How Land Use Regulation Affects County Governments’ Land Transfers and Public Service Provision

Abstract

As a populous country with limited per capita land area, China has implemented the strictest land use regulation to ensure food security. Yet quantitative assessments of how it shapes land use change and the subsequent economic impacts remain insufficient. Land use directly affects land supply for industry and services, thereby impacting local fiscal and tax revenues. Meanwhile, land transfer income serves as a major off-budget revenue source for local governments, with county fiscal capacity laying the foundation for national economic development and public welfare. Therefore, this study integrates county-level statistics with remotely sensed land use data and applies an Intensity Difference-in-Differences (Intensity DID) design to identify policy impacts. Specifically, it examines the effects of land use regulation on county governments’ land transfer activities, land use efficiency, as well as fiscal revenue and public service provision. Empirical results show that tighter land use regulation constrains the new supply of construction land by limiting cultivated land conversion. In response, local governments modify floor area ratios (FARs) and shorten construction cycles, which is conducive to improving land use efficiency. Nevertheless, the policy reduces the land transfer income, tax revenue, and general public budget revenue of county governments, weakening public service provision. Heterogeneity analysis indicates that major grain-producing counties are more severely affected by negative policy shocks than non-major ones. This study provides empirical evidence for optimizing the land use regulation system and offers policy implications for coordinating food security and balanced regional development through horizontal interest compensation in major grain-producing regions.

1. Introduction

Land is the core production factor supporting local economic development in China. After the 1994 tax-sharing reform, local governments faced significant fiscal imbalances [1,2,3]. Concurrently, official promotion was mainly linked to Gross Domestic Product (GDP) growth as one of the key performance indicators. Under these two institutional constraints, local governments engaged in two types of competition: vertical fiscal competition with the central government and horizontal development competition among peer governments [4]. This strengthened their motivation to pursue economic growth [5] and led to dependence on land finance [6]. Because existing land had high transaction costs, local governments preferred to expand the new land supply. This helped them secure extra-budgetary land transfer income and intra-budgetary industrial tax revenues [7,8]. This was done through two key approaches: first, “land monetization,” where local governments used their information and power advantages to acquire rural land at low costs for construction, then sold commercial land through market to gain substantial profits from price differences, offsetting losses from low-cost industrial land transfers [9]; second, “land-based investment attraction,” where land became a core competitive tool for attracting foreign direct investment (FDI), with local governments offering preferential terms such as low or even zero land prices [10]. Driven by decentralized competition, local governments’ growing demand for land expropriation has led to excessive conversion of farmland to non-agricultural use [11]. Data from China’s Third National Land Survey shows that the national cultivated land area decreased from 2.031 billion mu1 in 2009 to 1.918 billion mu in 2019, which brought it closer to the 1.8 billion mu cultivated land red line. The excessive occupation of cultivated land driven by rapid urbanization has heightened the central government’s concerns about food security [12,13], as it may break the dynamic balance between food security and economic development. How to strike a dynamic balance between food security and economic development through land planning is the core issue in China’s land resource allocation [14,15]. Against this background, China has adopted land use regulation centered on cultivated land protection, which has become a key institutional arrangement for balancing multiple land use needs.

Land use regulation is a globally critical instrument for achieving the efficient allocation and sustainable utilization of land resources [16,17,18], with different countries adopting differentiated regulatory measures based on their own institutional backgrounds. For instance, the United States classifies farmland into “white land” (designated for non-agricultural use) and “green spaces” (restricted to agricultural use only) to secure the supply of high-quality agricultural land [19]. Germany employs spatial planning to delineate restricted and unrestricted zones for land development into construction land. However, the scarcity of land resources inherently conflicts with competing demands across agricultural production, industrial development, and ecological conservation. On one hand, the positive externalities of cultivated land protection, such as food security and ecological security, are socially shared, while the direct costs are borne by local governments and farmers in cultivated land areas, leading to a spatial mismatch between costs and benefits [20]. On the other hand, as the primary providers of public services, local governments’ public service levels directly reflect regional well-being. Some studies argue that land finance2 has significantly promoted public service provision, encompassing both economic public services (e.g., infrastructure construction) and social public services (e.g., healthcare and education) [21,22,23]. However, this view remains contested. Other studies suggest that the deepening dependence on land finance has caused structural imbalances in public service provision. To obtain more fiscal revenue, local governments are more inclined to provide economic public services, crowding out non-economic public services [24,25,26]. Under China’s land use regulation, the development model of “land-driven development and revenue generation” is unsustainable, forcing the transformation of land finance [27], which in turn affects local public service provision. Consequently, the differentiated implementation of land use regulation may either channel resources toward sectors with comparative advantages or widen regional development disparities and welfare gaps.

This study takes China as a typical case to empirically analyze the impact of land use regulation on land use change. Counties are the basic administrative units exercising fiscal power in China. Article 9 of the “Interim Regulations on the Granting and Transfer of Urban State-Owned Land Use Rights” stipulates that the granting of land use rights is the responsibility of municipal and county people’s governments. Therefore, county governments have legally independent land transfer rights in principle. As the basic unit of the national economy, county economies are the intersection of urban and rural economies and play a vital role in promoting development and safeguarding people’s livelihoods. Existing studies have mostly evaluated the economic impacts of land use regulation at the inter-provincial level. However, China has a vast territory, and significant heterogeneity may exist within provinces. Few studies have examined the policy’s effects at the county level, yet such an analysis is essential for capturing local variations and informing targeted policy interventions. This study integrates county statistical data, remote sensing geographic information data, and micro land transaction data spanning 2006 to 2022. We use the difference between the 2020 cultivated land protection target specified in each city’s “Land Use Master Plan” for each county administrative region and the local actual cultivated land area as a measure of land use regulation intensity. Furthermore, the “Dual Guarantees Project” implemented by the Ministry of Land and Resources in 2010 serves as a form of central supervision over local land use regulation. Therefore, we take 2010 as the policy shock time point. The Intensity DID method is adopted to systematically examine how land use regulation drives county governments’ land transfer behavior and its heterogeneity impacts on local fiscal revenue and public service provision through the transmission mechanism of land use change.

The main contributions of this study are as follows: (1) It clarifies how land use regulation restricts cultivated land conversion to shape county governments’ construction land transfer scale and source structure. This study also reveals the policy’s dual impact on land supply volume and structure. (2) It explores how land use regulation affects local governments’ industrial land allocation strategies. This provides empirical support for evaluating the path to improving land use efficiency. (3) It quantifies the impact of land use regulation on intra-budgetary tax revenue, extra-budgetary land transfer income, and total local fiscal revenue. It uncovers heterogeneous effects between major grain-producing and non-major grain-producing counties. It also provides a basis for understanding the policy’s economic welfare implications. (4) By adopting the intensity DID method, this study empirically estimates how land use regulation affects county-level public service provision through fiscal revenue transmission. It analyzes regional heterogeneity in these effects. It verifies the policy’s impact on county-level livelihood welfare. (5) This study provides empirical evidence and policy implications for three goals: refining the land use regulation, balancing food security and regional development, and innovating cross-regional interest compensation mechanisms.

This paper is organized as follows: Section 2 elaborates on the Materials and Methods, including the policy background of China’s land use regulation, research hypotheses, data sources, and descriptive statistics, as well as the specification of the econometric model. Section 3 reports and analyzes empirical results. Section 4 discusses the key research findings. Section 5 concludes with policy recommendations.

2. Materials and Methods

2.1. Policy Background

China is a country with a large population and scarce per capita land resources. Therefore, land use is a strategic priority closely related to national food security and balanced regional development. This reflects the great importance and urgency of land use governance in China. Land use regulation is a widely adopted institutional arrangement to address land use challenges. It provides a standardized and reliable mechanism for the rational use of land resources and food security protection. China has undoubtedly implemented one of the world’s strictest regulatory systems, with clear rules on land use regulation enshrined in its laws. Article 4 of the Land Administration Law of the People’s Republic of China (hereinafter referred to as the Land Administration Law)3 explicitly stipulates that the state implements a land use regulation. Meanwhile, according to Articles 4 and 16 of the Land Administration Law, the connotation of land use regulation is summarized as follows: the main basis for land use regulation is the land use master plan, which regulates agricultural land, construction land, and unused land4. Its core goal is to protect cultivated land resources to ensure food security and strictly control the quantity of land for various uses through the formulation of land indicator plans, and the key focus of regulation is to strictly control the conversion of cultivated land to construction land and maintain the dynamic balance of total cultivated land. Based on the above, land use regulation in this study specifically refers to the control of the conversion of cultivated land to non-agricultural uses.

