Spatio-Temporal Analysis of the Redundancies of Construction Land in the Beijing-Tianjin-Hebei Region (2000–2020)

Abstract

Excessive redundancy of construction land in county-level units within the Beijing-Tianjin-Hebei region has become a significant obstacle to achieving high-quality development. The objective of this study is to discover the spatial and temporal patterns of redundancy of construction land, with a view to providing insights for promoting efficient land use. The study employs the SBM-DEA model, Markov transfer probability matrix analysis, and multiple regression analysis to analyze the spatial change characteristics, spatial differentiation, and influencing factors of construction land redundancy in this Beijing-Tianjin-Hebei county unit during the period of 2000–2020. The study shows that the Beijing-Tianjin-Hebei county unit has a serious oversupply of land and, combined with the reasons for redundancy in each sub-region, the degree of spatial redundancy has already formed a spatial lock-in effect. The degree of redundancy of construction land is affected by a variety of factors such as location, scale, economy, and facilities. Furthermore, the study puts forward suggestions for improving land use efficiency in Beijing-Tianjin-Hebei county units by adjusting the construction land supply and demand relationship, mechanisms to facilitate the flow of development factors, and strengthening land use supervision. These measures aim to reduce redundancy of construction land and support sustainable high-quality development in the region.

1. Introduction

During the past several decades, the rapid expansion of construction land has supported high-speed urbanization in China, contributing to a historic miracle in global urbanization processes [1,2]. Since China’s tax system reform in 1994, land transfers have emerged as a crucial source of revenue for local government, while the availability of relatively low-priced land has been instrumental in attracting industries [3]. Consequently, construction land has become a vital tool for wealth creation and economic stimulation [4], promoting local governments to actively pursue additional construction land for new projects.

This “land-based development” model has generated a substantial amount of inefficient and idle land [5,6], impeding sustainable development. The enthusiasm for land development has led to a situation where the pace of “land urbanization” significantly outstrips that of “population urbanization” [7]. Consequently, the insufficient emphasis on human development and land use quality has inflicted severe damage on ecological systems and agricultural landscapes [8].

By 2023, China’s urbanization rate had reached 66%; however, the prevailing model of urbanization driven by incremental construction land hardly underpins sustainable development in the future [9]. Moreover, as resource and environmental constraints become increasingly stringent, it is essential to enhance land use efficiency and curb extensive land use to achieve healthy urbanization in China’s new era [10].

In China, the central government implements a policy of total control of incremental construction land quotas, with both the long-term and annual quotas allocated in a top-down system across regions [11,12,13]. The quotas a region receives are determined by national or regional strategies, as well as local socioeconomic needs. Given the critical role that construction land plays in driving economic growth, these quotas also represent the “right to development”. To ensure equitable distribution, higher-level governments allocate construction land quotas based on population sizes, often providing preferential treatment to less developed regions [14]. Some studies also indicate that the construction land quotas obtained by local governments are influenced by their bargaining ability with higher-level authorities [15]. As a result, the supply and demand for construction land occasionally diverge, creating a coexistence of tension and inefficiency of land use, particularly in counties across China.

In order to ensure food and ecological security, the scale of incremental construction land is stringently controlled. Optimizing land allocation and usage, as well as improving land use efficiency, has become a critical topic of discussion. Existing studies have primarily focused on the measurement of land use efficiency, spatial differentiation, and the influencing factors and mechanisms. For instance, Bao et al. established an indicator system to evaluate land use efficiency [16]; Chen and Wu examined the economic efficiency of urban construction land and influential factors in Yangzte River Delta [17]; Liu et al. analyzed land use efficiency across cities in Hunan Province [18]; and Yang and Lang investigated land use efficiency in inland China [19]. These studies span multiple levels, including the national level, urban agglomeration, and provincial levels, with measurements mainly conducted at the provincial or prefectural-city level. For example, Liang et al. assessed the land-use efficiency of China [20] and Guo, Wu et al., and Chen et al. studied the land-use efficiency of the cities within urban agglomerations [21,22,23]. A variety of methodologies have been employed, such as indicator–based evaluation systems, principal component analysis, factor analysis [24], Malmquist-Luerberger index analysis [25], data envelopment analysis (DEA) [26,27], the super-efficiency SBM model [28], the stochastic frontier model [29], and the total factor productivity index method [30].

