Heterogeneous mechanisms of urban land price in China: a perspective of natural restrictions and strategic supply

The study addresses how natural restrictions and government interventions shape urban land prices in China, a critical issue for understanding the functioning of land markets and sustainable urbanization. With rapid economic development and urbanization, China’s land market has grown substantially, affecting housing affordability, industrial investment, and regional development. Existing work often emphasizes economic and governance perspectives but underexplores geographic factors such as natural landscapes (e.g., ecological redlines) that constrain land supply and create spatial heterogeneity. The research aims to identify determinants of urban land prices across land-use types (industrial, residential, commercial), integrating demand-side socioeconomic drivers and supply-side constraints, notably natural restrictions and strategic land supply by governments. It examines spatial-temporal dynamics between 2008–2015 (market formation and rapid growth) and 2015–2020 (booming market), revealing heterogeneity across time and land-use types and informing planners and developers.

The literature highlights that urban land prices reflect supply–demand dynamics and broader regional development patterns, with price increases typically originating in leading metropolitan areas before diffusing. Amenity accessibility (e.g., transit, parks) has become salient in western contexts, with emerging intra-urban studies in China. Socioeconomic factors (population, income, employment, GDP) are key determinants, and heterogeneity exists across land-use types (residential/commercial more demand-driven; industrial linked to macroeconomic conditions). However, assumptions of flexible land supply are challenged by natural and regulatory constraints; supply shortages have driven price increases globally, and research debates the relative roles of natural endowments (water bodies, terrain, wetlands) and policy restrictions (permits, growth boundaries, density limits, zoning) in shaping supply elasticity. Synergistic effects of natural and regulatory constraints can explain substantial interurban price differentials and elasticity variation, though impacts vary across space and scale. In China, incomplete market institutions and ambiguous property rights enable central and local government interventions via monopolized primary land supply and transaction method choices. Local governments pursue dual goals: subsidizing industrial land to attract investment and maximizing revenues via high-priced, limited residential/commercial land supply, leading to discriminatory pricing strategies and potential misallocation across land-use types and administrative hierarchies. Policies like the ecological redline may further restrict supply. The literature suggests spatial inequality in the land market is sensitive to local context, with limited understanding of how natural landscapes interact with strategic supply in China.

Study area: 31 provincial administrative units in mainland China, grouped into eastern, central, and western regions. Spatial scales: intra-city and inter-city. Data sources: City-level land-leasing data from China Land and Resources Statistical Yearbook (2009–2018) and China’s land market website (2019–2020); sociodemographic and macroeconomic data from China City Statistical Yearbooks (2009–2021). Parcel-level data (values, date, location, size, transaction method, usage) from China’s land market website, covering 1.6+ million transactions through 2020. Focused on industrial, residential, and commercial parcels in 2008, 2015, 2020. Transaction methods categorized as negotiated sales (non-market) vs. market (two-stage auctions/guapai, English auctions/paimai, sealed bids/zhaobiao). Outliers removed by excluding top and bottom 1% of recorded high/low transaction values. Variables: Dependent variable is the log unit price of land parcel (10,000 yuan/ha). Explanatory variables grouped as: (1) Natural restrictions: cultivated land per capita (log), % water, % green space, average elevation, average slope. (2) Strategic land supply: per capita leased industrial, residential, commercial land (log) during 2008–2020; change rates of industrial, residential, commercial land supply (2008–2020); transaction method dummy (NEGOS=1 negotiated). (3) Economic and social development: principal component (ECONOMIC) combining GDP, FDI, fiscal revenue per capita (>90% variance explained), population density (log), tertiary sector share. (4) Parcel-level controls: parcel size (log), new vs existing parcel (SOURCE), land grade dummies (1–5), city rank (CR: direct-controlled municipalities/sub-provincial/provincial capital vs prefectural), location in urban district vs county (URBAN). Inequality measures and spatial analysis: Computed spatial inequality in prices using Gini coefficient, Theil’s T, and coefficient of variation (CV) for 2008, 2015, 2020, by land-use type and region. Conducted Getis–Ord Gi* hot-spot analysis to map changing patterns. Econometric models: Estimated OLS models of log land price as a function of supply (S), economic/social factors (E), government/strategic variables (G), and parcel traits (L). Employed a spatial regime model allowing coefficients to vary across eastern, central, and western regions to capture heterogeneity; specified separate models for 2015 and 2020 to compare 2008–2015 vs 2015–2020 dynamics. Collinearity checked via VIF. Machine learning: Applied regression tree models (R packages rpart; Python sklearn) to assess non-linear relationships and variable importance, focusing on seven most important predictors for each land-use type in 2015 and 2020. Determinants assessed separately for industrial, residential, and commercial land.

