How to Evaluate the Green Utilization Efficiency of Cultivated Land in a Farming Household? A Case Study of Shandong Province, China

Abstract

The behavior of farming households is the most direct factor involved in the transition of cultivated land utilization from high-input/high-output to green and sustainable utilization mode. Improving farming households’ green utilization efficiency of cultivated land (GUECL) is of great significance in facilitating agricultural green development in China. However, there are few studies on GUECL based on the micro-perspective of farming households that cover the comprehensive benefits to the economy, ecology, and society. This paper builds a theoretical analysis framework of farming households’ green utilization of cultivated land and uses the super-efficiency EBM model and a questionnaire to conduct an empirical analysis of 952 farming households in Shandong Province to evaluate the green utilization efficiency of cultivated land. The results show that the GUECL of the farming households is generally not high, with an average value of 0.67, and can be further improved. The higher the GUECL, the lower the input and undesired output per unit yield and per unit output value. Tobit regression results show that a farming household’s per capita income is significantly positively correlated with the GUECL, while agricultural insurance, agricultural subsidies, cultivated land scale, cultivated land fragmentation, and regional economic level are significantly negatively correlated with the GUECL. In addition, recommendations can be made on promoting and innovating agricultural green development technology, popularizing and publicizing farming households’ thoughts on the green utilization of cultivated land, and ensuring and improving rural green life so as to provide a reference for promoting green transition of cultivated land utilization with diversified coordination and multiple measures.

1. Introduction

As the world’s population is increasing and real problems, such as climate change and scarcity of natural resources, are emerging, food security is becoming an increasingly important global issue [1]. Although the projections of food demand in 2050 vary widely, there is no denying that we are facing an impending increase in food demand. Moreover, after the outbreak of COVID-19, countries are aware of the urgent need to enhance food security [2]. Producing more food to feed the growing global population will require more cultivated land resources [3].

Cultivated land, as an important material basis of agricultural production, is important for ensuring national food security, ecological security, and social stability [4]. Therefore, effective cultivated land utilization is the key factor affecting regional sustainable development and food security [5]. The large-scale use of modern agricultural production chemicals, such as chemical fertilizers and pesticides, promotes increase in crop production, but also intensifies agricultural non-point source pollution. Due to limited cultivated land resources and severe food security challenges, human beings can no longer rely on the traditional cultivated land utilization mode of disorderly increase in input and must adopt more creative and technologically advanced methods to minimize the consumption of natural resources. Therefore, green transition of cultivated land utilization needs urgent attention. Especially for a populous country such as China, the transition is even more urgent.

In recent years, the core task for China in promoting green transition of cultivated land utilization is to reduce investment in chemicals such as pesticides and fertilizers, aiming to protect cultivated land and ecological environment, cut off the source of “unsafe” production factors [6], and effectively promote the transition of cultivated land in China from an economic output utilization basis to a green and sustainable one. Cultivated land utilization, as the core link in agricultural production, is essentially a composite system of land natural subsystem and land social economic subsystem, coupled with the human subsystem as the link and connection. Farming households’ behavior plays a key role in the system and is the most direct factor involved in the transition of cultivated land to a green and sustainable utilization mode. Farming households are the main body involved in agricultural production activities and the most basic micro-decision-making unit (DMU), and huge differences in cultivated land utilization exist in different farming households [7]. Farming households’ cultivated land utilization behavior is an important factor affecting the green development of cultivated land, and the key to promoting agricultural green development in China is to improve farming households’ green utilization efficiency of cultivated land (GUECL).

In recent years, much research has been carried out on the utilization efficiency of cultivated land at the macrolevel, mainly using panel data; selecting land, capital, and labor as input indicators [8] and agricultural gross output value as output indicators; analyzing regional differences [9,10]; and studying influencing factors, such as cultivated land endowment, economic conditions, natural conditions, and agricultural production conditions [11,12,13]. According to different research purposes, the methods adopted mostly involve stochastic frontier analysis (SFA) [14,15], data envelopment analysis (DEA) [16,17], etc. Since the 1990s, when the concept of eco-efficiency was put forward, researchers have applied eco-efficiency in different research fields, mostly industry and regional studies [18]. They have paid less attention to the eco-efficiency of agricultural land use. Some of the literature dealt with farm eco-efficiency [19]. Generally speaking, in the existing relevant literature, most index systems consider the economic and ecological benefits of the process of cultivated land utilization but do not pay attention to social benefit indices. In addition, there are few studies from the perspective of farming households. Some of the research results from the perspective of farming households are mainly analyzed considering the impact of a single factor on the utilization efficiency of cultivated land [5,20]. There is no final conclusion about the concept and implications of the green utilization of cultivated land; a few studies have focused on the macrolevel of 30 provinces [21] in China and the Yellow River Basin [22], but there are few studies on the GUECL at the microscale of farming households.

Accordingly, this paper aims to (1) build a theoretical analysis framework of the green utilization of cultivated land based on the micro-perspective of farming households and to construct an evaluation index system by comprehensively considering economic, social, and ecological benefits; (2) analyze the field survey data of 952 farming households in Shandong Province and use the super-efficiency EBM model based on undesirable output to measure farming households’ GUECL; (3) use the Tobit model to explore the influencing factors of farming households’ GUECL; and (4) draw recommendations, providing support for guiding farming households to use cultivated land in a reasonable and effective way, fundamentally reducing pollution, promoting the green transition of cultivated land utilization at the level of farming households, and improving the development of an ecological society.