In general, a scarcity of the literature specifically addresses land use efficiency at the county level. As counties lag significantly behind in terms of economic development, industrial structure, innovation, and population scales, they exhibit a higher dependence on land finance, with inefficient land being even more pronounced. However, existing studies usually put the county units and central districts of prefecture-level cities together when measuring land use efficiency, overlooking the hierarchical differences within the urban system structures in urban agglomerations. Additionally, total factor productivity is often utilized in the input–output model as a proxy for the output productivity of construction land, leading to inaccurate assessments of land use efficiency. To address this, this paper adopts the redundancy degree of construction land as an indicator of land productivity. In the input–output model, factor redundancy is commonly utilized to reflect the gap between the actual input and the optimal input under the scenario of full use of resources. Similarly, the redundancy of construction land refers to the difference between the supply and demand of construction land under optimal utilization, with the ratio of the difference to the ideal value being defined as the redundancy degree of construction land.

According to the Cobb-Douglas production function, labor, capital, and construction land constitute the primary inputs in the social and economic system. Optimizing the quantities of these inputs can maximize the utilization of each factor and generate high productivity (see Figure 1). Notably, the quantity of construction land can be adjusted by the government [31]. As mentioned above, the allocation of incremental construction land quota emerges from a balancing and bargaining process between superior and local governments. The primary objective of these quotas is to support both high construction quality and elevated economic performance. When the supply of construction land is appropriately calibrated, efficient land use is achieved. Conversely, when the total land supply significantly exceeds demand, inefficient land-use patterns may emerge, undermining the development goals.

During the past 30 years, local governments in China have often embarked on the development of new districts as part of national or regional development strategies [32,33,34]. In many instances, substantial amounts of construction land have been allocated to these areas, far exceeding the actual demand. Although development efficiency typically improves as these districts become operational, transitioning the land-use efficiency from initially low to high, excessive land supply can lead to issues such as land idling, with surpluses eventually evolving into long-term redundancy.

The Beijing-Tianjin-Hebei (BTH) region, one of China’s three major urban agglomerations, has undergone rapid urbanization over the past three decades, accompanied by a swift expansion of construction land [35,36]. In BTH, regional imbalances remain evident, implying that the actual demand for construction land varies significantly [37]. At the county level, the mismatch between land development and socioeconomic growth is particularly pronounced [38]. As such, regulating quotas for incremental construction land has become a critical strategy for enhancing land use efficiency [39].

This study analyzes the efficiency of construction land at the county level in the BTH region from the perspective of redundancy. Specifically, we assess construction land redundancy at five key time points, 2000, 2005, 2010, 2015, and 2020, to capture its evolution over the past two decades. The research focuses on addressing three key questions: (1) What are the spatio-temporal characteristics of construction land redundancy among counties in the BTH region from 2000 to 2020?; (2) Is there a spatial evolution pattern in the degree of redundancy across the region?; and (3) What are the main determinants influencing construction land redundancy at the county level?

Based on these findings, this study proposes recommendations to enhance construction land use efficiency, aiming to provide insights for promoting high-quality and integrated development in the BTH region.

2. Data and Methods

2.1. Study Area

The BTH region is located in northern China, encompassing the plains of the North China Plain and the foothills of the Yanshan Mountain. The BTH region spans roughly 218,000 square kilometers (Figure 2). The region comprises the municipalities of Beijing and Tianjin, along with 11 prefecture-level cities in Hebei Province. Except for those central districts in the prefecture-level cities, there are 144 county units in this region.

The BTH region exhibits a significantly high land development intensity. According to statistics, the regional land development intensity stands at 13.15%, more than three times the national average. Furthermore, the intensity has been increasing rapidly [40], with an average annual growth rate of nearly 2.2%. At the prefecture level, Beijing and Tianjin have the highest development intensity, while cities like Baoding and Handan exhibit lower levels of development intensity.

In the context of the rapid increase in construction land, this trend has not only led to extensive encroachment on ecological and arable land resources but also resulted in inefficient and unsustainable land use. In response, the Beijing-Tianjin-Hebei Coordinated Development Land Use Master Plan (2015–2020), issued in 2015, called for stringent controls on the expansion of construction land while emphasizing the need to unlock the potential of existing land resources [41].

2.2. Research Methods

2.2.1. SBM-DEA Model

Data Envelopment Analysis (DEA), first formalized by Charnes, Cooper, and Rhodes in 1978 [42], constitutes a non-parametric frontier estimation technique for evaluating relative efficiency. To overcome the deficiency of conventional DEA in handling slack variables, the Slacks-Based Measure (SBM-DEA) variant is developed fundamentally [43], enabling the precise measurement of proportional excess inputs required to attain frontier efficiency, which are particularly critical when analyzing heterogeneous resource utilization patterns.

