradient will decrease. Second, in contrast to the model without land market, the provision of local public good is no longer determined only inside the public sector; rather, it is now largely determined by the private sector—private firms and workers. Competition among local governments can be in the true sense of Tiebout Model (1956) rather than through administrative measures like local protectionism. Third, there is now a positive feedback mechanism: providing better local public goods will result in higher rent and consequently higher land leasing revenue, which in turn makes better provision of the public good possible. Local governments have stronger incentives to provide the public good and shift its effort away from managing SOEs. Fourth, local government has to allocate its effort among the three tasks in a way that maximizes fiscal revenue. Compared to the system without land leasing, government effort in land administration and public good provision will increase.

To simplify discussion, let’s assume (5) to be the following linear form:

02 aBea +⋅⋅= β Then we can have

ssaa ZdLXdNLdrherdLheTdeSda +=−−− ββ 111 22 (31) ( ) 02222 122121 =−−−− deehrldeehrldathdrheel aeaeaae βββ (32)

where ( ) βββ thehrLehhtehS aaaa

2211

21 22.2 −−−=

( ) thehhteT aa ββ 211 2.2 +=

( ) βββ thhehrLehhteX NaNN211

21 22.2 ++=

( ) LaLL hhetrLhehhteZ 211

21 22.2 βββ ++=

Furthermore, if SOEs are assumed to be unable to benefit much from the public good,

i.e., if public good a is not productive to SOEs and ha is close to zero, then ( ).1

hey

=∂∂

and

02

=∂∂ey

. Substituting them into (23) yields the following results:

( ) 0. =h ; 03

=∂∂

=eL

le (36)

This means that, if SOEs cannot benefit from better provision of public goods due to their inherent institutional problems, the optimal solution is to shut down SOEs and have local government devote optimal effort to maximize land leasing. In the real world, certainly there are additional (political and social) constraints for maintaining SOEs. But this result confirms current direction of economic reform and the enthusiasm of local governments towards public land leasing in China’s urban land reform.

13

VI. A PARETO EFFICIENCY TEST The two models analyze the equilibrium conditions of a two-sector, partially open urban

economy under two different institutions—with or without public land leasing. Although some general static-comparative analysis has indicated the efficiency of public land leasing relative to the old administrative allocation system, as summarized in Table 1, how can we formally evaluate which one is more efficient? To answer the question whether or not public land leasing is efficiency-improving in the traditional framework of equilibrium analysis, I propose a simple test of the Pareto optimality following Greenwald and Stiglitz (1986).

The basic question is to ask whether land leasing would leave urban household welfare unchanged and increase government revenue. I assume an initial equilibrium for the model without land leasing exists, with no land revenue to the local government (r = 0). If this original equilibrium is Pareto optimal, the problem

Max LrytB ⋅+⋅= s.t. ( ) 0;,, varwpv = , for workers in the private sector.18

has a solution at r = 0. Here 0v is the equilibrium utility level. Note this is a necessary but not sufficient condition. If the maximization problem does not have a solution at r = 0, then the equilibrium without land leasing is not Pareto optimal and any land leasing related to r is a welfare improvement.

Also note that in the system without public land leasing, the constraint from households’ welfare is not necessary (binding) for local public good at equilibrium. Therefore, wage and land cost are not important in determining public spending and local public services. Local government’s behavior has nothing to do with the market or urban households’ welfare; it is more determined by SOE output. Intuitively, any reform that links local government behavior to the welfare of urban households and the costs of private firms should be efficiency enhancing.

The first-order condition is then

drdL

rLdrdy

tdrdB

⋅++=

For the initial equilibrium to be optimal, dB/dr must be equal to zero at r = 0.

( ) 0. 11

0=+

∂∂

+∂∂

=+==

Lra

hehre

tLdrdy

tdrdB

ar

(37)

Intuitively, this condition means that the marginal revenue from collecting the first dollar of land rent should be able to offset any consequent loss of tax, which results from the impact on SOE's production. With r = 0, S, T, X and Z have all become simpler and (31) and (32) becomes:

Lhedrde

Tdrda

S a β11 2=− (38)

( )a

ae

thheel

drda β2122 −

= (39)

Obviously we can solve for de1/dr. Substituting the solution into (37) yields 18 Consistent with the treatment of SOEs workers in the model, I ignore them in the above formulation of the optimization problem because they are in a passive position due to the legacy of planned economy and segregated labor market in a partial approach to reform. Another way of justification is that their utility can also be treated as the same as those in the private sector because their wage is assumed to be indexed to the latter.

