The Determinants of Commercial Land Leases in the Non-Central Districts of a Large City in China: Data Analysis from the Government–Market Perspective

Abstract

Based on the data of the non-central districts in Shanghai, this paper investigates the determinants of the commercial land leases of district governments from the government–market perspective and how these determinants affect the price and area of commercial land leasing. A kernel density analysis is used to analyze the agglomeration degree and density distribution of commercial land leasing. The variables are considered as the factors impacting commercial land leases based on a literature review and land development in Shanghai. The mathematical models used for multiple linear regression for the leased price and area of the influencing factors of commercial land leases from the perspective of the government and market are proposed. The results show that Shanghai’s multi-center development strategy aims to optimize the city’s commercial layout by developing the key areas of non-central districts. The construction area and plot ratio of land; the distances from the land to the city center, district center, airports, the nearest middle schools, the nearest park, and the nearest industrial zone; and the quantity of subway stations and highways affect commercial land leases. Policies are proposed to improve commercial land lease efficiency, make more suitable land planning strategies, and optimize urban spatial structures.

1. Introduction

Commerce is very important for the economic development of a country. Commercial land as an urban function is helpful for promoting economic development and enhancing urban vitality. The buildings of retail, catering, office, hotel, and other service industries can be found on this land, which not only directly creates a large number of employment opportunities and tax revenue but also gathers people, logistics, and capital flows through commercial complexes, professional markets, and other formats, forming a regional economic vitality center. The scientific planning of commercial land layout can be used to optimize urban spatial structure, improve land resource utilization efficiency, and meet residents’ consumption and business activity needs. In addition, commercial land also promotes the integrated development of industry and city through organic integration with transportation hubs and residential areas, which is meaningful for enhancing the attractiveness and competitiveness of a city. Thus, it is essential to study the problem of commercial land leases.

The government plays an essential regulatory role in the lease of commercial land, and its decisions directly affect the quality and spatial structure of urban economic development. As a monopolistic supplier of land resources, the government can not only optimize the layout of commercial space and avoid resource waste caused by homogeneous competition by formulating differentiated lease strategies, but it can also guide industrial upgrading. The complete regulation of government land supply and price mechanisms can significantly improve the efficiency of land resource allocation. Therefore, figuring out how the government can balance economic and social benefits while pursuing land fiscal revenue is very important for sustainable urban development.

In the Chinese land lease market, government actions play a leading role and exhibit governance characteristics of multi-level government interaction [1]. The central and local governments have formed a clear division of functions in the housing and land market system [2]. The central government is responsible for designing top-level systems and promoting market-oriented reforms of land factors through policy tools [3]. As the main body of policy implementation, the core goal of local governments is to maximize the long-term benefits of land resources [4]. This institutional arrangement has given rise to specific incentive mechanisms [5,6]. Due to the central government’s delegation of land use rights which transfers authority to local governments, an objective competition pattern has been formed among local governments around investment attractiveness. The mechanism for distributing land lease profits under the fiscal decentralization system further strengthens the tendency of local governments to compete for capital factors through land policy tools [7,8]. In this context, the land lease decisions of local governments must take into account a dual goal: to optimize the efficiency of land resource allocation and systematically consider various influencing factors, such as location conditions, industrial demand, etc., that affect the realization of land value.

The process of land leasing is influenced by multidimensional factors, mainly including market factors such as land price fluctuations [9,10] and GDP growth levels [10]; institutional factors, involving government management systems [11,12,13] and statutory land lease methods, such as bidding, listing, etc. [7,14,15]; and political factors [16], especially the competition for promotion among local officials. These factors have dynamic impacts on the land and real estate markets. Among them, the decision-making behavior of the government and local officials, especially related to the adjustment of land lease strategies based on political promotion incentives, not only directly affects short-term urban economic growth but also has a profound impact on the long-term urbanization process [17,18]. It is worth noting that the transformation of land use patterns, such as the urbanization of agricultural land [19], not only promotes economic development [1,15] but is also constrained by land supply [20]. The close correlation between GDP growth and urban economic development further strengthens the two-way interaction between land leasing and regional economy [10,15]. The interaction of these diverse influencing factors provides an important perspective for understanding the intrinsic relationship between land resource allocation and economic development [7,15,16].

Accessibility-related factors have a systematic impact on the land market and urban development, and their mechanisms are mainly reflected in the following aspects. Firstly, in terms of transportation infrastructure, public transportation networks [21], especially urban rail transit systems [22,23], not only promote land development intensity by enhancing regional connectivity [24,25] but also have a significant impact on the premium of commercial housing prices, e.g., the premium rate around subway stations is significantly increased [26], which prompts local governments to use subway extensions as an important tool for enhancing suburban land value. Secondly, in terms of location characteristics, the distance from key nodes such as airports [27] and city centers is directly capitalized as land price differences. Thirdly, in terms of supporting public services, high-quality educational resources, such as key high schools, universities [28,29], and environmental comfort, such as the green coverage rate [21], are utilized through resident location selection behavior [30], which encourages local governments to improve the supply of public goods to enhance the attractiveness of their jurisdiction, thereby affecting the demand structure and lease strategies of the land market.