In this research, the SBM-DEA model is used to explore development efficiency, which can be reflected by the socioeconomic benefits generated per unit of input. At the county level, input factors, such as capital, land, and labor, are represented as X, while the social benefits (Y1) and economic benefits (Y2) are treated as outputs. The cost of generating unit benefits for each county can be calculated as X/Y1 (social benefit cost) and X/Y2 (economic benefit cost), with frontier projections identifying Pareto-efficient counties (A, C, E, F, G in Figure 3) versus inefficient units exhibiting input redundancy (B, D). For County B, the technical inefficiency measure OB derives from its orthogonal projection B’ on the frontier, such that

Redundancy Degree = (OB − OB′)/OB

(1)

where OB represents observed inputs and OB′ denotes efficient input levels under frontier technology. This slack-based decomposition reveals not only the magnitude but also the structural composition of resource misallocation.

In this study, capital, construction land, and labor are designated as input variables, with regional economic development and social development levels serving as output variables. Based on the availability, validity, and continuity of data, capital input is quantified through public budget expenditure; construction land input is represented by the scale of construction land in each county; and labor input is measured by the number of employees in the secondary and tertiary industries. For output indicators, public budget revenue and the GDP of the secondary and tertiary industries represent economic benefits, while the indicator of total deposits of urban and rural residents reflecting the family wealth is used to represent social benefits. After calculating the total factor productivity (TFP) of county-level units, the study further analyzes the redundancy of construction land.

2.2.2. Spatial Markov Model

The spatio-temporal dynamics of urban-rural construction land redundancy present a critical research question: how do these redundancies evolve temporally, and are there discernible patterns of spatial diffusion? To address this inquiry, this study conceptualizes the evolution of construction land redundancies as a Markovian process, employing a transition probability matrix to examine their dynamic evolution across temporal scales.

The Spatial Markov Model (SMM), which integrates spatial autocorrelation analysis with temporal transition probability analysis, is employed [44,45]. Our application of SMM incorporates a spatial lag term to quantify neighborhood effects, enabling rigorous examination of both temporal transitions and spatial spillover effects in county-level redundancy patterns.

In this study, redundancy levels are classified into four categories based on quartiles (0.25, 0.5, 0.75) for county-level units: low (k = 1), medium–low (k = 2), medium–high (k = 3), and high (k = 4). A lower redundancy level indicates higher construction land use efficiency. Based on the calculated results, a Markov transition probability matrix is constructed. In the matrix, P_ij represents the transition probability, which is the probability of being in category i at time t and transitioning to category j at time t + 1. The specific formula for calculating the transition probability is as follows:

P_ij = N_ij/N_i

(2)

Here, N_ij represents the number of county-level units that transitioned from type i at time t to type j at the next time period during the study period. N_i denotes the total number of county-level units classified as type i over the entire study period. When i = j, it indicates that redundancy remains unchanged, meaning the status quo is maintained. When i < j, it signifies an increase in redundancy, implying a decline in construction land use efficiency. When i > j, it indicates a decrease in redundancy, reflecting an improvement in construction land use efficiency.

The study utilizes the spatial Markov transfer matrix to investigate the spatial spillover effect caused by geographic proximity. The study introduces the spatial lag condition and decomposes the transfer probability matrix into four 4 × 4 conditional probability matrices by neighborhood type, where the element P_ij|k in the matrix represents the probability that the redundancy of a county unit will shift from type i to type j at the next moment under the condition that the spatial neighborhood type is k. The spatial lag value Lag_a for region a is a weighted average of the observations of the geographic units surrounding the region, calculated as

$$La{g}_a={\displaystyle {\sum }_{b=1}^n{Y}_b}{W}_{ab}$$

(3)

Here, Y_b is the observed value of region b; Lag_a is the spatial lag value of region a; n is the total number of counties; and the spatial weight matrix W_ab represents the spatial relationship between the domain of region a and region b [24], which is defined by the adjacency principle, i.e., the neighboring value of the region is 1; otherwise, it is 0.

To statistically verify whether the spatial spillover effect is significant—that is, whether changes in construction land redundancy are independent of each other or correlated with the neighborhood state types—the following formula was applied for testing:

$$Q_b=-2\log \left\{{\displaystyle \prod _{l=1}^k{\displaystyle \prod _{i=1}^k{\displaystyle \prod _{j=1}^k{\left[\frac{m_{ij}}{m_{ij}\left(S\right)}\right]}^{n_{ij}\left(S\right)}}}}\right\}$$

(4)

Here, k represents the number of redundancy types (k = 4); m_ij denotes the global Markov transfer probability; m_ij (S) denotes the probability of spatial Markov transfer with neighborhood type S; n_ij (S) indicates the number of counties with spatial Markov transfer with neighborhood state type S; and Q_b obeys a chi-square distribution with k (k − 1)² = 36 degrees of freedom.