In the 1980s, rapid industrialization and urbanization occupied a large amount of cultivated land, creating a demand for unified management of China’s land resources. In 1986, the Land Administration Law first stipulated that governments at all levels must formulate land use master plans. These plans require approval by higher-level governments before implementation, marking the legalization of land use master planning. In 1999, the State Council approved and issued the National Land Use Master Plan Outline (1997–2010). It explicitly required maintaining the national cultivated land area above 1.92 billion mu by 20105. In 2006, the State Council raised higher standards: “Formulate a historic, crisis-oriented, and strategic land use master plan.” It extended the planning period and emphasized upholding the 1.8 billion mu cultivated land red line without compromise. In 2008, the State Council approved the National Land Use Master Plan Outline (2006–2020) (hereinafter referred to as the Outline). It set national cultivated land protection targets of 1.813 billion mu (2010) and 1.805 billion mu (2020) and delegated these targets to provinces, autonomous regions, and municipalities directly under the Central Government. Under the Outline, restrictive cultivated land protection indicators were decomposed top-down through three administrative levels (provinces, cities, and counties). An accountability mechanism for government leaders was established and integrated into performance evaluations.

The 2008 subprime crisis slowed China’s economic growth. Against this backdrop, China launched a CNY 4 trillion infrastructure investment program to boost domestic demand and promote steady economic growth. As a result, while infrastructure investment across regions stimulated growth, it also triggered a sharp rise in the demand for construction land. This posed severe challenges to cultivated land protection and worsened tensions between land supply constraints and economic development. In 2009, the former Ministry of Land and Resources launched the “Guaranteeing Growth and Protecting the Red Line” Action. It aimed to balance economic development and cultivated land protection. In 2010, drawing on previous experience, the action was institutionalized. The “Two Guarantees Project” was set as a long-term task, forming a regular supervision mechanism. After 2020, the “Two Guarantees” concept was integrated into the delineation of “Three Zones and Three Lines6.” It emphasized the trinity protection of cultivated land in terms of quantity, quality, and ecology.

By the second half of 2009, the task of decomposing cultivated land protection targets to county-level administrative units (counties, districts, county-level cities, and banners) was substantially completed. Cultivated land protection targets for 2020 at the county level can be retrieved from the Land Use Master Plan (2006–2020) of each prefecture-level city via official provincial government websites. Variations in the magnitude of targets assigned by higher-level authorities have created heterogeneity in policy shock intensity. Combined with the supervision action time point of the central government’s “Two Guarantees Project” in 2010, it provides a basis for the subsequent econometric analysis.

2.2. Hypothesis

In the trade-off between protecting cultivated land and developing the economy, a game relationship has emerged between the central and local governments. Its core is conflicting objectives when addressing negative externalities. From the central government’s perspective, decisions are based on overall national interests and long-term development. Its key goal is to safeguard national food security, ecological security, and sustainable development. This goal is achieved by curbing the negative externalities from excessive conversion of cultivated land to non-agricultural use [28,29]. For developing countries, rapid urbanization drives regional economic growth but also creates cross-regional and intergenerational negative externalities such as cultivated land loss and environmental degradation. These directly threaten national food security and long-term financial and social stability [30,31,32]. For local governments, the central government mobilizes local enthusiasm through a hierarchical power structure and allocates resources based on economic performance rankings. This institutional design encourages local governments to prioritize economic growth and fiscal revenue enhancement during their tenure [33,34]. The negative externalities of excessive cultivated land occupation are borne at the national level, while the direct benefits of land development accrue to local governments. This cost–benefit asymmetry further strengthens local governments’ short-termist tendencies. As a scarce production factor, land supply scale and structure are determined by the balance between marginal cost and marginal benefit. Without rigid regulatory constraints, cultivated land is mostly located in suburban areas with low development costs and obvious locational advantages. It naturally becomes the main source of new construction land. At this stage, the total land supply curve is flat, with a large transfer volume. The implementation of land use regulation raises the threshold for converting cultivated land to construction land in two key ways. First, it directly increases economic costs. The “balance between occupation and compensation” system requires local governments to convert an equivalent area of unused land to cultivated land within their jurisdiction when developing cultivated land for construction. This involves additional investments in topsoil stripping, reclamation, and land leveling. Second, it strengthens political accountability costs. Cultivated land protection has been fully integrated into local governments’ performance evaluation and official accountability systems. Illegal conversion of cultivated land results in administrative sanctions and restricted resource allocation. Since cultivated land conversion is the core source of new construction land, reduced supply shifts the total land supply curve leftward. In the short term, land demand driven by urbanization and industrial development remains relatively stable. This ultimately reduces the total land transfer volume. Although local governments can supplement supply by revitalizing existing construction land, renovating and redeveloping such land requires additional investments in demolition compensation and site preparation. Moreover, the scarcity of high-quality existing land makes it hard to fully offset the supply gap from restricted cultivated land conversion.

Hypothesis 1.

Land use regulation curbs total land transfers by raising cultivated land conversion costs, with the effect mainly falling on new construction land transfers due to insufficient existing construction land.

Hypothesis 2.

Through the substitution effect, land use regulation restructures new construction land sources and reduces local land-related revenues by decreasing the land transfer volume.

To mitigate fiscal losses, local governments are forced to shift from extensive to intensive land use. They offset revenue shortfalls by improving land use efficiency.

Hypothesis 3.

Land use regulation improves land use efficiency with heterogeneous paths across land types and imposes a dual impact on local finance through land transfer behavior adjustments.

On one hand, reduced land transfer volume directly lowers land transfer income, and higher per-unit land revenue cannot offset the negative impact of lower total transfers. On the other hand, insufficient construction land supply restricts regional industrial and commercial development, thereby reducing tax sources. Under the current fiscal and taxation system, local governments, especially those at the county level, lack sufficient alternative revenue streams to buffer the decline in land-related income. Mounting fiscal balance pressure ultimately leads to reduced public service expenditures.

Hypothesis 4.

Land use regulation reduces local governments’ land transfer income, tax revenue, and total fiscal revenue. Such fiscal contraction further inhibits public service-related expenditures.

Land use regulation’s policy effects show significant regional heterogeneity. This is rooted in the mismatch between local governments’ administrative responsibilities and fiscal powers. Major grain-producing counties are required to retain more cultivated land to meet national grain output targets. Low grain production returns and restricted secondary and tertiary industry development lead to generally low fiscal self-sufficiency rates in these counties. This results in a more pronounced crowding-out effect on public service provision.

Hypothesis 5.

Land use regulation has regionally heterogeneous negative effects on fiscal revenue and public service provision, with major grain-producing counties facing more severe adverse impacts than non-major ones.

2.3. Data Sources and Description

2.3.1. Data Sources

Land-Related Data Sources:

(1)

The 2020 cultivated land protection target for each county-level administrative region was retrieved from the Land Use Master Plan (2006–2020) of its corresponding prefecture-level city via official provincial government websites.

(2)

2006–2022 China land use geographic information data were obtained from the Annual China Land Cover Dataset (CLCD) (Institute of Remote Sensing and Information Processing, Wuhan University). This dataset provides 30 m resolution annual land use data covering China, including land types such as cultivated land, forest land, grassland, water bodies, and unused land. By comparing land type identifier changes for the same county grid between the previous year and the current year, this study constructs a land conversion matrix and calculates the proportion of new construction land derived from different land type conversions at the county level.

(3)

Land transaction data were sourced from land transaction result announcements on the China Land Market Network (2006–2022). The dataset includes over 3 million micro-records covering transaction details such as land supply area, location, use type, transaction price, floor area ratio constraints, and construction cycle. For this study, data were aggregated by county-level administrative divisions to form county-level indicators of land transfer and utilization efficiency.

Sources of Other Variables:

(1)

The list of major grain-producing counties was manually compiled from official websites of municipal governments and finance bureaus, delineating counties responsible for national food production strategic tasks.

(2)

County-level economic indicators (e.g., per capita GDP) and public service indicators (e.g., number of beds in medical and health institutions) were obtained from the China County Statistical Yearbook (2006–2022).

(3)

Population density data were sourced from the LandScan Global Population Dynamic Statistical Analysis Database. Using the national unified 1:4 million administrative region vector data layer, resident population data within each county’s administrative boundary were extracted and divided by the administrative area to calculate the average county-level population density. Table 1 reports descriptive statistics for relevant variables.

Table 1. Descriptive statistics.

