Land Use Efficiency in the Yellow River Basin in the Background of China’s Economic Transformation: Spatial-Temporal Characteristics and Influencing Factors

Abstract

Rapid urbanization has led to the increasing scarcity of land resources in China. Exploring the spatial-temporal characteristics and influencing factors of urban land use efficiency (LUE) is of great significance for optimizing the allocation efficiency of land resources and promoting regional sustainable development. In this study, the Super-SBM model was used to calculate the urban LUE of the Yellow River Basin from 2009 to 2018. The regional differences and agglomeration characteristics of LUE in the Yellow River Basin were analyzed. Moreover, a panel regression model was used to analyze the influencing factors of LUE. The results showed that the LUE in the Yellow River Basin experienced a process of fluctuation decline during the study period. The regional difference of LUE in the Yellow River Basin was as follows: upper reaches > middle reaches > lower reaches. The hot and cold spots of LUE were relatively stable in spatial distribution during the study period. The hot spots were mainly distributed in Ordos in the upper reaches and Yulin in the middle reaches, while the cold spots were mainly distributed in Henan Province in the lower reaches. Globalization had a positive impact on LUE in the lower reaches. Marketization had a positive impact on LUE in the whole basin and lower reaches, and a negative impact on LUE in the middle reaches. Decentralization had a positive impact on the LUE of the whole basin and the upper reaches, and a negative impact on the LUE of the lower reaches.

1. Introduction

The United Nations predicts that by the middle of this century, more than 70% of the world’s population will be concentrated in urban areas [1]. The excessive growth of urban population leads to the continuous expansion of urban built-up areas, but the expansion scope of urban built-up areas is limited. The disorderly expansion of urban built-up areas and extensive land management model have damaged the ecological environment around the city, which is not conducive to achieving the goal of global urban sustainable development [2,3]. How to balance the relationship between urban population growth and land use has become an important global issue [4]. Intensive land use is conducive to coordinating the relationship between urban population and land resources, which is of great significance for promoting the healthy and sustainable development of cities [5].

In China, the reform and opening up policy has accelerated the process of urbanization. According to relevant statistics, from 1978 to 2018, China’s urbanization rate increased rapidly from 17.9% to 59.58%, which strongly promoted China’s rapid economic growth and social development. However, rapid urbanization has caused the rapid expansion of urban construction land into surrounding areas, leading to the encroachment of many rural and agricultural lands. Moreover, in order to maximize regional economic benefits in the short term, some local governments have intensified cooperation with local land developers, resulting in a large number of urban undeveloped land and rural land being converted into urban construction land [6]. Relevant research shows that in the next 20 years in China, about 14 million rural people will enter urban areas every year, which means that the area of urban construction land will be further expanded [7]. With more and more rural people moving in, urban areas demand for construction land is increasing, which will further aggravate the contradiction between urban population growth and limited land resources. Therefore, local governments should constantly optimize land development policies, and improve urban LUE [8].

At present, scholars have conducted a series of studies on the concept and connotation of urban LUE. Some scholars used the economic output per unit of urban land area [9,10], the ratio of added value of secondary and tertiary industries to land area [11] to characterize urban LUE. With the concept of sustainable development being put forward, more and more people begin to pay attention to the social and ecological benefits of urban development. Therefore, some scholars put forward that LUE should fully reflect the economic, social, and ecological benefits of land use, and they calculated urban LUE by building a comprehensive indicator system [12,13]. As China’s economy has entered the stage of high-quality development, the development of urbanization has changed from pursuing “high speed” to “high quality”. Therefore, the evaluation of urban LUE should emphasize not only the economic output of land, but also coordinate the social and ecological benefits of urban land use [14].

In the 1990s, smart growth theory [15] and compact development theory [16] were put forward one after another to guide the rational development and utilization of urban land resources. The resource allocation theory provided an important theoretical support for promoting the rational allocation of urban land resources and improving the efficiency of urban land use [17]. The Urban Growth Boundary Policy (UGB) had played an important role in controlling the continuous expansion of urban land and promoting the intensive use of urban land in the United States [18]. Some European countries control the continuous spread of land in built-up areas by improving the utilization level of infrastructure and reusing abandoned areas around cities [19,20]. In China, with the rapid development of urbanization, the central government began to attach importance to the intensive use of urban land. In order to improve the efficiency of urban land allocation, the central government implemented the urban land transfer system in 2001 [6]. Moreover, more and more scholars began to study the efficiency of urban land use in China and made some research results. Relevant achievements mainly focus on the following aspects: first, the quantitative assessment of urban LUE. One was to evaluate LUE based on economic output of unit land area [21,22,23]. The other was to calculate the LUE by constructing the input-output index system and using DEA and its extended model [24,25,26,27,28]. Second, the analysis of spatial-temporal characteristics of urban LUE. Spatial analysis and other methods were used to explore the spatio-temporal evolution characteristics of urban LUE [14,29,30]. Third, the assessment of influencing factors of urban LUE. Statistical analysis methods were used to explore the effects of economic development, population size, industrial structure, and other factors on LUE [31,32,33]. In general, in terms of LUE evaluation, current scholars mainly use Super-SBM model to evaluate LUE by building a comprehensive evaluation index system. In terms of research methods, most scholars analyze the spatio-temporal evolution of LUE based on ArcGIS spatial analysis technology, and the research areas mainly focus on the national level, provincial level, or single cities [34,35]. There are relatively few studies on regional differences of LUE in large river basin areas. In the aspect of influencing factors research, the traditional influencing indicators such as industrial structure and economic development level were mainly selected for regression analysis of influencing factors. However, it is relatively insufficient to study the influencing factors of LUE by building a theoretical analysis framework.

At present, China’s economic development has entered a transitional period [36]. Economic transformation had not only released great vitality for China’s urban socio-economic activities, but also had a profound impact on urban land development and utilization [37]. In September 2019, the Yellow River Basin became a major national strategic area in China [38]. However, the development and construction activities aimed at economic growth and the long-term extensive land use model lead to the Yellow River Basin face the dual pressure of waste of land resources and ecological environment damage. Moreover, compared with the Yangtze River Basin, the economic development and opening up of the Yellow River Basin are relatively lagging behind, and the regional differences in land use status are large. Therefore, there may be obvious regional differences in the impact of economic transformation on LUE in the Yellow River Basin.

