From Growth-Oriented to Sustainability-Oriented: How Does the Transformation of Development Goals Reshape Urban Land Supply? An Analysis Based on a Spatial General Equilibrium Model

Abstract

Following the launch of the Sustainable Development Goals (SDGs) process at the Rio+20 Summit, China has progressively strengthened sustainability-oriented considerations in development target setting and administration cadre performance assessment, which provides an institutional window to examine how the transformation of development goals reshapes urban land supply patterns. This study develops a spatial general equilibrium model and uses panel data for 286 prefecture-level cities in China from 2007 to 2021 to examine how the transformation of development goals affects urban land supply patterns. The results show that higher economic growth targets significantly expand total land supply, raise the ratio of industrial to residential land supply, and tighten floor-area-ratio (FAR) regulation. “Soft constraint” wording dampens the effect on land supply scale but strengthens the effects on land supply structure and FAR regulation, while the degree of vertical and horizontal target escalation generates substantial heterogeneity in these relationships. Moreover, after governance shifted from growth-oriented to sustainability-oriented objectives, the marginal effectiveness of using land supply structure and FAR regulation to deliver predetermined growth targets declined significantly. This study provides empirical evidence and policy-relevant insights for improving sustainability-oriented target accountability systems and urban governance incentive mechanisms.

1. Introduction

For decades, local governments in China have been the monopolistic suppliers in urban land markets [1,2,3]. In practice, they can secure additional land quotas through mechanisms such as the “increase–decrease linkage” of urban–rural construction land indicators and through negotiations with the central government, thereby partially relaxing binding quota constraints [4,5]. Their objective is to convert as much land as possible into urban construction land, although these channels are subject to significant institutional and practical limits [6]. Consequently, local governments must operate largely within constrained quotas and deploy a range of land supply instruments to allocate scarce land resources. Land supply patterns shape urban form and the associated environmental and social pressures; therefore, they are closely linked to urban sustainability [7]. Historically, however, China’s development agenda has prioritized economic growth targets; under land-based fiscal incentives in particular, land supply has served as a central lever for local economic expansion. While this growth-centered approach supported rapid development in the short run, it also contributed to excessive resource depletion, environmental degradation, and rising social inequality [8,9]. With the advance of the global sustainable development agenda, China’s development objectives have increasingly shifted away from growth alone toward a more balanced emphasis on economic, social, and environmental outcomes, reflecting a broader commitment to sustainable development. Therefore, a key question is how the transition in development goals from growth-oriented to sustainability-oriented will systematically reshape local governments’ land supply toolkits and decision-making logic.

Government development objectives encapsulate local developmental priorities and intentions and serve as key instruments for building consensus, mobilizing resources, and allocating public assets. They therefore provide a useful lens for interpreting urban land supply patterns [10,11,12]. Goal setting is shaped by both higher-level directives and local development needs. Under the long-standing dominance of land-based fiscal incentives, these objectives functioned as an important mechanism through which local governments steered urbanization trajectories [13,14]. However, as the central government places greater emphasis on sustainable development goals, the model of relying solely on land supply to drive economic growth has revealed significant negative effects, posing challenges to Sustainable Development Goal 11. In response, local governments have begun to incorporate more multidimensional considerations into their development objectives—such as social equity, environmental protection, and resource conservation—to better align with rising domestic and global expectations for sustainable development.

Economic growth targets are a central component of goal-based governance and shape government behavior in the socio-economic field [15,16,17]. At the macro level, most predetermined growth targets were met between 2003 and 2013, nearly 80% of prefecture-level cities achieved their planned growth objectives, and the average realized growth exceeded the target by more than 30% [18]. However, the prior research also suggests that excessively high growth targets can lead local governments to prioritize new investment while neglecting other activities such as innovation. In addition, such target signals may also disrupt market-based credit allocation and distort factor prices—particularly in labor markets—thereby aggravating factor misallocation and, ultimately, undermining sustainable development [19,20,21]. Land is a key input into economic growth and one of the few policy instruments that local governments can directly manipulate. Under growth-target constraints—and given that local fiscal revenue in China relies heavily on tax income and land-related revenue—local governments tend to adjust land supply instruments, exploiting differences across land uses to pursue dual objectives of revenue generation and economic growth [18,22,23].

In 2012, the Rio+20 Summit helped launch the process leading to the global Sustainable Development Goals (SDGs) agenda and prompted reflection in China on reorienting its development objectives [24,25]. In 2013, the Central Organization Department issued the “Notice on Improving the Performance Evaluation of Local Party and Government Leadership Teams and Leading Cadres”, stressing that evaluations of local officials should take overall performance into account. The Notice explicitly discouraged treating regional GDP and its growth rate as primary indicators, prohibited growth-rate rankings, and cautioned against using GDP growth performance as the basis for evaluation grades. At the same time, it elevated a broader set of objectives—including livelihood improvement, ecological conservation, and resource conservation—alongside economic development, thereby encouraging adjustments in local development models. As the Party’s highest authority for cadre appointments and performance evaluation, the Central Organization Department’s policy signal marked a major shift away from a GDP-centered assessment system. This shift implies that indicators closely related to public welfare—such as livelihood outcomes and environmental governance—have become increasingly salient in shaping local leaders’ incentives. This policy reduces the weight of GDP in evaluating officials’ performance while strengthening accountability for multidimensional constraints such as resource conservation, environmental protection, and public welfare. Consequently, it diminishes the marginal benefit that local governments derive from achieving predetermined growth targets through land supply, thereby altering the marginal incentive for growth targets to drive land supply. Therefore, local governments may no longer pursue higher land transfer prices as the sole objective in their land supply strategies. Instead, they may focus on optimizing the structure of land supply and balancing social demand with the carrying capacity of the ecological environment. More broadly, the gradual incorporation of sustainability-oriented objectives into China’s policy framework provides an institutional context for analyzing changes in urban land supply patterns and carries important implications for sustainable urban development.

This study proceeds in three steps. First, it develops an analytical framework based on a quantitative spatial general equilibrium model to examine the interaction between government development objectives and urban land supply patterns. Second, using a panel dataset of Chinese prefecture-level cities, it estimates the effects of government economic growth targets on land-supply-related decisions. Finally, treating the 2013 reform of the government’s official performance appraisal mechanism as a nationwide institutional policy shock, it examines whether urban land supply patterns have changed significantly under the transition toward sustainable development goals. Overall, this study makes three contributions. (1) By analyzing urban land supply patterns through the lens of government development objectives, it offers new insights into the role of local governments in urban land markets. (2) It integrates development objectives and land supply patterns within a unified quantitative spatial general equilibrium framework, helping to strengthen the theoretical foundations for their relationship while incorporating China’s distinctive land institutions and labor-mobility frictions. (3) It provides evidence that the shift toward sustainability-oriented objectives is associated with systematic changes in urban land supply patterns, thereby extending the literature on how transformations in goal-based governance translate into local policy behavior.

2. Theoretical Analysis

Building on the analytical frameworks of Monte et al. and Hsieh et al. [26,27], this study constructs a spatial general equilibrium model and, on this basis, incorporates land institutional arrangements with Chinese characteristics as well as frictions in labor migration. Specifically, first, the utility function considers the effect on workers’ utility of the level of per capita public services provided by each city and introduces the efficiency of urban public service expenditure as an additional factor. Second, given the hukou (household registration) system, labor migration across regions entails an additional utility loss, and the model therefore further incorporates workers’ migration decisions across regions. Third, land is introduced as a factor of production into the analytical framework, and governmental regulation of the land market is explicitly modeled by incorporating the effects on the housing market of the total quota of construction land, the structure of land supply, and regulatory restrictions on FAR. The model comprises M heterogeneous cities and N individuals. Each city hosts its own representative production sector, which produces a homogeneous consumption good and provides employment opportunities for workers, and each worker has idiosyncratic preferences over cities and chooses a place of residence subject to migration frictions. To improve readability, the main text reports only the key steps in the model setup and derivations. Full derivations and technical details are provided in Supplementary S1.