From the existing literature, it can be seen that the government plays an important role in China’s land lease behavior, and the influencing factors of land leasing are diverse, such as political influence, land lease methods, economic development, accessibility, etc. At present, there are relatively few studies on the government’s lease of commercial land from the government and market perspective. Few studies have conducted comprehensive analyses on the influencing factors of land leasing from multiple perspectives, with most research focusing on one aspect. However, for megacities in China, the lease of commercial land is conducted by district governments; thus, it is necessary to study the government’s lease of commercial land based on the data of districts. Also, there are multiple aspects of the influence factors to be considered; thus, a comprehensive analysis on the influencing factors of land leasing is essential.

As the most representative international metropolis in China, Shanghai has undergone the complete process of land system reform, from free allocation in the planned economy period to market-oriented transfer and then to the current stage of refined management development. Secondly, as a city regarded as an economic center, Shanghai has a mature commercial land market and a well-established mechanism for land price formation. The core area’s commercial land prices have long been among the top in the country, providing a reference for studying the value patterns of commercial land. Furthermore, Shanghai has continuously innovated policy tools in the lease of commercial land, such as mixed use land models, which have largely been used for optimizing land resource allocation. In addition, Shanghai’s high participation of foreign investment and deep marketization of commercial land market characteristics make its research results have broader reference value. Meanwhile, the land use rights in large cities, such as Shanghai, belong to the district government. Thus, it is very necessary to study the commercial land lease behavior of district governments.

Based on mathematical models and data analysis, this paper investigates the determinants of the commercial land leasing of district governments from the perspective of the government and market in non-central districts and how these determinants affect the leased price and area of commercial land. A kernel density analysis is used to analyze the agglomeration degree and density distribution of commercial land leases. The variables are considered as the factors impacting commercial land leases based on a literature review and land development in Shanghai. The mathematical models used for multiple linear regression for the leased price and area of the influencing factors of commercial land leases from the perspective of the government and market are proposed. Policies are proposed to improve commercial land lease efficiency, make more suitable land planning strategies, and optimize urban spatial structures.

2. Methodology

2.1. Kernel Density Analysis

Kernel density analysis is a non-parametric statistical method used to estimate the probability density function of a random variable, which reveals the spatial or numerical distribution characteristics of data by smoothing discrete sample points. In this paper, this method is applied to analyze the agglomeration degree and density distribution of commercial land lease. The formula is

$$f(x)={\displaystyle \frac{1}{n{h}^2}}{\displaystyle \sum _{i=1}^{n}K\left(\frac{d_i}{h}\right)},$$

(1)

where $f(x)$ is the density estimation at point x, n is the number of commercial land leases, $d_i$ is the distance from x to the i-th commercial land lease, K is a kernel function which is a Gaussian function used to calculate distance weights, and h is the bandwidth.

2.2. Variables

The district government must carefully evaluate various factors when supplying commercial land to ensure the best outcomes and sustainable growth. However, many governments struggle to balance these factors in practice. This study examines what influences the needs and provision of land parcels for district governments and the market in terms of commercial land leases. Based on existing research and policies, this paper puts forward three key assumptions.

2.2.1. Land Variables

Land variables include area, plot ratio, and lease method. The land area serves as the fundamental attribute that directly affects both the transaction process and developer returns. The plot ratio plays a crucial role in determining land value and influencing residential living quality. Additionally, the lease method, whether through tenders, auctions, or public listings, significantly impacts the land transaction process and final pricing outcomes [16].

2.2.2. District-Development Variables

District-development variables mainly include GDP and the district mayor’s term. GDP growth influences land supply [10], while land lease methods also affect economic growth [15]. Since land lease revenue contributes to each district’s GDP and reflects regional economic development, this paper considers GDP as a key factor in commercial land leases. Additionally, political incentives and local officials’ actions significantly impact regional economies [31,32]. In China, GDP growth is closely tied to officials’ career advancement, meaning that their political prospects may affect land transfer policies. In Shanghai, district mayors hold the highest authority in the government. Therefore, this study considers the district mayor’s term length as a potential factor influencing land leases.

2.2.3. Accessibility Variables

Accessibility variables primarily include the distance from the land parcel to the city center, district center, airports, Shanghai Railway Station, the nearest subway, the nearest expressway, the nearest university, the nearest park, and the nearest industrial zone. Transportation accessibility significantly influences land transactions. Subway stations enhance commercial vitality and urban development while positively impacting commercial land values [25,26]. Major transport hubs like airports and railway stations similarly affect land prices, with studies showing shorter distances to highway interchanges correlating with higher rental values [33,34]. Educational resources also play a key role, as high accessibility to prestigious universities or high-quality schools elevates land prices [29]. Green spaces demonstrate notable externalities, where closer access to parks increases property values [35,36]. The industrial zone significantly impacts urban land use planning and government decision-making regarding land leases [37]. The spatial location of land determines China’s land pricing mechanisms [38] and consequently influences governments’ land transaction behaviors [10].