First of all, the paper used the Super-SBM model to calculate the urban LUE of the Yellow River Basin from 2009 to 2018. Secondly, from the perspective of economic transformation, the paper analyzed the factors affecting the LUE of the Yellow River Basin by building a theoretical framework for the relationship between economic transformation and land use (Figure 1). It is of great significance to improve the regional land use model and promote regional sustainable development. Compared to the existing research, this paper is advanced in three aspects: first, the theoretical relationship framework between economic transformation and LUE was constructed. The theoretical relationship framework between economic transformation and LUE provided theoretical support for analyzing the influencing factors of LUE. Second, taking the Yellow River Basin as the research area was typical and valuable. The Yellow River Basin region is a major national strategic region of China. It is of great significance to study the regional differences and influencing factors of LUE to guide the rational development and utilization of land resources in the Yellow River Basin. Third, Super-SBM model was introduced to evaluate LUE, which not only allows the efficiency evaluation result to be greater than 1, but also solves the relaxation problem of traditional DEA model, making the evaluation result more scientific and reliable.

The rest of this paper was arranged as follows: Section 2 constructed the theoretical analysis framework of the relationship between economic transformation and LUE. Section 3 mainly introduced the overview of the study area and the establishment of the model method. Section 4 was the main results and analysis. Section 5 discussed the research results. Section 6 concluded the research results and put forward some policy suggestions.

2. How Does Economic Transformation Affect LUE? An Analytical Framework

Since the reform and opening up, China’s economic development has experienced a transformation process from closed to open, from planned economy to market economy, and fully integrated into economic globalization [36]. The process of globalization, marketization, and decentralization is generally defined as economic transformation [39]. Since the 21st century, more and more scholars have gradually realized the important role of economic transformation in urban land use [37], urban economic growth [40], and urban ecological environment [41], and have carried out some preliminary studies.

Since China’s accession to WTO in 2001, the process of China’s economic integration into the world economic globalization has been greatly accelerated. Globalization has promoted the integration and redistribution of the world’s factors of production, and has produced scale effects, technological effects, and structural effects on LUE through foreign direct investment [42,43]. First of all, globalization is conducive to attracting foreign enterprises, stimulating the expansion of urban construction land, increasing the economic output of urban land, and improving the efficiency of land use, which forms the scale effect of urban land use. Secondly, while attracting advanced foreign-funded enterprises, globalization will also bring advanced production technology and business philosophy to the region, which is conducive to promoting the increase of the output value of the secondary and tertiary industries, and improving the economic benefits of urban land use, which forms the technological effect of urban land use [44]. Finally, globalization is conducive to the introduction of advanced production equipment, promoting the transformation of backward production capacity into a high-tech industry with low energy consumption and high output, and improving the efficiency of land use, which forms the structural effect of urban land use [45].

Marketization is a process from planned economy to market economy [46]. Land is a very valuable economic resource. In the process of land development, local governments signed land transfer agreements with land developers, forming a process of land marketization [47,48]. In China, Land marketization has not only become the main source of local public finance revenue, but also an important force to stimulate local economic growth [49]. The process of land resource marketization reflects the transfer and circulation of land resources in a region. First of all, land marketization can stimulate domestic demand, effectively stimulate consumption and investment, and promote urban economic growth, which is conducive to improving urban LUE [50]. Secondly, land marketization is conducive to promoting the transfer of rural population to urban areas, promoting the transfer of rural population employment from the primary industry to the secondary and tertiary industries, increasing regional economic output. Finally, land marketization is conducive to attracting domestic and foreign enterprises to invest in development and construction, increasing the added value of regional secondary and tertiary industries, and realizing efficient use of land resources, which is conducive to improving urban LUE [51].

Since the implementation of the tax system reform in China in 1991, the central government has delegated economic and financial power to local governments [52], which has effectively promoted local socio-economic development. GDP is taken as the core assessment standard of local government performance, which enables some local governments to directly participate in the competition for regional economic growth in order to achieve good performance [53]. On the one hand, decentralization is conducive to improving urban LUE. Local governments have more financial autonomy under the decentralized system, which is conducive to improving the efficiency of government allocation of local land resources [54]. For instance, local governments promote local economic growth and improve urban LUE by attracting investment from foreign and local enterprises [55,56]. On the other hand, decentralization may lead to the decline of urban LUE. In order to achieve good political achievements, some local governments may attract investment by lowering the land price to achieve the increase of local fiscal revenue and economic growth in the short term, which may lead to the increase of urban idle land and the decline of LUE [57].

From the perspective of China’s economic transformation, taking into account the economic development and land use status of the Yellow River Basin, and referring to relevant research [14,29], this study added industrial structure, urbanization, industrial scale, economic development, and population size as basic factors, and established a theoretical framework for the relationship between economic transformation and LUE (Figure 2).

3. Study Area and Methods

3.1. Study Area

The Yellow River Basin in this study mainly includes eight provinces or autonomous regions, such as Shandong, Henan, Shanxi, Shaanxi, Inner Mongolia, Ningxia, Gansu, and Qinghai [38]. Hulunbeir, Xing’an, Tongliao, and Chifeng in the east of Inner Mongolia belong to the northeast region, they were not considered in the study area [58]. The ecological environment in the Yellow River Basin is fragile, with serious water and soil loss and poor carrying capacity of resources and environment. Meanwhile, the economic development in the Yellow River Basin is extremely uneven, and the problem of extensive land use is particularly prominent. According to the physical geography and economic development characteristics of the Yellow River Basin, this paper defined Shandong and Henan as the lower reaches, Shanxi and Shaanxi as the middle reaches, and Qinghai, Gansu, Inner Mongolia, and Ningxia as the upper reaches (Figure 3).