2.2.3. Indicators

To identify the key factors influencing the redundancy of construction land in counties and explore the intrinsic mechanisms of spatial evolution, this study draws on the relevant literature [46,47,48,49], taking into account the location conditions, endowment foundation, economic level, and facilities and services of the county units (see

Table 1. Selection and calculation of the indicators influencing redundancies.

Dimension	Indicators	Calculation
Location	Distance from regional centers (SC)	Calculation of the natural logarithm of the minimum distance from Beijing and Tianjin
	Distance from the prefecture-level city to which it belongs (SD)	Calculation of the natural logarithm of the distance from the prefecture-level city to which it belongs
Endowment base	Whether it is in mountainous area (ST)	0: Plain; 1: Mountainous area
	Population Size (POP)	Population size of county units, normalized
Economic development	Economic development of the prefecture-level city to which it belongs	GDP per capita of the prefecture-level city to which it belongs, normalized
	Economic development of the county (EC)	GDP per capita of the county, normalized
	location quotient of the secondary industry (ESE)	ESE = (Output value of the secondary industry of the county/GDP of the county)/(Output value of the secondary industry of the region/GDP of the region)
	location quotient of the tertiary industry (ETE)	ETE = (Output value of the tertiary industry of the county/GDP of the county)/(Output value of the tertiary industry of the region/GDP of the region)
Public facilities	Educational service	Teacher–student ratio, normalized
	Medical care service	Number of beds in health institutions per 10,000 population, normalized

). Ordinary Least Squares (OLSs) regression analysis is then employed to examine the differences in the primary influencing factors across different periods.

From the perspective of location, this study selects the shortest distance from the geometry center of the county to the geometry center of Beijing or Tianjin, as well as the distance to the geometry center of the local prefecture-level city as explanatory variables. It is hypothesized that Beijing and Tianjin, as central cities within the urban agglomeration, and the prefecture-level cities, as regional central cities, exert functional spillover effects on surrounding areas. The shorter the distance, the greater the intensity of the spatial radiation received, which in turn influences land use efficiency by affecting land functions and structure.

From the perspective of the endowment base, this study considers the topographical conditions and population size of county-level units as explanatory variables. Topographical conditions are assessed based on whether the county is mountainous, as mountainous counties typically face limitations in external transportation accessibility and have complex terrain, which hinders development and construction, thereby restricting the allocation of incremental construction land. Population size is closely related to the functional capacity and centrality of a county. Additionally, there are significant differences in the ability to acquire incremental construction land and the marginal benefits of expansion based on population size.

From the perspective of economic levels, this study incorporates the per capita GDP of the prefecture-level city and the county and the location entropy of the secondary and tertiary industries as explanatory variables. The per capita GDP of the prefecture-level city reflects the economic development level in the BTH region. The higher the regional development level, the more development opportunities the county-level unit is likely to receive, thus enhancing its development quality. The per capita GDP of the county, along with the location entropy of the secondary and tertiary industries, indicates the competitive advantages of local industrial development. These variables are used to examine the interaction between economic development level and land use efficiency.

For the public facilities, the teacher–student ratio of schools and the number of beds per capita in hospitals are selected to characterize the level of facilities and services, to test whether a high level of services and facilities promote the improvement in land use efficiency.

2.3. Data Sources

Land use data in the study area were derived from the China Multi-Period Remote Sensing Land Use Monitoring Dataset (CNLUCC), produced by the Institute of Geomatics Science and Natural Resources Research, Chinese Academy of Sciences, with a resolution of 30 m (http://www.resdc.cn (accessed on 1 November 2023)), and we extracted the data of land use at five time points: 2000, 2005, 2010, 2015 and 2020. Using the software of Arcgis 10.8 and the administrative boundaries, we analyzed the changes in construction land. The statistical data, including the population scale, employee numbers, GDP, public budget revenue, total deposits of urban and rural residents, and public facilities in 2000, 2005, 2010, 2015, and 2020 were obtained from the Beijing Statistical Yearbook, Tianjin Statistical Yearbook, Hebei Statistical Yearbook, China County Statistical Yearbook, and the Statistical Communique on National Economic and Social Development of the counties.

3. Results

3.1. The Redundancies of the Counties from 2000 to 2020

The SBM-DEA model is utilized to measure the construction land redundancy across counties in the BTH region at five time points from 2000 to 2020, enabling the analysis of the spatio-temporal evolution of construction land redundancy (see Figure 4 and Figure 5).

Between 2000 and 2020, the redundancy level in the county-level areas of the BTH region remained relatively high. At each research time point, redundancy exceeded 34%, indicating that, under ideal conditions, the proportion of urban and rural construction land in county-level units that could be reduced was greater than 34% (see Figure 6).