The three proposed categories of variables cover the key factors affecting commercial land leases from different perspectives, i.e., land, district development, and accessibility aspects.

2.3. Mathematical Models and Regression

The land lease price and lease area correspond to the bidirectional interaction between the reflection of market and district government actions in needing and providing land parcels. Under the public ownership of land in China, governments are the sole suppliers of construction land, and the area leased directly reflects the will of administrative regulation. The price reflects the developer’s, i.e., demander’s, evaluation of the land value.

The mathematical model of the factors affecting commercial land lease prices from the market perspective is established as

$$\ln L{P}_{it}=a+{\displaystyle \sum _{m=1}^M{\alpha }_m{A}_{mit}}+{\displaystyle \sum _{j=1}^J{\beta }_j{D}_{jit}}+{\displaystyle \sum _{k=1}^K{\gamma }_k{L}_{kit}}+DU{M}_1+DU{M}_2+u_{it},$$

(2)

where $L{P}_{it}$ is the price of the commercial land lease in the district $i$ in the year $t$, $A_{mit}$ is the land variable $A_m$ of the district $i$ in the year $t$, $D_{jit}$ is the district development variable $D_j$ of the district $i$ in the year $t$, $L_{kit}$ is the accessibility variable $L_k$ of the district $i$ in the year $t$, $M$ is the number of $A_{mit}$, $J$ is the number of $D_{jit}$, $K$ is the number of $L_{kit}$, $DU{M}_1$ is the dummy variable of year, $DU{M}_2$ is the dummy variable of district, $a$ is a constant, and $u_{it}$ is the error. The detailed variable definitions are shown in

Table 1. Variable definitions and summary of statistics.

Category	Variable	Description	Observation	Mean	Standard Deviation	VIF
	$\ln L{P}_{it}$	Land lease price (CNY)	672	3.03 $\times$ 108	5.92 × 108
$A_{1it}$	$\ln L{A}_{it}$	Construction area of leased land (m2)	672	30,715.83	40,173.51	1.22
$A_{2it}$	$\ln (FA{R}_{it})$	Plot ratio of land	672	2	1.25	1.52
$A_{3it}$	$MO{D}_{it}$	Land lease method: =1 is listing; =2 is tender; =3 is auction.	672	1.04	0.19	1.2
$D_{1it}$	$\ln (GD{P}_{it})$	GDP of district (CNY)	672	1.99 $\times$ 1011	2.18 × 1011	7.47
$D_{2it}$	$D{M}_{it}$	District mayor’s term (Year)	672	3.31	1.77	2.62
$L_{1it}$	$\ln (SP{S}_{it})$	Distance to city center (m)	672	26,570.8	14,000.46	6.1
$L_{2it}$	$\ln (GO{V}_{it})$	Distance to district center (m)	672	9778.11	6791.74	1.66
$L_{3it}$	$\ln (SU{B}_{it})$	Distance to the nearest subway station (m)	672	9072.62	10,824	3.98
$L_{4it}$	$\ln (HIG{H}_{it})$	Distance to the nearest expressway entrance and exit (m)	672	2286.81	1829.6	1.71
$L_{5it}$	$\ln (UN{I}_{it})$	Distance to the nearest university (m)	672	9756.85	6602.44	2.38
$L_{6it}$	$\ln (PA{R}_{it})$	Distance to the nearest park (m)	672	5772.48	3351.05	1.38
$L_{7it}$	$\ln (I{P}_{it})$	Distance to the nearest industrial zone (m)	672	2759.44	3874.64	1.21
$L_{8it}$	$\ln (H{Q}_{it})$	Distance to Hongqiao Airport (m)	672	25,464.66	14,673.54	6.55
$L_{9it}$	$\ln (P{D}_{it})$	Distance to Pudong Airport (m)	672	44,764.2	17,275.03	4.46
$L_{10it}$	$\ln (R{S}_{it})$	Distance to Shanghai Railway Station (m)	672	26,908.05	14,445.51	9.87
$L_{11it}$	$\ln (M{S}_{it})$	Distance to the nearest key middle school (m)	672	11,449.51	11,132.21	4.85
	$\ln A_{it}$	Total area of leased land in a district (m2)	130	179,364.9	197,265.8
$H_{1it}$	$\ln (SP{S}_{it})$	Distance of district to the nearest city center (m)	130	16,842.95	11,741.03	4.48
$H_{2it}$	$\ln (UN{I}_{it})$	Distance of district to the nearest university (m)	130	3703.48	7358.98	1.45
$H_{3it}$	$\ln (M{S}_{it})$	Distance of district to the nearest key middle school (m)	130	4077.15	5799.49	1.84
$H_{4it}$	$\ln (S{N}_{it})$	Quantity of subway stations in a district	130	9.56	16.83	2.74
$H_{5it}$	$\ln (H{N}_{it})$	Quantity of expressway entrances and exits in a district	130	61.81	48.11	2.35

.