2.1. Workers’ Utility

Workers can choose in which city to work and reside, and this choice depends on the overall utility they obtain in each city. Workers are homogeneous in all economic dimensions except for the migration cost incurred when moving from their place of household registration to a different city; whether a worker chooses to migrate is determined by the level of utility associated with living in each city. Upon migrating to a city, a worker allocates income to the consumption of a generic composite good and housing and derives utility from the public services provided by the city. In addition, workers are assumed to have intrinsic idiosyncratic preferences for certain cities. Following Monte et al. and Ahlfeldt et al., the utility function of workers in city $j$ is specified as follows [26,28]:

$$\begin{array}{c}{U}_{ij}={\epsilon }_{ij}{g_j}^{\eta }{\left({\displaystyle \frac{c_{ij}}{\theta }}\right)}^{\theta }{\left({\displaystyle \frac{h_{ij}}{1-\theta }}\right)}^{1-\theta }\end{array}$$

(1)

In Equation (1), $U_{ij}$ denotes the utility level of worker $i$ in city $j$; ${\epsilon }_{ij}$ captures worker $i$’s idiosyncratic preference for city $j$. When an individual’s preference for a given city is very weak, the worker will not choose to migrate there even if that city offers relatively high labor income or high-quality public services. The term ${\epsilon }_{ij}$ is assumed to be independently and identically distributed according to a Fréchet distribution with cumulative distribution function $F_{\epsilon }=exp\left(-x^{-\propto }\right)$, where $\propto >0$ measures the dispersion of the Fréchet distribution [29]. The variable $g_j$ represents the quality of per capita public services provided by city $j$; $\eta \in \left[0, 1\right]$ measures the efficiency of urban public service expenditure, that is, the extent to which public services raise residents’ utility. The variable $c_{ij}$ denotes the quantity of the generic composite good consumed by worker $i$ in city $j$, and $\theta$ is the share of labor income allocated to the consumption of this composite good. The variable $h_{ij}$ denotes the housing area consumed by worker $i$ in city $j$, and ($1-\theta$) $\in \left[0,1\right]$ is the share of labor income allocated to housing consumption.

Since workers’ income in each city is entirely derived from wages, their budget constraint can be written as:

$$\begin{array}{c}{p}_j^c{c}_{ij}+p_j^h{h}_{ij}\le {\xi }_j{w}_j\end{array}$$

(2)

In Equation (2), $p_j^c$ denotes the price of the generic composite good in city $j$; this price is identical for all workers and therefore does not carry the subscript $i$. The price of the composite good is normalized to one. The term $p_j^h$ represents the housing price in city $j$. The parameter ${\xi }_j$ denotes the share of workers in the total population of city $j$, which also reflects the scale of employment opportunities in the city and directly affects the city’s equilibrium wage level; its value lies in the interval [0, 1]. Finally, $w_j$ denotes the wage level in city $j$.

Solving the labor utility maximization problem yields the following indirect utility function for each worker in the city:

$$\begin{array}{c}{V}_{ij}={\epsilon }_{ij}{g_j}^{\eta }{\left({\xi }_j{w}_j\right)}^{\theta }{\left({\displaystyle \frac{{\xi }_j{w}_j}{p_j^h}}\right)}^{1-\theta }\end{array}$$

(3)

In Equation (3), $V_{ij}$ denotes the indirect utility of worker $i$ in city $j$.

2.2. Migration Decisions

By integrating out the idiosyncratic preference term in Equation (3), we obtain the following expression for the average indirect utility in each city:

$$\begin{array}{c}{V}_j={g_j}^{\eta }{\left({\xi }_j{w}_j\right)}^{\theta }{\left({\displaystyle \frac{{\xi }_j{w}_j}{p_j^h}}\right)}^{1-\theta }\end{array}$$

(4)

In Equation (4), $V_j$ denotes the average indirect utility of workers in city $j$. Accordingly, workers’ utility in city $j$ depends on local employment opportunities, wage income, housing prices, and both the quality of public services provided and the extent to which these services enhance individual utility.

Following Tombe et al. [30], let $m_{kj}$ denote the share of workers who migrate from region $k$ to city $j$, where region $k$ is the workers’ place of household registration. Then:

$$\begin{array}{c}{{\displaystyle \sum }}_j{m}_{kj}=1\end{array}$$

(5)

Workers incur an additional loss of utility when migrating across regions. Owing to the hukou (household registration) system, the costs of working and living outside one’s place of registration are heightened, primarily because migrants cannot fully enjoy equal access to urban public services, employment opportunities, and related rights. We therefore assume that the cost for a worker to migrate to a new city is $1/{\pi }_{kj}$, where a larger ${\pi }_{kj}$ implies a lower migration cost and hence a smaller loss of utility; ${\pi }_{kj}$ takes values in the interval [0, 1]. When choosing a destination, workers compare the expected utility $E\left({\pi }_{kj}{\epsilon }_{ij}{V}_j\right)$ across regions and select the region that yields the highest expected utility as their future place of work or residence. This expected utility incorporates the average indirect utility $V_j$ of city $j$, the worker’s idiosyncratic preference ${\epsilon }_{ij}$ for city $j$, and the potential utility loss associated with migrating to that city, captured by $1/{\pi }_{kj}$. By the law of large numbers, the share $m_{kj}$ of workers who migrate from region $k$ to city $j$ can be expressed as the probability that a worker residing in region $k$ chooses to migrate to city $j$.

$$\begin{array}{c}{m}_{kj}=Pr\left\{E\left({\pi }_{kj}{\epsilon }_{ij}{V}_j\right)\ge ma{x}_{j^{\prime }\ne j}\left[E\left({\pi }_{k{j}^{\prime }}{\epsilon }_{i{j}^{\prime }}{V}_{j^{\prime }}\right)\right]\right\}\end{array}$$

(6)

In Equation (6), $\mathrm{Pr}$ denotes the probability density function governing workers’ migration across regions.

Under the Fréchet assumption, evaluating the integral yields:

$$\begin{array}{c}{\mathrm{m}}_{\mathrm{kj}}={\left({\mathsf{\pi }}_{\mathrm{kj}}{\mathrm{V}}_{\mathrm{j}}\right)}^{\propto }/{\displaystyle \sum _{{\mathrm{j}}^{\prime }\ne \mathrm{j}}}{\left({\mathsf{\pi }}_{\mathrm{k}{\mathrm{j}}^{\prime }}{\mathrm{V}}_{{\mathrm{j}}^{\prime }}\right)}^{\propto }\end{array}$$

(7)

Equation (7) shows that workers’ migration decisions are influenced by migration costs, the average utility level in the destination city, and the degree of dispersion in the distribution of idiosyncratic preferences.

2.3. Production Technology

The goods consumed by workers in each city are produced by the production sector. Each city hosts a representative production sector that produces a homogeneous good, using technology, capital, labor, and productive land as inputs. Output corresponds to the city’s gross domestic product, and the product market is perfectly competitive. The production function of the representative firm in each city is specified as follows:

$$\begin{array}{c}{Y}_j=A_j{K}_j^{1-\alpha -\beta }{\left({\xi }_j{N}_j\right)}^{\alpha }{L}_{I,j}^{\beta }\end{array}$$

(8)

In Equation (8), $Y_j$ denotes the gross domestic product of city $j$; $A_j$ represents total factor productivity in city $j$. Although production technologies can be identical across cities, there are differences in production efficiency and factor input levels. The variable $K_j$ denotes the capital input used in production in city $j$; ${\xi }_j{N}_j$ denotes the labor input employed in production in city $j$; and $L_{I,j}$ denotes the quantity of productive land used in production in city $j$. The parameters $\alpha$, $\beta$, and $1-\alpha -\beta$ are the output shares of labor, land, and capital, respectively, each taking values in the interval [0, 1].