2. Materials and Methods

2.1. Data Sources

The data used in this paper came from a series of questionnaires answered by Shandong farming households. The survey was carried out by the Qufu Normal University Research Group in July–August 2017, 2018, and 2020. The survey mainly included 5 cities (Dongying, Taian, Rizhao, Linyi, and Jining) in Shandong Province, involving different geomorphic types (Figure 1). We randomly selected 1 or 2 counties (urban areas), to make a total of 8 counties, from each city; 2 or 3 natural villages from each county; and a number of farming households from each village in order to conduct questionnaire interviews. The questionnaire included questions on farming households’ family structure, agricultural production activities, environmental cognition, and social security and other aspects, which fully reflect the land utilization and rural development of the farming households in the study area. A total of 985 questionnaires were distributed. After sorting out and excluding invalid questionnaires, a total of 952 valid questionnaires were obtained (46 in Linyi, 342 in Rizhao, 121 in Taian, 110 in Dongying, and 333 in Jining), with a questionnaire effectiveness rate of 96.65%.

2.2. Analysis Framework and Research Method

2.2.1. Analysis Framework

The green utilization of cultivated land is not only to carry out input–output activities according to the traditional cultivation mode or to achieve more agricultural output in terms of quantity and quality for less input of land, water, nutrients, energy, labor, or capital [23]. Rather, it is to further combine the technology currently in use with modern green technology, optimize the allocation of elements’ input, thoroughly implement the concept of green development, reduce the input of external harmful substances and the generation of agricultural wastes, ensure the process of cultivated land utilization without environmental pollution and excessive use of natural resources, and improve the utilization efficiency of cultivated land resources and the safety level of agricultural products. The green utilization of cultivated land not only pays attention to ecological benefits rather than economic benefits but also emphasizes the unity of economic, social, and ecological benefits and balanced and coordinated development. Sustainable development theory [24], circular economy theory [25], ecological agriculture theory [26], land-use-related research [27], etc., all provide reference ideas for a cultivated land green utilization mode and its evaluation.

Theories of farming households’ behavior can be divided into the following three schools:

(1)

The rational smallholder school represented by Schultz [28], which pursues the maximization of profits.

(2)

The moral smallholder school represented by Chayanov [29]. The main purpose of farming households’ production is to maintain their livelihood, that is, to pursue the minimum risk in production rather than maximize the benefits.

(3)

The comprehensive smallholder school represented by Huang [30]. This school synthesizes the three views: rational smallholder, moral smallholder, and the class smallholder of Marx. It considers farming households as semi-proletarian agricultural producers, not entirely maximum-profit pursuers in the sense of Schultz, nor livelihood producers in the sense of Chayanov.

Due to the different research objects, research methods, and historical stages, their conclusions are also different. Some researchers believe that, with the continuous transformation of China’s social economy, farming households’ behaviors are also changing. Rational and irrational production behaviors coexist, resulting in farming households following different cultivated land utilization approaches [31].

Farming households’ behaviors are responses to external economic signals when they make choices within the rural social and economic environment [32]. As farming households are in a fluctuating market environment, their production decisions will also change along with production costs, product prices, policy evolution, etc. Under the joint action of various constraints, farming households constantly revise and improve their own business objectives to maximize interests and minimize their losses. From society’s perspective, farming households’ cultivated land utilization for production and life is also related to social stability. Moreover, farming households produce a large amount of environmental pollution wastes, or pesticide and fertilizer residues and volatilization in the process of agricultural production. If farming households’ ecological protection awareness and relevant policy publicity are not improved, serious environmental pollution problems will be caused. Therefore, the decision-making regarding green utilization of cultivated land is a choice for farming households in order to avoid risks as much as possible and to pursue the maximization of benefits according to their own conditions, while considering the existing resources and family conditions, relying on their existing cognition, and this will be comprehensively affected by internal and external elements. The behavior of farming households regarding green utilization of cultivated land is an input–output activity produced by farming households’ decision-making, with the main goal of improving family income, maintaining social stability and reducing environmental pollution (Figure 2).

2.2.2. Variable Selection

From the perspective of input–output activity, the input index of cultivated land utilization mainly includes three factors: land, capital and labor. Specifically, this paper selected the cultivated land area operated by farming households as the measurement index of land input; the agricultural production expenditure per unit of cultivated land area of farming households interviewed in that year as the measurement index of capital input (capital expenditure of pesticides, fertilizers, seeds, irrigation water, agricultural machinery lease); and the actual labor input in production as the measurement index of labor input. Output indicators included desirable output and undesirable output. Desirable output involved both economic and social aspects. As farming households produce different varieties of agricultural products, and their agricultural income can better reflect their agricultural output benefits, this paper selected the agricultural income of farming households as the desirable output variable. All money values are uniformly converted to 2019 values. From society’s perspective, the green utilization of cultivated land is related to the living standard of farming households and social stability. To ensure food security, grain yield was also included in the desirable output index. In addition, considering the negative impact of cultivated land utilization on the environment, the undesirable output index included agricultural carbon emission and agricultural non-point source pollution emission. Agricultural carbon emission refers to the greenhouse gas emission directly or indirectly caused by farming households during agricultural production and livestock breeding. The carbon emission estimation formula is

E = ∑E_i = ∑T_i·δ_i,

where E is the total carbon emission of agriculture, E_i is the carbon emission of each carbon emission source, T_i is the quantity of each carbon emission source, and δ_i is the carbon emission coefficient of each carbon emission source; refer to the research of [33] for the carbon emission coefficient value of each carbon emission index.

Agricultural non-point source pollution emission is mainly caused by the excessive use of chemical fertilizers, pesticides, and agricultural film. This paper used the amount of chemical fertilizers, pesticides, and agricultural film to estimate the level of agricultural non-point source pollution. The amount of chemical fertilizer pollution is calculated as

Amount of chemical fertilizer application × (1-chemical fertilizer utilization rate).