Equation (2) can be written as

$$\ln L{P}_{it}=a+\mathit{a}\mathit{A}+u_{it},$$

(3)

where

$$\mathit{a}=({\alpha }_1,{\alpha }_2,\dots ,{\alpha }_M,{\beta }_1,{\beta }_2,\dots ,{\beta }_J,{\gamma }_1,{\gamma }_2,\dots ,{\gamma }_K,1,1),$$

(4)

$$\mathit{A}={(A_{1it},A_{2it},\dots ,A_{Mit},D_{1it},D_{2it},\dots ,D_{Jit},L_{1it},L_{2it},\dots ,L_{Kit},DU{M}_1,DU{M}_2)}^{\mathrm{T}}.$$

(5)

Then, the value of mathematic expectation and variance is zero, i.e.,

$$\mathrm{E}(\ln L{P}_{it}-a-\mathit{a}\mathit{A})=0,$$

(6)

$$\mathrm{E}(A_j(\ln L{P}_{it}-a-\mathit{a}\mathit{A}))=0,$$

(7)

where $A_j$ is a component of the vector $\mathit{A}$.

For the given data sample, we can obtain the estimated value $\widehat{\mathit{a}}$ of the coefficient vector $\mathit{a}$ as

$$\widehat{\mathit{a}}=({\widehat{a}}_1,{\widehat{a}}_2,\dots ,{\widehat{a}}_M),$$

(8)

From Equations (6) and (7), we have

$${\displaystyle \frac{1}{N}}{\displaystyle \sum _{n=1}^N(\ln L{P}_{itn}-\widehat{a}-\widehat{\mathit{a}}{\mathit{A}}_n)}=0,$$

(9)

$${\displaystyle \frac{1}{N}}{\displaystyle \sum _{n=1}^N{A}_{jn}(\ln L{P}_{itn}-\widehat{a}-\widehat{\mathit{a}}{\mathit{A}}_{\mathit{n}})}=0,$$

(10)

where $N$ is the number of data sample.

Let

$$\overline{y}={\displaystyle \frac{1}{N}}{\displaystyle \sum _{n=1}^N\ln L{P}_{itn}},$$

(11)

from Equation (9), it is obvious that

$$\overline{y}=\widehat{a}+\widehat{\mathit{a}}{\overline{\mathit{A}}}_n.$$

(12)

Then,

$$\widehat{a}=\overline{y}-\widehat{\mathit{a}}{\overline{\mathit{A}}}_n,$$

(13)

where

$${\overline{\mathit{A}}}_n={\displaystyle \frac{1}{N}}{\displaystyle \sum _{n=1}^N{\mathit{A}}_n}.$$

(14)

Due to the arbitrariness of the number of data samples, it follows from Equation (10) that

$${\displaystyle \sum _{n=1}^N{A}_{jn}(\ln L{P}_{itn}-\widehat{a}-\widehat{\mathit{a}}{\mathit{A}}_n)}=0.$$

(15)

Substituting Equation (12) into Equation (15), we can obtain

$${\displaystyle \sum _{n=1}^N{A}_{jn}(\ln L{P}_{itn}-(\overline{y}-\widehat{\mathit{a}}{\overline{\mathit{A}}}_n)-\widehat{\mathit{a}}{\mathit{A}}_n)}=0;$$

(16)

then, we have

$${\displaystyle \sum _{n=1}^N{A}_{jn}(\ln L{P}_{itn}-\overline{y})}=\widehat{\mathit{a}}{\displaystyle \sum _{n=1}^N{A}_{jn}({\mathit{A}}_n-{\overline{\mathit{A}}}_n)}.$$

(17)

Solving Equation (17), we can obtain the estimated value $\widehat{\mathit{a}}$ as

$$\widehat{\mathit{a}}=\left[{\displaystyle \sum _{n=1}^N{A}_{jn}(\ln L{P}_{itn}-\overline{y})}\right]{\left[{\displaystyle \sum _{n=1}^N{A}_{jn}({\mathit{A}}_n-{\overline{\mathit{A}}}_n)}\right]}^{-1}.$$

(18)

The mathematical model of the factors affecting commercial land lease area from the government perspective established is

$$\ln L{A}_{it}=a+{\displaystyle \sum _{k=1}^K{\gamma }_k{H}_{kit}}+DU{M}_1+DU{M}_2+u_{it},$$

(19)

where $\ln A_{it}$ is the area of the commercial land lease in the district $i$ in the year $t$, $H_{kit}$ is the accessibility variable $H_k$ of the district $i$ in the year $t$, and $K$ is the number of $H_{kit}$. Table 1 presents the variable definitions.