3.2. Methods

3.2.1. Super-SBM Model

Charnes et al. first proposed the DEA model [59] by building an effective production boundary to evaluate the relative efficiency of decision making units (DMUs) under the conditions of multiple inputs and multiple outputs. DEA model does not need to manually determine the weight, thus eliminating the impact of subjective factors, and has strong objectivity. However, the traditional DEA model is radial and angular, which does not take into account the input/output relaxation problem, so that the efficiency value is usually overestimated. Tone proposed a nonradial SBM model, which directly solved the relaxation problem [60]. However, no matter DEA or SBM model, the efficiency value of multiple effective DMU may be 1 in the calculation results, which makes it impossible to distinguish them effectively. Later, Tone proposed the Super-SBM model, which not only allows the efficiency evaluation result to be greater than 1, but also solves the relaxation problem of the traditional DEA model [61], making the evaluation result more scientific and reliable. Therefore, this paper used the Super-SBM model to evaluate the LUE of cities in the Yellow River Basin. This assessment result can more scientifically reflect the current situation of urban land use in the Yellow River Basin. The calculation formula of the model is as follows:

$$Min\rho =\frac{1+\frac{1}{m}{\displaystyle {\sum }_{i=1}^m\frac{s_i^{-}}{x_{i0}}}}{1-\frac{1}{s}{\displaystyle {\sum }_{r=1}^s\frac{s_r^{+}}{y_{r0}}}}$$

(1)

$$S.T.\{\begin{array}{l}{\displaystyle \sum _{j=1,j\ne 0}^n{x}_{ij}{\beta }_j-s_i^{-}\le x_{i0} (i=1,2,\cdots ,m)}\\ {\displaystyle \sum _{j=1,j\ne 0}^n{y}_{rj}{\beta }_j+s_i^{+}\ge y_{r0} (r=1,2,\cdots ,s)}\\ {\displaystyle \sum _{j=1}^n{\beta }_j=1; {\beta }_j\ge 0; \beta , s^{-}, s^{+}\ge 0}\end{array}$$

(2)

where ρ is the value of LUE, β is the weight vector coefficient, $x_{i0}$ and $y_{r0}$ are input index and output index, respectively. m and s are the number of input index and output index, respectively, and $s_i^{-}$ and $s_r^{+}$ are the relaxation variables of input index and output index, respectively.

3.2.2. Exploratory Spatial Data Analysis

Exploratory spatial data analysis is a collection of a series of spatial data analysis methods and technologies, which can explore the spatial clustering characteristics of elements and the spatial relationships between elements [62]. In this paper, the global spatial autocorrelation model was selected to explore the overall spatial correlation characteristics of urban LUE in the Yellow River Basin. The calculation formula of the model is as follows:

$$I=\frac{n{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{w}_{ij}(x_i-\overline{x})}(x_j-\overline{x})}}{{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{w}_{ij}}}{\displaystyle \sum _{i=1}^n{(x_i-\overline{x})}^2}}$$

(3)

where I is the global Moran index, and its value range is (−1,1); n is the number of spatial units; x_i and x_j is the observation value of space unit i and j, respectively; $\overline{x}$ is the average of all space unit observations; w_ij is the spatial weight matrix composed of space unit i and j.

Getis-Ord Gi* index [63] was selected to analyze the local spatial agglomeration characteristics of urban LUE in the Yellow River Basin. The calculation formula of the model is as follows:

$$G_i{}^{*}=\frac{{\displaystyle \sum _{j=1}^n{W}_{ij}(d)X_j}}{{\displaystyle \sum _{j=1}^n{X}_j}}(i\ne j), Z(G_i^{*})=\frac{G_i-E(G_i^{*})}{\sqrt{Var(G_i^{*})}}$$

(4)

where W_ij is the spatial weight and $E(G_i^{*})$ and $Var(G_i^{*})$ are the mathematical expectation and variance of $G_i^{*}$, respectively. If the value of $Z(G_i^{*})$ is positive and significant, it indicates that the surrounding values of space unit i are high, then the unit is determined as the hot spot in the region; if the value of $Z(G_i^{*})$ is negative and significant, it indicates that the surrounding values of space unit i are low, then the unit is determined as the cold spot in the region.

3.2.3. Panel Regression Model

This study is based on the urban panel data of the Yellow River Basin from 2009 to 2018. Therefore, the panel regression analysis model [64] was constructed to analyze the influencing factors of LUE in the Yellow River Basin. Firstly, the fixed effects model and random effects model regression of panel data were carried out by Stata software. Secondly, the optimal interpretation model was selected by Hausmann test. The test results showed that the fixed effect model is suitable for the analysis of influencing factors in this study. The model was constructed as follows:

$$y_{i,t}={\mu }_0+{\displaystyle \sum _{k=1}^n{\beta }_k{x}_{i,t}}+{\mu }_i+{\epsilon }_{i,t}$$

(5)

where y_i,t refers to the LUE of city i in year t, β refers to the coefficient to be estimated of the explanatory variable, µ₀ refers to the intercept term, x_i,t refers to the impact factor of LUE, i represents the number of samples, t represents the year, k represents the number of explanatory variables, µ_i represents individual difference, and ε_i,t represents residual.

3.3. Input and Output Indicators

This study combined results from previous studies [65,66,67], selected the area of construction land, the number of employees in the secondary and tertiary industries, and the total investment in fixed assets as input indicators, and the added value of the secondary and tertiary industries, greening coverage rate of built-up areas and the average wage of urban workers as output indicators to build an input output indicator system for LUE (

Table 1. Descriptive statistics of input and output data.

	Indicators	Unit	Max	Min	Mean	Std. Dev.
Input	Construction land area (Municipal district)	km2	658	6	102.9288	95.9702
	Number of employees in secondary and tertiary industries	104 persons	195.8304	1.2788	23.4134	28.5485
	Total investment in fixed assets	108 yuan	8398.94	40.5400	1359.9676	1287.9142
Output	Value added of secondary and tertiary industries	108 yuan	11,614.61	63.7800	1597.7372	1608.2038
	Greening coverage rate of built-up areas (Municipal district)	%	55.17	2.75	30.16	34.2817
	Average salary of urban employees	yuan	128,242	4792	48,222.5083	15,445.5548

).

3.4. Dependent Variable and Independent Variable

The input–output index system of LUE was constructed from three aspects of society, economy and ecology, which can not only reflect the socio-economic output of urban land, but also reflect the benefit balance between land development and ecological environment in the Yellow River Basin. In this study, LUE was selected as the dependent variable.