The amount of pesticide pollution is calculated similarly. The amount of agricultural film pollution refers to the residual amount of agricultural film. Combined with existing research [34,35], the chemical fertilizer utilization rate, the pesticide pollution rate, and the agricultural film residual rate were calculated at 35%, 50%, and 10%, respectively (

Table 1. Variable definition and descriptive statistics (n = 952).

Item	Variable	Meaning and Measurement	Mean	Standard Deviation
Inputs	Cultivated land input	Cultivated land area operated (mu)	15.388	72.533
	Capital investment	Expenditure on fertilizers, pesticides, agricultural film, seeds, machinery, and irrigation (yuan/mu)	955.480	4033.952
	Labor input	Family laborer (person/mu)	0.498	0.547
Outputs	Agricultural output value	Annual agricultural income of farming households (yuan/mu)	2343.202	9200.242
	Grain yield	Yield of food crops (kg/mu)	848.169	892.935
	Carbon emission	Carbon emissions from fertilizers, pesticides, agricultural film, plowing and irrigation (kg)	1461.500	5009.913
	Pollution emission	Pollution emission from fertilizers, pesticides, plastic film, etc. (kg)	967.929	3432.726
Individual characteristics of farming households	Age	Actual age (years)	59.220	10.25
	Gender	Male = 1; female = 0	0.639	0.481
	Education level	0 = illiterate; 1 = primary school; 2 = junior high school; 3 = high school; 4 = secondary school; 5 = university	1.431	1.019
Family characteristics of farming households	Household income per capita	Ratio of total household income to total population (yuan)	20,009.108	39,221.409
	Proportion of agricultural income to the total household income	Ratio of agricultural income to total household income	0.428	0.405
Policy factors	Agricultural subsidies	0 = no; 1 = yes	0.644	0.479
	Agricultural insurance	0 = no; 1 = yes	0.592	0.492
Cognitive characteristics of farming households	Cognition of environmental pollution by fertilizers	1 = no; 2 = it doesn’t matter; 3 = yes	2.070	0.956
	Cognition of environmental pollution by pesticides	1 = no; 2 = it doesn’t matter; 3 = yes	2.617	0.761
Cultivated land conditions	Scale of cultivated land	Cultivated land area (mu)	15.388	72.533
	Transfer scale of cultivated land	Ratio of the transferred cultivated land area to the total cultivated land area	0.228	0.677
	Degree of fragmentation	Number of family cultivated land plots (blocks)	4.767	6.204
Regional economic characteristics	Regional economic development level	1 = Linyi City; 2 = Taian City; 3 = Jining City; 4 = Rizhao City; 5 = Dongying City	3.373	0.999

).

Referring to the relevant literature [5,36], this paper divided the factors that affect farming households’ GUECL into six variable groups:

(1)

Farming households’ individual characteristic variables: as the main body of cultivated land utilization, farming households’ characteristics have a direct impact on the level of GUECL. In this paper, the characteristics of a farming household were the gender, age, and education level of the head of the household, which have different effects on the household’s production decision-making behavior. The ability of male farming households to accept new things was expected to be higher than that of female farming households and it was considered that gender has a positive impact on farming households’ GUECL. The younger the subject and the higher their education level, the stronger the ability to accept new technologies and new factors, and the bigger the capability of improving the GUECL.

(2)

Family characteristic variables: in this paper, the characteristics of farming household families were the proportions of per capita income and agricultural income to the total income of the families. In general, the higher a farming household’s per capita income, the more able the household is to increase investment in cultivated land, for example, by purchasing large-scale agricultural machinery and mechanizing operations. Thus, it is expected that a farming household’s per capita income has a positive impact on the GUECL. The larger the proportion of farming households’ agricultural income, the more the family livelihood depends on cultivated land resources. Thus, farming households will cherish their cultivated land more, paying more attention to its sustainable development and green utilization, so we expected that the proportion of agricultural income to total household income would have a positive impact on the GUECL.

(3)

Policy factor variables: the influence of policy factors on the GUECL is also important. These variables included agricultural subsidies and agricultural insurance. Agricultural subsidies include direct subsidies for grain, high-class-seed and agricultural machinery, and comprehensive direct subsidy for purchasing agricultural supplies. In theory, agricultural subsidies and agricultural insurance increase farming households’ enthusiasm for planting crops, so it was expected that these two variables would play a positive role in improving the GUECL

(4)

Farming households’ cognitive characteristic variables: farming households’ cognition of chemical fertilizers and pesticides is also an important factor affecting their GUECL. This paper reflected farming households’ cognition through a survey of their perceptions of chemical fertilizers and pesticides pollution. The question asked was “Do you think chemical fertilizers/pesticides will pollute the environment?” In general, the clearer a farming households’ cognition of pollution by such chemical products and the better their understanding of the consequences of pollution, the higher the GUECL.

(5)

Cultivated land condition variables: in addition to individual and family characteristics, farming households’ cultivated land production efficiency may also be affected by cultivated land conditions, such as the cultivated land scale, cultivated land transfer scale and cultivated land fragmentation degree. In this paper, the cultivated land conditions of farming households were represented by cultivated land area, the ratio of the transferred cultivated land area to the total cultivated land area, and the number of household cultivated land plots. The cultivated land scale, cultivated land transfer scale, and cultivated land fragmentation degree were expected to have a negative impact on the GUECL.

(6)

Regional economic characteristic variables. From the perspective of cultivated land resources, the level of regional economic development has a close relationship with the production efficiency of cultivated land in the region. This paper selected the per capita GDP ranking to measure the level of regional economic development. Compared with the total GDP, the per capita GDP is more representative of the real economic development level. Due to the difference in economic development levels among different regions, the allocation of agricultural production resources and the green utilization of cultivated land are also different. The economic development level was expected to have a positive impact on the GUECL.