Let $r_k$, $w_j$, and $r_{I,j}$ denote, respectively, the rate of return to capital in city $j$, the wage of labor, and the unit price of productive land. Capital is assumed to be freely mobile across cities, so that the rate of return to capital is equalized across cities at a constant level ${\overline{r}}_k$. The total cost faced by the production sector in city $j$ is therefore given by:

$$\begin{array}{c}{C}_j={\overline{r}}_k{K}_j+w_j{\xi }_j{N}_j+r_{I,j}{L}_{I,j}\end{array}$$

(9)

By combining Equations (8) and (9) and solving the producer’s cost-minimization problem, the wage of labor, the rate of return to capital, and the price of productive land can be derived as:

$$\begin{array}{c}{w}_j=\alpha A_j{K}_j^{1-\alpha -\beta }{\left({\xi }_j{N}_j\right)}^{\alpha -1}{L}_{I,j}^{\beta }\end{array}$$

(10)

$$\begin{array}{c}{\overline{r}}_k=\left(1-\alpha -\beta \right)A_j{K}_j^{-\alpha -\beta }{\left({\xi }_j{N}_j\right)}^{\alpha }{L}_{I,j}^{\beta }\end{array}$$

(11)

$$\begin{array}{c}{r}_{I,j}=\beta A_j{K}_j^{1-\alpha -\beta }{\left({\xi }_j{N}_j\right)}^{\alpha }{L}_{I,j}^{\beta -1}\end{array}$$

(12)

2.4. Housing Supply

China’s land supply market has long been monopolized by the government, and governmental regulation of land directly shapes the functioning of the housing market. Under relevant land administration regulations, government regulation of land mainly operates through three interrelated stages. The first is the acquisition of construction land quotas. Based on the national master plan for land use, provincial governments and lower-level governments prepare their own regional land-use master plans, in which they determine the amount of construction land that may be used over a relatively long planning horizon. This aggregate quota constraint is largely determined by natural and geographical conditions, such as requirements for the protection of cultivated land and the availability of developable land resources within the administrative region and is therefore relatively exogenous from the perspective of local governments. The second is the land supply stage. Because the total amount of construction land quotas is strictly controlled, local governments must regulate the process of land supply when releasing the limited land available, including differentiated decisions regarding the scale of land transfer, the composition of land supply, and the mode of conveyance, over which they enjoy substantial discretion. The third is the land development and utilization stage. Although land is ultimately developed and used in the secondary market by production sectors and housing developers, local governments can impose specific use restrictions on each individual parcel at the plot level, primarily in the form of FAR regulations.

This paper classifies local governments’ operational land policy tools along three dimensions. (1) Land supply scale: determined by the total stock of land available for construction in a city (construction land indicators/quotas), which characterizes an expansionary land supply. (2) Land supply structure: captured by the share of residential construction land in total supply (or the residential-to-productive land ratio), which reflects the “productive–residential” allocation orientation. (3) FAR regulation: reflected in intensity constraints that govern the conversion of residential land into effective housing supply. The subsequent comparative static analysis examines how changes in the “growth weight” and “welfare weight” in the government’s objective function modify the marginal benefits and costs of these three policy tools and, in turn, induce systemic adjustments in land supply patterns.

Accordingly, suppose that the total amount of construction land available in city $j$ is constrained by ${\overline{L}}_j$. As defined above, $L_{I,j}$ denotes productive construction land, so non-productive construction land (which in the model refers specifically to land used for housing construction) is given by $L_{H,j}={\overline{L}}_j-L_{I,j}$. These two types of land together constitute the total stock of construction land available for urban development. We further specify the following relationship:

$$\begin{array}{c}{L}_{H,j}={\delta }_j{\overline{L}}_j\end{array}$$

(13)

In Equation (13), ${\delta }_j$ denotes the share of residential construction land in the total amount of land supply quotas in city $j$. This ratio also reflects the local government’s developmental orientation: in principle, if the local government places greater emphasis on promoting local economic growth, it will reduce the share of residential construction land and increase the amount of productive land; conversely, if the local government intends to stabilize regional housing prices, it will raise the share of residential construction land in total land supply.

In addition, at the land development and utilization stage, the conversion of residential construction land in the land market into housing units in the housing market is further influenced by the stringency of the FAR regulation, ${\kappa }_j$, imposed by the local government. This parameter likewise reflects the local government’s developmental orientation.

Further derivation yields the conditions for equilibrium in the housing market as follows:

$$\begin{array}{c}\left(1-\theta \right)w_j{\xi }_j{N}_j=p_j^h{\tau }_j{\kappa }_j^{-1}{\delta }_j{\overline{L}}_j\end{array}$$

(14)

In Equation (14), ${\tau }_j$ represents the efficiency of effective housing supply in city $j$. It captures the component of housing prices that cannot be explained by factors such as the scale of the labor force, wage levels, the amount of residential construction land, and the FAR.

2.5. Local Government Behavior

Local governments use land conveyance revenues to finance urban infrastructure construction and improvements in public services, thereby enhancing regional economic development and residents’ welfare. Land conveyance revenues account for a substantial share of local fiscal budgetary revenues; in 2020, this share reached 84%. Drawing on related studies [31], the model is simplified by assuming that all fiscal expenditures on public services or infrastructure construction are financed entirely by land conveyance revenues, so that the local government faces the following budget constraint:

$$\begin{array}{c}{G}_j={\sigma }_j{p}_j^h{L}_{H,j}+r_{I,j}{L}_{I,j}\end{array}$$

(15)

In Equation (15), $G_j$ denotes public expenditure in city $j$, and ${\sigma }_j$ represents the share of the price of residential construction land in the housing price in city $j$.

Next, we introduce the local government’s objective function. On the one hand, local governments are strongly committed to promoting regional economic development; on the other hand, they also care about the welfare of residents and seek to enhance their sense of well-being. Accordingly, the local government’s objective function can be specified as follows:

$$\begin{array}{c}ma{x}_{G_j}{Y_j}^{{\gamma }_j}{\left(N_j\int U_{ij}d{F}_{\epsilon }\right)}^{1-{\gamma }_j}\end{array}$$

(16)

In Equation (16), $Y_j$ is determined by the production function in Equation (15), and $U_{ij}$ is determined by the indirect utility function in Equation (3). The parameter ${\gamma }_j\in \left[0, 1\right]$ denotes the weight placed on city $j$’s GDP in the local government’s objective function. That is, the government’s development orientation is captured by ${\gamma }_j$: a higher ${\gamma }_j$ implies greater emphasis on short-term economic output and observable growth performance, whereas a lower ${\gamma }_j$ implies a stronger focus on resident utility, including housing, public services, and environmental quality. $1-{\gamma }_j$ represents the weight assigned to residents’ utility, capturing the importance the local government attaches to local welfare.

In the workers’ utility specification, $g_j$ denotes the quality of per capita public services provided by city $j$. After introducing the government sector, this specification can be further elaborated as follows:

$$\begin{array}{c}{g}_j={\displaystyle \frac{G_j}{{N_j}^{\lambda }}}\end{array}$$

(17)

In Equation (17), $\lambda$ represents the elasticity of the quality of public services with respect to city $j$’s population size, that is, the extent to which a given level of public services can be enjoyed by multiple individuals simultaneously without mutually diminishing each other’s benefits. The parameter $j$ takes values in the interval [0, 1], with a larger value indicating a higher degree of excludability (or rivalry) in the provision of public services.

Further solving the local government’s objective function maximization problem yields the following land supply structure arrangement for two land use categories: residential construction land and productive land:

$$\begin{array}{c} {\displaystyle \frac{L_{H,j}}{L_{I,j}}}={\displaystyle \frac{\left(1-\theta \right)\left(\alpha +\beta \right)\left(1-{\gamma }_j\right)}{\beta \left[{\gamma }_j+\eta \left(1-{\gamma }_j\right)\right]}}\end{array}$$

(18)

As Equation (18) shows, the allocation ratio between the two types of land is affected by factors such as the share of housing expenditure in urban residents’ total spending, the output shares of labor and land, the efficiency of public service expenditure, and the local government’s developmental orientation. Holding other parameters constant, a higher weight on GDP in the local government’s objective function implies a lower ratio of residential construction land to productive land, indicating a stronger governmental preference for supplying productive land. When ${\gamma }_j=1$, meaning that the local government cares only about economic development and pays no attention to residents’ welfare, we have ${\displaystyle \frac{L_{H,j}}{L_{I,j}}}=0$, and the local government allocates all construction land quotas to productive uses. Conversely, when ${\gamma }_j=0$, meaning that the local government is concerned only with residents’ welfare and disregards GDP growth, we have ${\displaystyle \frac{L_{H,j}}{L_{I,j}}}={\displaystyle \frac{\left(1-\theta \right)\left(\alpha +\beta \right)}{\beta \left({\gamma }_j+\eta \right)}}$. This implies that even if the local government’s sole objective is to maximize residents’ welfare, it will still allocate a certain proportion of construction land quotas to productive uses, which is consistent with intuitive reality: in an economic system without production, residents cannot obtain the basic goods necessary for life, let alone experience improvements in welfare. In addition, as the efficiency of public service expenditure increases, the ratio of residential to productive construction land begins to decline, because improvements in the effectiveness of public services in raising residents’ utility induce local governments to rely more on high-output production to sustain high levels of public spending.