14

( ) ( ) ( )[ ]222121

21 2.1 hhtehheelLthe aaea −+−+ βββ =0 (40)

Note that this condition does not contain any parameters from (1) and (2), the two basic equilibrium conditions for rent and wage capitalization. Again, for simplicity, if ha is assumed to be very small or close to zero, we can obtain a simpler condition:

)(2 012 aaee −= (41) This condition suggests that, if there were an equilibrium with zero rent, the ratio

between government effort devoted to public goods provision and that devoted to SOEs needs to be proportional to the public good provided (over the basic level). Put in another way, rent capitalization should not have any effect on the provision of public good. However, when ha is zero, the result in (36) shows that it is optimal to shut down SOEs, implying that government effort in managing SOEs (e1) should be zero or close to zero. Then, obviously, as long as government needs to input effort into providing local public good, i.e., e2 ≥ 0, condition (41) does not hold. This means that the initial equilibrium without land leasing cannot be Pareto optimal. It is not surprising that the system without public land leasing, which relies only on SOEs for local public finance, cannot be efficient. This system loses both sources of fiscal revenue since it is also optimal to shut down SOEs if they are not able to benefit from local public goods. The efficiency of public land leasing lies in the fact that it brings private firms into the determinants of public goods provision through rent capitalization.

This simple test on institutional change confirms the intuitions from the success of China’s urban land reform and re-orientation of local government functions. Table 1 summarizes how the introduction of land leasing affects government effort allocation and wage and rent gradients. Before urban land reform, a lot of anecdotal evidence suggests that local governments were much more interested in the production and expansion of SOEs instead of providing public goods. For example, before the economic reform, many cities set up small steel mills, petro-chemical plants, and cigarette and liquor factories in spite of the fact that these small factories are inefficient due to their small production scales. Local governments also tried various ways to protect local products from outside competitors. The motivation was simply to generate revenue from these high-value-added industries for local public finance. Another widespread phenomenon in Chinese cities at that time was that investments in SOEs had been increasing while public infrastructure had been deteriorating due to dwindling investment in local public goods. Although this was certainly related to the ignorance of consumer welfare in socialist planned economy (Kornai 1980), the old system of local public finance worsened this trend.

In addition to demonstrating the efficiency of market in allocating land resource, the success of public land leasing in China supports the theory of rent capitalization, especially its role in the endogenous evolution of urban institutions. In a compelling analysis based on review of empirical evidence for rent capitalization, Fischel (2001) argued that it explains the persistence of property tax for local public finance in America. This benefit view of property tax regards institutions in local public finance and urban land use as endogenous response to rent capitalization. This paper is in the same vein of thinking. Introducing public land leasing into an urban economy of two segregated sectors creates a positive feedback loop, or benefit link, among public goods provision, rent capitalization and local public finance. The vicious cycle of inefficient local public finance in traditional planned economy was broken. This becomes especially important after private firms start to grow. The fact that local governments have often been leading the central government in urban land reform is clear evidence of their enthusiasm for public land leasing.

15

Another merit of public land leasing is that it helps to shift government effort from managing SOEs to providing better public goods. Before economic reform, local government is more interested in building their factories even though they could only survive under local protectionism. This is certainly an inefficient misalignment of incentives not only because local governments are not good at managing firms (the “fatal conceit”, according to Hayek) but also because local protectionism slow down the formation of national market and the realization of scale economy. The introduction of public land leasing helps to mitigate this problem. It aligns local government’s interest in fiscal revenue with public goods provision. In this sense, it helps to re-orient government function out from the legacy of planned economy to the provision of public goods.