This paper constructed a theoretical framework for analyzing the relationship between economic transition and LUE, and took globalization, marketization, and decentralization as the core explanatory variables to explore the influencing factors of LUE. On the basis of existing research [5,11,68], based on the principles of representativeness, rationality, and operability, this study selected industrial structure, economic development level, urbanization level, industrial scale and population size as explanatory variables to explore the influencing factors of LUE. The selected indicators of relevant variables and their meanings were shown in

Table 2. Explanation of LUE and impact indicators.

Variables	Indicators	Definition	Codes
Dependent variable	LUE	Land use efficiency value	lue
Explaining variable	Globalization	FDI/urban population	glo
	Marketization	Non-state-owned economy/Total industrial output value	mar
	Decentralization	Urban per capita fiscal expenditure/(Urban per capita fiscal expenditure + Provincial per capita fiscal expenditure + National per capita fiscal expenditure)	dec
Control variable	Industrial structure	The proportion of secondary industry in GDP	str
	The level of economic development	GDP per capita	pgdp
	Urbanization level	The ratio of the urban population in the total population	urb
	Industrial scale	Industrial output value/Urban area	agg
	Population size	Urban population/Urban area	pop

.

Industrial structure: Since the reform and opening up, industrialization has promoted the adjustment of industrial structure, which has strongly promoted the process of urbanization in China. The upgrading of industrial structure can promote the transfer of labor force from the primary industry to the secondary and tertiary industries, increase local economic output, and improve urban LUE [69]. The proportion of heavy chemical enterprises in the Yellow River Basin is relatively high, and the secondary industry plays an important role in promoting regional economic growth. Therefore, the proportion of the output value of the secondary industry in GDP was selected to characterize the industrial structure.

The level of economic development: The economic growth of a region means that the economic output of the region increases. If other conditions are unchanged, increasing the economic output per unit land area can effectively improve the LUE [70]. The improvement of economic development level may improve the urban LUE of the Yellow River Basin [71]. Therefore, GDP per capita was selected to characterize the level of economic development.

Urbanization level: Since the reform and opening up, China’s urbanization process has been rapidly promoted, which has strongly promoted the development and construction of urban land [14]. On the one hand, urbanization can trigger rural population to gather in cities, promote efficient use of urban land and increase economic output, which is conducive to improving the efficiency of urban land use. On the other hand, rapid urbanization may lead to blind expansion of urban built-up areas and extensive land management, leading to the decline of LUE. Therefore, the ratio of the urban population in the total population was selected to characterize the level of urbanization.

Industrial scale: Heavy chemical industry is widely distributed in the Yellow River Basin, and the gross industrial product accounts for a high proportion in the total regional economy. On the one hand, the increase in the scale of industrial enterprises is conducive to promoting regional economic growth, increasing the economic output per unit of land, and improving the urban LUE. On the other hand, the excessive increase of industrial enterprises may lead to waste of land resources and damage to the local ecological environment, leading to the decline of urban LUE. Therefore, industrial output value/urban area was selected to characterize the industrial scale.

Population size: Population size can have an important impact on regional land use [8]. On the one hand, population is the object of land resources development and use. The increase of population size is conducive to promoting urban development and construction, attracting foreign enterprises to settle in, and promoting the intensive use of urban land and local economic growth. On the other hand, the excessive increase of population size may lead to blind expansion of urban construction land, compression of urban green space area, waste of land resources, and decline of ecological benefits, which may lead to the decline of urban LUE. Therefore, urban population/urban area was selected to characterize population size.

3.5. Data Source

Owing to the data of some cities in the Yellow River Basin cannot be obtained, 80 cities in the Yellow River Basin were finally determined as the research samples. The data used in this study were mainly from China City Statistical Yearbook, China Statistical Yearbook and the statistical yearbooks of various provinces and regions in 2010–2019. A few missing data in some years were supplemented by interpolation.

4. Results

4.1. Temporal Evolution Characteristics of LUE

There were long-term and stable regional differences in LUE in the Yellow River Basin (Figure 4). From 2009 to 2018, the average value of LUE in the Yellow River Basin was generally as follows: the upper reaches > the middle reaches > the lower reaches. From the perspective of evolution trend, the evolution process of the average value of LUE of the whole basin from 2009 to 2018 can be roughly divided into three stages. The first stage was the dramatic evolution stage from 2009 to 2013, and the LUE showed an M-shaped evolution trend of first rising, then falling, then rising, and then falling. The second stage was the stable evolution stage from 2013 to 2016. The third stage was the drastic evolution stage from 2016 to 2018, and the LUE showed an inverted V evolution trend of first rising and then declining. On the whole, the LUE of the Yellow River Basin had experienced a development process from violent shock to stable development and then to violent shock, and the average value of LUE of the whole basin reached the highest value of 0.7762 in 2010 and dropped to the lowest value of 0.6380 in 2018. The evolution trend of LUE in the middle and lower reaches was basically consistent with the whole basin, which had gone through three stages. The average value of LUE in the upper reaches had been relatively stable, which may be attributed to the fact that the upper reaches is located in the inland region, the degree of opening to the outside world is relatively low, and the restricted development area accounts for a large proportion. Moreover, the local government has more strict control measures for the development of local land resources, and the intensive level of land use is high, which has kept the LUE in a relatively stable state.

In order to further explore the dynamic evolution characteristics of LUE in the Yellow River Basin, four years (2009, 2012, 2015, and 2018) were selected for nuclear density analysis of LUE in the Yellow River Basin with the help of stata software (Figure 5). From 2009 to 2018, the nuclear density curve of LUE in the whole basin was relatively stable, showed the coexistence of “one high and one low” double peaks, indicating that the LUE in the Yellow River Basin was polarized at both ends. The nuclear density curve in the upper reaches showed a “double peak” phenomenon, and the center of the curve moved to the right, indicating that the overall level of LUE in the upper reach was constantly improving, and the regional gap was constantly reducing. The nuclear density curve in the middle reaches showed that the main peak moved to the left and the wave peak height gradually flattens, indicating that the overall level of LUE in the middle reaches decreased and the regional gap became larger. The nuclear density curve in the lower reaches showed that the main peak moved to the left, and the single peak phenomenon gradually highlighted, indicating that the overall level of LUE in the lower reaches was reduced, and the regional gap was smaller.

