Quantifying Ecological Well-Being Loss under Rural–Urban Land Conversion: A Study from Choice Experiments in China

Abstract

Rural–urban land conversion has led to the degradation of agricultural system ecological services, and therefore human ecological well-being. There is a need to transform the non-marketed value of ecosystem services provision into a monetary loss of ecological well-being in rural–urban land conversion, which could serve as a basis for ecological compensation. In this paper, a choice experiment method is adopted to investigate the willingness-to-pay (WTP) of rural and urban residents in six cities of three provinces selected from different regions in China. The results reveal that the attributes reflecting the ecological well-being of rural and urban residents are experiencing different degrees of decline. Two attributes, health and security, show the most obvious decline among all ecological well-being attributes for urban residents. In view of stakeholders, rural residents are facing a greater decline in ecological well-being than urban residents, which is mainly driven by their different linkages and interactions with the agro-ecosystem. In terms of regional comparisons, residents in the central region (Hubei Province) of China are subject to the sharpest decline in ecological well-being, followed by those living in the western region (Guizhou Province) and the eastern region (Guangdong Province). These differences are basically determined by their land resource conditions and socioeconomic circumstances. This paper argues that it is pressing to establish an ecological compensation mechanism to regulate rural–urban land conversion and maintain human ecological well-being.

1. Introduction

Ecosystems are multifaceted sources of well-being for humankind [1,2]. As one of the important components of ecosystems, the agro-ecosystem is deliberately selected for crops and livestock, and it also generates various ecosystem services (ES) for human communities [3]. Examples of agricultural ES particularly important for mankind are the maintenance of the genetic diversity essential for successful crop and animal breeding, biological control of pests and diseases, erosion control and sediment retention, air and water regulation, and nutrient cycles [4,5].

Land use/cover change (LUCC) has been highlighted as a key human-induced effect on agricultural ES [6]. As a type of LUCC, rural–urban land conversion is defined as agricultural land in rural areas being converted into developed urban land [7]. With the process of conversion, agricultural land’s original soil texture, biological community, moisture migration, and surface structure are destroyed [8]. It not only causes fundamental changes in biological and environmental elements but also interrupts the regular operation of ecological and physical processes. Ultimately, the capability of the agro-ecosystem to provide various ES to human beings is gradually diminished, with the potential to further reduce human well-being [1,9,10,11].

Globally, it is one of the major challenges for policy-makers to keep a trade-off between economic development and agro-ecosystem protection. In most developing countries, especially in southern Africa and Asia, rapid urbanization and increasing human population are driving massive land transformation [12,13,14]. With the largest population and relatively scarce land resources, China has witnessed a significant growth in urbanization [15,16]. In recent years, ecological risks caused by rural–urban land conversion have attracted growing attention. The Central Government of the Communist Party of China (CPC) put forward the strategy of ecological civilization to reconcile the relationship between human and nature, taking “optimization of the spatial layout”, “effective regulation of development scale”, and “improvement of the ecology system” as main guidelines of future development mode. It is increasingly presenting not only as a response to eco-environment degradation in China but as a vision for our global future [17].

Researchers have been sharing growing concerns about the impacts of rural–urban land conversion on agricultural ES and their impacts on human well-being. Several studies have adopted physical assessment methods [18,19,20] to uncover the relations between agricultural ES and landscape features describing the land cover structure and composition. Value assessment methods are often employed to identify stakeholders’ preferences for some particular types of agricultural ES. Kvakkestad et al. analyzed Norwegian farmers’ attitudes towards various aspects of multifunctional agriculture and their evaluation of different forms of agricultural payments [21]. Barrena et al. concluded that agricultural heritage conversation influences the welfare not only of local populations where this ES is generated but also of distant stakeholders [22]. However, the linkage between the agricultural ES decrease and human well-being decline has not been sufficiently discussed. Although some scholars have tried to estimate the impact of land-use change on residents’ well-being (especially farmers) or social equity [15,23,24], the loss of well-being caused by rural–urban land conversion is often expressed in economic dimension, falling short in capturing ecological dimension. A few pieces of research have called for constructing ecological compensation for human well-being loss based on the non-market value and externality theory [25,26], while empirical evidence for quantifying the loss of residents’ well-being loss needs to be explored to clarify the negative impacts of rural–urban land conversion.