2.2.3. Super-Efficiency EBM Model and Tobit Model

In this paper, data envelopment analysis (DEA) was used to measure the green utilization efficiency of cultivated land of the farming households interviewed in Shandong Province, and the Tobit model was used to analyze the influencing factors on green utilization efficiency of cultivated land of farming households.

Classical data envelopment analysis (DEA) models can be divided into two types, the radial model (the CCR model [37]) and the BCC model [38]. However, because the radial measurement model does not include slack variables in the inefficiency measurement of the decision-making unit (DMU), Tone [39] also proposed another type of model, the non-radial model; this, however, minimizes the measured efficiency value and has some other shortcomings. Therefore, Tone and Tsutsui [40] constructed a hybrid model, the Epsilon-based measure (EBM) model, which contains both radial and non-radial distance functions. This model is divided into three types of oriented models: input oriented, output oriented, and non-oriented. The input oriented EBM has been proposed as the following linear programming model:

$$\begin{gathered} {\rho }^{*}=min \theta -{\epsilon }_x{\displaystyle \sum }_{i=1}^m\frac{w_i^{-}{s}_i^{-}}{x_{io}} \\ \mathrm{s}.\mathrm{t}. {\theta x}_{iO}-{{\displaystyle \sum }}_{j=1}^n{x}_{\mathrm{ij}}{\lambda }_j-s_i^{-}=0 i=1,2,\dots \dots m \\ {{\displaystyle \sum }}_{j=1}^n{y}_{rj}{\lambda }_j\ge y_{ro} r=1,2,\dots \dots s \\ {\lambda }_1+{\lambda }_2+\dots \dots +{\lambda }_n=1 \\ {\lambda }_j\ge 0,s_i^{-}\ge 0 \end{gathered}$$

(1)

where ${\rho }^{*}$ is the efficiency value for the EBM model in which there are n DMUs to be measured; $\theta$ is the radial programming parameter; $x_{ij},y_{rj}$ represent the ith input and rth output of DMU_j; m, s denote number of inputs and outputs, respectively; $w_i^{-}$ represents the weight of input i and meets the condition ${{\displaystyle \sum }}_{i=1}^n{w}_i^{-}$ = 1; ${\epsilon }_x$ is the significant parameter in range (0, 1), which determines whether the model pertains to the radial model (${\epsilon }_x$ = 0) or non-radial model (${\epsilon }_x$ = 1); $s_i^{-}$ represents the slack variable of input i; $\lambda$ is the DMU radial combination coefficient; and the subscript “o” denotes the DMU to be measured. The EBM model specification considering the undesirable output relies on the work of Du et al. [41] and Fan et al. [42] as follows:

$$\begin{gathered} {\gamma }^{*}=min \theta -{\epsilon }_x{\displaystyle \sum }_{i=1}^m\frac{w_i^{-}{s}_i^{-}}{x_{io}} \\ \mathrm{s}.\mathrm{t}. {{\displaystyle \sum }}_{j=1}^n{x}_{ij}{\lambda }_j+s_i^{-}={\theta x}_{iO} (i=1,2,\dots \dots m) \\ {{\displaystyle \sum }}_{j=1}^n{y}_{rj}{\lambda }_j\ge y_{ro} (r=1,2,\dots \dots s) \\ {{\displaystyle \sum }}_{j=1}^n{b}_{pj}{\lambda }_j=b_{po} (p=1,2,\dots \dots q) \\ {\lambda }_1+{\lambda }_2+\dots \dots +{\lambda }_n=1 \\ {\lambda }_j\ge 0,s_i^{-}\ge 0 \end{gathered}$$

(2)

where ${\gamma }^{*}$ is the efficiency when considering undesirable outputs; $b_{pj}$ represents the pth undesirable output of DMU_j; and q denotes the number of undesirable outputs.

Since the effective unit efficiency value measured by the EBM model is 1, it is difficult to further analyze the efficiency differences of the effective evaluation units. Therefore, Andersen et al. [43] proposed a super-efficiency model to solve this problem. The super-efficiency EBM model not only overcomes the shortcomings of the traditional DEA model and its derivative methods, but also inherits the advantages of these models. Therefore, the super-efficiency EBM model can reflect more real and effective cultivated land utilization information. This paper used an input-oriented super-efficiency EBM model that includes undesired outputs to measure the green utilization efficiency of cultivated land of farming households, and the efficiency value can be greater than 1.

The Tobit model is suitable for analyzing the influencing factors of efficiency, and can help identify ways to improve it. The general expression is:

$$\begin{gathered} Y_i^{*}=\beta X_i+{\epsilon }_i,i=1,2,\dots \dots ,n \\ {Y}_i=Y_i^{*}, if Y_i^{*}>0 \\ {Y}_i=0, if Y_i^{*}\le 0 \end{gathered}$$

(3)

where $Y_i$ represents the green utilization efficiency of cultivated land of farming households, $X_i$ is the independent variable of influencing factors, $\beta$ is the regression coefficient, ${\epsilon }_i$ is the random error term, and $Y_i^{*}$ is the latent variable.