By combining Equations (8) and (16), it can be seen that when ${\gamma }_j=1$, meaning that the local government cares only about economic development and not about residents’ welfare, the government’s objective function simplifies to $ma{x}_{G_j}{Y}_j$, which is independent of the stringency of FAR regulation on residential land. Conversely, when ${\gamma }_j=0$, meaning that the local government is concerned only with residents’ welfare and disregards GDP growth, the objective function can be written as $ma{x}_{G_j}{N}_j\int U_{ij}d{F}_{\epsilon }=ma{x}_{G_j}{\left({\displaystyle \frac{G_j}{{N_j}^{\lambda }}}\right)}^{\eta }{\left({\displaystyle \frac{C_j}{\theta }}\right)}^{\theta }{\left({\displaystyle \frac{H_j}{1-\theta }}\right)}^{1-\theta }$. Further derivation yields ${\displaystyle \frac{d{G}_j}{d{\kappa }_j^{-1}}}<0$ and ${\displaystyle \frac{d{H}_j}{d{\kappa }_j^{-1}}}>0$. Thus, the relationship between the local government’s objective preference and the stringency of FAR regulation is not a simple linear one; rather, it depends on the relative contributions of local public services and housing supply to improvements in residents’ utility.

3. Research Design

3.1. Model Specification

Based on the theoretical analysis above, this study specifies the following regression model to empirically examine the impact of economic growth targets on urban land supply decision-making:

$$\begin{array}{c}{Y}_{it}={\beta }_0+{\beta }_1{X}_{it}+{\beta }_2{Z}_{it}+{\mu }_i+{\delta }_t+{\epsilon }_{it}\end{array}$$

(19)

In Equation (19), $Y_{it}$ is the dependent variable, capturing city $i$’s land supply scale, land supply structure, and FAR regulation decisions in year $t$. Specifically, land supply scale is proxied by the total area of land supplied, and the number of land parcels supplied is used as an alternative measure for robustness checks. Land supply structure is proxied by the ratio of industrial land area to residential land area, and the ratio of the number of industrial land parcels to the number of residential land parcels is used as an additional robustness measure. FAR regulation is defined as in the previous section and is proxied by the stringency of FAR regulation on residential land, with FAR regulation on industrial land and the citywide average FAR regulation used as further supporting indicators. The variable $X_{it}$ is the independent variable of interest and denotes the economic growth target of city $i$ in year $t$. The coefficient ${\beta }_1$ captures both the direction and magnitude of the impact of economic growth targets on urban land supply decisions. The vector $Z_{it}$ contains city-level control variables, including natural geographic endowments, socio-economic development conditions, and characteristics of the city’s chief administrative official. Because the current year’s land supply plan is typically influenced by the previous year’s socio-economic conditions, all control variables are included with a one-year lag, which also helps to alleviate endogeneity concerns to some extent. The coefficient vector ${\beta }_2$ reflects the effects of these control variables on urban land supply decisions. The terms ${\mu }_i$ and ${\delta }_t$ denote city fixed effects and year fixed effects, respectively, used to control for unobserved factors that are time-invariant at the city level and city-invariant over time. The term ${\epsilon }_{it}$ is the idiosyncratic error term. The model is estimated using a high-dimensional fixed-effects regression approach with city-level clustered robust standard errors.

3.2. Variable Definitions

3.2.1. Dependent Variables

The dependent variables in this study capture urban land supply–related decision-making, specifically along three dimensions: land supply scale, land supply structure, and FAR regulation. First, micro-level land transaction records are aggregated to the prefecture-level city. Second, samples located in Xinjiang, Tibet, and Hong Kong, Macao, and Taiwan are excluded, as are observations belonging to autonomous prefectures, leagues, and administrative bureaus; only observations for the period 2007–2021 are retained. Finally, based on the available variables, three indicators are constructed: land supply scale, measured as the logarithm of total land supply area; land supply structure, measured as the ratio of industrial land area to residential land area; and FAR regulation, measured as the average statutory FAR regulation on residential land. The construction of the FAR regulation measure is described in detail in Supplementary S2.

3.2.2. Independent Variable

The independent variable in this study is the annual economic growth target at the prefecture-level city. This indicator is derived from local Government Work Reports. In cases where the economic growth target in the work report is specified as an interval, the midpoint of the upper and lower bounds is taken as the economic growth target for that year, which is consistent with common practice in the literature.

3.2.3. Control Variables

The control variables are defined at the city level. Following related studies, this study selects a range of factors that may influence local governments’ urban land supply decisions, including administrative area, population density, level of economic development, industrial structure, capital input, labor input, human capital, infrastructure investment, and fiscal autonomy. All continuous variables are expressed in logarithmic form to mitigate the impact of heteroskedasticity. In addition to city characteristics, a large body of research shows that local chief officials exert an important influence on regional economic and social development, particularly that the tenure of local officials significantly affects land conveyance behavior [32,33]. This study therefore further controls for the tenure of the municipal Party secretary. The data on local officials’ characteristics include only the month and year of appointment and the month and year of departure from office, from which the length of tenure is calculated. It is worth noting that if a local official assumes office in the second half of a given year, the tenure is counted as beginning in the following year.

The main variables and their definitions are presented in

Table 1. Selection and Definition of Main Variables.

Variable Type	Variable Name	Unit	Definition
Dependent variable	Land supply scale	hectares	Logarithm of total land supply area
	Land supply structure	%	Ratio of industrial land area to residential land area
	FAR regulation	—	FAR regulation for residential land
Independent variable	Economic growth target	%	Target GDP growth rate
Control variable	Administrative area	square kilometers	Logarithm of administrative land area
	Population density	persons/sq. km	Logarithm of population density
	Level of economic development	yuan	Logarithm of GDP per capita
	Industrial structure	%	Share of secondary industry value added in GDP
	Capital input	%	Ratio of total fixed asset investment to GDP
	Labor input	10,000 persons	Logarithm of the average number of on-the-job employees
	Human capital level	10,000 persons	Logarithm of the number of university students in the city
	Infrastructure investment	km/person	Logarithm of per capita road mileage
	Fiscal autonomy	%	(Budgetary revenue − budgetary expenditure)/GDP
	Promotion incentive	years	Tenure of the municipal Party secretary

.

3.3. Data Sources

The data used in this study comprises panel data from 286 prefecture-level cities in China spanning 2007 to 2021. Specifically, it includes land supply data, economic growth targets and binding indicator data, urban statistical data, and local official characteristic data. After collecting and processing individual datasets, they were matched and merged to form the final database adopted for subsequent regression analysis.

(1)

Land Supply Data

At present, officially published city-level statistics on land supply are mainly drawn from the Land and Resources Statistical Yearbook, which reports information such as annual land conveyance area and land-use structure but does not contain data on FAR, the core variable required in this study. This study therefore relies on an official micro-level parcel transaction dataset obtained from the China Land Market website, which has been widely used in recent academic research and records every parcel of land supplied by local governments in the land market since 2000. The dataset includes detailed information on each parcel’s administrative location (county/district level), electronic supervision number (a unique parcel identifier), project name (which reflects the specific location of the parcel), area of the transferred parcel, land source (including newly added construction land and stock construction land), land use (two-level land-use classification), mode of land supply (allocation, “bid–auction–listing” and negotiated transfer, leasing, and contribution of land as equity, among others), land-use term, associated industry, land grade (reflecting the locational value of the parcel), transaction price, installment payment arrangements, land-use right holder, agreed FAR (i.e., the statutory FAR set by the local government, including upper and lower bounds), and key time nodes related to transaction and development. Based on the raw data extraction, the dataset is cleaned by reconciling changes in city names and merged parcels, reclassifying land uses, reclassifying land sources and supply modes, removing outliers in land area and transaction price, eliminating duplicate records, and reprocessing the agreed FAR, resulting in a final sample of more than 2.2 million observations. Finally, the city-level land supply data aggregated from the micro dataset are compared with the statistics reported in the Land and Resources Statistical Yearbook in terms of both total land supply area and total land transaction value, thereby indirectly validating the reliability of the city-level land supply measures constructed from the micro data.