The coefficient vector $\mathit{a}$ in Equation (19) can be obtained as

$$\widehat{\mathit{a}}=\left[{\displaystyle \sum _{n=1}^N{A}_{jn}(\ln L{A}_{itn}-\overline{y})}\right]{\left[{\displaystyle \sum _{n=1}^N{A}_{jn}({\mathit{A}}_n-{\overline{\mathit{A}}}_n)}\right]}^{-1},$$

(20)

where

$$\mathit{a}=({\gamma }_1,{\gamma }_2,\dots ,{\gamma }_K, 1, 1),$$

(21)

$$\mathit{A}={(H_{1it},H_{2it},\dots ,H_{Kit},DU{M}_1,DU{M}_2)}^{\mathrm{T}},$$

(22)

$$\overline{y}={\displaystyle \frac{1}{N}}{\displaystyle \sum _{n=1}^N\ln L{A}_{itn}}.$$

(23)

Similarly to the processes of Equations (2)–(18), we can obtain the estimated value of the coefficients in Equation (19).

The variance inflation factor (VIF) is used to test the multi-collinearity of the variables of Equations (2) and (3), and the results show that the VIFs of Equation (2) are less than 10, and those of Equation (3) are less than 5. The results of the VIF calculations are shown in Table 1.

3. Data Collection

This paper mainly studies commercial land leasing in non-central districts, including suburban districts and counties. According to Shanghai Overall Planning, the districts of Minhang, Baoshan, Jiading, Nanhui, Jinshan, Songjiang, Qingpu, and Pudong New Area are suburban districts, and Chongming District is a county. Based on the current situation of commercial land leases in various districts of Shanghai and the relevant data collected on commercial land leases, the land resources in central districts are scarce, the number of commercial land leases is relatively low, and the proportion of commercial land leases in suburban districts and counties is relatively high.

This paper collects data on districts from nine non-central districts in Shanghai from 2004 to 2015, which is the period that the real estate market of China developed rapidly. In recent years, the development of China’s real estate sector has been sluggish. This paper examines the period of rapid real estate development, aiming to assist local governments in understanding the influencing factors behind such growth in order to formulate better real estate development policies. The data of GDP and district mayor tenure are collected from Shanghai Statistical Yearbook and Shanghai District Statistical Yearbook for each district.

Data on commercial land in the non-central districts of Shanghai from 2004 to 2015, including the lease price, area, plot ratio, and lease method of commercial land, are collected from the website of the Shanghai Municipal Bureau of Planning and Natural Resources. Figure 1 shows the leases of commercial land in non-central districts from 2004 to 2015.

The latitude and longitude of location data are collected from Baidu Map. ArcGIS is used to calculate the distance from each land to the city center, district center, Shanghai Railway Station, Hongqiao Airport, Pudong Airport, the nearest subway station, the nearest expressway entrance and exit, the nearest university, the nearest park, and the nearest industrial zone. A total of 32 key middle schools, 32 major universities, 50 main parks, and 98 industrial zones are collected. Figure 2 shows the distribution of subway stations and expressway entrances and exits in 2004 and 2015. Figure 3 shows the distribution of key middle schools, major universities, and parks. Figure 4 shows the distribution of industrial zones. A summary of the statistics of the data is shown in Table 1.

4. Results and Discussion

4.1. Spatial Distribution and Characteristics of Commercial Land Leases

Based on the data collection, Figure 5 shows the spatial distribution of commercial land leases according to the construction area and transaction price.

From Figure 5a, the construction areas of commercial land leases in Qingpu, Songjiang, and Minhang District are relatively large. This is mainly due to Shanghai’s implementation of the multi-center urban development strategy, which promotes the relocation of population and industries to suburban districts. Minhang has become a business center relying on the Hongqiao hub, while Songjiang and Qingpu have benefited from the extension of rail transit and expressways, such as the Shanghai Hangzhou high-speed railway, rail transit Lines 9 and 17, new city construction, and population introduction. Songjiang New City was developed in 2001, and Qingpu New City accelerated its expansion after 2004. This is a key development area that requires a large amount of commercial land to improve urban functions. Large commercial complexes, such as Qingpu Outlets and Songjiang Wanda Plaza, and supporting service industries have been developed, while suburban land prices are relatively low, and land supply is sufficient, further promoting the large-scale lease of commercial land. Qingpu Outlets opened in 2006, and Wanda Plaza opened in 2013, becoming city-level commercial nodes.