2.3. Descriptive Statistics

Table 1 gives the descriptive statistical results of each variable. According to the survey results, the minimum age of the interviewees was 29 years and the maximum age was 85 years, with the average age being 59 years. Of these, 32.53% were 51–60 years old and 47.18% were over 60 years old. The ratio of men to women was about 1.8:1. The education level of the interviewees was generally low (mainly in primary and junior high schools), accounting for 69.98% of the total interviewees, and only 10.88% of the interviewees had a high school degree or above. There were more farming households with an annual per capita income below 10,000 yuan and between 10,000 and 25,000 yuan, accounting for 41.39% and 38.03% of the total, respectively. The mean value of ratio of agricultural income to total household income was 42.76%, reflecting that agricultural income is still an important part of most households. About half of the farming households thought that fertilizers would pollute the environment, while 78.80% thought that pesticides would pollute the environment. Compared with awareness of pesticide pollution, farming households’ awareness of fertilizer pollution was weak. The sample of households’ cultivated land fragmentation was common, with an average of 4.76 pieces of cultivated land per household. Only 170 farming households operated one piece of cultivated land, accounting for 17.83% of the total; most operated two to five pieces of cultivated land; the largest number of pieces of cultivated land was 100. Regarding agricultural subsidies, 64.32% of the farming households enjoyed agricultural subsidies and 59.18% of farming households had purchased agricultural insurance.

3. Results

3.1. Analysis of the Farming Households’ GUECL

This paper applied MaxDEA software to calculate the GUECL based on the survey data of farming households. The maximum GUECL value of the sample farming households was 3.12, the minimum was 0.24, and the average was 0.67. The efficiency value was generally not high. Among the sample farming households, 37.29% had a higher GUECL than the average level, indicating that most of the farming households’ input structure is unreasonable and the green utilization of cultivated land has not been realized.

Efficiency is a relative concept. At present, there is no consistent classification standard in academic circles. To observe the distribution characteristics and laws of efficiency of the calculated results more clearly, according to the level of GUECL of sample farming households, the measured GUECL of 952 farming households was divided into three levels: >=0 and <0.5, low-GUECL group; >=0.5 and <1, medium-GUECL group; and >1, high-GUECL group (

Table 2. Distribution of the GUECL of farming households.

GUECL Level	Low-GUECL Group	Medium-GUECL Group	High-GUECL Group
GUECL value interval	0 ≤ γ* < 0.5	0.5 ≤ γ* < 1	γ* ≥ 1
Number of farming households	107	769	76
Proportion	11.24	80.78	7.98

). Table 2 shows that 11.24% of the farming households had a low GUECL, with room for improvement in the green utilization of cultivated land; 80.78% had a medium GUECL, with 45.51% having a GUECL between 0.5 and 0.6; and 7.98% had a high GUECL, with relatively reasonable cultivated land utilization, which can realize the green utilization of cultivated land.

Results showed that the average GUECL of farming households in Rizhao is the largest (0.72), followed by Linyi, Taian, Jining, and Dongying (0.71, 0.66, 0.63, and 0.61, respectively). To better observe the green utilization of cultivated land of farming households in each of the surveyed cities, the proportion of the number of people at each GUECL level among the total surveyed people in the city was used for statistics (Figure 3).

As can be seen from Figure 3, medium-GUECL farming households in Dongying account for 76.36% of the total surveyed farming households in Dongying, and high-GUECL farming households account for 4.55%; the number of farming households in Tai’an with high GUECL, accounting for 10.74% of the total number surveyed in the city; in Rizhao, 85.47% and 9.88% of sampled farming households had a medium and a high GUECL, respectively; in Linyi, 9.09% farming households had a high GUECL; and in Jining, 15.62% farming households had a low GUECL. Overall, in all surveyed cities, the proportion of medium-GUECL farming households is the highest (more than 75%). Specifically, high-GUECL farming households in Tai’an account for a relatively high proportion, at more than 10%; Linyi has a relatively low proportion of low-GUECL farming households; and the proportion of high-GUECL farming households in Rizhao and Linyi is higher than that of low-GUECL farming households.

The GUECL can be understood as the maximum economic and ecological benefits generated by the lowest-cost input in the process of cultivated land utilization. From the average value of the required input and undesired output for each group of unit output per unit yield (

Table 3. Input and undesired output of yield per unit and output value per unit (mean).

	Yield per Unit			Output Value per Unit
	0 ≤ γ* < 0.5 n = 107	0.5 ≤ γ* < 1 n = 769	γ* ≥ 1 n = 76	0 ≤ γ* < 0.5 n = 107	0.5 ≤ γ* < 1 n = 769	γ* ≥ 1 n = 76
Cultivated land input	0.06060	0.01454	0.00523	0.02601	0.00585	0.00119
Capital investment	1.63765	0.94768	1.62382	0.70283	0.38099	0.36932
Labor input	0.00050	0.00062	0.00049	0.00021	0.00025	0.00011
Carbon emission	5.30890	1.45265	0.47130	2.27843	0.58401	0.10719
Pollution emission	3.50791	0.96792	0.28976	1.50550	0.38913	0.06590

), for farming households with a high green utilization efficiency of cultivated land, the capital investment is between low and medium efficiency for farming households, and the cultivated land input, labor input, carbon emissions, and non-point source pollution emissions are relatively low. The reason may be that farming households purchase agricultural machinery by increasing capital investment. On the one hand, this facilitates mechanized farming, reduces idle waste of cultivated land resources, and intensively uses cultivated land. On the other hand, the use of mechanization can reduce labor input and increase labor productivity. Farming households with a medium green utilization efficiency of cultivated land have lower capital investment and higher labor input. The reason may be that medium-GUECL farming households, on the one hand, have a low level of mechanization, and, on the other hand, are older, and the elderly are in relatively poor physical condition, so they spend more time in farming. Farming households with a low green utilization efficiency of cultivated land have relatively high capital input, carbon emission, and non-point source pollution emission. The reason may be that low-GUECL farming households buy and use a lot of chemical fertilizers and pesticides in order to increase crop yields, which, on the one hand, increases the cost of farming, resulting in a large capital investment, and, on the other hand, leads to environmental pollution.