V. DISCUSSION

Other Forms of Property Rights? A legitimate question is: why doesn’t China simply adopt fee simple land ownership that

is more popular and more common in the world? In terms of model specification in this paper, land sale is not easily distinguishable from land leasing. The answer to this question must go beyond rent capitalization. Che and Qian (1998) argued that the widespread presence of Chinese local government ownership of firms is due to their ability to hide revenue from the central government and the predatory threat from government to private property rights. Ambiguous property rights arrangement also facilitates government’s preference for ex post negotiation instead of ex ante commitment (Li 1996). In this way, institutional ambiguity of property rights may be able to provide some leeway for a changing society (Ho 2001). Deng (2002a) pointed out it may be efficient to integrate landowner and collective goods provider in face of uncertainty so that the political hold-up problem could be mitigated. This argument can also be applied to China where uncertainty may come from a predatory state. Hence, fee simple ownership may not be a feasible or efficient choice in China’s special social and political circumstance, at least in the short term. Although the positive effects of public land leasing based on these arguments certainly exist under China’s special social and political constraints, they do not come without a price. Deng (2003) point out many problems inside China’s public land leasing.

Why Not Tax? It is apparent that the model specification in this paper can also be applied to property tax

or land use fees. The latter is now also collected by Chinese government. In practice, land use fee or tax is regarded as nominal when compared to the land-leasing premium charged up front. Theoretically, we can ask why doesn’t Chinese local government rely on land use fee or property tax instead of land leasing. A related question is why couldn’t they rely on taxing private firms.

The answer to the second question is partially positive. Tax from private firms is becoming more and more important in China, given their growing size and importance in the economy. But, two points need to be clarified. First, back in 1980s or even 1990s, it is not easy for the government to tax private firms. On one hand, even SOEs and local governments are good at hiding revenue in an economy that relies heavily on cash and anonymous bank accounts (Qian and Weingast 1997), let alone private firms. On the other hand, local governments have been providing various tax incentives to attract private investment. They are more interested in

16

private firms’ indirect impact on local economy than their direct contribution to tax. Second, even if it is not so difficult to collect tax from private firms, it may still not be the best choice for local public finance given the link between land rent and public goods provision. Even in developed countries, property tax persists in local public finance because of rent capitalization (Fischel 2001).

In a word, the difficulty with tax collection in combination with political merits of public land ownership determines property tax was not the first choice for Chinese local governments, at least during the 1980s and 1990s. Public land leasing becomes an efficient choice, at least for the transition purpose. With the gradual dilution of public landownership, property tax may become more and more attractive.

VI. CONCLUSION China’s urban land reform, which is based on public land leasing, has been transforming

its urban landscape at a huge scale. By incorporating the two driving forces behind China’s economic reform, namely fiscal decentralization and the entry of private firms, I construct a simple general equilibrium model to provide another perspective on the economic rationale of public land leasing. In addition to the conventional theory of privatization, I especially emphasize the role of rent capitalization in the evolution of urban institutions and property rights arrangement. Through the link of the capitalization of local public goods into land rent, public land leasing aligns the interest of landowner and collective goods provider, both of which are among the multiple roles of local government, and helps to shift government effort from managing SOEs to providing public good.

The model is based on the assumption of two segregated sectors: SOEs and private firms. Without the segregation, a partial and gradual reform strategy is doomed to failure (Murphy, Shleifer and Vishny 1992). SOEs are assumed to be producing with fixed factors and are more or less protected from the product market competition by administrative measures. In contrast, workers of SOEs are highly mobile across cities. Since Chinese local government is simultaneously the landowner, the provider of public good and the owner of SOEs, an important feature of the model is a multi-task local government that internally allocates its effort to maximize fiscal revenue.

The findings from the model show that, first, inside the old system without land leasing, local governments are more interested in increasing the revenue from SOEs and the effort devoted to the provision of public goods is kept at minimum, which is also determined solely within the public sector. After the introduction of land leasing, local governments shift their efforts from the management of SOEs to the provision of public good, due to the positive feedback mechanism based on rent capitalization. Private firms are now inside the loop of local public finance. Second, local public goods are completely capitalized into wage before land reform. With SOEs shrinking over time, inefficient local public finance will lead to higher wage cost and ultimately drive away private firms. In other words, once private firms enter the economy, the old system of local public finance starts a vicious cycle unless private firms cannot free ride on local public goods. Last, a simple test of Pareto efficiency shows that public land leasing is an efficient institutional change.