(2)

Economic Growth Targets and Binding Indicator Data

Data on economic growth targets and binding indicators are primarily drawn from annual Government Work Reports. Each year, the government work report is typically delivered by the mayor to the local people’s congress at the beginning of the year and submitted for deliberation, and the full text is archived under the “government information disclosure” section on local government websites. This study manually collects and compiles annual Government Work Reports at both the provincial and prefecture-level city. A preliminary reading of these reports indicates that, at both levels, the content is generally structured into a review of the previous year’s work and an arrangement of work for the coming year; in the opening year of each Five-Year Plan, the reports also summarize the achievements of the previous five years and outline the work plan for the next five years. In the section on work arrangements for the new year, the first item is typically a list of major expected targets, which includes the core indicator in this study—the economic growth target—as well as the binding indicators used in the subsequent analysis. The position and wording of these two types of indicators are largely consistent across years within a given report, and the expressions used across provinces, and prefecture-level cities are also highly similar. Accordingly, this study uses Python 3.8-based text analysis to extract economic growth targets and binding indicators from provincial and municipal Government Work Reports. It then compiles annual economic growth targets and binding indicators at both the provincial and prefecture-level city levels using additional tools.

(3)

City-Level Statistical Data

City-level statistical data are mainly obtained from the China City Statistical Yearbook and the China Urban–Rural Construction Statistical Yearbook.

(4)

Local Officials’ Characteristics Data

Data on local officials’ characteristics are mainly obtained from the local leaders’ database on People.cn and other publicly available media sources. From these sources, this study extracts information on the names and terms of office of municipal Party secretaries at the prefecture-level city. The municipal Party secretary, rather than the mayor, is taken as the representative of the local chief executive because, in practice, the Party secretary is the top decision-maker in local socio-economic governance [34,35] and continues to wield decisive influence over urban land supply decisions. Accordingly, the dataset on local officials focuses on the characteristics of municipal Party secretaries.

Table 2. Descriptive statistics of the main variables.

Variable Name	Observations	Mean	Std. Dev	Min	Max
Land supply scale	4217	6.219	0.979	0.011	9.027
Land supply structure	4091	2.879	4.663	0.145	29.312
FAR regulation	4024	0.412	0.243	0	1
Economic growth target	4158	10.268	3.211	1	31
Administrative area	4217	9.364	0.805	7.016	12.474
Population density	4217	5.741	0.902	1.792	7.882
Level of economic development	4217	10.436	0.717	4.605	13.056
Industrial structure	4217	46.954	11.046	8.850	85.640
Capital input	4217	81.660	44.320	4.814	451.149
Labor input	4217	3.488	0.803	0.693	6.650
Human capital level	4217	4.693	0.992	0.693	7.166
Infrastructure investment	4217	3.595	0.574	0.600	6.925
Fiscal autonomy	4217	−11.428	10.318	−138.328	6.393
Promotion incentive	4217	2.750	1.666	1	10

reports the descriptive statistics of the main variables. Three points merit attention. First, the standard deviations of some variables are relatively large. For example, the minimum value of urban land supply structure is 0.15, whereas the maximum reaches 29.31. The internal variation in economic growth targets is also substantial, with the maximum target in the sample being 31% and the minimum only 1%. An inspection of the raw data indicates that most of the relatively low economic growth targets appear during the COVID-19 pandemic period. Second, both the mean and the minimum of fiscal autonomy are negative. Given its definition as (budgetary revenue − budgetary expenditure)/GDP, this suggests that most local governments operate under fiscal deficits. Finally, although the tenure of municipal Party secretaries ranges from 1 to 10 years, the majority of secretaries leave office within three years, a pattern that is consistent with findings in the existing literature [36].

4. Empirical Results

4.1. Benchmark Model Results

The baseline regression results are reported in

Table 3. Effects of economic growth targets on urban land supply strategies.

Variable	(1) Land Supply Scale	(2) Land Supply Structure	(3) Residential FAR Regulation	(4) Industrial FAR Regulation	(5) Average FAR Regulation
Economic growth target	0.047 ***	0.135 **	0.022 *	0.008	0.021 **
	(0.007)	(0.065)	(0.011)	(0.006)	(0.009)
Administrative area	0.064	0.051	0.148	0.009	0.163 *
	(0.251)	(1.779)	(0.099)	(0.063)	(0.085)
Population density	0.473	2.293	0.088	0.093	0.133
	(0.536)	(2.475)	(0.103)	(0.064)	(0.085)
Level of economic development	0.509 ***	−0.080	−0.338 ***	−0.008	−0.288 ***
	(0.127)	(0.441)	(0.088)	(0.050)	(0.068)
Industrial structure	0.006	−0.005	0.004	0.000	0.002
	(0.004)	(0.021)	(0.004)	(0.002)	(0.003)
Capital input	0.002 ***	−0.004	−0.000	−0.000	−0.000
	(0.000)	(0.003)	(0.001)	(0.000)	(0.001)
Labor input	0.105	−0.174	−0.024	0.088	−0.005
	(0.073)	(0.447)	(0.102)	(0.064)	(0.088)
Human capital level	0.035	0.544 *	0.002	−0.050 **	0.004
	(0.030)	(0.278)	(0.041)	(0.025)	(0.037)
Infrastructure investment	−0.216	0.947	−0.315 ***	−0.076	−0.197 **
	(0.179)	(0.782)	(0.116)	(0.068)	(0.094)
Fiscal autonomy	−0.015 ***	−0.007	−0.000	0.002	0.001
	(0.004)	(0.020)	(0.004)	(0.002)	(0.003)
Promotion incentive	−0.009	−0.027	0.001	−0.007	−0.005
	(0.006)	(0.052)	(0.008)	(0.005)	(0.007)
City fixed effects	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes
Observations	4149	4023	3957	3633	4137
Adjusted R-squared	0.777	0.233	0.143	0.228	0.126

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The values in parentheses denote robust standard errors.

. Column (1) focuses on land supply scale and shows a significantly positive coefficient on the economic growth target, indicating that higher growth targets are associated with a larger land supply area. Specifically, a one-percentage-point increase in a city’s economic growth target is associated with a 4.7% increase in land supply area. Column (2) examines land supply structure and also yields a significantly positive coefficient on the economic growth target, suggesting that higher growth targets are associated with a higher industrial-to-residential land supply ratio. Quantitatively, a one-percentage-point increase in the growth target is associated with a 0.14-percentage-point increase in the land supply structure measure. Column (3) reports the results for residential FAR regulation, controlling for city-level characteristics as well as region and year fixed effects. The coefficient on the economic growth target is significantly positive, indicating that higher growth targets are associated with tighter controls over residential FAR. Column (4), which likewise includes city characteristics and city and year fixed effects, shows that the coefficient on economic growth targets is not statistically significant, implying that economic growth targets do not have a significant effect on FAR regulation for industrial land. By contrast, column (5) reports that the coefficient on economic growth targets is positive and statistically significant when the dependent variable is the citywide average FAR regulation, indicating that economic growth targets significantly increase the overall stringency of FAR regulation at the city level.