For farming households, cultivated land utilization’s main purpose is to obtain economic benefits, and when economic efficiency meets the expectations of farming households, they are more likely to maintain or further optimize existing cultivated land utilization behavior and management methods, so as to promote improvement of the GUECL. Therefore, when comprehensively considering farming households’ judgments and choices regarding market economy demand, according to the required input and the undesired output per unit of output value, farming households with high green utilization efficiency of cultivated land have relatively low input and undesired output. High-GUECL farming households not only have a higher yield but also show a more diverse planting structure, usually a combination of food crops and cash crops, and have better control of market demand.

3.2. Analysis of Influencing Factors

The bootstrap method proposed by Simar and Wilson [44] can overcome the limitations of the Tobit model, thus making the regression results of influencing factors more reliable. Therefore, this paper used Stata 15.0 software and the bootstrap method to estimate by regression the influencing factors on the GUECL of farming households. The number of iterations selected was 500, and the model estimation results are shown in

Table 4. Tobit regression results of influencing factors of farming households’ GUECL (n = 952).

Variable	Observed Coefficient	Bootstrap Standard Error	z	p > |z|	Normal-Based (95% Confidence Interval)
Constant term	0.8608	0.0726	11.85	0.000	(0.7184279,1.003162)
Age	−0.0009	0.0007	−1.30	0.194	(−0.0023065,0.000468)
Gender	−0.0235	0.0156	−1.51	0.132	(−0.0541075,0.0070772)
Education level	−0.0064	0.0070	−0.93	0.354	(−0.0200943,0.0071958)
Annual household income per capita **	0.0000	0.0000	2.26	0.024	(0.000000362,0.0000051)
Proportion of agricultural income in the total household income	−0.0085	0.0290	−0.29	0.769	(−0.0653237,0.048303)
Agricultural subsidies ***	−0.0881	0.0180	−4.88	0.000	(−0.1234202,−0.0526843)
Agricultural Insurance ***	−0.0804	0.0138	−5.98	0.000	(−0.1073987,−0.0533469)
Cognition of environmental pollution by fertilizers	0.0076	0.0062	1.22	0.221	(−0.0045769,0.0198004)
Cognition of environmental pollution by pesticides	0.0106	0.0066	1.62	0.105	(−0.0022254,0.0235245)
Scale of cultivated land **	−0.0014	0.0006	−2.22	0.027	(−0.0026168,−0.0001612)
Transfer scale of cultivated land	0.0036	0.0105	0.35	0.729	(−0.0169991,0.0242839)
Degree of fragmentation ***	−0.0040	0.0013	−3.18	0.001	(−0.0064673,−0.0015345)
Regional economic development level ***	−0.0187	0.0067	−2.79	0.005	(−0.0318421,−0.0055677)

***, ** represent the significance levels of 10%, 5%, respectively.

.

In terms of farming household characteristic variables, the regression coefficient of the household per capita income was positive and the significance level of the statistical test was less than 5%, indicating a significant positive correlation between the household per capita income and the GUECL when other conditions are unchanged. The greater the increase in the household per capita income level, the more funds are used to learn green production technology and improve agricultural production conditions, and the more conducive the environment to promoting the improvement of the green production technology level, which can promote the GUECL to a certain extent.

In terms of policy factor variables, the regression coefficients of agricultural subsidies and agricultural insurance variables were both negative at the significance level of 1%, indicating a significant negative correlation between the variables and the GUECL. The increase in government subsidies for agricultural materials, such as pesticides, fertilizers, and plastic films, reduces the cost of purchasing. The excessive use of chemicals, such as pesticides and fertilizers, by farming households intensifies agricultural non-point source pollution and affects farming households’ green utilization of cultivated land. Farming households that purchase agricultural insurance are more risk-averse and more worried about crop yield; they often choose to apply more polluting inputs, such as chemical fertilizers and pesticides, which is not conducive to the green utilization of cultivated land.

In terms of cultivated land condition variables, the cultivated land scale and cultivated land fragmentation degree were significant, at levels of 5% and 1%, respectively, and the regression coefficient was negative, indicating that the cultivated land scale and cultivated land fragmentation degree are significantly negatively correlated with the GUECL. Farming households with a larger scale of cultivated land may ignore its green utilization. With the expansion of the cultivated land scale, farming households that rely on agricultural management to make a living are more willing to apply more fertilizers to improve yield and benefit. The expansion of the cultivated land scale does not refer to an increase in the number of cultivated land plots, because such a simple increase does not necessarily mean the cultivated land is centralized and continuous. A high degree of fragmentation makes high-efficiency mechanized operation difficult and also causes a certain degree of waste of resources, which is not conducive to farming households’ GUECL.

For regional economic characteristic variables, the regression coefficient was negative at the significance level of 1%, indicating a significant negative correlation between the regional economic level and the GUECL. In economically developed areas, chemical fertilizers and pesticides are used more, which has a greater impact on the environment and correspondingly reduces farming households’ GUECL.

4. Discussion

At present, storing grain in the ground is an important strategy in land use in China, and also an important aspect to promote agricultural green development. The core link of agricultural green development is the greening of the agricultural production mode. As the micro-actors and direct participants in agricultural production activities, farming households with a green production mode determine the level of agricultural green production to a large extent. At present, Chinese farming households’ enthusiasm for capital investment is often higher than that for labor investment and the investment level in fertilizers and pesticides is relatively high, which will inevitably have a certain impact on the ecological environment. The green utilization of cultivated land should mainly focus on saving cultivated land resources and improving the ecological environment considering the coordinated development of the economy, society, and environment as the goal. The green utilization behavior of cultivated land of farming households is related to the unity of economic, social, and ecological benefits. The investigation and analysis of farming households’ GUECL and its influencing factors will help guide farming households to use cultivated land in a reasonable and effective way, significantly reduce pollution, and provide a reference for sustainable and intensive cultivated land utilization.