4.2. Spatial Evolution Characteristics of LUE

In order to explore the spatial evolution characteristics of LUE in the Yellow River Basin, this study used ArcGIS software to conduct spatial visualization of urban LUE in the Yellow River Basin from 2009 to 2018 (Figure 6). In general, the regional differences of LUE in the Yellow River Basin were as follows: the upper reaches > the middle reaches > the lower reaches. The high value areas of LUE were mainly concentrated in Ordos and its surrounding areas, as well as Qingdao and Jinan, while the low value areas were mainly concentrated in Shaanxi and Henan.

Specifically, the areas with high LUE were mainly distributed in 11 cities in the upper reaches in 2009, including Yinchuan, Shizuishan, Wuzhong, Guyuan, and Ordos; Yangquan, Jincheng, Shuozhou, Luliang, and Shangluo in the midstream region; and Qingdao, Zibo, Dongying, Jining, Binzhou, and Xinyang in the downstream region, all of which had LUE values higher than one. In 2012, the areas with high LUE were mainly distributed in 10 cities in the upper reaches, such as Haidong, Yinchuan, Shizuishan, Wuzhong, and Ordos; Yangquan, Jincheng, Luliang, Yan’an, Yulin, and Shangluo in the middle reaches; and Jinan, Qingdao, Zibo, and Heze in the lower reaches. In 2015, the areas with high LUE were mainly distributed in 12 cities, including Shizuishan, Wuzhong, Guyuan, and Zhongwei; Luliang, Yan’an, Yulin, and Shangluo in the middle reaches; Jinan, Qingdao, Zibo, Weifang, Heze, and Sanmenxia in the lower reaches. In 2018, the areas with high LUE were mainly distributed in 13 cities in the upper reaches, including Haidong, Longnan, Qingyang, and Jinchang; Taiyuan, Shuozhou, Linfen, Luliang, Yulin, and Shangluo in the middle reaches; Jinan, Qingdao, and Yantai in the lower reaches. The analysis was carried out in combination with the global trend surface from 2009 to 2018 (Figure 7). In the east–west direction, the LUE of the Yellow River Basin had evolved from a “U” in 2009 to a gradient spatial pattern of “high in the west and low in the east” in 2018. In the north–south direction, the LUE had evolved from an inverted “U” in 2009 to a gradient spatial pattern of “high in the north and low in the south” in 2018. This showed that the regional differences of LUE in the Yellow River Basin have changed. The northwest region had become a high value area of LUE, and the southeast region had become a low value area of LUE.

4.3. Spatial Correlation Analysis of LUE

This paper analyzed the overall spatial autocorrelation characteristics of urban LUE in the Yellow River Basin by calculating the global Moran index (

Table 3. Global Moran’s I Index of LUE in the Yellow River Basin.

Year	Global Moran’s I	p-Value	Z-Score
2009	−0.0387	0.3740	−0.3815
2010	−0.0107	0.4300	0.0932
2011	0.0523	0.1320	1.0779
2012	−0.0031	0.3980	0.1907
2013	0.0297	0.0320	2.6454
2014	0.0974	0.0420	2.7886
2015	0.1414	0.0140	2.4717
2016	0.2467	0.0020	4.3127
2017	0.2147	0.0020	3.7763
2018	0.1701	0.0100	2.9519

). From 2009 to 2012, the overall Moran index failed to pass the significance test, indicating that there was no spatial correlation in the LUE of the Yellow River Basin in this period. From 2013 to 2018, the overall Moran index was positive and passed the significance test, indicating that the LUE of the Yellow River Basin generally had a significant positive spatial correlation. The urban LUE in the Yellow River Basin had changed from overall spatial irrelevance to overall spatial positive correlation, indicating that the spatial relationship of LUE in the Yellow River Basin had changed from weak to strong.

The Global Moran Index can only reflect the overall spatial correlation characteristics of LUE, but cannot reflect the local spatial correlation characteristics. Therefore, Getis-Ord G_i^* was selected in this study to analyze the local spatial agglomeration characteristics of urban LUE in the Yellow River Basin (Figure 8). From 2009 to 2018, the hot spots and cold spots of urban LUE in the Yellow River Basin were generally distributed stably. The hot spots were mainly distributed in Ordos in the upper reaches and Yulin in the middle reaches, while the cold spots were mainly distributed in Henan and the southwest of Shandong in the lower reaches. Specifically, Jiayuguan and Jiuquan were added to the hot spots from 2009 to 2012. From 2012 to 2015, Baiyin, Jiayuguan, and Jiuquan were reduced as hot spots, and Wuzhong, Yinchuan, and Dongying were added, while the distribution of cold spots was reduced. From 2015 to 2018, the distribution of hot spots was relatively stable, mainly concentrated in the upper reaches, while the distribution of cold spots expanded to the surrounding areas.

4.4. Influencing Factors of LUE in the Yellow River Basin in the Background of Economic Transformation

4.4.1. Stability Test

In order to prevent the pseudo regression phenomenon in the process of variable regression, panel unit root method was used to test the robustness of data. The results showed that all variable panel data have passed the significance and stability tests (

Table 4. Stability test results of panel data.

Variable	ADF	LLC	IPS	PP
lue	3.642 ***	−1.420 ***	−2.147 ***	3.451 ***
glo	4.360 ***	−2.761 ***	−6.318 ***	5.230 ***
mar	7.423 ***	−5.472 ***	−4.105 ***	3.259 ***
dec	4.330 ***	−3.426 ***	−3.224 ***	5.124 ***
str	3.103 ***	−2.017 ***	−4.291 ***	4.144 ***
pgdp	10.296 ***	−7.302 ***	−7.145 ***	6.028 ***
urb	12.437 ***	−8.304 ***	−9.183 ***	5.781 ***
agg	2.030 ***	−1.201 ***	−3.156 ***	4.049 ***
pop	4.552 ***	−2.003 ***	−4.321 ***	3.154 ***

Note: *** refers to the 1% significance levels.

). The results of F test showed that there was significant fixed effect in the model.

4.4.2. Influencing Factors of LUE in the Yellow River Basin

This study used Stata software to conduct random effect and fixed effect model regression on panel data of the Yellow River Basin from 2009 to 2018. The regression results of the model were shown in

Table 5. Regression results of the influence factors of the LUE in the Yellow River Basin.