In some sense, land leasing is an approximate form of property tax under special institutional environment. It is also like an excise tax that helps to internalize externalities that arise due to the entry of private firms into urban economy and the inability of traditional local

17

public finance to include them. Rent capitalization and its impact on local public finance are obviously essential to the success of public land leasing.

18

REFERENCES

Au, Chun-Chung and Vernon Henderson. 2002. "How Migration Restrictions Limit Agglomeration and Productivity in China?" NBER Working Paper 8707

Bertaud, Alain and Renaud, Bertrand. 1994. Cities without Land Markets: Lessons of the Failed Socialist Experiment. Washington, DC: World Bank.

Bewley, Truman F. 1981. "A Critique of Tiebout's Theory of Local Public Expenditures," Econometrica 49 (3): 713-740

Brennan, Geoffrey and James M. Buchanan. 1980. The Power to Tax: Analytical Foundations of a Fiscal Constitution. Cambridge, London: Cambridge University Press.

Brueckner, Jan. 1983. "Property Value Maximization and Public Sector Efficiency," Journal of Urban Economics 14 (1): 1-15

Byrd, William A. 1987. "The Impact of the Two-Tier Plan/Market System in Chinese Industry." Journal of Comparative Economics 11:295-308.

Chan, Kam W. 1994. Cities With Invisible Walls: Reinterpreting Urbanization in Post-1949 China. Hong Kong: Oxford University Press.

Che, Jiahua and Yingyi Qian. 1998. "Insecure Property Rights and Government Ownership of Firms." The Quarterly Journal of Economics CXIII:467-96.

Cheng, Teijun and Mark Selden. 1994. "The Origins and Social Consequences of China's Hukou System." The China Quarterly 139:644-68.

Deng, F. Frederic. 2002. "Ground Lease-Based Land Use System versus Common Interest Development," Land Economics 78 (2): 190-206

Deng, F. Frederic. 2003. "The Political Economy of Public Land Leasing in Beijing, China. " In Steven C. Bourassa and Yu-Hung Hong ed., Public Leasehold: Policy Debates and International Experiences. Cambridge, MA: Lincoln Institute of Land Policy.

Fischel, William A. 2001. The Homevoter Hypothesis: How Home Values Influence Local Government Taxation, School Finance, and Land-Use Policies. Cambridge, MA: Harvard University Press.

George, Henry. 1879. Progress and Poverty. New York, NY: The Robert Schalkenbach Foundation.

Glaeser, Edward L. 1996. "The Incentive Effects of Property Taxes on Local Governments," Public Choice 89: 93-111

Gordon, Peter and Harry W. Richardson. 2001. "Hayek and Cities: Guidelines for Regional

19

Scientists. " In Ronald E. Miller and Michael E. Lahr ed., Perspectives in Regional Science: A Festschrift in Memory of Benjamin H. Stevens. New York: Elsevier.

Gordon, Roger H. and Wei Li. 1999. "Government As a Discriminating Monopolist in the Financial Market: The Case of China." NBER Working Paper.

Granick, David. 1990. Chinese State Enterprises: A Regional Property Rights Analysis. Chicago, IL: The University of Chicago Press.

Greenwald, Bruce C. and Joseph E. Stiglitz. 1986. "Externalities in Economics With Imperfect Information and Incomplete Markets." The Quarterly Journal of Economics 101:229-64.

Hamilton, Bruce W. 1976. "Capitalization of Intrajurisdictional Differences in Local Tax Prices," American Economic Review 66: 743-753

Hayek, Friedrich A. v. 1994. The Road to Serfdom. Chicago, IL: University of Chicago Press.

Ho, Peter. 2001. "Who Owns China's Land? Policies, Property Rights and Deliberate Institutional Ambiguity," The China Quarterly 166: 394-421

Huang, Youqin and William A. V. Clark. 2002. "Housing Tenure Choice in Transitional Urban China: A Multi-Level Analysis," Urban Studies 39 (1): 7-32

Knight, John and Lina Song. 1999. "Employment Constraints and Sub-Optimality in Chinese Enterprises." Oxford Economic Papers 51(2):284-99.