The concept of eco-efficiency was put forward in the 1990s, which is consistent with the goals and implications of the GUECL mentioned in this paper. Eco-efficiency involves two aspects, ecology and economy. Some studies have proposed that farm eco-efficiency be measured as a ratio of economic value added to an aggregated indicator of environmental pressure [45]. Considering farming households’ green utilization efficiency of cultivated land studied in this paper, it focused on the micro-perspective of farming households to study the behavior and process of farming households’ cultivated land utilization, and pursued the coordinated development of the economy, society, and environment.

Currently, more attention has been paid to research on the influencing factors of the utilization efficiency of cultivated land. However, most of research is based on the regional perspective, and research from the perspective of farming households is relatively rare, with the investigation of, and empirical studies on, farming households’ GUECL, especially in Shandong Province, a large agricultural province, being even rarer. Based on the survey data of farming households in Shandong Province, this paper selected the following factors influencing farming households’ GUECL: individual characteristics, family characteristics, policy factors, cognitive characteristics, cultivated land conditions, and regional economic characteristics, and analyzed the significance level of each influencing factor. Except for individual characteristics and cognitive characteristics, all other factors impact farming households’ GUECL. Specifically, for family characteristics, some studies have shown that the higher the farming household per capita income, the stronger the ability of the household to obtain green production information, and thus the higher the knowledge literacy regarding green production [46]. Our results also show that household per capita income is significantly positively correlated with the GUECL.

Regarding policy factors, some studies [5] show that agricultural subsidies can not only enhance the upper limit of farming households’ capital investment in cultivated land but also promote cultivated land utilization efficiency and significantly enhance farming households’ enthusiasm and initiative in agricultural production. However, our results show that agricultural subsidies have a significant negative impact on the GUECL. To ensure food security and promoting the rapid development of agriculture and a stable increase in grain output, China has provided various subsidies to agricultural production for a long time. However, most of the agricultural subsidies fail to consider the impact on the environment and consider economic benefits rather than ecological environment protection as the goal, negatively affecting the GUECL. The aim of the green utilization of cultivated land is not to blindly pursue higher yield but to focus on resource conservation, environmental friendliness, and quality safety. For a long time, the objectives of agricultural development policies focused mostly on promoting agricultural development, increasing agricultural output, and ensuring food security and other aspects. In 2016, the Ministry of Finance and the Ministry of Agriculture jointly issued the Reform Plan for Establishing a Green Ecology-Oriented Agricultural Subsidy System, but this is still in the exploration, innovation, and experience summary stage and the improvement of policies needs to be continuously adjusted.

Regarding agricultural insurance, on the one hand insured farming households worried about their output, therefore increasing the use of chemical fertilizers, pesticides, and other chemicals in farming, affecting the GUECL. China’s agricultural insurance experiences some problems, such as inadequate claims and low level of security. In the process of investigation, only 59.5% of farming households have agricultural insurance. When uninsured farming households were asked for reasons for not having insurance, some said that they had had insurance before but had not obtained claims. Some farming households said that the amount of agricultural insurance coverage was too low and the diversification of products was insufficient to meet the needs of risk transfer. On the other hand, due to the insurance protection, some insured farming households may reduce their efforts in agricultural production and reduce disaster prevention and loss prevention measures. To be specific, farming households may change the farming system at will, not manage crops according to business norms, blindly introduce new varieties, select low-quality seeds or plant insufficiently, fertilize improperly, reduce investment in the prevention and control of crop diseases and insect disasters, and fail to harvest crops on time [47], indicating that agricultural insurance may be an obstacle to improving the GUECL.

Regarding cultivated land conditions, land in general is conducive to eco-efficiency as a larger utilized agricultural area facilitates the implementation of environmentally friendly investments (through a plow-less cultivation system, use of precision fertilization, etc.) [48]. However, our results show that the cultivated land scale is significantly negatively correlated with the GUECL, which is basically consistent with previous research [49,50]. A study [50] on the impact of farm scale on resource utilization efficiency showed that smaller farms have higher utilization efficiency, mainly because larger farms waste resources due to land fragmentation. The larger the cultivated land scale, the more the farming households’ livelihood inclined towards pure agriculture, the more the attention paid to agricultural economic output, and the more environmental pollution caused by increasing chemical input is ignored. This also reflects that, when mechanized large-scale management improves output, environmental benefits cannot be ignored simply to improve economic output. In addition, our results show that the cultivated land fragmentation degree is significantly negatively correlated with the GUECL. Rahma and Rahman [51] collected data by visiting farming households and found that land fragmentation significantly reduces technical production efficiency. Xu et al. [52] found that, for large- and middle-scale farming households, the impact of cultivated land fragmentation on its utilization efficiency is opposite to that for small-scale farming households; the former had a negative impact, while the latter had a positive impact. Small-scale farming households invest too much labor in replacing other factors, and they also ignore their own labor costs. However, the average area of plots of small-scale farming households is small and the plots are scattered, fragmentation hinders the development of new agricultural technology and mechanization, and the technical efficiency is obviously lower than that of middle- and large-scale farming households.

Regarding regional economic characteristics, some studies have shown that a higher regional economic level has a significant positive impact on the cultivated land utilization efficiency because more developed regions are relatively stronger economically and can put more funds into the improvement of agricultural production conditions. However, this paper showed that a higher regional economic level is significantly negatively correlated with the GUECL. The reason for this difference may be that this paper considered the undesirable output in cultivated land utilization. A high chemical fertilizer input [53] and a high undesirable output, such as carbon emission, in economically developed areas affect the GUECL.