Variable	All		Upper		Middle		Lower
	fe	re	fe	re	fe	re	fe	re
glo	−0.0124 (−0.14)	0.0539 (0.69)	−0.1395 (−0.60)	0.1098 (0.52)	−0.0757 (−0.20)	0.0516 (0.21)	0.0758 * (1.87)	0.0619 (0.75)
mar	0.3194 * (1.96)	0.3603 *** (3.96)	0.0617 (1.04)	−0.1907 (0.82)	−0.1626 ** (−2.66)	0.3421 (1.15)	0.2173 * (1.64)	0.1049 (0.33)
dec	0.1982 ** (2.87)	−0.1754 (−1.18)	0.3243 ** (2.93)	−0.1517 (−0.64)	−0.1053 (−0.26)	0.3021 (1.02)	−0.4062 ** (−2.39)	−0.0411 (−0.15)
str	0.0942 * (1.71)	0.0556 (0.76)	0.0095 (0.07)	0.0486 (0.49)	−0.2835 ** (2.81)	0.1194 (1.08)	0.2459 (1.40)	0.04509 (0.47)
pgdp	0.2357 (0.69)	0.2443 ** (2.11)	−0.1070 (−0.38)	0.0789 (0.32)	0.5560 (1.54)	0.2734 (1.12)	0.3316 (1.51)	0.1009 (0.57)
urb	−0.2867 *** (−2.81)	−0.2213 *** (−2.89)	−0.0251 (−0.12)	−0.1031 (−0.62)	−0.3172 * (−1.67)	−0.5308 (−2.05)	−0.3538 *** (−2.68)	−0.2627 ** (−2.35)
agg	−0.0530 (−0.54)	−0.0974 (−1.23)	1.5092 (0.89)	1.2633 (0.32)	−0.0174 (−0.02)	−0.2825 (−0.41)	−0.0832 (−0.76)	−0.0899 (−0.97)
pop	−0.2173 * (−1.64)	−0.3704 (−0.97)	3.0040 (0.79)	1.4028 (−0.51)	−0.2931 ** (−2.35)	−0.1472 (−2.48)	−0.1073 * (−1.40)	−0.0227 (−0.23)
Cons	0.3097 *** (4.55)	0.5729 *** (3.68)	0.8568 *** (3.40)	0.5607 *** (3.78)	0.5178 (0.97)	0.1803 (1.11)	0.3129 (1.26)	0.1054 (0.96)
R2	0.7629	0.6732	0.8196	0.6986	0.6984	0.5850	0.7969	0.5926
F	8.69	—	8.26	—	4.74	—	8.11	—

Note: ***, **, and * refer to the 1%, 5%, and 10% significance levels, respectively.

. Hausman test also showed that the fixed effect model is appropriate. Therefore, the regression results of the fixed effect model were finally selected to analyze the factors affecting LUE.