Variable	Definition	Mean	Std. Dev	Min	Max
${\mathit{T}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$	ln(Total Transferred Land Area (ha))	4.463	0.966	2.085	5.942
${\mathit{I}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$	ln(Transferred area of newly added construction land (ha))	4.050	1.104	0.949	5.827
${\mathit{S}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$	ln(Transferred area of newly added construction land (from the land stock bank) (ha))	2.299	1.485	−1.902	5.178
${\mathit{N}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$	ln(Transferred area of existing construction land (ha))	2.346	1.536	−2.303	5.204
${\mathit{I}\mathit{C}\mathit{L}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{t}}$	ln(Transferred area of newly added industrial land (ha))	2.877	1.303	−0.899	5.193
${\mathit{I}\mathit{C}\mathit{L}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{t}}$	ln(Transferred area of newly added commercial and residential land (ha))	2.433	1.272	−0.753	4.837
${\mathit{F}\mathit{a}\mathit{r}\mathit{m}\mathit{l}\mathit{a}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{t}}$	Proportion of newly added construction land converted from cultivated land (%)	0.790	0.243	0.000	1.000
${\mathit{F}\mathit{o}\mathit{r}\mathit{e}\mathit{s}\mathit{t}}_{\mathit{i},\mathit{t}}$	Proportion of newly added construction land converted from forest land (%)	0.038	0.075	0.000	0.704
${\mathit{G}\mathit{r}\mathit{a}\mathit{s}\mathit{s}}_{\mathit{i},\mathit{t}}$	Proportion of newly added construction land converted from grassland (%)	0.104	0.200	0.000	1.000
${\mathit{W}\mathit{a}\mathit{t}\mathit{e}\mathit{r}\mathit{s}}_{\mathit{i},\mathit{t}}$	Proportion of newly added construction land converted from water bodies (%)	0.055	0.101	0.000	1.000
${\mathit{U}\mathit{n}\mathit{u}\mathit{s}\mathit{e}\mathit{d}}_{\mathit{i},\mathit{t}}$	Proportion of newly added construction land converted from unused land (%)	0.015	0.064	0.000	1.000
${\mathit{L}\mathit{T}\mathit{F}}_{\mathit{i},\mathit{t}}$	ln(Total revenue from land transfers (CNY 10 k))	9.992	1.627	5.599	13.535
${\mathit{T}\mathit{a}\mathit{x}}_{\mathit{i},\mathit{t}}$	ln(Total tax revenue of local government (CNY 10 k))	10.976	1.091	8.366	13.666
${\mathit{F}\mathit{i}\mathit{s}\mathit{c}\mathit{a}\mathit{l}\mathit{r}\mathit{e}\mathit{v}}_{\mathit{i},\mathit{t}}$	ln(Total fiscal revenue of local government (CNY 10 k))	11.032	1.097	8.399	13.609
${\mathit{L}\mathit{F}\mathit{A}\mathit{R}}_{\mathit{l},\mathit{t}}$	Minimum Floor Area Ratio (FAR)	0.794	0.767	0.000	4.420
${\mathit{U}\mathit{F}\mathit{A}\mathit{R}}_{\mathit{l},\mathit{t}}$	Maximum Floor Area Ratio (FAR)	2.089	1.285	0.130	6.800
${\mathit{A}\mathit{C}\mathit{D}}_{\mathit{l},\mathit{t}}$	ln(Agreed construction period of the land parcel (days))	6.484	0.466	4.796	6.999
${\mathit{T}\mathit{e}\mathit{a}\mathit{c}\mathit{h}\mathit{s}\mathit{t}\mathit{u}}_{\mathit{i},\mathit{t}}$	Student–teacher ratio in regular primary and secondary schools	13.490	5.448	1.788	31.524
${\mathit{H}\mathit{e}\mathit{a}\mathit{l}\mathit{t}\mathit{h}\mathit{c}\mathit{a}\mathit{r}\mathit{e}}_{\mathit{i},\mathit{t}}$	Number of beds in medical and health institutions per 10,000 population	39.984	24.675	9.394	154.643
${\mathit{S}\mathit{o}\mathit{c}\mathit{i}\mathit{a}\mathit{l}\mathit{c}\mathit{a}\mathit{r}\mathit{e}}_{\mathit{i},\mathit{t}}$	Number of beds in various social welfare and adoptive institutions per 10,000 population	32.624	26.118	1.011	141.000
${\mathit{R}\mathit{e}\mathit{g}\mathit{u}\mathit{l}\mathit{a}\mathit{t}\mathit{i}\mathit{o}\mathit{n}}_{\mathit{i},\mathit{t}}$	${\mathit{T}\mathit{i}\mathit{g}\mathit{h}\mathit{t}}_{\mathit{i}}\times {\mathit{P}\mathit{o}\mathit{s}\mathit{t}}_{\mathit{t}}$	−0.107	0.140	−0.516	0.111
${\mathit{T}\mathit{i}\mathit{g}\mathit{h}\mathit{t}}_{\mathit{i}}$	(2020 Cultivated Land Reservation Target − 2009 Actual Cultivated Land Area)/Administrative Region Area	−0.135	0.146	−0.545	0.121
${\mathit{P}\mathit{o}\mathit{s}\mathit{t}}_{\mathit{t}}$	1 = Year 2010 and later; 0 = Otherwise	0.802	0.399	0.000	1.000
${\mathit{M}\mathit{G}\mathit{P}\mathit{C}}_{\mathit{i}}$	1 = Major grain-producing county; 0 = Otherwise	0.539	0.499	0.000	1.000
${\mathit{I}\mathit{n}\mathit{v}\mathit{e}\mathit{s}\mathit{t}}_{\mathit{i},\mathit{t}}$	Total Fixed Asset Investment/GDP	0.925	0.626	0.105	3.805
${\mathit{P}\mathit{G}\mathit{D}\mathit{P}}_{\mathit{i},\mathit{t}}$	ln(Per Capita GDP)	10.368	0.796	8.584	12.383
${\mathit{S}\mathit{u}\mathit{f}\mathit{f}\mathit{i}\mathit{c}\mathit{i}\mathit{e}\mathit{n}\mathit{c}\mathit{y}}_{\mathit{i},\mathit{t}}$	Fiscal revenue/Fiscal expenditure	0.364	0.249	0.044	1.282
${\mathit{N}\mathit{a}\mathit{g}\mathit{r}}_{\mathit{i},\mathit{t}}$	Added Value of the Secondary and Tertiary Industries/GDP	0.814	0.116	0.446	1.019
${\mathit{W}\mathit{a}\mathit{g}\mathit{e}}_{\mathit{i},\mathit{t}}$	ln(Natural Logarithm of the Wages of Urban Employees)	10.674	0.670	9.399	12.968
${\mathit{P}\mathit{d}\mathit{e}\mathit{n}\mathit{s}\mathit{i}\mathit{t}\mathit{y}}_{\mathit{i},\mathit{t}}$	Population Density (persons/km2)	3.021	3.423	0.025	21.890
${\mathit{A}\mathit{u}\mathit{c}\mathit{t}\mathit{i}\mathit{o}\mathit{n}}_{\mathit{i},\mathit{t}}$	1 = Land is transferred via bidding, auction, and listing (BAL) methods; 0 = Otherwise	0.654	0.476	0.000	1.000
${\mathit{L}\mathit{a}\mathit{n}\mathit{d}\mathit{l}\mathit{e}\mathit{v}\mathit{e}\mathit{l}}_{\mathit{i},\mathit{t}}$	1 = high-grade land; 0 = Otherwise	0.446	0.497	0.000	1.000

2.3.2. Description

This study constructs a county-level panel dataset covering 1969 county-level administrative units in China from 2006 to 2022. In the variable notation, subscript $i$ denotes a county-level administrative unit, subscript $j$ denotes a land parcel, and subscript $t$ denotes the year.

The dependent variables are classified into five categories:

(1)

Local government land transfers, including the total area of transferred land, the transferred area from newly added construction land (i.e., the increment over the previous year), the transferred area from newly added construction land in the land stock bank, and the transferred area from existing construction land. In addition, it also includes industrial land and commercial–residential land transferred from newly added construction land.

(2)

Land conversion, measured as the share of newly added construction land in the current year converted from cultivated land in the previous year. The conversion shares of forestland, grassland, water areas, and unused land are calculated in the same way.

(3)

Fiscal indicators, including land transfer revenue, total tax revenue, and total fiscal revenue of the county government per year.

(4)

Construction land use efficiency, including the minimum FAR, maximum FAR, and construction period of land parcels.

(5)

Public service provision, specifically covering education, healthcare, and social services.

The second section of Table 1 presents the independent variables (focusing on land use regulation) and the heterogeneous grouping variable (whether a county is a major grain-producing county, as defined in the Results and Discussion section). The third section lists the control variables. Detailed definitions of all variables are provided in the second column of Table 1.

2.4. Model Specification

In the baseline regression, referring to the practices of Tang and Shao (2022) [35] and Xie and Zhang (2024) [36], the Intensity DID method is adopted as the identification strategy to evaluate the impact of land use regulation on county governments’ land transfer behavior, fiscal revenue and expenditure, and public service provision. Spatially, due to differences in cultivated land protection targets and actual cultivated land area among various districts and counties, the intensity of land use regulation shocks varies significantly at the district and county levels, providing a natural experiment condition for constructing differentiated land use regulation intensity at the district and county levels.