Human well-being is a multi-dimensional concept that reflects the extent to which individuals satisfy with their life generally and with different aspects of their life [27,28]. According to the definition given by the Millennium Ecosystem Assessment (MA), human well-being is a human experience that includes the sense of security, basic materials for a good life, freedom of choice and action, health, and good social relations [29]. The level of well-being is strongly dependent on the specific cultural, geographical, and historical context in which different human societies develop, and is determined by cultural–socioeconomic processes as well as by the provision of ES [30]. Existing studies commonly adopted generalized definition of human well-being, which is too broad to clearly reflect the ecological impacts of ES changes on human well-being. From these concerns, the concept of human well-being discussed in this paper is confined to those elements closely related to agroecosystem and defined as “human ecological well-being”.

Rural–urban land conversion involves multiple stakeholders [31], who have diverse relations with the ES framework [32,33]. The spatial distribution of land use is closely related to the provision of ES because it is affected by the ecological processes and human disturbances at a wide range of spatial scales [34]. Therefore, stakeholders’ heterogeneity and spatial characteristics should be taken into account in such land use analysis. The aim of this study is to evaluate the human ecological well-being loss in rural–urban land conversion, especially uncover its differences among regions and groups. This paper is organized as follows. Section 2 presents the measurement methods and experiment design; Section 3 is empirical study; Section 4 discusses the difference in attributes, stakeholders and regions of ecological well-being loss, as well as policy implications from results; Section 5 summarizes research findings and limitations.

2. Materials and Methods

2.1. Study Area

China is a developing country with fierce competition in land use and significant pressure on ecological conservation. There are significant differences in land resources endowment, social and economic development conditions, and development goals among the east, the middle, and the west of China. Cities in the east of China, as the most developed economic region, have basically completed the urbanization process (e.g., the urbanization rate of Guangzhou is more than 85%). The rapid economic growth is at the cost of a large number of rural–urban land conversions, which leads to prominent conflict between economic growth and ecological protection. The central region, which has China’s main grain production bases, is also an important growth pole of China’s economy in past 15 years. Abundant cultivated land plays a great role in the production and life of residents. While in recent years, the central region has become a highly concentrated area for rural-urban land conversion driven by the implementation of the Rise of the Central Region Plan. The western region, which is dominated by plateau mountainous landforms, is the key ecological functional area of China. The ecosystem here is extremely fragile and sensitive. Rural–urban land conversion is strictly restricted under land use planning, and its economic development is slow.

To spatially regulate agricultural land and protect its ecological functions, in 2011, China introduced the Major Function Oriented Zoning Division (MFOZD). The MFOZD divides nationwide land into a series of regional units and designates a specified major function for each unit, including priority development zones (PDZ), key development zones (KDZ), restricted development zones (RDZ), and forbidden development zones (FDZ). This paper selects study areas synthetically considering the traditional geographical division of the eastern, central, and western regions of China, the division of MFOZD, as well as the natural conditions (topography) and social development levels (urbanization rate). The study area covers three districts: the eastern region (represented by Guangdong Province), the central region (represented by Hubei Province) and the western region (represented by Guizhou Province), from which one provincial capital city and one prefecture-level city are selected (see Figure 1 and

Table 1. Research area.

Research Area		Region It Belongs To	City Levels	MFOZD	Topography	Urbanization Rate in 2015 (%)
Sample Provinces	Sample Cities
Guangdong Province	Guangzhou	Eastern region	Provincial capital	PDZ	Plains/hills	85.53
	Shaoguan		Prefecture-level	KDZ	Mountains/hills	52.27
Hubei Province	Wuhan	Central region	Provincial capital	KDZ	Plains	79.41
	Shiyan		Prefecture-level	RDZ	Mountains/hills	51.60
Guizhou Province	Guizhou	Western region	Provincial capital	RDZ	Plateau mountains	64.63
	Duyun		Prefecture-level	RDZ/FDZ	Plateau mountains	44.40

Data reference: Urbanization rates are taken from the statistical yearbook of each city in 2016.