From Figure 5b, the land transaction prices in Minhang, Pudong, and Baoshan District are higher, and these districts are closer to the central districts. This phenomenon is mainly due to the superior location conditions and policy factors. Pudong, as a national strategic new area, gathers finance technology industries in core areas such as Lujiazui and Zhangjiang; Minhang relies on the Hongqiao hub to develop into a business center, and mature communities enhance land value; Baoshan has integrated into the main city through rail transit extension and the transformation of Wusong Industrial Zone. The industrial agglomeration effect, seen in locations such as Pudong Free Trade Zone and Minhang Zizhu High tech Zone, of the three districts is significant. At the same time, due to the proximity of the three districts to the city center, the available land resources are limited, especially for commercial and office use. The fierce competition among developers has jointly increased land prices, making them significantly higher than those of other suburbs. In terms of policies, the comprehensive supporting reform in Pudong was implemented in 2005. Minhang and Baoshan were included in the expansion scope of the main urban area in the overall urban planning process, and the government and developers increased their capital investment.

Based on Equation (1) solved in ArcGIS, the result of the kernel density of commercial land leases is shown in Figure 6. As seen in Figure 6, the concentration of commercial land leases in Jiading, Pudong, and Minhang District is higher. Firstly, due to the strategic location advantage, Pudong relies on Lujiazui, Zhangjiang, and Minhang to benefit from the Hongqiao hub, and Jiading leverages the Shanghai–Nanjing development axis to form a core commercial agglomeration area. Secondly, driven by major infrastructure projects, the opening of rail transit lines, such as Lines 11 and 9, significantly enhances regional value, and the surrounding areas of stations have become commercial hotspots. The third aspect to consider is the agglomeration of industrial population, with strong commercial demand brought by industries such as the finance, automobile, and high-tech industries, which, coupled with rapid population growth, promote the centralized supply of commercial land.

The spatial distribution reflects the idea that Shanghai’s multi-center development strategy aims to optimize the city’s commercial layout by developing the key areas of non-central districts, and there are resource agglomeration effects in the core area of the urban expansion process.

4.2. Market Perspective

Based on the regression of the mathematical models of Equation (2) through Stata 12.0,

Table 2. The results of the model of Equation (2).

Variable	Coefficient	Robust Standard Error
$\ln (Are{a}_{it})$	0.9351 ***	0.02421
$\ln (FA{R}_{it})$	0.60599 ***	0.05256
$MO{D}_{it}$	−0.03491	0.11292
$\ln (GD{P}_{it})$	0.4358 **	0.2053
$D{M}_{it}$	−0.01122	0.01876
$\ln (SP{S}_{it})$	−1.08787 ***	0.36141
$\ln (GO{V}_{it})$	−0.06156 **	0.03271
$\ln (SU{B}_{it})$	−0.02663	0.0295
$\ln (HIG{H}_{it})$	−0.02974	0.03234
$\ln (M{S}_{it})$	−0.1021 ***	0.04573
$\ln (UN{I}_{it})$	−0.06628	0.05002
$\ln (PA{R}_{it})$	−0.14636 ***	0.04057
$\ln (I{P}_{it})$	0.02822 ***	0.00999
$\ln (H{Q}_{it})$	−0.19468 ***	0.07148
$\ln (P{D}_{it})$	0.39242 ***	0.08933
$\ln (R{S}_{it})$	0.51294	0.43029
Constant	5.17339	5.27317
Dummy variable (district)	Yes
Dummy variable (year)	Yes
Observation	672
R2	0.8834

Note: ***: significant at the 0.01 level; **: significant at the 0.05 level.

presents the results of the factors affecting commercial land lease prices.

Table 2 demonstrates that the land area and plot ratio of commercial land have a statistically significant impact on district government land leases. Regression analysis shows that the positive coefficient of land area is 0.94, indicating that for every 1% increase in land area, land prices rise by 0.94%. This strong area premium incentivizes local authorities to prioritize the allocation of larger commercial land parcels, maximizing fiscal revenue through land transaction fees.

In addition, a significant plot ratio effect with a coefficient of 0.61 was also identified. This indicates that for every 1% increase in plot ratio, land prices rise by 0.61%, reflecting the economic value of development intensity. From the perspective of land economics, higher plot ratios enable developers to achieve larger building volumes, thereby obtaining higher development returns and enhancing their bidding competitiveness. This market dynamic is consistent with the municipal government’s fiscal optimization strategy, creating mutual incentives for both parties to support higher-density commercial land transactions.

District economic development measured by GDP has a statistically significant positive impact on commercial land leases. The coefficient is 0.44, indicating that for every 1% increase in district GDP, commercial land prices rise by 0.44%. From a fiscal perspective, in high-GDP regions where the capitalization effects of land value are more pronounced, local governments are more inclined to provide commercial land to maximize municipal revenue. Secondly, the economic multiplier effect of commercial development creates an effective cycle, where the allocation of commercial land stimulates the growth of the tertiary industry, thereby improving GDP performance.