Specifically, this study constructs a proxy variable for land use regulation intensity (${Tight}_i$) through the following formula:

$${Tight}_i={\displaystyle \frac{{2020CultivatedLandReservationTarget}_i-{2009ActualCultivatedLandArea}_i}{{AdministrativeRegionArea}_i}}$$

Among them, the numerator reflects the constraint intensity of the cultivated land protection target by taking the difference between the 2020 cultivated land protection target specified in the plan and the actual cultivated land area in 2009 (when the cultivated land protection quota allocation was completed). The denominator is standardized by the area of the administrative region to eliminate the impact of scale differences on policy intensity.

Based on this, the Intensity DID model is set as follows:

$${\mathrm{Y}}_{it}=\alpha +\beta {Regulation}_{it}+\gamma X_{it}+{\mu }_i+{\sigma }_t+{\epsilon }_{it}$$

(1)

where ${\mathrm{Y}}_{it}$ denotes the respective dependent variables of interest in this paper. $i$ represents the county-level administrative region, and $t$ denotes the year. To enforce the strictest farmland protection policies and promote land conservation, China’s Ministry of Natural Resources launched the “Protect Economic Development, Safeguard Farmland Red Line” initiative in 2010. Consequently, the period from 2010 onward is designated as the policy shock period, with a time dummy variable ${Post}_t$ (0 for $t<2010$, and 1 for $t\ge 2010$). The core explanatory variable ${Regulation}_{it}$ captures the interaction term between land use intensity constraints $\left({Tight}_i\right)$ and the dummy variable (${Post}_t$). $\beta$ is the estimated coefficient of the DID, reflecting the net effect of land use regulation. Control variables $X_{it}$ include socioeconomic indicators such as fixed-asset investment, per capita regional GDP, and industrial structure (see Table 1). ${\mu }_i$ represents the county fixed effect, controlling for unobservable time-invariant characteristics at the county level, year fixed effects by ${\sigma }_t$, and the residual term by ${\epsilon }_{it}$, with standard errors clustered at the county level.

3. Results and Discussion

3.1. Impact on County Land Transfer Scale

Table 2 reports the impact of land use regulation on the annual county-level land transfer scale. Column (1) of Table 2 shows that land use regulation significantly inhibits the total transferred land area (${TCL}_{i,t}$) at the 10% significance level. Economically, a 1-unit increase in land use regulation intensity reduces the natural logarithm of the total transferred land area by 0.146, corresponding to an approximately 13.6% decrease in actual total land transfer area. Column (2) indicates that the regression coefficient of land use regulation on the transferred area of newly added construction land (${ICL}_{i,t}$) is −0.377, significant at the 1% level, suggesting a stronger inhibitory effect. This is consistent with the policy’s core goal of restricting cultivated land conversion to construction land. Column (3) shows that land use regulation significantly increases the new construction land transfer area from the land stock bank7 (${SCL}_{i,t}$) (coefficient = 0.527, p < 0.01), indicating that the policy forces local governments to revitalize stock land to alleviate the gap in construction land supply. Column (4) reveals that land use regulation has no significant impact on the existing construction land transfer area (${NCL}_{i,t}$), as the transfer of existing construction land only involves the transfer of land use rights and does not occupy new cultivated land, thus avoiding direct policy constraints. The ${TCL}_{i,t}$ equals the sum of the ${ICL}_{i,t}$, the ${SCL}_{i,t}$, and the ${NCL}_{i,t}$; therefore, the policy’s impact on the total land transfer area is only significant at the 10% significance level. The above analysis verifies Hypothesis 1.

Table 2. Impact of land use regulation on land transfer scale.

	(1)	(2)	(3)	(4)
	${\mathit{T}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$	${\mathit{I}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$	${\mathit{S}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$	${\mathit{N}\mathit{C}\mathit{L}}_{\mathit{i},\mathit{t}}$
${Regulation}_{i,t}$	−0.146 *	−0.377 ***	0.527 ***	0.092
	(0.078)	(0.120)	(0.195)	(0.157)
${Invest}_{i,t}$	−0.009 *	0.003	−0.044 ***	−0.018 ***
	(0.005)	(0.006)	(0.006)	(0.004)
${PGDP}_{i,t}$	0.042 *	0.062 **	0.042	0.098 **
	(0.022)	(0.027)	(0.051)	(0.039)
${Sufficiency}_{i,t}$	0.120 **	0.234 ***	0.024	0.083
	(0.052)	(0.056)	(0.068)	(0.080)
${Nagr}_{i,t}$	0.239 **	0.093	0.610 **	0.435 *
	(0.121)	(0.114)	(0.239)	(0.228)
${Wage}_{i,t}$	−0.061 ***	0.005	−0.088 *	−0.118 ***
	(0.022)	(0.028)	(0.050)	(0.039)
${Pdensity}_{i,t}$	−0.005	−0.021 **	0.039 ***	−0.011
	(0.007)	(0.008)	(0.014)	(0.012)
_cons	4.440 ***	3.212 ***	2.252 ***	2.253 ***
	(0.330)	(0.407)	(0.742)	(0.569)
Year	yes	yes	yes	yes
County	yes	yes	yes	yes
N	22,136	20,802	15,112	20,365
R2	0.477	0.434	0.308	0.422

Notes: The first row reports the regression coefficients, and the second row (in parentheses) shows the standard errors (SE). Standard errors in parentheses are clustered at the county level. * p < 0.1, ** p < 0.05, *** p < 0.01. Differences in sample sizes are due to missing values of some control variables. When interpreting the log-linear coefficients in the regression results, the reported percentage effects of land use regulation on land transfer area are based on standard logarithmic approximations (i.e., $\mathsf{\Delta }\mathrm{l}\mathrm{n}\left(y\right)\approx \mathsf{\Delta }y/y$).

3.2. Impact on the Source Structure of County New Construction Land

Table 3 reports the impact of land use regulation on the source structure of new construction land. Column (1) of Table 3 shows that land use regulation significantly reduces the proportion of new construction land derived from cultivated land (${Farmland}_{i,t}$)—for each 1-unit increase in regulation intensity, the proportion decreases by 3.9% (p < 0.01). This inhibitory effect is achieved through three mechanisms: strict approval of cultivated land conversion, increased economic costs of conversion (e.g., purchasing land requisition-compensation balance indicators), and strengthened performance evaluation of cultivated land protection. Columns (2)–(5) indicate that land use regulation significantly increases the proportion of new construction land derived from forest land (${Forest}_{i,t}$), grassland (${Grass}_{i,t}$), water bodies (${Waters}_{i,t}$), and unused land (${Unused}_{i,t}$)—reflecting a non-cultivated land substitution effect. Among them, the conversion of forest land and grassland effectively compensates for the land supply gap caused by reduced cultivated land conversion, ensuring the land demand for county economic development. The substitution effect in Hypothesis 2 is verified.

Table 3. Impact of land use regulation on the source structure of new construction land.

	(1)	(2)	(3)	(4)	(5)
	${\mathit{F}\mathit{a}\mathit{r}\mathit{m}\mathit{l}\mathit{a}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{t}}$	${\mathit{F}\mathit{o}\mathit{r}\mathit{e}\mathit{s}\mathit{t}}_{\mathit{i},\mathit{t}}$	${\mathit{G}\mathit{r}\mathit{a}\mathit{s}\mathit{s}}_{\mathit{i},\mathit{t}}$	${\mathit{W}\mathit{a}\mathit{t}\mathit{e}\mathit{r}\mathit{s}}_{\mathit{i},\mathit{t}}$	${\mathit{U}\mathit{n}\mathit{u}\mathit{s}\mathit{e}\mathit{d}}_{\mathit{i},\mathit{t}}$
${Regulation}_{i,t}$	−0.039 ***	0.012 ***	0.020 ***	0.004 *	0.004 ***
	(0.006)	(0.002)	(0.004)	(0.002)	(0.001)
${Invest}_{i,t}$	−0.001	0.001 *	−0.000	−0.000	0.000
	(0.001)	(0.000)	(0.000)	(0.000)	(0.000)
${PGDP}_{i,t}$	0.004 *	0.002 **	0.004 **	−0.007 ***	−0.002 ***
	(0.002)	(0.001)	(0.002)	(0.002)	(0.001)
${Sufficiency}_{i,t}$	−0.005 *	0.006 ***	0.004 **	−0.007 ***	0.001
	(0.003)	(0.001)	(0.002)	(0.002)	(0.001)
${Nagr}_{i,t}$	0.002	0.001 **	−0.000	−0.001	−0.002 ***
	(0.002)	(0.001)	(0.001)	(0.001)	(0.001)
${Wage}_{i,t}$	0.000	0.000	−0.005 ***	0.008 ***	−0.004 ***
	(0.002)	(0.001)	(0.001)	(0.002)	(0.001)
${Pdensity}_{i,t}$	−0.000	−0.000 **	0.001 ***	−0.000	−0.000 **
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
_cons	0.747 ***	0.013	0.114 ***	0.049 **	0.080 ***
	(0.031)	(0.010)	(0.024)	(0.024)	(0.011)
Year	yes	yes	yes	yes	yes
County	yes	yes	yes	yes	yes
N	29,992	29,892	29,884	29,988	29,321
R2	0.882	0.796	0.890	0.645	0.765

Notes: The first row reports the regression coefficients, and the second row (in parentheses) shows the standard errors (SE). Standard errors in parentheses are clustered at the county level. * p < 0.1, ** p < 0.05, *** p < 0.01. Differences in sample sizes are due to missing values of some control variables.