4.2. Robustness Checks

To demonstrate that the baseline regression results are robust and credible, this study conducts robustness checks from two perspectives. First, because the sample period 2007–2021 covers major shocks such as the global financial crisis and the COVID-19 pandemic, the first robustness exercise excludes observations for the years affected by the financial crisis and the pandemic and then re-estimates the baseline specifications on the restricted sample to avoid contamination by these exceptional years. Second, given that economic growth targets have a significant impact on urban land supply decisions, it is plausible that local governments’ land supply behavior may in turn influence the setting of economic growth targets, raising concerns about reverse causality and related endogeneity. To address this issue as far as possible, this study employs an instrumental variables (IV) approach. The robustness check results are reported in Supplementary S3. Both checks support the robustness of the benchmark regression.

4.3. Heterogeneity Analysis

The wording of economic growth targets in Government Work Reports suggests that many local governments specify these targets as binding indicators while using expressions such as “around,” “within a range,” “strive to achieve,” or an explicit fluctuation band. This study treats such wording as indicating a soft-constraint target. These soft constraints capture the degree of flexibility in target setting and may affect urban land supply decisions even under the same stated growth target. Accordingly, this study constructs a binary indicator for the presence of soft constraints and then conducts subsample regressions to examine heterogeneity. The construction of the soft-constraint indicator is described in detail in Supplementary S4.

The subsample regression results based on whether economic growth targets are accompanied by soft constraints are reported in

Table 4. Economic growth targets with and without “soft constraints”.

Variable	Land Supply		Land Supply Structure		FAR Regulation
	(1) No Soft Constraint	(2) With Soft Constraint	(3) No Soft Constraint	(4) With Soft Constraint	(5) No Soft Constraint	(6) With Soft Constraint
Economic growth target	0.050 *** (0.008)	0.024 * (0.014)	0.138 * (0.082)	0.179 ** (0.089)	0.011 (0.010)	0.048 ** (0.021)
City fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Observations	2661	1262	2586	1213	2535	1225
Adjusted R-squared	0.763	0.839	0.224	0.266	0.144	0.155

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The values in parentheses denote robust standard errors.

. The table is divided into three panels, with the dependent variables being land supply scale, land supply structure, and FAR regulation, corresponding to columns (1)–(2), (3)–(4), and (5)–(6), respectively. All specifications control for city characteristics as well as city and year fixed effects. Columns (1) and (2) show that, regardless of whether soft constraints are present, the coefficient on the independent variable—economic growth targets—is positive and statistically significant, indicating that economic growth targets exert a significant positive effect on land supply scale in both subsamples. However, when soft-constraint indicators are absent, the estimated coefficient on the economic growth target is larger and more statistically significant. This pattern suggests that soft constraints weaken the rigidity of target attainment, thereby reducing the marginal pressure on local governments to rely on land supply scale to meet their targets. Columns (3) and (4) indicate that, with or without soft constraints, the coefficient on economic growth targets remains positive and statistically significant, implying that economic growth targets have a significant positive effect on land supply structure in both cases. In contrast to the scale results, the coefficient is smaller and less significant when soft constraints are absent, suggesting that the presence of soft constraints strengthens the effect of a given growth target on land supply structure. Columns (5) and (6) show that when soft constraints are present, the coefficient on economic growth targets is positive and statistically significant, whereas it becomes statistically insignificant when soft constraints are absent. This pattern indicates that attaching soft constraints to economic growth targets amplifies the influence of a given target value on the stringency of FAR regulation. The two groups in the table differ markedly in sample size. This imbalance reflects common target-setting language in Government Work Reports: point targets (without qualifiers such as “around” or “within a range”) are the prevailing standardized form, whereas softening expressions—such as “around”, “within a range”, “between”, or “within a band”—are relatively rare.

To capture the phenomenon of “layer-by-layer escalation” embedded in economic growth targets, this study further distinguishes between the degrees of “vertical escalation” and “horizontal escalation” in the heterogeneity analysis. First, under China’s top-down administrative contracting system, targets set by higher-level governments are naturally amplified at each successive lower level. At the same time, promotion competition among local officials is more intense within the same province [37], and officials seek to demonstrate their competitive advantage to higher-level authorities by improving economic performance in their own jurisdictions. This dynamic encourages mutual competition among lower-level governments and leads them to ratchet up the economic development targets handed down from above [38], thereby generating both top-down “vertical escalation” and “horizontal escalation” among peer governments. Vertical escalation captures heterogeneity in the intensity of pressure transmission across multi-tiered governance structures. Accordingly, when cities face different accountability requirements and performance weightings for the same growth targets, they respond with distinct combinations of land policy instruments. Second, although there is a clear hierarchical difference between provincial departments and prefecture-level cities in the vertical dimension, prefecture-level cities also compete with one another within the same province and therefore tend to take the provincial government’s economic growth target as an important reference when setting their own targets. Finally, the heterogeneity in cities’ land policy responses to identical growth targets and performance metrics reflects not only institutional design—such as soft versus hard language and vertical versus horizontal pressures—but also pre-existing population density, migration pressures, and regional inequality [39]. Using only the city’s own economic growth target as the independent variable would make it difficult to net out the influence of such underlying factors on the dependent variables. For this reason, this study employs the ratio of a prefecture-level city’s economic growth target to the corresponding provincial growth target, which not only captures the degree of vertical escalation but also helps to filter out the impact of city-specific endowment differences on urban land supply decisions. A binary variable is then constructed based on whether the degree of vertical escalation is above the median, and subsample regressions are conducted for the heterogeneity analysis.

The subsample regression results based on the degree of vertical escalation of economic growth targets are reported in

Table 5. Degree of “vertical escalation” in economic growth targets.

Variable	Land Supply Scale		Land Supply Structure		FAR Regulation
	(1) Below 50%	(2) Above 50%	(3) Below 50%	(4) Above 50%	(5) Below 50%	(6) Above 50%
Economic growth target	0.031 ** (0.014)	0.045 *** (0.010)	0.007 (0.129)	0.200 ** (0.099)	0.010 (0.022)	0.032 ** (0.015)
City fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Observations	1951	2185	1875	2133	1861	2096
Adjusted R-squared	0.818	0.773	0.229	0.265	0.165	0.157

Note: *** and ** indicate significance at the 1% and 5% levels, respectively. The values in parentheses denote robust standard errors.

. The table is divided into three panels corresponding to the three dimensions of urban land supply decisions—land supply scale, land supply structure, and FAR regulation—which are presented in columns (1)–(2), (3)–(4), and (5)–(6), respectively. All specifications control for city characteristics as well as city and year fixed effects. Columns (1) and (2) show that, regardless of whether the degree of vertical escalation is high or low, the coefficient on the independent variable—economic growth targets—is positive and statistically significant, indicating that economic growth targets exert a significant positive effect on land supply scale in both subsamples. However, in the subsample where the degree of vertical escalation exceeds the median, the coefficient on economic growth targets is larger and more statistically significant, suggesting that a higher degree of vertical escalation strengthens the impact of a given target value on land supply scale. Columns (3) and (4) indicate that, in the subsample with a degree of vertical escalation above the median, the coefficient on economic growth targets is positive and statistically significant, implying that economic growth targets have a significant positive effect on land supply structure. By contrast, in the subsample with a degree of vertical escalation below the median, the coefficient on economic growth targets is positive but not statistically significant, suggesting that vertical escalation amplifies the influence of a given growth target on land supply structure. Columns (5) and (6) show a similar pattern: when the degree of vertical escalation is above the median, the coefficient on economic growth targets is positive and statistically significant, indicating a significant positive effect on FAR regulation, whereas in the subsample with a lower degree of vertical escalation, the coefficient is positive but not statistically significant. This again suggests that vertical escalation strengthens the effect of a given growth target on the stringency of FAR regulation. Overall, the deeper the degree of vertical escalation, the greater the impact of local governments’ economic growth targets on urban land supply decisions.

To measure horizontal intensification, this study follows established approaches in the literature [40]. First, it constructs an intra-provincial spatial weight matrix based on geographic distance by retaining only the K closest cities to each city within the province and assigning zero weights to all other cities; the matrix is then row-normalized to compute, for each prefecture-level city, the average economic growth target of its neighboring cities within the same province. Next, this study uses the ratio of a prefecture-level city’s economic growth target to the average target of its neighboring prefecture-level cities within the province as a proxy for horizontal intensification. A binary indicator is then defined as one if horizontal intensification exceeds the median and zero otherwise. Finally, heterogeneity is examined via subsample regressions. This study focuses on intra-provincial neighbors, rather than cross-provincial neighbors or broader metropolitan reference groups, because cadre assessment and promotion competition are primarily organized along provincial boundaries and because information diffusion is more likely to operate within these institutional boundaries, consistent with the related literature [36,40].