Kornai, Janos. 1980. Economics of Shortage. New York, NY: North-Holland Pub. Co.

Li, David D. 1996. "A Theory of Ambiguous Property Rights in Transition Economies: The Case of the Chinese Non-State Sector," Journal of Comparative Economics 23: 1-19.

Li, Ling Hin. 1997. “The Political Economy of the Privatisation of the Land Market in Shanghai,” Urban Studies. 34:321-335.

Murphy, Kevin, Andrei Shleifer, and Robert W. Vishny. 1992. "The Transition to a Market Economy: Pitfalls of Partial Reform." Quarterly Journal of Economics 107:889-906.

Naughton, Barry. 1995. Growing Out of the Plan: Chinese Economic Reform, 1978-1993. Cambridge, MA: Cambridge University Press.

Niskanen, William A. 1971. Bureaucracy and Representative Government. Chicago and New York: Aldine-Atherton.

Oates, Wallace. 1969. "The Effects of Property Taxes and Public Spending on Property Values: An Empirical Study of Tax Capitalization and the Tiebout Hypothesis," Journal of Political Economy 77: 957-971

Oates, Wallace E. and Robert M. Schwab. 1988. "Economic Competition among Jurisdictions: Efficiency Enhancing or Distortion Inducing?" Journal of Public Economics 35: 333-354

20

Qian, Yingyi and Barry R. Weingast. 1997. "Federalism As a Commitment to Preserving Market Incentives." Journal of Economic Perspectives 11(4):83-92.

Roback, Jennifer. 1982. "Wages, Rents, and the Quality of Life." Journal of Political Economy 90:1257-78.

Romer, Thomas and Howard Rosenthal. 1982. "Median Voters or Budget Maximizers: Evidence from School Expenditure Referenda," Economic Inquiry XX: 556-578

Sivitanidou, Rena and W. Wheaton. 1992. "Wage and Rent Capitalization in the Commercial Real Estate Market." Journal of Urban Economics 31:206-29.

Sonstelie, Jon C. and Paul R. Portney. 1978. "Profit Maximizing Communities and the Theory of Local Public Expenditure," Journal of Urban Economics 5: 263-277

Starrett, David A. 1981. "Land Value Capitalization in Local Public Finance," Journal of Political Economy 89 (2): 306-27

Stiglitz, Joseph E. 1977. "The Theory of Local Public Goods. " In M. S. Feldstein and R. P. Inman ed., Economics of Public Services. London: The MacMillan Press Ltd.

Tang, Wing-Shing. 1994. "Urban Land Development under Socialism: China between 1949 and 1977," International Journal of Urban and Regional Research 13: 392-415

Tiebout, Charles M. 1956. "A Pure Theory of Local Expenditure," Journal of Political Economy 64: 416-424

Walder, Andrew G. 1994. "Corporate Organization and Local Government Property Rights in China." In Vedat Milor ed., Changing Political Economies: Privatization In Post-Communist And Reforming Communist States. Boulder, Colorado and London: Rienner. pp. 53-66

Wang, Shaoguang. 1997. The Bottom Line of Decentralization. Beijing, P. R. China: China Planning Press.

World Bank. 1990. China: Revenue Mobilization and Tax Policy. Washington, D.C.: World Bank.

———. 1993. China: Urban Land Mangement in an Emerging Market Economy. World Bank.

Zhu, Jieming. 1999. The Transition of China's Urban Development: From Plan-Controlled To Market-Led. Westport, CT: Praeger Publishers.

21

Table 1: Summary of Different Scenarios

(Exogenous changes: p

N s

∂∂

<0, pLs

∂∂

<0, 0v

N s

∂∂

<0, 0v

Ls

∂∂

<0)

ha e1 e2 e3 ∂w/∂a ∂r/∂a

> 0 > 0 > 0 0 < 0 0 Without land leasing ≤ 0 1 0 0 0 0 > 0 > 0 > 0 > 0 ?

(decreasing) > 0

(increasing) With land leasing

≤ 0 0 > 0 > 0 ? > 0