3.3. Impact on the Efficiency of Industrial Land Transfer and Utilization

3.3.1. Impact on Industrial Land Transfer and Utilization Efficiency

Industrial development is the core driver of county economic growth. As the primary carrier of county industrial development, industrial land use efficiency has a direct impact on the quality of industrial development and the efficiency of land resource allocation. Table 4 reports the impact of land use regulation on the scale and utilization efficiency of county industrial land transfer. Column (1) shows that land use regulation significantly reduces the area of newly transferred industrial land (${ICL\_ind}_{i,t}$) (coefficient = −0.525, p < 0.01), reflecting the policy’s constraint on industrial land supply. Columns (2)–(3) indicate that land use regulation significantly raises the minimum floor area ratio of industrial land (${LFAR\_ind}_{i,l}$). A 1-unit increase in regulation intensity increases the minimum FAR by 0.049–0.055. Given the average industrial land FAR of 0.8, this improves land use efficiency by approximately 6.1–6.9% (p < 0.01), helping to avoid land waste from low-density development. Columns (4)–(5) find no significant effect of land use regulation on the maximum FAR of industrial land. This is because industrial land development intensity is mainly limited by production processes and safety standards, with no rigid constraints on the maximum FAR from local governments. Columns (6)–(7) show an insignificant negative impact of land use regulation on the agreed construction period of industrial land. This suggests the policy tends to shorten the development cycle, but the effect is not statistically significant.

Table 4. Impact of land use regulation on county industrial land transfer and utilization efficiency.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	${\mathit{I}\mathit{C}\mathit{L}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{t}}$	${\mathit{L}\mathit{F}\mathit{A}\mathit{R}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{l}}$	${\mathit{L}\mathit{F}\mathit{A}\mathit{R}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{l}}$	${\mathit{U}\mathit{F}\mathit{A}\mathit{R}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{l}}$	${\mathit{U}\mathit{F}\mathit{A}\mathit{R}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{l}}$	${\mathit{A}\mathit{C}\mathit{D}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{l}}$	${\mathit{A}\mathit{C}\mathit{D}\_\mathit{i}\mathit{n}\mathit{d}}_{\mathit{i},\mathit{l}}$
${Regulation}_{i,t}$	−0.525 ***	0.049 ***	0.055 ***	0.007	0.031	−0.002	0.016
	(0.138)	(0.009)	(0.018)	(0.013)	(0.022)	(0.008)	(0.018)
${Invest}_{i,t}$	0.001	−0.008 **	−0.060 ***	−0.012 **	−0.053 ***	−0.016 ***	−0.041 ***
	(0.006)	(0.004)	(0.010)	(0.006)	(0.014)	(0.004)	(0.008)
${PGDP}_{i,t}$	0.106 ***	−0.008	−0.116 ***	0.026 **	−0.184 ***	−0.001	0.007
	(0.039)	(0.009)	(0.013)	(0.013)	(0.018)	(0.008)	(0.012)
${Sufficiency}_{i,t}$	0.142 **	−0.038 ***	−0.001	−0.022 ***	0.010	0.061 ***	0.078 ***
	(0.069)	(0.006)	(0.011)	(0.008)	(0.012)	(0.004)	(0.006)
${Nagr}_{i,t}$	0.225	−0.040 ***	0.016	−0.050 ***	0.046 ***	0.042 ***	0.015
	(0.160)	(0.007)	(0.011)	(0.011)	(0.016)	(0.007)	(0.010)
${Wage}_{i,t}$	−0.039	0.217 ***	−0.114 ***	0.024	−0.296 ***	−0.095 ***	−0.332 ***
	(0.039)	(0.035)	(0.042)	(0.043)	(0.058)	(0.030)	(0.035)
${Pdensity}_{i,t}$	−0.021 *	−0.011 ***	0.020 ***	0.008 ***	0.029 ***	−0.008 ***	−0.004
	(0.012)	(0.002)	(0.003)	(0.002)	(0.003)	(0.001)	(0.002)
${Auction}_{i,l}$		0.106 ***	0.134 ***	0.189 ***	0.118 ***	0.089 ***	0.106 ***
		(0.007)	(0.009)	(0.010)	(0.013)	(0.007)	(0.009)
${Landlevel}_{i,l}$		0.022 ***	0.027 ***	0.015 **	0.018	−0.017 ***	−0.004
		(0.004)	(0.008)	(0.006)	(0.012)	(0.004)	(0.007)
Land Source		Newly Added	Non-Newly Added	Newly Added	Non-Newly Added	Newly Added	Non-Newly Added
_cons	1.949 ***	1.126 ***	1.850 ***	1.742 ***	3.284 ***	6.026 ***	6.405 ***
	(0.582)	(0.124)	(0.180)	(0.183)	(0.247)	(0.112)	(0.162)
Year	yes	yes	yes	yes	yes	yes	yes
County	yes	yes	yes	yes	yes	yes	yes
N	18,519	77,381	34,509	77,381	34,509	76,636	32,785
R2	0.395	0.468	0.522	0.609	0.608	0.489	0.543

Notes: The first row reports the regression coefficients, and the second row (in parentheses) shows the standard errors (SE). Standard errors in parentheses are clustered at the county level. * p < 0.1, ** p < 0.05, *** p < 0.01. Differences in sample sizes are due to missing values of some control variables. The “Land Source” row indicates the sources of industrial land, “Newly-added” means land converted from other uses in the previous year to construction land in the current year, while “non-newly-added” refers to existing construction land or construction land from the land stock bank.

3.3.2. Impact on Commercial–Residential Land Transfer and Utilization Efficiency

Table 5 reports the impact of land use regulation on the scale and utilization efficiency of county commercial–residential land transfer. Column (1) shows that land use regulation has a significant negative effect on the transfer scale of new commercial–residential land (${ICL\_comm}_{i,t}$) (coefficient = −0.483, p < 0.05). This reduction in commercial–residential land supply directly affects local governments’ land transfer income and related tax revenue. Columns (2)–(3) indicate that land use regulation increases the minimum FAR of commercial–residential land (${LFAR\_comm}_{i,l}$). Based on the average commercial–residential land FAR of 1.5, for each 1-unit increase in average regulation intensity, land use efficiency is improved by approximately 3.9–6.4% (p < 0.01), which helps avoid low-density development and enhances the carrying capacity of infrastructure. Different from industrial land, Columns (4)–(5) reveal that land use regulation reduces the maximum FAR of new commercial–residential land (${UFAR\_comm}_{i,l}$) (coefficient = −0.138, p < 0.01). This aims to control the development intensity of commercial–residential projects, prevent excessive population density, and balance living comfort with intensive land use. Columns (6)–(7) show that land use regulation shortens the agreed construction period of commercial–residential land (${ACD\_comm}_{i,l}$). This may be due to real estate developers accelerating construction to recover funds or local governments imposing stricter requirements on the development cycle to improve land turnover efficiency.