During the study period, the redundancy degree underwent an initial increase followed by a subsequent decrease. From 2000 to 2005, redundancy rose rapidly from 37.1% to 47.1%, then gradually decreased to 45.9% by 2010. After 2015, the redundancy of construction land remained relatively stable. These changes suggest that, following the rapid construction land expansion in the early 2000s, county-level areas in the BTH region have shifted toward a more intensive development model. Additionally, the gap in redundancy between county-level units has gradually narrowed, indicating that land use efficiency is moving toward a more balanced development stage.

Based on the zoning scheme in the document “The county-level territorial spatial Planning guidelines in Hebei Province”, the province is divided into four sub-regions: Beijing and Tianjin surrounding area, Northwest Hebei, Central and Southern Hebei, and Coastal area. In this research, we incorporate the counties and districts in Beijing and Tianjin into the zoning scheme and analyze the redundancies of construction land by the sub-regions (see Figure 7). The results exhibit the different features in the four subregions: high redundancy in northwest Hebei, low redundancy in Central and Southern Hebei and surrounding Beijing-Tianjin, and considerable variation in coastal areas. However, due to different factors driving redundancy at various time points, regional redundancy levels also display fluctuation trends (see Figure 8).

The high redundancy in Northwest Hebei is primarily characterized by low output, which can be attributed to relatively low socioeconomic development, resulting in low construction land productivity. Throughout the study period, the average redundancy level in county-level units in this region remained high, though it showed a consistent downward trend, decreasing from 64.79% in 2000 to 47.70% in 2020. The temporary decline in redundancy observed in 2005 was mainly due to passive reduction, as redundancies in other regions temporarily increased. As land use efficiency improved in other areas, redundancy levels in Northwest Hebei rose again.

In the areas surrounding Beijing and Tianjin, construction land is primarily characterized by low redundancies with balanced development, indicating that land supply and socio-economic development are generally well-coordinated. Throughout the study period, redundancies in this region remained relatively low with minimal fluctuation, reflecting stable and high land use efficiency. From 2000 to 2020, the lowest redundancy level occurred in 2015, at 17.86%, while the highest was recorded in 2010, at 33.67%.

In the Central and Southern Hebei region, the redundancy level exhibited unstable fluctuations, with temporary periods of high redundancy, followed by a return to lower levels as socio-economic progress continued. These fluctuations were primarily due to short-term excess land supply. The trend in redundancy changes shows that between 2000 and 2010, redundancy increased from 25.58% to 51.37%, indicating that during this period, construction land supply exceeded demand, and the economic growth was mainly driven by mass construction land. From 2010 to 2020, redundancy in this region gradually decreased, as output efficiency gradually offset the redundancy created by construction land expansion, reflecting the positive effects of land-based urbanization in driving population urbanization.

Redundancy levels in coastal areas vary significantly, with some regions exhibiting extremely high redundancy, a typical example of high-supply redundancy caused by ample construction land supply. Coastal areas have long been strategic focal points for Hebei Province, leading to a substantial allocation of construction land quotas. However, these quotas far exceeded actual demand, resulting in persistently high levels of land use redundancy. In 2005, the average redundancy in these areas peaked at 86.04%, and although it has declined since then, it remained relatively high at 55.36% in 2020. Two representative cases in this region are Caofeidian District and Huanghua City, both located in Hebei’s strategic zones: the Caofeidian New Area and the Bohai New Area. Due to excessive construction land quota allocation, their redundancy levels before 2010 were the highest in the BTH region, reflecting a severe oversupply of construction land. With the shift in development modes, Caofeidian’s redundancy level underwent a significant transition, and its land use efficiency reached the efficiency frontier. In contrast, while Huanghua’s redundancy level has declined, it remains the highest in the region, resulting in persistent long-term redundancy.

3.2. The Spatial Evolution of the Redundancies

The results of the transition probability matrix reveal the following characteristics (see

Table 2. The result of the Markov transfer matrix and spatial Markov transfer matrix.