The subsample regression results based on the degree of horizontal escalation of economic growth targets are reported in

Table 6. Degree of “horizontal escalation” in economic growth targets.

Variable	Land Supply Scale		Land Supply Structure		FAR Regulation
	(1) Below 50%	(2) Above 50%	(3) Below 50%	(4) Above 50%	(5) Below 50%	(6) Above 50%
Economic growth target	0.048 *** (0.012)	0.045 *** (0.011)	0.080 (0.116)	0.325 ** (0.154)	0.030 * (0.018)	0.081 *** (0.020)
City fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Observations	2007	2062	1933	2015	1918	1989
Adjusted R-squared	0.806	0.770	0.194	0.267	0.160	0.145

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The values in parentheses denote robust standard errors.

. The table is divided into three panels corresponding to the three dimensions of urban land supply decisions—land supply scale, land supply structure, and FAR regulation—which are presented in columns (1)–(2), (3)–(4), and (5)–(6), respectively. All specifications control for city characteristics as well as city and year fixed effects. Columns (1) and (2) show that, regardless of whether the degree of horizontal escalation is high or low, the coefficient on the independent variable—economic growth targets—is positive and statistically significant, and the magnitudes of the coefficients are very similar. This indicates that economic growth targets exert a significant positive effect on land supply scale, with little difference between the high and low horizontal escalation subsamples. Columns (3) and (4) reveal that, in the subsample where the degree of horizontal escalation exceeds the median, the coefficient on economic growth targets is positive and statistically significant, implying that economic growth targets have a significant positive effect on land supply structure. By contrast, in the subsample with a degree of horizontal escalation below the median, the coefficient on economic growth targets is positive but not statistically significant, suggesting that a higher degree of horizontal escalation strengthens the impact of a given growth target on land supply structure. Columns (5) and (6) show a similar pattern for FAR regulation: when the degree of horizontal escalation is above the median, the coefficient on economic growth targets is positive and statistically significant, indicating a significant positive effect on the stringency of FAR regulation. In the subsample with a lower degree of horizontal escalation, the coefficient remains positive but its level of statistical significance declines, again indicating that horizontal escalation amplifies the influence of a given growth target on FAR regulation. Overall, the deeper the degree of horizontal escalation, the stronger the effect of economic growth targets on land supply structure and FAR regulation, whereas the impact on land supply scale does not differ significantly across subsamples.

5. Extended Analysis

For a long time, higher-level governments in China have primarily evaluated lower-level governments on economic performance, with GDP serving as the core metric. Within this readily observable assessment framework, local officials have been strongly incentivized to promote local economic growth. China is also the only country that releases economic growth rates at four administrative levels—national, provincial, municipal, and county—which further underscores the centrality of growth targets, as reflected in the fact that annual Government Work Reports routinely specify economic growth targets. However, a development model that has long prioritized growth rates has also led local officials to neglect demands in areas such as people’s livelihoods. Against this backdrop, the performance evaluation system for local officials underwent a major adjustment in 2013 [41]. At the Third Plenary Session of the 18th Central Committee of the Communist Party of China in November 2013, General Secretary Xi Jinping repeatedly stressed that officials should not be judged simply by the speed of economic growth, nor should they be preoccupied with rankings based on growth rates. In December of the same year, the Organization Department of the CPC Central Committee issued the “Notice on Improving the Performance Evaluation of Local Party and Government Leading Groups and Leading Cadres”, emphasizing that evaluations should focus on officials’ overall work rather than relying primarily on regional GDP and its growth rate, and that improvements in people’s livelihoods should be placed on an equal footing with economic development, with greater emphasis on diversified objectives such as social security, scientific and technological innovation, and public health. These high-level statements and the issuance of formal documents signaled a major adjustment to the previous performance evaluation system. When economic growth targets are no longer the sole core of local officials’ assessments, local governments’ policy choices and decision-making are expected to adjust accordingly [41]. This institutional change also provides an opportunity for this study to examine how urban land supply decisions evolve under a transformation of goal-oriented incentives.

5.1. Econometric Model Specification

To examine how the effect of economic growth targets on urban land supply decisions changes under the transformation of local governments’ goal-oriented incentives, this study follows related work [41] and specifies the following regression model:

$$\begin{array}{c}{Y}_{it}={\beta }_0+{\beta }_{1}trea{t}_{it}+{\beta }_{2}trea{t}_{it}\times post+{\beta }_3{Z}_{it}+{\mu }_i+{\delta }_t+{\epsilon }_{it}\end{array}$$

(20)

In Equation (20), $Y_{it}$ is the dependent variable, capturing city $i$’s land supply decisions in year $t$ in terms of land supply scale, land supply structure, and FAR regulation. Specifically, land supply scale is proxied by land supply area; land supply structure is proxied by the ratio of industrial land area to residential land area; and FAR regulation is proxied by residential FAR regulation. The variable $trea{t}_{it}$ is a dummy variable indicating whether the degree of horizontal escalation for city $i$ in year $t$ is above the sample median: it equals 1 if the degree of horizontal escalation exceeds 50%, and 0 otherwise. The variable $post$ is a time dummy that equals 1 for years after 2013 and 0 for 2013 and earlier. The vector $Z_{it}$ contains city-level control variables, including cities’ natural geographic endowments, socio-economic development conditions, and the tenure of the local chief official; all controls are included with a one-year lag, which also helps to alleviate endogeneity concerns to some extent. The terms ${\mu }_i$ and ${\delta }_t$ denote city fixed effects and province-year fixed effects, respectively. These terms control for unobservable factors that are time-invariant across cities and city-invariant over time, and they mitigate the influence of contemporaneous province-level policy shocks—such as land use quotas, real estate regulations, and industrial and environmental constraints—on the estimation results. The term ${\epsilon }_{it}$ is the idiosyncratic error term. The model is estimated using a high-dimensional fixed-effects regression approach with city-level clustered robust standard errors. The coefficient ${\beta }_1$ captures the effect of economic growth targets on municipal land supply decisions for prefecture-level cities with high horizontal intensification before 2013. ${\beta }_2$ captures the change in this effect after 2013, reflecting differential impacts across groups with different levels of horizontal intensification around the institutional turning point. This pattern is consistent with a shift in the role of economic growth targets in officials’ performance evaluations after 2013.

5.2. Regression Results

The regression results based on the model specified in the previous section are reported in

Table 7. Urban land supply decisions under the transformation of goal-oriented incentives.

Variable	Land Supply Scale		Land Supply Structure		FAR Regulation
	(1)	(2)	(3)	(4)	(5)	(6)
treat	0.082 ** (0.034)	0.079 ** (0.032)	0.396 * (0.221)	0.385 * (0.224)	0.027 * (0.015)	0.033 * (0.011)
Treat × post	−0.022 (0.040)	−0.010 (0.039)	−0.695 ** (0.323)	−0.612 * (0.322)	−0.026 * (0.012)	−0.046 * (0.028)
City fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Province-year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Observations	4209	4209	4083	4083	4015	4015
Adjusted R-squared	0.752	0.772	0.229	0.230	0.666	0.670

Note: ** and * indicate significance at the 5% and 10% levels, respectively. The values in parentheses denote robust standard errors.

. The table is organized into three panels corresponding to the three dependent variables—land supply scale, land supply structure, and FAR regulation—which are presented in columns (1)–(2), (3)–(4), and (5)–(6), respectively.

Column (1) reports the results from a specification that excludes all control variables and regresses the dependent variable only on the independent variable, while still controlling for city and year fixed effects. Column (2) adds city-level control variables to the specification in column (1), and the two sets of estimates are broadly consistent. The coefficient on the treatment variable is significantly positive at the 1% level, suggesting that before 2013, local governments facing higher horizontal escalation were more inclined to expand land supply in pursuit of predetermined economic growth targets. The coefficient on the interaction term treat × post is not statistically significant, indicating that this relationship did not change materially after 2013. This pattern implies that, even as the incentive structure for local officials evolved, expanding urban land supply remained an important channel through which local governments sought to meet predetermined growth targets.