Table 5. Impact of land use regulation on county commercial–residential land transfer and utilization efficiency.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	${\mathit{I}\mathit{C}\mathit{L}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{t}}$	${\mathit{L}\mathit{F}\mathit{A}\mathit{R}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{l}}$	${\mathit{L}\mathit{F}\mathit{A}\mathit{R}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{l}}$	${\mathit{U}\mathit{F}\mathit{A}\mathit{R}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{l}}$	${\mathit{U}\mathit{F}\mathit{A}\mathit{R}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{l}}$	${\mathit{A}\mathit{C}\mathit{D}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{l}}$	${\mathit{A}\mathit{C}\mathit{D}\_\mathit{c}\mathit{o}\mathit{m}\mathit{m}}_{\mathit{i},\mathit{l}}$
${Regulation}_{i,t}$	−0.483 **	0.058 ***	0.096 ***	−0.138 ***	−0.002	−0.032 **	−0.091 ***
	(0.202)	(0.020)	(0.018)	(0.044)	(0.040)	(0.013)	(0.013)
${Invest}_{i,t}$	0.015	0.019 ***	−0.019 ***	0.005	0.036 ***	−0.001	−0.004 *
	(0.014)	(0.004)	(0.005)	(0.005)	(0.007)	(0.002)	(0.002)
${PGDP}_{i,t}$	0.064	0.045 ***	0.028 ***	−0.055 ***	−0.114 ***	−0.019 ***	0.008 *
	(0.045)	(0.009)	(0.009)	(0.014)	(0.014)	(0.005)	(0.005)
${Sufficiency}_{i,t}$	0.240 ***	0.051 ***	−0.036 **	0.045 **	−0.039 *	0.046 ***	0.064 ***
	(0.071)	(0.014)	(0.015)	(0.020)	(0.021)	(0.006)	(0.008)
${Nagr}_{i,t}$	0.849 ***	0.062 ***	0.051 ***	−0.065 ***	−0.018	0.030 ***	0.001
	(0.251)	(0.008)	(0.008)	(0.012)	(0.013)	(0.005)	(0.005)
${Wage}_{i,t}$	0.000	−0.156 ***	0.074 ***	−0.263 ***	0.229 ***	−0.053 **	−0.313 ***
	(0.047)	(0.042)	(0.016)	(0.061)	(0.036)	(0.024)	(0.013)
${Pdensity}_{i,t}$	−0.007	−0.018 ***	−0.035 ***	−0.006	−0.023 ***	−0.005 ***	−0.001
	(0.016)	(0.003)	(0.003)	(0.004)	(0.004)	(0.001)	(0.001)
${Auction}_{i,l}$		−0.126 ***	−0.018 ***	0.063 ***	0.079 ***	0.168 ***	0.230 ***
		(0.005)	(0.004)	(0.007)	(0.007)	(0.003)	(0.003)
${Landlevel}_{i,l}$		0.124 ***	0.091 ***	0.247 ***	0.319 ***	0.009 ***	0.012 ***
		(0.005)	(0.004)	(0.007)	(0.008)	(0.002)	(0.003)
Land Source		Newly Added	Non-Newly Added	Newly Added	Non-Newly Added	Newly Added	Non-Newly Added
_cons	0.782	−0.026	0.110	3.617 ***	3.621 ***	6.355 ***	6.432 ***
	(0.690)	(0.126)	(0.117)	(0.193)	(0.192)	(0.073)	(0.067)
Year	yes	yes	yes	yes	yes	yes	yes
County	yes	yes	yes	yes	yes	yes	yes
N	18,360	171,940	218,141	17,194	218,141	169,720	179,862
R2	0.377	0.305	0.432	0.332	0.339	0.436	0.480

Notes: The first row reports the regression coefficients, and the second row (in parentheses) shows the standard errors (SE). Standard errors in parentheses are clustered at the county level. * p < 0.1, ** p < 0.05, *** p < 0.01. Differences in sample sizes are due to missing values of some control variables. The “Land Source” row indicates the sources of commercial–residential land, “Newly-added” means land converted from other uses in the previous year to construction land in the current year, while “non-newly-added” refers to existing construction land or construction land from the land stock bank.

Overall, land use regulation has restricted the land transfer scale while pushing local governments to improve land use efficiency via policy constraints—realizing a shift from extensive expansion to intensive, efficient land use. This efficiency gain is mainly reflected in two aspects: FAR adjustments and shorter construction cycles, with differentiated regulation for industrial and commercial–residential land. For industrial land, the focus is on raising the minimum FAR to ensure intensive use. For commercial–residential land, lowering the maximum FAR (while balancing height controls) helps reconcile intensive use with living quality, and shortening construction cycles boosts land turnover efficiency. This differentiated regulation not only satisfies the normal land demand for industrial development and urbanization but also protects residents’ living standards. The improvement in land use efficiency driven by the policy can offset the negative impact of reduced land supply on fiscal revenue, forming a critical buffer for balancing cultivated land protection and economic development. These efficiency gains partially offset the adverse impacts of lower land supply and support sustainable county-level economic growth by raising output per unit of land. The heterogeneous improvement paths of land use efficiency in Hypothesis 3 are verified.

3.4. Impact on Local Finance

Land conveyance revenue and tax revenue are the two core sources of fiscal revenue for county governments. By influencing the scale and efficiency of land conveyance, land use regulation exerts an impact on county-level fiscal performance through both direct and indirect channels. This impact is reflected not only in changes in the total volume of fiscal revenue but also in significant regional heterogeneity. Table 6 presents the impact of land use regulation on county-level fiscal performance and the corresponding heterogeneity analysis. The results in Column (1) show that the regression coefficient of land use regulation on total land conveyance revenue (${LTF}_{i,t}$) is −0.294, which is significant at the 1% significance level. For each one-unit increase in the intensity of land use regulation, the actual land conveyance revenue decreases by approximately 25.7%. As a non-tax revenue obtained by county-level governments through the conveyance of land use rights, the scale of land conveyance revenue directly depends on the area of conveyed land and the corresponding transaction prices. By curbing the scale of land conveyance, land use regulation directly leads to a reduction in land conveyance revenue. The regression results in Column (2) indicate that land use regulation significantly reduces county-level tax revenue (${LTF}_{i,t}$). Tax revenue is closely linked to the level of industrial development in a county. By restricting the supply of land for industrial and commercial-service sectors, land use regulation indirectly slows down the extensive expansion of related industries, thereby leading to a short-term decline in tax revenue. A tighter supply of industrial land may slow down new projects and restrict the expansion of production scales of existing enterprises, resulting in a downward trend in growth. This, in turn, reduces taxes such as value-added tax and corporate income tax. A decrease in the supply of commercial and residential land directly affects the supply of commercial housing. The supply-demand imbalance in the real estate market may drive up housing prices, but it will also suppress the scale of real estate transactions. Meanwhile, it will impact the development of related service industries such as commerce, catering, and retail, leading to a reduction in real estate-related taxes (e.g., deed tax and real estate tax) and service industry taxes. As noted earlier, land use regulation promotes the improvement of land use efficiency and the increase in industrial added value per unit of land, which may boost certain types of taxes. However, this positive effect is smaller than the inhibitory effect caused by the reduction in land supply. Therefore, the overall tax revenue still shows a downward trend. The regression results in Column (3) confirm that land use regulation exerts a significant impact on the general public budget revenue of county-level governments (${Fiscalrev}_{i,t}$).

Table 6. Impact of Land use Regulation on County Finance and Heterogeneity Analysis.

	(1)	(2)	(3)	(4)	(5)	(6)
	${\mathit{L}\mathit{T}\mathit{F}}_{\mathit{i},\mathit{t}}$	${\mathit{T}\mathit{a}\mathit{x}}_{\mathit{i},\mathit{t}}$	${\mathit{F}\mathit{i}\mathit{s}\mathit{c}\mathit{a}\mathit{l}\mathit{r}\mathit{e}\mathit{v}}_{\mathit{i},\mathit{t}}$	${\mathit{L}\mathit{T}\mathit{F}}_{\mathit{i},\mathit{t}}$	${\mathit{T}\mathit{a}\mathit{x}}_{\mathit{i},\mathit{t}}$	${\mathit{F}\mathit{i}\mathit{s}\mathit{c}\mathit{a}\mathit{l}\mathit{r}\mathit{e}\mathit{v}}_{\mathit{i},\mathit{t}}$
${Regulation}_{i,t}$	−0.294 ***	−0.439 ***	−0.222 ***	−0.541 ***	0.112	−0.014
	(0.108)	(0.055)	(0.034)	(0.150)	(0.071)	(0.036)
${Regulation}_{i,t}$ $\times {MGPC}_i$				−0.648 ***	−0.209 ***	−0.201 ***
				(0.160)	(0.075)	(0.045)
${MGPC}_i$				0.842 ***	0.568 ***	0.509 ***
				(0.024)	(0.012)	(0.008)
${Invest}_{i,t}$	0.069 ***	−0.007	0.014 ***	−0.126 ***	−0.135 ***	−0.132 ***
	(0.019)	(0.007)	(0.004)	(0.019)	(0.009)	(0.007)
${PGDP}_{i,t}$	0.017	0.112 ***	0.103 ***	−0.033	0.136 ***	0.049 ***
	(0.032)	(0.013)	(0.008)	(0.020)	(0.010)	(0.008)
${Sufficiency}_{i,t}$	0.411 ***	0.874 ***	1.902 ***	2.467 ***	2.252 ***	2.722 ***
	(0.076)	(0.032)	(0.033)	(0.060)	(0.034)	(0.028)
${Nagr}_{i,t}$	2.475 ***	1.973 ***	1.434 ***	2.445 ***	2.306 ***	1.635 ***
	(0.193)	(0.088)	(0.051)	(0.121)	(0.061)	(0.045)
${Wage}_{i,t}$	−0.016	−0.023 *	−0.020 ***	−0.007	0.051 ***	0.044 ***
	(0.032)	(0.013)	(0.007)	(0.021)	(0.011)	(0.009)
${Pdensity}_{i,t}$	0.016	0.016 ***	−0.009 **	0.063 ***	0.026 ***	0.027 ***
	(0.016)	(0.004)	(0.004)	(0.004)	(0.002)	(0.001)
_cons	7.703 ***	8.113 ***	8.344 ***	6.847 ***	6.090 ***	7.496 ***
	(0.474)	(0.193)	(0.107)	(0.284)	(0.154)	(0.116)
Year	yes	yes	yes	yes	yes	yes
County	yes	yes	yes	no	no	no
N	27,135	27,576	27,648	27,153	27,593	27,665
R2	0.750	0.886	0.970	0.470	0.620	0.808

Notes: The first row reports the regression coefficients, and the second row (in parentheses) shows the standard errors (SE). Standard errors in parentheses are clustered at the county level. * p < 0.1, ** p < 0.05, *** p < 0.01. Differences in sample sizes are due to missing values of some control variables.