Spatial Type	Type	1	2	3	4	n
Global	1	69.01%	21.83%	6.34%	2.82%	142
	2	19.70%	50.76%	22.73%	6.82%	132
	3	7.46%	23.88%	45.52%	23.13%	134
	4	3.29%	5.26%	25.66%	65.79%	152
Neighborhood (k = 1)	1	46.15%	23.08%	15.38%	15.38%	13
	2	0.00%	100.00%	0.00%	0.00%	3
	3	0.00%	33.33%	33.33%	33.33%	3
	4	0.00%	50.00%	0.00%	50.00%	2
Neighborhood (k = 2)	1	72.31%	21.54%	4.62%	1.54%	65
	2	25.93%	42.59%	24.07%	7.41%	54
	3	6.12%	26.53%	40.82%	26.53%	49
	4	6.06%	9.09%	33.33%	51.52%	33
Neighborhood (k = 3)	1	66.67%	25.49%	5.88%	1.96%	51
	2	12.73%	50.91%	27.27%	9.09%	55
	3	7.41%	16.67%	55.56%	20.37%	54
	4	3.08%	4.62%	24.62%	67.69%	65
Neighborhood (k = 4)	1	84.62%	7.69%	7.69%	0.00%	13
	2	25.00%	65.00%	10.00%	0.00%	20
	3	10.71%	32.14%	35.71%	21.43%	28
	4	1.92%	1.92%	23.08%	73.08%	52

). First, the probabilities along the diagonal are higher than those off the diagonal, indicating that counties with different redundancy levels are more likely to maintain their current status in the next stage. This suggests the presence of a “club convergence” phenomenon in redundancy levels. Particularly at the extremes—counties with either low or high redundancy—the likelihood of maintaining their current state is significantly higher, reaching 69.01% and 65.79%, respectively, resulting in spatial lock-in effects. In contrast, counties with medium redundancy levels, such as medium-low or medium-high redundancy, have a relatively lower probability of remaining stable and are more likely to transition to neighboring states. The probabilities of upward and downward transitions are roughly equivalent.

Second, the likelihood of leapfrog transitions in redundancy levels between consecutive time periods is low, with the probability of significant upward or downward shifts being less than 8%. Under the influence of major strategic initiatives, strategic focal areas may initially experience high redundancy at the beginning of construction. As land is developed and land productivity increases rapidly, redundancy levels decrease quickly, with the potential for leapfrog transitions, as observed in Caofeidian. However, when land allocation far exceeds actual demand, long-term redundancy may occur, as seen in Huanghua City.

The spatial Markov transition matrix was used to compare the conditional transition probabilities of county-level units under different neighborhood conditions (i.e., k = 1,2,3, Table 2). To mitigate the disturbance effects caused by major regional strategies, units designated as provincial-level new areas during the study period were excluded. Specifically, Caofeidian District (Caofeidian New Area) and Huanghua City (Bohai New Area) were removed from the analysis. The remaining 142 county-level units were analyzed using the spatial Markov framework, and the results are presented in

Table 3. Influencing factors affecting the degree of redundancy in county-level units.

	Influencing Factors	2000		2005		2010		2015		2020
		Coefficient	VIF	Coefficient	VIF	Coefficient	VIF	Coefficient	VIF	Coefficient	VIF
Location	SC	−0.073	1.379	0.026	1.651	0.024	1.353	0.155 *	1.489	−0.231 **	1.737
	SD	−0.01	1.239	0.039	1.212	0.008	1.153	0.000	1.134	0.017	1.124
Endowment base	ST	0.469 ***	1.487	0.117	1.165	0.18 **	1.351	0.167 **	1.253	0.13	1.353
	POP	0.282 ***	1.339	0.236 ***	1.489	0.448 ***	1.323	0.404 ***	1.161	0.052	1.38
Economic development	EM	−0.285	1.998	−0.18 **	1.609	−0.134 *	1.905	−0.078	2.115	0.14	2.251
	EC	−0.133 *	2.337	0.068	2.085	−0.195 **	2.631	−0.249 **	2.398	−0.287 ***	2.297
	ESE	−0.433	1.675	−0.541 ***	2.094	−0.541 ***	2.451	−0.341 ***	2.37	−0.348 **	2.789
	ETE	−0.289	1.995	−0.171 *	1.479	−0.359 ***	2.174	−0.289 ***	2.166	−0.431 ***	2.266
Public facilities	FE	0.157 **	2.161	0.05	1.651	0.100	1.43	0.260 ***	1.271	0.098	1.234
	FM	−0.14	1.615	−0.153	1.212	−0.182 **	1.481	−0.062	1.35	−0.114	1.389
Modified R2		0.491		0.306		0.473		0.356		0.152
F-value		14.3		7.083		13.2		8.638		3.475

*** indicates significance at the 99% confidence level; ** indicates significance at the 95% confidence level; * indicates significance at the 90% confidence level.

.

The result of the spatial spillover effect shows that Q_b = 29.6 < X² (36) = 47.212. Therefore, under a 90% confidence level, the original hypothesis cannot be rejected, indicating that the degree of neighborhood redundancy does not affect the changes in construction land redundancy in adjacent units. The spatial Markov analysis indicates that county-level units in the Beijing-Tianjin-Hebei region lack connections and functional spillovers. Low-efficiency land units are unable to obtain development factors, advanced technologies, and institutional support from neighboring high-efficiency land units, resulting in an insignificant regional synergy effect.