Accessibility determines the factors affecting commercial land price from the market perspective. Land close to city centers shows a strong economic premium, with prices increasing by 1.09% (p < 0.01) for every 1% decrease in distance from People’s Square, reflecting classic bidding theory and the center position effect. Similarly, proximity to the district government shows a price premium of 0.06% (p < 0.05) for every 1% reduction in distance, indicating that there exists a secondary centrality pattern. Airport accessibility presents different impacts. Due to Hongqiao’s mature commercial hinterland, for every 1% reduction in distance, Hongqiao’s proximity increases by a 0.19% premium, while for every 1% increase in Pudong’s proximity, it decreases by a 0.39% discount, reflecting its surrounding location and underdeveloped environment.

Parks generate a 0.15% price premium for every 1% increase in proximity to the park, indicating the capitalization of environmental amenities. On the contrary, for every 1% increase in the proximity of industrial zones, there is a price drop of 0.03%, highlighting the negative externalities of manufacturing zones. Key middle schools have a significant impact on the lease prices of commercial land, with a 1% reduction in land distance from middle schools and a 0.1% increase in prices. Key middle schools attract high-income families to form a stable and high-quality consumer group, driving the upgrading of surrounding commercial formats. Secondly, high-quality educational resources enhance the overall value of the region and reduce investment risk expectations through the process of spatial capitalization. Also, education-related businesses have a stronger ability to pay premiums and a more stable consumer base.

The results explain the district government’s preference for transporting commercial plots with superior centrality, higher environmental quality, and farther distance from industrial areas, as these factors increase land value through agglomeration economy and environmental externalities.

4.3. Government Perspective

From

Table 3. The results of the model of Equation (3).

Variable	Coefficient	Robust Standard Error
$\ln (SP{S}_{it})$	3.02231 ***	0.93012
$\ln (S{N}_{it})$	4.55727 *	2.71607
$\ln (H{N}_{it})$	2.6943 ***	0.49352
$\ln (M{S}_{it})$	−0.23395 **	0.10488
$\ln (UN{I}_{it})$	0.17347 **	0.08859
Constant	−33.31071 ***	11.51635
Dummy variable (district)	Yes
Dummy variable (year)	Yes
Observation	130
R2	0.4134

Note: ***: significant at the 0.01 level; **: significant at the 0.05 level; *: significant at the 0.1 level.

, which shows the results of Equation (3), it is indicated that the quantity of subway stations and expressway entrances and exits and the distance from the district to the city center and the nearest university have positive effects on the total area of commercial land leases. The distance from the district to the nearest key middle school has a negative impact on the total area of commercial land leases.

The farther the distance from the district to the city center, the more total commercial land area to be leased. This phenomenon mainly stems from the land supply mechanism under the ‘multi-center’ spatial development strategy in Shanghai. As the distance from the city center increases, the government cultivates secondary commercial centers by increasing the supply of commercial land to alleviate the excessive agglomeration pressure in the core area. For example, the cost of land acquisition in peripheral areas is relatively low, and there are conditions for large-scale land supply. Secondly, the extension of rail transit drives the increase in land development value along the line. Meanwhile, the manufacturing industry relocates and transfers existing land to commercial development areas. The overall urban planning process guides the rational distribution of population and industries through land index regulation.

The more subway stations or expressway entrances and exits in the district, the more total area to be leased, which is due to the catalytic effects of transportation infrastructure on the development of commercial land. Subway stations and expressway entrances and exits generate economic effects by enhancing location accessibility, significantly increasing the potential development value of commercial land and enabling peripheral areas to accommodate the relocation of commercial functions. Secondly, the government follows the concept of ‘Transit-Oriented Development’ (TOD) and actively increases the supply of commercial land along subway lines and highway interchanges, stimulating market development willingness through incentive policy tools.

The farther the distance from the district to the nearest university, the more total area to be leased, which demonstrates the spatial mismatch between higher education resources and commercial land supply. In the non-central districts of Shanghai, due to the concentration of high-quality universities in central areas, as the distance from university campuses increases, district governments tend to compensate for educational resource disadvantages by increasing the supply of commercial land and cultivating alternative commercial growth poles.

The closer the district is to the nearest key middle school, the more total area that is leased. This reflects the agglomeration effects of high-quality educational resources on the development of commercial land. Driven by consumer demand, the stable flow of people, such as parents accompanying students, brought by schools creates sustained commercial demand, stimulating the supply of land for retail, catering, and other purposes. In addition, the upgrading effects of supporting facilities often synchronously improve the level of regional infrastructure and enhance the feasibility of commercial development through the construction of educational facilities.

To verify the results of Equations (2) and (3), the variables of the distance to the nearest subway station and the quantity of subway stations are removed from the models, and the results show that the significant variables and sign of all the variables do not change. In addition, the variable of the distance to the nearest university is removed from two models, and the results show that the significant variables and sign of all the variables do not change either. Therefore, the results of both models are considered valid.

Table 4. The results of the validation of the models.