To address challenges such as the rigid growth in grain demand, intensified resource constraints, and volatility in the international market, the CPC Central Committee and the State Council released the National Plan for Increasing Grain Production Capacity by 50 Million Tons (2009–2020) in 2009. This document first proposed classifying all counties and districts nationwide into major grain-producing counties (MGPCs) and non-major grain-producing counties. It clearly specified enhancing grain production capacity via measures such as building core production zones and providing targeted policy support.

As the core regions responsible for implementing the national food security strategy, major grain-producing counties are subject to higher targets for cultivated land protection and stricter constraints from land use regulation. Therefore, by introducing a dummy variable ${MGPC}_i$ into the model, we further analyze the regional heterogeneity of land use regulation’s impact on local fiscal performance. The coefficients of the interaction term (${Regulation}_{i,t}\times {MGPC}_i$) in Columns (4) to (6) of Table 6 indicate that land use regulation exerts a more significantly negative impact on the land conveyance revenue, tax revenue, and general public budget revenue in MGPCs compared with non-MGPCs.

MGPCs are dominated by the agricultural sector in their industrial structure, with relatively low shares of industry and services and a narrow tax base. For these counties, land conveyance is not only an important supplement to county governments’ fiscal revenue but also a crucial means of attracting investment to drive industrial development. Hence, they are more vulnerable to the negative shocks of land use regulation. MGPCs already have a relatively low fiscal self-sufficiency rate due to their industrial structure dominated by agriculture. Land use regulation further erodes their independent fiscal capacity, while the growth of transfer payments from higher-level governments is insufficient to fill this gap. This exacerbates the predicament of the “grain-finance imbalance,” a unique fiscal dilemma in China where counties bearing heavier responsibilities for national grain production face disproportionately greater fiscal distress.

With the acceleration of urbanization, counties have witnessed a sustained increase in demand for public services such as infrastructure construction, education, medical care, and elderly care, leading to a continuous expansion of fiscal expenditure scales. The dual pressures of declining fiscal revenue and rising expenditure will result in an expansion of fiscal deficits and a surge in debt risks for MGPCs. Such fiscal pressure may induce short-sighted speculative behaviors among local governments, such as illegal land conveyance and the protection of cultivated land in terms of quantity rather than quality. Ironically, these behaviors may hinder the achievement of the core objective of land use regulation, namely, safeguarding national food security. The revenue effects in Hypotheses 2, 3 and 4 are verified.

3.5. Impact on Public Service Provision

The fiscal capacity of county-level governments constitutes the core foundation for ensuring public service provision, and access to public services is a key determinant of residents’ quality of life. The economic impacts of land use regulation are transmitted from the government to residents. This study measures public service provision across three dimensions: education, medical care, and social welfare.

The student–teacher ratio in regular primary and secondary schools (${Teachstu}_{i,t}$) measures educational services. It is a negative indicator: a higher ratio indicates a relative shortage of teaching resources. The number of beds in medical and health institutions per 10,000 people (${Healthcare}_{i,t}$) serves as the core indicator for medical service provision. It assesses the accessibility and adequacy of medical resources. The number of beds in social welfare and adoptive institutions per 10,000 people (${socialcare}_{i,t}$) measures social welfare. These two are positive indicators.

Table 7 presents the impact of land use regulation on county-level public service provision and the corresponding heterogeneity analysis. Columns (1)–(3) verify the public service provision effect of Hypothesis 4. In Column (1), the regression coefficient of ${\mathrm{R}\mathrm{e}\mathrm{g}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\mathrm{i},\mathrm{t}}$ on ${\mathrm{T}\mathrm{e}\mathrm{a}\mathrm{c}\mathrm{h}\mathrm{s}\mathrm{t}\mathrm{u}}_{\mathrm{i},\mathrm{t}}$ is 0.563, significant at the 1% level, indicating that land use regulation exacerbates the shortage of educational resources. On the one hand, the reduction in fiscal revenue makes it difficult for local governments to increase educational input, which restricts the recruitment and training of teachers. On the other hand, constrained industrial development has an adverse impact on the regional wage level, leading to the outflow of young and middle-aged labor, and a subsequent increase in the number of left-behind children. Fiscal constraints represent one of the most critical factors. Ultimately, real spending on education and the delivery of public services depend on governments’ policy priorities and fiscal expenditure structure. The regression results in Column (2) show that land use regulation leads to a significant decline in the capacity of medical service provision. There are three main reasons for this: first, reduced fiscal revenue results in insufficient medical investment, making it difficult for local governments to build new hospitals or expand existing ones, thus causing delays in the construction of medical facilities; second, inadequate land supply limits the provision of land for medical use, meaning new hospitals lack construction land and cannot increase the number of beds; third, the agglomeration effect of medical resources is weakened—tight land supply makes it hard to attract medical talents, and the quality of medical services therefore fails to improve. Column (3) shows that the regression coefficient of ${\mathrm{R}\mathrm{e}\mathrm{g}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\mathrm{i},\mathrm{t}}$ on ${\mathrm{S}\mathrm{o}\mathrm{c}\mathrm{i}\mathrm{a}\mathrm{l}\mathrm{c}\mathrm{a}\mathrm{r}\mathrm{e}}_{\mathrm{i},\mathrm{t}}$, is −8.529 ***, with the largest absolute value among the three indicators. This suggests that among the three types of public service provision, land use regulation exerts the strongest inhibitory effect on social welfare. The reason is that social welfare services, due to their public welfare nature, are more dependent on government fiscal input. When fiscal revenue decreases, local governments will first cut the scale of non-compulsory social welfare expenditures.

Table 7. Impact of land use regulation on county public services and heterogeneity analysis.

	(1)	(2)	(3)	(4)	(5)	(6)
	${\mathit{T}\mathit{e}\mathit{a}\mathit{c}\mathit{h}\mathit{s}\mathit{t}\mathit{u}}_{\mathit{i},\mathit{t}}$	${\mathit{H}\mathit{e}\mathit{a}\mathit{l}\mathit{t}\mathit{h}\mathit{c}\mathit{a}\mathit{r}\mathit{e}}_{\mathit{i},\mathit{t}}$	${\mathit{S}\mathit{o}\mathit{c}\mathit{i}\mathit{a}\mathit{l}\mathit{c}\mathit{a}\mathit{r}\mathit{e}}_{\mathit{i},\mathit{t}}$	${\mathit{T}\mathit{e}\mathit{a}\mathit{c}\mathit{h}\mathit{s}\mathit{t}\mathit{u}}_{\mathit{i},\mathit{t}}$	${\mathit{H}\mathit{e}\mathit{a}\mathit{l}\mathit{t}\mathit{h}\mathit{c}\mathit{a}\mathit{r}\mathit{e}}_{\mathit{i},\mathit{t}}$	${\mathit{S}\mathit{o}\mathit{c}\mathit{i}\mathit{a}\mathit{l}\mathit{c}\mathit{a}\mathit{r}\mathit{e}}_{\mathit{i},\mathit{t}}$
${Regulation}_{i,t}$	0.563 ***	−5.182 ***	−8.529 ***	0.033	2.055	−1.535
	(0.200)	(1.203)	(1.586)	(0.355)	(1.278)	(1.768)
${Regulation}_{i,t}$ $\times {MGPC}_i$				1.334 ***	−7.309 ***	−19.332 ***
				(0.442)	(1.536)	(2.441)
${MGPC}_i$				2.914 ***	−5.275 ***	−1.255 ***
				(0.079)	(0.279)	(0.414)
${Invest}_{i,t}$	−0.146 ***	−0.554 ***	−1.005 ***	−0.405 ***	−0.115	−0.093
	(0.032)	(0.177)	(0.238)	(0.054)	(0.255)	(0.284)
${PGDP}_{i,t}$	−0.571 ***	5.236 ***	2.026 ***	−2.461 ***	7.025 ***	5.975 ***
	(0.055)	(0.320)	(0.454)	(0.066)	(0.268)	(0.318)
${Sufficiency}_{i,t}$	−0.281 **	3.356 ***	−2.957 ***	1.135 ***	−0.165	5.068 ***
	(0.118)	(0.852)	(1.043)	(0.167)	(0.984)	(0.928)
${Nagr}_{i,t}$	−0.552	19.644 ***	8.087 ***	8.372 ***	11.453 ***	9.062 ***
	(0.355)	(1.832)	(2.733)	(0.398)	(1.614)	(1.747)
${Wage}_{i,t}$	−0.190 ***	−2.382 ***	0.756 *	−0.772 ***	−0.356	0.724 **
	(0.059)	(0.320)	(0.439)	(0.073)	(0.298)	(0.356)
${Pdensity}_{i,t}$	0.305 ***	−0.274 *	−0.497 ***	0.257 ***	0.268 ***	0.032
	(0.021)	(0.150)	(0.150)	(0.010)	(0.068)	(0.058)
_cons	21.325 ***	−5.360	−0.190	38.237 ***	−36.671 ***	−47.392 ***
	(0.807)	(4.514)	(6.596)	(0.921)	(4.029)	(4.767)
Year	yes	yes	yes	yes	yes	yes
County	yes	yes	yes	no	no	no
N	27,563	27,634	27,634	27,580	27,651	27,651
R2	0.886	0.838	0.665	0.179	0.336	0.143