• Globalization had a positive impact on LUE in the lower reaches, while it was not significant in the whole basin, upper reaches, and middle reaches. Specifically, the impact coefficient of globalization on LUE in the lower reaches of the Yellow River Basin was 0.0758, and had passed the significance test at the level of 10%. It showed that every 1% increase in per capita foreign direct investment will promote the LUE in the lower reaches to increase 0.0758%. In the lower reaches of the Yellow River Basin, the improvement of globalization was conducive to improving regional LUE. However, in the upper and middle reaches of the Yellow River Basin, the impact of globalization on regional LUE was not obvious. The lower reaches in the Yellow River Basin is the region with the highest degree of openness to the outside world, and also the region with the highest degree of globalization. Globalization is conducive to attracting foreign enterprises to the lower reaches, promoting the expansion of urban construction land, improving the economic output of urban land, and promoting the LUE in the lower reaches. Moreover, globalization can not only promote the introduction of advanced foreign-funded enterprises in the lower reaches, but also bring advanced production technology and business philosophy, promote the increase of output value of secondary and tertiary industries, and improve the economic benefits of land use in the lower reaches. Finally, globalization is conducive to the introduction of advanced production equipment in the lower reaches, so that backward production capacity can be transformed into a high-tech industry with low energy consumption and high output, thus promoting the improvement of LUE in the lower reaches. The impact of globalization on LUE was not significant in the whole basin, the upper reaches, and the middle reaches. The middle and upper reaches of the Yellow River Basin were located in China’s inland regions, with relatively backward economic development and relatively low degree of opening to the outside world, leading to the insignificant impact of globalization on LUE. Moreover, the middle and upper reaches of the Yellow River Basin are one of the regions with fragile ecological environment and the most serious water and soil loss in China, and also the regions with a high proportion of restricted and prohibited development areas in China. Therefore, the local government has a strong awareness of land resources protection and strict control over land development, which is not conducive to attracting foreign enterprises to carry out local investment, leading to the insignificant impact of globalization on LUE.
• Marketization had a positive impact on the LUE of the whole basin and the lower reaches and a negative impact on the LUE of the middle reaches, but had no significant impact on the LUE of the upper reaches. In the lower reaches of the Yellow River Basin, the improvement of globalization was conducive to improving regional LUE. In the middle reaches of the Yellow River Basin, the increase of globalization was not conducive to improving regional LUE. In the upper reaches of the Yellow River Basin, the impact of globalization on regional LUE was not obvious. Specifically, the impact coefficients of marketization on LUE in the whole basin and lower reaches were 0.3194 and 0.2173, respectively, and both had passed the significance test at the level of 10%, which showed that every 1% increase in the degree of marketization will promote the LUE in the whole basin and lower reaches to increase 0.3194% and 0.2173%, respectively. First of all, marketization can stimulate domestic demand, effectively stimulate local consumption and investment, promote regional economic growth, and thus promote the improvement of LUE in lower reaches and the whole basin. Secondly, land marketization is conducive to promoting the transfer of rural population to urban areas, increasing the number of secondary and tertiary industry employees, thereby increasing urban economic output, and improving the LUE of lower reaches and the whole basin. Finally, land marketization is conducive to attracting domestic and foreign enterprises to invest in development and construction in the Yellow River Basin, promoting regional economic growth and efficient use of land resources, and thus promoting the improvement of LUE in the Yellow River Basin. The impact coefficient of marketization on LUE in the middle reaches was −0.1626, and had passed the significance test at the level of 5%. It showed that every 1% increase in the degree of marketization will reduce the LUE in the middle reaches by 0.1626%, which may be attributed to the extensive development and utilization of land resources in the middle reaches during the process of marketization, and the lack of attention to ecological environment protection, leading to the reduction of LUE in the middle reaches. The impact of marketization on LUE in the upper reaches was not significant, which may be attributed to the relatively lagging economic development in the upper reaches and the low degree of marketization. Moreover, local land development pays more attention to ecological environment protection, resulting in the insignificant effect of marketization on LUE.
• Decentralization had a positive impact on the LUE of the whole basin and the upper reaches, a negative impact on the LUE of the lower reaches, and an insignificant impact on the LUE of the middle reaches. In the upper reaches of the Yellow River Basin, the improvement of decentralization was conducive to improving regional LUE. In the lower reaches of the Yellow River Basin, decentralization was not conducive to improving regional LUE. In the middle reaches of the Yellow River Basin, the effect of decentralization on regional LUE was not obvious. Specifically, the impact coefficients of decentralization on the LUE of the whole basin and the upper reaches were 0.1982 and 0.3243, respectively, and both had passed the significance test at the level of 5%, indicating that every 1% increase in the degree of decentralization will promote the LUE of the whole basin and the upper reaches to increase by 0.1982% and 0.3243%, respectively. The central government has delegated its power to local governments. Under the decentralized system, local governments have more financial autonomy, attract investment from domestic and foreign enterprises through investment promotion and other ways, promote regional population employment and economic growth, and improve the LUE in the whole basin and upper reaches. The impact coefficient of decentralization on LUE in the lower reaches was −0.4062, and it had passed the significance test at the level of 5%. It showed that every 1% increase in the degree of decentralization will reduce the LUE in the lower reaches by 0.4062%. This may be attributed to the fact that Shandong and Henan, which are located in the lower reaches of the Yellow River, are the major economic provinces in the Yellow River Basin. In order to achieve rapid local economic growth in the short term, the local government has promoted investment attraction by lowering land prices and other ways, resulting in the increase of idle land and inefficient use of land resources, which has led to the reduction of LUE in the lower reaches of the Yellow River Basin. Decentralization had no significant impact on LUE in the middle reaches. This may be attributed to the fact that Shanxi and Shaanxi, which are located in the middle reaches, are big coal resource provinces. In the process of economic transformation, local governments not only have to face the economic transformation dilemma of long-term over dependence on local resources for economic development, but also face the difficulty of finding new driving forces that can promote local green and efficient development. When faced with land resource development, local governments often show a cautious attitude, leading to the effect of decentralization on LUE is not significant in the middle reaches.
• Control variable analysis. The industrial structure had a positive impact on the LUE of the whole basin and the middle reaches. The proportion of heavy chemical enterprises in the middle reaches is relatively high. The upgrading of industrial structure can promote the transfer of labor force from the primary industry to the secondary and tertiary industries, increase local economic output, and improve the urban LUE. The impact of industrial structure on LUE in the upper reaches and the lower reaches was not significant. The impact of economic development level on LUE in the Yellow River Basin was not significant, which may be attributed to the fact that regional economic development had not only promoted regional economic growth, but also caused waste of land resources and destruction of ecological environment, leading to the insignificant impact of economic development on LUE. Urbanization had a negative impact on the LUE of the whole basin, the middle reaches and the lower reaches, which may be attributed to the blind expansion of urban built-up areas and extensive land management caused by rapid urbanization, leading to the decline of urban LUE. The impact of urbanization on LUE in the upper reaches was not significant, which may be attributed to the relatively low level of urbanization in the upper reaches. The impact of industrial scale on LUE in the Yellow River Basin was not significant, which may be attributed to the small impact of the average industrial output value on urban land use. The population size had a negative impact on the LUE of the whole basin, the middle reaches, and the lower reaches. This may be attributed to the dense population distribution in the middle and lower reaches, the increase of population size, which is more likely to lead to blind expansion of urban construction land, the compression of urban green space area, resulting in waste of land resources and reduction of ecological benefits. The impact of population size on LUE in the upper reaches was not significant, which may be attributed to the small population distribution in the upper reaches and the high proportion of restricted development zones, resulting in a small impact of population size on LUE in the upper reaches.

5. Discussion

5.1. Regional Differences of LUE in the Yellow River Basin

From 2009 to 2018, the regional difference of LUE in the Yellow River Basin was as follows: the upper reaches > the middle reaches > the lower reaches, which was consistent with the research results of Jiang et al. [6]. The hot spots and cold spots of urban LUE in the Yellow River Basin were generally distributed stably. The hot spots were mainly distributed in Ordos in the upper reaches and Yulin in the middle reaches, while the cold spots were mainly distributed in Henan and the southwest of Shandong in the lower reaches. The average value of LUE in the upper reaches was the highest, while that in the lower reaches was the lowest. The average value of LUE in the middle reaches was the closest to that in the whole basin. This was not consistent with the regional differences of economic development in the Yellow River Basin. This may be attributed to the fact that in the evaluation of LUE, the output indicators not only take into account the economic benefits of the region, but also fully consider the environmental and social benefits of the region. The upper reaches of the Yellow River Basin are relatively vulnerable to ecological environment. The local government pays attention to ecological environment protection and has strict control over land resource development, which makes the ecological benefits of land use in the upper reaches higher. In addition, although the total GDP of the upper reaches is not high, the population distribution is sparse, the per capita GDP is high, leading to better social benefits in the region [27,35]. For instance, in 2018, the per capita GDP of Ordos in the upper reaches reached 160,200 yuan, far higher than the national average of 65,500 yuan. Henan and Shandong, which are located in the lower reaches, are big agricultural and industrial provinces, with a high proportion of the primary and secondary industries. Owing to the primary and secondary industries needing to consume a lot of labor and the product added value not being high, the regional economic benefits are not high. Moreover, the large distribution of secondary industry in the lower reaches, which is mainly heavy chemical enterprises, the construction of enterprise sites needs to occupy a large number of rural land resources, resulting in low ecological benefits of land resources in the lower reaches.