- Inequality dynamics: Industrial and commercial land price inequality rose markedly from 2008 to 2015 and declined from 2015 to 2020, especially in the eastern region; residential land price inequality continued to increase throughout 2008–2020. - Quantitative inequality metrics (Table 3): For industrial land (China), CV increased from 0.53 (2008) to 1.54 (2015) then declined to 1.35 (2020); Gini from 0.28 to 0.36 to 0.39. For residential land (China), CV 1.21→1.40→1.56; Gini 0.49→0.49→0.53. For commercial land (China), CV 1.33→2.08→1.54; Gini 0.52→0.54→0.55. - Regional land market evolution (Table 2): Total leased land area rose from 1.6586 million ha (2008) to 2.2489 million ha (2015) to 3.255 million ha (2020); revenues increased from RMB 1.03 trillion (2008) to 3.12 trillion (2015) to 8.18 trillion (2020). Market transactions (Zhaopaigua) dominated and expanded; negotiated sales volumes fluctuated but their revenue share in the east dropped notably by 2020. - Discriminatory pricing strategies intensified in the east post-2015: Governments favored low-priced industrial land (often via negotiated sales) to attract investment, while keeping residential/commercial land supply tighter with higher prices. - Natural restrictions: Cultivated land per capita, water, green space, elevation, and slope generally exert significant negative effects on land prices (industrial, residential, commercial), indicating scarcity constraints; the negative impact on industrial land prices diminished after 2015. - Strategic land supply effects: 2008–2015 models show increasing residential land supply associated with higher residential land prices (contrary to standard supply–demand), consistent with government interventions to maximize revenue and stimulate markets; industrial and commercial prices decreased with increased respective supply, aligning with subsidized pricing strategies. - Transaction method effects (NEGOS): For 2008–2015, NEGOS is significantly negative for industrial land (negotiated sales reduce industrial prices) and positive for residential and commercial land (negotiated sales associated with higher prices). After 2015, the NEGOS coefficient for industrial land becomes positive, consistent with strengthened discriminatory pricing policies. - Market forces gained importance post-2015: ECONOMIC and POPDEN effects strengthened across land-use types; regression tree analyses show rising importance of socioeconomic variables and declining importance of cultivated land for industrial prices after 2015. - City–county convergence: The importance of URBAN (city vs county) declines after 2015, indicating narrowing price gaps, especially for industrial and commercial land, reflecting development in rural and county regions. - Parcel-level effects: Evidence of a plattage effect (smaller parcels command price premiums). Existing residential parcels tended to be more valuable than new ones, with the gap widening after 2015. - Model performance: Spatial regime models achieved R2 above 0.2 for all land-use types; for example, R2≈0.247 (industrial 2015), 0.428 (residential 2015), 0.331 (commercial 2015); 0.253, 0.546, 0.316 respectively in 2020.

The findings demonstrate that both natural restrictions and strategic government interventions shape urban land prices in China, with their relative influence changing over time and varying by land-use type and region. During 2008–2015, natural constraints—particularly cultivated land availability—played a dominant role, especially for industrial land, reflecting the strong influence of geography on supply elasticity. Post-2015, the diminishing effect of natural restrictions and the heightened significance of socioeconomic demand indicators (economic development, population density) indicate a shift toward more market-driven pricing dynamics. Government strategies produce heterogeneous effects: industrial land is often priced lower (especially in eastern regions) to attract investment, while residential land—central to land finance—exhibits pricing outcomes consistent with revenue-maximizing behavior and managed supply, contributing to persistent and rising inequality. The narrowing city–county price differentials suggest broader spatial diffusion of development benefits and a maturing land market beyond core urban districts. Policy initiatives such as the ecological redline and territorial space planning likely further constrain supply, reinforcing the need to balance environmental protection with market functioning and equitable access to land for different uses.

This study maps the spatial-temporal evolution of China’s urban land market (2008–2020) and disentangles heterogeneous mechanisms across industrial, residential, and commercial land. It shows that inequality in industrial and commercial land prices rose pre-2015 and declined thereafter, while residential inequality continued to rise, linked to land finance reliance. Natural restrictions (cultivated land, water, green, terrain) significantly constrain prices, but their influence—particularly on industrial land—declined after 2015 as market forces strengthened. Strategic supply and transaction methods play distinct roles by land-use type: discriminatory pricing strategies on industrial land intensified post-2015, while governmental control and parcel traits largely shape residential prices. City–county price gaps narrowed, reflecting the growing role of county and rural markets. Policy implications include the need for market-oriented land-use planning in western regions, improved efficiency in industrial land allocation in the east, and optimized residential land supply under ecological and territorial planning constraints to sustain housing market vitality. The research contributes to understanding how natural and institutional factors jointly structure land price dynamics and spatial inequality in transitional economies and provides guidance for balancing land supply, pricing, and sustainable growth.