Notes: The first row reports the regression coefficients, and the second row (in parentheses) shows the standard errors (SE). Standard errors in parentheses are clustered at the county level. * p < 0.1, ** p < 0.05, *** p < 0.01. Differences in sample sizes are due to missing values of some control variables.

Columns (4) to (6) of Table 7 present the heterogeneity in the impact of land use regulation on public service provision between major grain-producing counties and non-major grain-producing counties. The coefficients of the interaction term (${Regulation}_{i,t}\times {MGPC}_i$) indicate that, compared with non-major grain-producing counties, land use regulation exerts more pronounced negative effects on the educational resources, medical services, and social welfare of major grain-producing counties. Hypothesis 5 is verified.

The regional imbalance in public services caused by land use regulation not only hurts the quality of life of residents in major grain-producing counties (MGPCs) but also introduces a range of long-term negative impacts: First, it worsens urban–rural and inter-regional development gaps. MGPCs are mostly rural, and insufficient public service provision will further widen the urban–rural divide—directly contradicting the strategic goal of rural revitalization. Second, it causes talent loss. Backward education and medical services in MGPCs drive high-quality talent away, further limiting county economic development and even trapping these areas in a vicious circle. Third, it weakens the sustainability of the national food security strategy. Farmers in MGPCs bear the costs of cultivated land protection but lack equal access to public services. This may reduce their enthusiasm for protecting cultivated land, undermining the foundation of the national food security strategy. However, the above analysis only shows how land use regulation affects fiscal constraints in public service provision. It does not mean that reduced public services are an inevitable outcome of land use regulation. Local governments can offset the impact of lower fiscal revenue on public services by adjusting their fiscal expenditure structure, increasing transfer payments, and adopting other policy tools. This is also the core of the policy recommendations put forward in this paper.

4. Conclusions

Based on county-level statistical data, remote sensing geographic information data, and micro land transaction data from 2006 to 2022, this study adopts the Intensity Difference-in-Differences (Intensity DID) method to systematically analyze the impact mechanism and heterogeneous characteristics of land use regulation on county land transfer, fiscal revenue, and public service provision. The core findings are as follows:

Land Transfer: Through the binding constraint of cultivated land protection targets, land use regulation has significantly curbed the scale of land transfer: the total land transfer area decreased by approximately 13.6%, and the new construction land transfer area decreased by approximately 31.7%. Meanwhile, the policy has forced local governments to revitalize existing construction land to offset the gap in construction land demand. In terms of land source structure, the policy has promoted a shift in new construction land sources from cultivated land to forest land, grassland, water bodies, and unused land, forming a distinct non-cultivated land substitution effect. Additionally, the regulation has effectively driven local governments to improve land use efficiency by increasing the minimum FAR for industrial and commercial–residential land, reducing the maximum FAR for commercial–residential land, and shortening the construction cycle, thereby realizing the transformation of land use from extensive expansion to intensive and efficient utilization.

Fiscal Impact: Land use regulation has exerted a dual impact on county finance. It directly reduces extra-budgetary land transfer income and indirectly decreases tax revenue by inhibiting industrial development, ultimately leading to a decline in general public budget revenue. Compared with non-major grain-producing counties, major grain-producing counties have experienced more severe negative impacts from the policy, primarily due to their stricter cultivated land protection intensity, single industrial structure, low fiscal self-sufficiency rate, and higher dependence on land transfer income.

Public Service Provision: Land use regulation has indirectly led to insufficient public service provision by suppressing fiscal revenue. Specifically, the shortage of educational resources has increased by approximately 4.2%, medical service provision capacity has decreased by approximately 12.96%, and social welfare service provision capacity has decreased by approximately 26.14%. The public service shock in major grain-producing counties is more pronounced: the shortage of educational resources has increased by approximately 10.14%, medical service provision capacity has decreased by approximately 23.42%, and social welfare service provision capacity has decreased by approximately 64%, exacerbating regional disparities in public service levels. It is important to note that the inadequate provision of public services in some regions is a combined result of fiscal constraints and local governments’ fiscal expenditure decisions, rather than a direct and inevitable consequence of land use regulation itself.

5. Policy Recommendations

To coordinate the food security strategy with balanced regional development, promote equal access to public services, and ultimately achieve the sustainable development of land resources, we put forward the following policy recommendations to optimize the current land use regulation system based on the above research findings:

Strengthen the combined vertical and horizontal interest compensation mechanism for major grain-producing areas. Vertically, link fiscal support closely to cultivated land protection: optimize the central transfer payment formula for major grain-producing counties by including core indicators such as cultivated land area and grain output, ensuring funding matches protection responsibilities; adjust the transfer payment structure to increase general transfer payments, granting local governments greater autonomy in fund use. Meanwhile, boost special investment in public services—central and provincial finances should prioritize rural education and medical infrastructure, attract talents via targeted training and subsidies, and build cross-regional resource allocation mechanisms to enable urban high-quality resources to support rural areas through remote collaboration. Horizontally, implement the “who benefits, who compensates” principle: major grain-consuming areas should compensate producing areas based on grain import volumes to offset development opportunity costs and fiscal losses. The central government may set up a special fund for additional subsidies, with provincial pilot programs preceding nationwide promotion to alleviate the “grain-finance inversion” dilemma.

Implement differentiated land use regulation and land market allocation optimization strategies. For major grain-producing counties, on the premise of strictly abiding by the cultivated land red line and food security bottom line, moderately relax non-cultivated land development restrictions and optimize the market-oriented allocation of stock construction land. Support them in revitalizing existing land through FAR incentives and compatible land use policies, and allocate a higher proportion of land transfer income to public services. For non-major grain-producing counties, strengthen rigid constraints on cultivated land protection, improve land use efficiency, and strictly control low-density development. Guide them to develop high-value-added industries, reduce reliance on land transfer income, and enhance fiscal autonomy through industrial upgrading and increased tax revenue. Land use regulation is an essential policy instrument. The combination of rigid farmland protection and flexible land market allocation holds the key to balancing food security and economic development, as well as achieving sustainable growth.

Improve the incentive and constraint mechanisms for intensive land use. On the incentive side, offer tax reductions and fiscal subsidies to industrial, commercial, and residential projects with FAR above standards and short construction cycles. Encourage local governments to revitalize idle land through urban–rural construction land increase–decrease linkage and reclaiming of abandoned industrial/mining land, boosting effective construction land supply. On the constraint side, incorporate indicators such as FAR, construction cycle, and per-unit land output into local government performance evaluations. Reduce new construction land quotas for counties with low land use efficiency to force optimization of land use methods.

Promote industrial structure optimization and upgrading in major grain-producing areas. In modern agriculture and agricultural product processing, increase agricultural science and technology investment to develop high-efficiency and protected agriculture, and support deep-processing enterprises with land and tax incentives to extend industrial chains and increase agriculture-related tax revenue. In characteristic service industries, leverage local resource endowments to develop rural tourism, agricultural sightseeing, and rural e-commerce, fostering new growth points. Strengthen infrastructure construction, improve the investment environment, attract external capital, and promote coordinated industrial and service sector development to reduce reliance on land transfers.