Column (3) reports the results from a specification that excludes all control variables and regresses the dependent variable only on the independent variable, while still controlling for city and year fixed effects. Column (4) further adds city-level control variables to the specification in column (3), and the two sets of estimates are broadly consistent. The regression results show that the coefficient on treat is positive and statistically significant at the 10% level. This suggests that, prior to 2013, local governments with higher levels of horizontal escalation were more likely to pursue predetermined economic growth targets by increasing the industrial-to-residential land supply ratio. The coefficient on the interaction term treat × post is negative and statistically significant at the 5% level, implying that the above effect weakened markedly after 2013.

Column (5) reports the results from a specification that excludes all control variables and regresses the dependent variable only on the independent variable, while still controlling for city and year fixed effects. Column (6) further adds city-level control variables to the specification in column (5), and the two sets of estimates are broadly consistent. The regression results show that the coefficient on the treatment variable is positive and statistically significant at the 10% level, suggesting that prior to 2013, local governments with higher levels of horizontal escalation were more likely to pursue predetermined economic growth targets by tightening FAR regulation. The coefficient on the interaction term treat × post is negative and statistically significant at the 10% level, indicating that this relationship weakened after 2013. This pattern suggests that, as officials’ target incentives shifted, the use of FAR regulation as a channel for meeting predetermined growth targets has changed.

5.3. Robustness Test

This section further conducts robustness checks for the extended analysis. The study first conducts a placebo test. (1) For each year, we randomly assign half of the prefecture-level cities in the sample to the treatment group (treat = 1) and assign the remaining cities to the control group (treat = 0), thereby randomizing horizontal intensification at the city-year level. (2) Using the simulated treatment assignment, we replicate the regression procedure from the previous section to estimate the coefficients of interest. (3) We repeat steps (1)–(2) 500 times and plot the empirical distributions of the coefficients on treat and treat × post when the dependent variable is land supply scale. In addition, this study conducts a parallel trends test. The robustness check results are reported in Supplementary S5. Both checks support the robustness of the benchmark regression.

6. Conclusions and Discussion

6.1. Conclusions

The setting of development goals is a central component of local government target governance and has important implications for urban land supply patterns and sustainable development. Using data for 286 prefecture-level cities in China from 2007 to 2021, this study employs high-dimensional fixed-effects estimators to examine how government economic growth targets affect urban land supply. Exploiting the 2013 adjustment to the official performance evaluation mechanism as a policy shock, the analysis further assesses whether urban land supply patterns changed as government objectives shifted toward sustainability. The results indicate that higher economic growth targets are associated with a larger land supply scale, a higher industrial-to-residential land supply ratio, and tighter FAR regulation. The heterogeneity analysis shows that “soft constraint” phrasing in target formulation attenuates the effect on land supply scale but amplifies the effects on land supply structure and FAR regulation. In addition, target intensification generates significant heterogeneity in urban land supply responses. Finally, the evidence suggests that, following the shift from growth-oriented to sustainability-oriented development goals, the marginal role of land supply structure and FAR regulation in facilitating the achievement of predetermined economic growth targets has diminished.

6.2. Discussion

The shift toward sustainability-oriented development is increasingly central to understanding China’s national governance. Governance objectives embody developmental priorities and intentions and function as key instruments through which governments build consensus, mobilize resources, and steer allocation decisions—thereby shaping urban land supply. This study finds that higher economic growth targets significantly expand land supply and tilt allocations toward industrial land to meet short-run growth pressures. At the same time, under ambitious growth targets, local governments tighten FAR regulation for residential land, seeking to stabilize housing prices and to secure land conveyance revenues and property-related tax receipts. These revenues can, in turn, offset losses associated with low-priced industrial land transfers and support infrastructure investment. While these land resource allocation strategies may have alleviated local fiscal pressures and boosted regional economic development in the short run, they may also lead to urban sprawl, weaken environmental carrying capacity, and reduce resident welfare in the long run [42,43,44]. Higher economic growth targets that induce an expansionary land supply scale and structural bias may affect the pace of suburban development and the spatial structure of land supply. When residential FAR regulation tightens, effective housing supply may be constrained, which can increase housing costs and intensify demand pressures for low-cost living space. As Wu et al. (2012) note, even relatively moderate changes in local government incentives or governance approaches can restructure housing supply patterns, residential density, and the spatial distribution of population through land supply and development-intensity controls, with substantial implications for urban form and real estate market functioning [45].

As China advances economic restructuring and sustainable development goals, the target responsibility assessment system should shift from a singular focus on economic growth toward a multidimensional approach that balances constraints and incentives. This transition is intended to reduce local governments’ excessive reliance on economic growth targets in land supply and to promote a more rational and sustainable allocation of land resources. This study finds that after the adjustment to the performance evaluation mechanism, local governments continue to promote economic growth through land supply. At the same time, the reduced weight placed on economic growth targets has encouraged local governments to pursue more coordinated economic, social, and environmental development by optimizing land supply structure and increasing FAR [46,47]. In many market-oriented land systems, local governments have weaker direct control over land supply volume and pricing, which may make the “growth target–land supply tool” transmission chain less direct than in China [48]. However, in local governance systems characterized by strong planning controls, fiscal decentralization pressures or performance-oriented approaches, more attention should be paid to substitution across land policy tools, as well as potential tool migration and unintended consequences arising from strengthened sustainability targets [49]. Governments should place greater emphasis on performance metrics aligned with global sustainable development goals. Growth targets should shift from rigid rankings and single-factor weighting to range-based or floor constraints, with ecological constraints, public services, and housing affordability incorporated into a unified assessment framework to curb excessive land supply and structural biases. As development objectives shift from growth-oriented to sustainability-oriented, attention should be paid to local governments’ substitution across policy tools under multi-objective constraints. Complementary mechanisms in finance, housing, and public services should be implemented to mitigate side effects arising from misaligned incentives.

In optimizing the government’s target responsibility assessment system, the judicious use of “soft constraint” phrasing can, in principle, provide buffers and adjustment space for achieving government objectives, thereby facilitating a more rational allocation of land resources. This study finds that “soft constraints” partly attenuate the effect of economic growth targets on urban land supply scale while amplifying their influence on land supply structure adjustments and FAR regulation. Accordingly, softening the phrasing of growth targets may facilitate the reallocation of land resources. At the same time, the “layered escalation” of economic growth targets may hinder sustainable land resource allocation. Under the combined effects of top-down administrative contracting and horizontal promotion competition, higher-level targets are amplified during vertical transmission and inflated through peer comparisons, making local governments more likely to adopt land supply strategies aligned with short-term performance metrics. Against this background, an effective oversight and accountability mechanism is needed to curb excessive escalation and prevent resource allocation distortions driven by tiered target amplification.

Future research directions include three avenues. (1) Dynamic responses under policy instrument substitution and multi-objective governance: This study finds that economic growth targets affect not only land supply scale but also interact with land supply structure and FAR regulation. Future research could identify the dynamic substitution pathways local governments adopt within broader policy toolkits as sustainability imperatives intensify, and examine whether such substitutions generate unintended consequences and cross-regional spillovers. (2) The full transmission chain from land policy to urban outcomes: This study focuses on responses at the land policy stage. Future work could link changes in land policies to urban outcomes—such as housing affordability, industrial spatial restructuring, commuting patterns and carbon emissions, public service provision, and welfare distribution—to assess net effects and distributional impacts of shifts in governance orientation for urban sustainability. (3) Cross-national institutional comparison and generalizability: Although the institutional context is specific to China, the analytical framework—“goal-oriented governance → policy tool selection → spatial outcomes”—may apply to other settings characterized by strong planning controls, fiscal decentralization pressures, or performance-driven governance. Future research could develop comparable indicator systems for cross-national analyses to distinguish mechanisms with broader applicability from those contingent on particular land systems and governance structures.