5.2. The Impact of Economic Transformation on LUE

Globalization had a positive impact on LUE in the lower reaches, which was consistent with the research results of Zhang et al. [42]. Located in the lower reaches of the Yellow River Basin, Shandong, as a coastal province in eastern China, is highly open to the outside world. Qingdao and Yantai in Shandong Province are among the first coastal open cities in China, playing an important role in promoting economic globalization. In addition, Shandong is close to Japan and South Korea, which is conducive to attracting Japanese and South Korean enterprises to invest, promoting local economic and social development, and thus helping to improve the economic and social benefits of land resources. Marketization had a positive impact on the LUE of the whole basin and lower reaches, which was consistent with some previous research conclusions [21,46]. In 1978, after the reform of the economic system, China’s economy experienced a major transformation from a planned economy to a market economy. Market economy can promote the marketization of land resources, which can stimulate regional consumption and investment, improve residents’ lives, promote economic growth, and thus realize the economic and social benefits of land use. Marketization had a negative impact on LUE in the middle reaches, which was consistent with the research results of Jiang et al. [6]. Shanxi and Shaanxi, which are located in the middle reaches of the region, are large provinces of coal resources. They have long relied on coal resources to develop local economy, with a high proportion of state-owned economy. Under the background of marketization, the development of state-owned economy is relatively slow, and the economic benefits generated by local land development are low, leading to the negative effect of marketization on LUE. Decentralization had a positive impact on the LUE of the whole basin and the upper reaches, which was consistent with the research results of Mu [51]. Under the decentralized system, local governments have more financial autonomy, which is conducive to improving the efficiency of government allocation of local land resources [50]. The upper reaches are located in the inland region and has a low degree of opening to the outside world. Local governments play an important role in promoting regional economic and social development, which is conducive to balancing the economic, social and ecological benefits generated by land use. Decentralization had a negative impact on LUE in the lower reaches, which was consistent with the research results of Li and Zhou [53]. The lower reaches are highly globalized and market-oriented. In order to attract more domestic and foreign enterprises, local governments have weak control over land resource development. While seeking economic benefits of land resources, local governments often neglect social and ecological benefits, leading to a decline in LUE.

6. Conclusions and Suggestions

6.1. The Main Conclusions

The improvement of LUE can promote the intensive use of land resources, which is conducive to promoting the ecological protection and high-quality development of the Yellow River Basin. This study employed a Super-SBM model to calculate the LUE of cities in the Yellow River Basin from 2009 to 2018, and to analyze their temporal evolution characteristics and regional differences. In addition, the influencing factors of LUE in the Yellow River Basin were analyzed by building a theoretical framework for the relationship between economic transformation and land use. The main conclusions were as follows:

• From 2009 to 2018, the LUE of the Yellow River Basin showed an evolutionary trend of fluctuation and decline. The evolution of LUE in the Yellow River Basin could be divided into three stages. The first stage was the dramatic evolution stage from 2009 to 2013, and the LUE showed an M-shaped evolution trend. The second stage was the stable evolution stage from 2013 to 2016. The third stage was the dramatic evolution stage from 2016 to 2018, and the LUE showed an inverted V evolution trend.
• From 2009 to 2018, the regional difference of LUE in the Yellow River Basin was generally shown as follows: the upper reaches > the middle reaches > the lower reaches. The northwest of the Yellow River Basin was a high value area of LUE, while the southeast was a low value area of LUE. The spatial distribution of hot spot and cold spot of LUE was relatively stable. The hot spots were mainly distributed in Ordos in the upper reaches and Yulin in the middle reaches, while the cold spots were mainly distributed in Henan and the southwest of Shandong in the lower reaches.
• Globalization had a positive impact on LUE in the lower reaches, while it had no significant impact on the whole basin, the upper reaches, or the middle reaches. Marketization had a positive impact on the LUE in the whole basin and the lower reaches and a negative impact on the LUE in the middle reaches, but had no significant impact on the LUE in the upper reaches. Decentralization had a positive impact on the LUE in the whole basin and the upper reaches, a negative impact on the LUE in the lower reaches, and no significant impact on the LUE in the middle reaches. The industrial structure had a positive impact on the LUE in the whole basin and the middle reaches. Urbanization had a negative impact on LUE in the whole basin, the middle reaches, and the lower reaches. The population scale had a negative impact on the LUE in the whole basin, the middle reaches, and the lower reaches. Economic development level and industrial scale had no significant impact on LUE in the Yellow River Basin.

6.2. Policy Suggestions

According to the research results of LUE in the Yellow River Basin, this study proposes some suggestions are as follows:

• For local governments, they should optimize land development policies and improve the efficiency of land resource allocation. Local governments should change the “spread the pie” mode of urban development, optimize land development policies, and promote the transformation of land development mode from “incremental expansion” to “improving the quality of stock”. For instance, they can control the continuous spread of land in built-up areas by improving the utilization level of infrastructure and reusing abandoned areas around cities [34]. Moreover, local governments should give full play to the macro-control role of the land market, optimize the land use structure, and improve the allocation efficiency of land resources.
• For land developers, they should give full play to the leading role of the market and promote the intensification of land use. Land developers should give full play to the leading role of the market in land development, scientifically and rationally develop and use land resources, and maximize the economic benefits of land. Moreover, when developing land, land developers should strengthen the awareness of ecological environment protection, strictly implement land development policies and standards, and promote the rational and efficient use of land resources.
• For the upper reaches, local governments should strictly enforce land development policies and standards, optimize the land use structure, improve the allocation efficiency of land resources, and coordinate the economic, social, and ecological benefits of land resource development. For the middle reaches, they should get rid of the long-term dependence on local resources to develop the economy, and strive to find new drivers that can promote local green and efficient development, so as to smoothly realize the regional economic transformation. For the lower reaches, they should make full use of the positive role of globalization and marketization to improve the urban LUE. Meanwhile, local governments should play a macro-control role in the land market and coordinate the economic, social, and ecological benefits of land development.

In the long run, we can try to improve from the following aspects. First, urban LUE is affected not only by local socio-economic factors, but also by natural factors such as terrain, hydrology, and climate. We can further explore the impact of natural conditions on LUE. Second, the Yangtze River Basin and the Yellow River Basin are both major national strategic regions, and a comparative study of LUE can be carried out for these two regions in the future. Third, local policies have an important impact on urban land use, and subsequent studies can fully take local policies into account.