3.3. The Influencing Factors of the Redundancies of the Counties

A multiple regression model is employed using the Ordinary Least Squares (OLSs) method to estimate the parameters and identify the key factors influencing construction land redundancy in county-level units across five historical periods. The statistical analysis results (see Table 3) indicate no multicollinearity issues among the selected indicators during any period, and the model demonstrated reasonable explanatory power for redundancy levels.

The results indicate that the factors influencing construction land redundancy vary across different stages of urbanization. First, location conditions did not significantly affect the construction land redundancy of neighboring county-level units in 2000–2010. After 2015, the interconnection between regional central cities and the redundancy levels of neighboring counties became more pronounced.

During the period of 2000–2010, central cities were primarily in a phase of resource aggregation, which limited their influence on surrounding county-level units. As the functional and economic levels of central cities improved, their “trickle-down” effects began to stimulate development in neighboring counties. Furthermore, local governments increasingly prioritized strategic development in areas surrounding central cities by allocating more land resources to these regions.

In contrast, prefecture-level cities did not exhibit a significant impact on the redundancy levels of neighboring county units throughout the study period. This suggests that prefecture-level cities have yet to establish an effective radiative influence on the development of adjacent county-level areas.

Second, county development conditions exert a significant influence on construction land redundancy. Among these factors, topographic conditions have a particularly notable impact. Counties located in mountainous areas face disadvantages such as poor transportation accessibility and challenging land-use conditions, which hinder economic development and result in lower land output, leading to higher land redundancy. In 2005, the influence of topographic conditions on land redundancy was insignificant, primarily due to the widespread and large-scale allocation of construction land across regions. During this period, the counties in the plains developed more construction land, leading to high redundancies with an oversupply of construction land. Meanwhile, the redundancies of mountainous counties were mainly caused by the low outcome. Despite the causes being different, the redundancies did not demonstrate a discernible spatial differentiation pattern. However, since 2010, the incremental construction land has been controlled strictly. The redundancies of plain counties declined quickly with the rapid development of social and economic development, while the counties in mountainous areas still had high redundancies due to the low social and economic development levels. Therefore, the impact of topographic conditions regained significance in 2010.

Population size also demonstrated a significant relationship with construction land redundancy, except in 2020, when its impact was statistically insignificant. In other years, population size was significant at the 99% confidence level. This finding reinforces the notion that counties with larger populations tend to secure more construction land quotas. However, it also highlights that in many populous counties within the BTH region, socioeconomic output efficiency has not kept pace with the construction land increase. As a result, there is a positive correlation between population size and land redundancy: the larger the population size is, the higher the redundancy level is.

Third, economic development levels play a critical role in determining the construction land redundancies. The economic development level of the prefectural city had a negative impact on land redundancy in 2005 and 2010, indicating that during this period, higher regional economic levels provided more opportunities for counties, thereby enhancing productivity.

The economic level of the county, along with the location quotient of the secondary and tertiary industries, also showed a negative correlation with redundancies. Moreover, the absolute values of the standardized coefficients were relatively high, suggesting that higher economic development levels, particularly advantages in secondary industry development, significantly reduce construction land redundancy and improve land output efficiency.

Fourth, service facilities have a relatively minor impact on county-level construction land redundancy; however, in certain periods, higher service levels positively influence land use efficiency. The level of educational facilities showed a negative correlation with construction land redundancy in 2000 and 2015, whereas the level of healthcare facilities exhibited a positive correlation in 2010. These varying patterns suggest distinct underlying mechanisms. One possible explanation is that the educational resource indicator—the student–teacher ratio—is closely associated with population outflow at the county level. A high student–teacher ratio often reflects lower economic development levels or a lack of local employment opportunities, which in turn are linked to lower land output efficiency. In contrast, counties with better healthcare facilities tend to experience higher levels of social and economic development, which enhances land use efficiency. Although the direct influence of service facilities on redundancy is limited, improved social service infrastructure can act as a positive incentive for socio-economic development and population retention, thereby indirectly improving construction land efficiency.

4. Discussion

The construction land redundancy at the county level mainly arises from the mismatch between the supply and demand for construction land, a challenge exacerbated by the transition from an expansionary to an intensive development model. Addressing the imbalance requires fundamental adjustments in both supply and demand.