Variable	Coefficient	Robust Standard Error	Coefficient	Robust Standard Error
Model of Equation (2)
$\ln (Are{a}_{it})$	0.93687 ***	0.02408	0.93703 ***	0.02398
$\ln (FA{R}_{it})$	0.6087 ***	0.05264	0.60841 ***	0.05257
$MO{D}_{it}$	−0.0375	0.11295	−0.02479	0.11092
$\ln (GD{P}_{it})$	0.46386 **	0.20354	0.43954 **	0.20645
$D{M}_{it}$	−0.01269	0.01857	−0.00998	0.01887
$\ln (SP{S}_{it})$	−1.08102 ***	0.35758	−0.91482 ***	0.34542
$\ln (GO{V}_{it})$	−0.06302 **	0.03266	−0.0624 **	0.0329
$\ln (SU{B}_{it})$	Removed		−0.02349	0.02907
$\ln (HIG{H}_{it})$	−0.02943	0.03217	−0.02831	0.03241
$\ln (M{S}_{it})$	−0.10944 **	0.04513	−0.10801 **	0.04538
$\ln (UN{I}_{it})$	−0.06348	0.04917	Removed
$\ln (PA{R}_{it})$	−0.14818 ***	0.04042	−0.14387 ***	0.04085
$\ln (I{P}_{it})$	0.02886 ***	0.00993	0.02978 ***	0.00989
$\ln (H{Q}_{it})$	−0.1916 ***	0.07061	−0.17267 ***	0.06808
$\ln (P{D}_{it})$	0.3865 ***	0.08827	0.42265 ***	0.08923
$\ln (R{S}_{it})$	0.47151	0.42029	0.28564	0.3984
Constant	5.17339	5.27317	4.500932	5.278768
Dummy variable (district)	Yes		Yes
Dummy variable (year)	Yes		Yes
Observation	672		672
R2	0.8833		0.883
Model of Equation (3)
$\ln (SP{S}_{it})$	2.04316 ***	0.58399	3.44634 ***	0.92891
$\ln (S{N}_{it})$	Removed		5.14076 *	2.76408
$\ln (H{N}_{it})$	2.65691 ***	0.48335	2.44336 ***	0.471238
$\ln (M{S}_{it})$	−0.23141 **	0.10581	−0.27231 **	0.105408
$\ln (UN{I}_{it})$	0.19783 **	0.08718	Removed
Constant	−19.81908 ***	6.724647	−35.99339 ***	11.66591
Dummy variable (district)	Yes		Yes
Dummy variable (year)	Yes		Yes
Observation	130		130
R2	0.3923		0.3941

Note: ***: significant at the 0.01 level; **: significant at the 0.05 level; *: significant at the 0.1 level.

shows the results of the validation of the models.

5. Conclusions

Based on the data on districts in Shanghai from 2004 to 2015, this paper investigates the determinants of commercial land leases by district governments from the perspective of the government and market in non-central districts and how these determinants affect the price and area of commercial land leases.

In conclusion, the results show that Shanghai’s multi-center development strategy aims to optimize the city’s commercial layout by developing the key areas of non-central districts. Multiple factors significantly affect the pricing of commercial land from the market perspective. The key determining factors include land variables (land area and plot ratio) and location variables (distance to the city center, district government, Hongqiao Airport, and Pudong Airport and proximity to parks, industrial zones, and middle schools). The observed price elasticity reflects the rational economic behavior of district governments, which strategically prioritize the transportation of high-value land parcels and maximize fiscal revenue through land leases. Regarding the government aspect, the quantity of subway stations and expressway entrances and exits and the distance from the district to the city center and the nearest university have positive effects, and the distance to the nearest key middle school has a negative impact.

Specifically, the government has shown a preference for commercial land, especially because of the following: (1) the potential for larger-scale and higher-density development; (2) convenience in accessing the central business district and secondary centers; (3) proximity to mature commercial centers, educational services, and environmental facilities while avoiding plots near negative externalities (industrial areas) or underdeveloped outskirts; and (4) the principle of prioritizing location value, which is reflected in its significant preference for transportation node areas and the active cultivation of educational resource economic circles. The results provide empirical evidence for optimizing urban land supply strategies for economic benefits and urban development quality.

Based on the conclusions, it is recommended that the government adopt the following land policies.

(1) A differentiated pricing system on the demand side of the market that includes external factors, such as transportation and education, can be established, and plot ratio rewards for areas near subway stations and key school districts can be implemented.

(2) The supply of transportation hubs and commercial land around universities on the supply side of the government can be prioritized, and industrial buffer zones can be set up to control development intensity.

(3) The implementation of mixed use land models and cross-regional balance mechanisms can be synchronized, and policy effects through digital platforms can be dynamically monitored.

These policy implications can not only improve the efficiency of land resource allocation but also promote the coordinated development of functions such as commerce, transportation, and education, ultimately achieving a virtuous cycle of optimizing urban spatial structure and increasing fiscal revenue. Other cities can learn from the above policy recommendations, such as using modular tools such as the flexible adjustment of the plot ratio and negative list management, but need to carry out adaptive transformation according to the local development stage, spatial structure, and land system.