In addition, some studies show that the elderly are attached to the countryside [54], and they rely more on cultivated land for their lives, so they cherish cultivated land more. Other studies show that older farmers are often known to have a limited vision and demonstrate less interest in current technology implementation [55].Young people with higher education, good health and training are better able to accept new factors [56], master green agricultural production technologies [57], and are more likely to choose green production methods. As for the other influencing factors on individual characteristic variables, education level does not play an important role in explaining the GUECL. This may be due to the fact that highly educated farming households have little connection with agricultural activities and that farming knowledge is a secondary source of their income. In addition, high education level does not mean that they have strong practical ability. Green utilization of cultivated land, as an agricultural production activity, needs practical experience rather than theoretical knowledge. Some studies have found that farming households’ cognition of the negative impact of pesticide application is the most important factor affecting the choice of green vegetable production behavior [58]. Farming households are the main body for production activities and the most basic decision-making unit, and the green development of a region cannot be separated from their support and cooperation. Green sustainable development is a rational behavior that requires people to have a higher level of cognition and consciousness. Farming households’ cognition of the environmental pollution caused by fertilizers and pesticides is low, resulting in a lack of high-level rational thinking in the process of cultivated land utilization and deviation from the overall goal of agricultural green development. In addition, some farming households are aware of the harm caused by fertilizers and pesticides but ignore this for short-term benefits [59]. In this paper, farming households’ individual characteristic and cognition of fertilizer and pesticide pollution were also selected as influencing factors, but the influence was not significant. For significance analysis of the influencing factors, in this paper only explanatory variables that have a significant impact on farming households’ GUECL were obtained from the estimation results of the Tobit model. However, the insignificant influence of other factors does not mean that they do not affect the GUECL, but simply that their effects cannot be currently confirmed by Tobit model analysis.

Therefore, here are some suggestions for promoting farming households’ GUECL:

(1)

Farming households’ thoughts: it is important to improve farming households’ awareness of green and sustainable development and strengthen their cognitive level, consciousness, and initiative in the green utilization of cultivated land. One way to do this is to publicize their thoughts on the green utilization of cultivated land through television, radio, newspapers, the Internet, and other media, which will improve other farming households’ awareness and help them thoroughly implement the concept of green development, use standardized cultivated land utilization, maintain a moderate business scale, prevent and control agricultural non-point source pollution, and rationally use chemicals, such as chemical fertilizers and pesticides. It is also important to improve the agricultural green subsidy policy and give subsidies to farming households who realize green production, so as to stimulate their enthusiasm to participate in green production and the green utilization of cultivated land.

(2)

Agricultural green development technology: it is necessary to increase investment in agricultural green development technology, promote the adjustment and change of cultivated land utilization patterns, develop in the direction of intensive and efficient use, and continuously improve the output rate of cultivated land and the utilization rate of resources so as to reduce environmental pollution while ensuring output. This will also speed up and increase the R&D of green technologies in agricultural production, help agricultural households absorb and learn from the agricultural principles and technical systems of traditional culture, and strengthen technical support, helping agricultural households explore a suitable mode of agricultural green development. The promotion of agricultural green development technology and setting up of green technology promotion teams will strengthen agricultural green development technology training and guidance and guide farming households toward green production.

(3)

Rural life: it is important to promote the construction of an ecological civilization in rural areas and implement various colorful publicity and education activities on environmental culture, ecological civilization, circular economy, and cleaner production. Rural areas in all localities must be encouraged to formulate plans for the protection of the cultivated land ecosystem according to local conditions; policy support and technical guidance must be provided for the ecological construction of cultivated land, guiding farming households to carry out the ecological construction in a scientific way. This will consolidate the foundation of agricultural production capacity and improve support policies for agricultural mechanization equipment, agricultural water-saving equipment, high-standard cultivated land construction, and cultivated land protection.

Compared with the related research, differences in research perspectives, methods, and sample selection may have certain influences on the results of this study. This paper applies the survey data of farming households in some areas of Shandong Province to carry out empirical analysis with certain typicality and representativeness, but it may not be comprehensive in the selection of influencing factors, and the research methods and data processing need to be further explored to obtain more stable and practical analysis results.

5. Conclusions

Based on the theoretical analysis framework of cultivated land green utilization, this paper constructed an evaluation system covering economic, ecological and social comprehensive benefits, and measured the GUECL using the super-efficiency EBM model. For the study, 952 farming households in Shandong Province were given a questionnaire. The super-efficiency EBM model effectively overcomes and optimizes the shortcomings of the traditional DEA model (e.g., inability to deal with undesired output and not considering slack variables). At the same time, it has the advantages of being able to distinguish the efficiency of effective DMU so as to reflect farming households’ cultivated land utilization information more truly and effectively. On the basis of evaluating the GUECL of farming households, the Tobit model is further used to comprehensively explore the factors affecting the GUECL from both the macro- and micro-aspects and then optimize the regulation suggestions for cultivated land green utilization. The following conclusions were drawn:

(1)

The GUECL of sample farming households is generally not high, with an average value of 0.67. 80.78% farming households had a medium GUECL, and farming households’ green utilization of cultivated land can be further improved. The higher the GUECL, the lower the input and undesired output per unit yield and per unit output value.

(2)

Farming households’ per capita income, agricultural insurance, agricultural subsidies, cultivated land scale, cultivated land fragmentation, and regional economic level have different degrees and directions of correlation on farming households’ GUECL. The household per capita income is significantly positively correlated with the GUECL, while the remaining five factors are significantly negatively correlated with the GUECL.

(3)

To promote the green transition of cultivated land utilization and improve the green utilization efficiency of farming households’ cultivated land with diversified coordination and multiple measures, this paper puts forward some suggestions on promoting and innovating agricultural green development technology, popularizing and publicizing farming households’ thoughts on the green utilization of cultivated land, and ensuring and improving rural green life.