On the supply side, it is crucial to establish an allocation model for construction land quotas based on “real demand”. In regions with high supply-based redundancy, targeted interventions are necessary in order to reduce construction land through ecological remediation of inefficient land. In areas with low-output redundancy, it is essential to strictly control the incremental land quotas, while releasing land potential and promoting spatial integration.

On the demand side, leveraging resource endowments is key to establishing competitive advantages in the secondary and tertiary industries, which will lower redundancies. The government should facilitate stronger synergistic linkages between high-level cities, such as Beijing, Tianjin, and the surrounding prefecture-level cities with county-level units, to enhance the spillover effects from the regional centers.

A land quota trading system, like the Transfer of Development Rights (TDR) system in the US and some parts of Europe, should be established, with the government setting transaction rules while allowing the market to play a determinant role in resource allocation. Competitive bidding can direct land quotas to regions with low redundancies of construction land. Underdeveloped regions with high redundancy can benefit by selling surplus land quotas, effectively converting land resources into capital.

It is essential to strengthen the driving role of central cities and economically advanced counties to enhance the synergistic improvement in land use efficiency at the county level. In alignment with an open regional economy, neighboring areas can benefit from development opportunities, advanced technologies, and institutional frameworks. Knowledge transfer and imitation will further facilitate the improvement in land use efficiency across regions.

Strengthened regulation and dynamic assessments ensure the optimal utilization of incremental construction land. A periodic evaluation and monitoring system for redundancy of construction land should be established to guide land allocation and performance assessments. Counties with high redundancy of construction land should have their annual incremental land quotas reduced. In key strategic areas, redundancy thresholds should be established to dynamically adjust the land supply scale, which could trigger a halt in the supply of new construction land when redundancy exceeds the specific threshold. This phased and proportional system will help prevent the accumulation of long-term land use redundancy.

Based on the land redundancy characteristics of the subregions in the BTH region, the study proposes targeted strategies to improve efficiency in the following areas:

(1)

Northwest Hebei

Leveraging the region’s ecological endowments, the focus should be on developing ecological tourism and green industries to create a unique regional development advantage. Industries such as tourism, sports industries, and new energy industries (wind and solar energy industries) have become key economic drivers. Strict controls should be placed on the expansion of newly designated projects, and the government should encourage the reduction in construction land from ecological spaces;

(2)

Beijing-Tianjin Surrounding Area

Continued promotion of functional and technological spillover from Beijing and Tianjin to peripheral county regions is essential for enhancing the industrial development in these counties. Against the backdrop of the capital’s functional decentralization, labor-intensive industries should be relocated to surrounding counties. These counties should improve infrastructure to attract industries and populations. Construction land supply should be project-oriented, focusing on timing and scale to prevent long-term land redundancy;

(3)

Central and Southern Hebei

Promoting industrialization to enhance the land output in counties is key. The government should foster local industries, particularly in manufacturing, where there is a solid foundation. However, it should be noted that there are a number of industrial development zones with considerable idle land. New industries should be encouraged to utilize existing construction land to decrease the low efficient land. Prefecture-level cities such as Shijiazhang, Xingtai, Hengshui, and Cangzhou should share advanced technologies and systems with the counties. Meanwhile, it is important to avoid chaotic competition resulting from selling land at low prices to non-viable enterprises;

(4)

Coastal Area

The coastal area is often a focal point for both provincial and local government, which has led to dispersed development. Recently, compiled spatial plans have strictly controlled construction space, altering the land use landscape. The construction quotas must also be regulated and a moratorium on new quotas in regions with high redundancy ought to be imposed. For already developed but inefficient areas, renewal initiatives or ecological decommissioning measures should be implemented to optimize land use structures and curb land expansion.

5. Conclusions

The emergence of construction land redundancy reflects a paradigm shift in urban development from extensive expansion to intensive utilization. Reducing land redundancy and enhancing land efficiency are critical for China’s pursuit of high-quality development. This study investigates the spatio-temporal evolution of construction land redundancy across counties in the BTH region from 2000 to 2020, identifying overarching trends, regional differentiation patterns, and key influencing factors. Our key conclusions are as follows:

(1)

Construction land redundancy remains relatively high at the county level in the BTH region, accompanied by issues of extensive land use. Notably, the characteristics of types of redundancy vary across subregions;

(2)

Spatial redundancy among county-level units exhibits a “club effect” and spatial lock-in, showing insufficient intra-county coordination to enhance construction land efficiency across counties;

(3)

The county’s endowment base and economic development level of the county have consistently played a pivotal role in determining the redundancy of construction land. After 2015, additional factors, location, and public facilities have had impacts on construction land redundancy;

(4)

In order to enhance land use efficiency, the supply–demand relationships should be optimized by establishing robust supporting mechanisms and setting different land use strategies across subregions.