).

In this study, the respondents were composed of two groups: rural residents and urban residents, which are defined according to China’s census registration system. “Rural residents” refer to those registered in rural areas and are generally qualified to be assigned to agricultural land for work and livelihood. “Urban residents” refers to the residents registered as urban permanent residents and living in urban areas. There are differences between the two groups in the way of obtaining living materials from agro-ecosystem. As a direct cultivator, most rural residents in China can cultivate food and feed themselves (direct access), while urban residents need to purchase food from the urban market (indirect access).

2.2. Choice Experiment

The complex impact of rural–urban land conversion on residents’ ecological well-being takes the form of externalities, which is difficult to directly measure by market price. The choice experiment (CE) is increasingly applied to environmental value assessments aimed at ecological policy design [35,36,37]. The theoretical basis of CE derives from the characteristics theory [38] and random utility theory [39]. In the characteristics theory, the utility of goods or services for a typical consumer can be decomposed into various attributes or observable characteristics (such as price, appearance, and functions). The utility function of respondent n choosing combination scheme i can be expressed as follows:

$$U_{ni}=V_{ni}\left(X_{ni},S_n\right)+{\epsilon }_{ni}$$

(1)

where U_ni represents the total utility of combination scheme i selected by respondents n. V_ni is the observable fixed utility component of scheme i selected by respondent n, which is determined by the condition vector X_in of each attribute selected by respondents and the personal characteristic S_n of respondents. ε_ni represents the unobserved utility of the random error item.

The random utility theory premises that respondents always make their choices based on the principle of utility maximization [40]. For the given choice set c, when U_ni > U_nj (i≠ j), rational respondents will choose scheme i. The probability that respondent n choosing i over another option j from the complete choice sets C is given by

$$P\left(\frac{i}{C}\right)=P\left(U_{ni}>U_{nj}, \forall i\ne j\right)=P\left[\left(V_{ni}+{\epsilon }_{ni}\right)>\left(V_{nj}+{\epsilon }_{nj}\right), \forall i,j\in C, i\ne j\right]$$

(2)

In general, the random error term ε_ni is assumed to be independent and identically distributed and of type I extreme value distribution [40]. The probability of respondent n choosing scheme i with maximum utility is P (ni) as follows:

$$P\left(ni\right)=exp\left(\mu V_{ni}\right)/{\displaystyle \sum }_{j\in C}exp\left(\mu V_{n\mathrm{j}}\right)$$

(3)

where μ refers to the proportional parameters that are subject to the 0–1 distribution, which is used to specify the true parameters of the variance in the unobserved utility component. The utility obtained by respondent n from scheme i can be transformed into a linear form:

$$V_{ni}=ASC+{\displaystyle \sum }_i{\beta }_{ni}{Z}_{ni}$$

(4)

where Z_i refers to the attribute characteristics of the non-market value of environmental goods to be evaluated, β_i is the coefficient to be estimated for the attribute characteristics of environmental goods, and ASC refers to the alternative specific constant for options. In addition, the introduction of individual socio-economic characteristic variables can improve the model’s fit and thoroughly explain the impacts on different groups. In this case, the utility function can be expressed as

$$V_{ni}=ASC+{\displaystyle \sum }_i{\beta }_{ni}{Z}_{ni}+{\displaystyle \sum }_i{\alpha }_n{S}_n$$

(5)

where S_n refers to the socio-economic variable of individuals; and α_n is its estimated coefficient. When the individual’s utility reaches its maximum at dv_i = 0, the marginal value of a change within a single attribute can be estimated using

$$MWT{P}_i=\frac{dT}{d{Z}_i}=-\frac{\partial v_i}{\partial Z_i}/\frac{\partial v_i}{\partial T}=-\frac{{\beta }_i}{{\beta }_T} \left(i=1, 2, 3, \dots \right)$$

(6)

where β_i refers to the coefficient of the environmental goods attribute to be estimated, and β_T is the coefficient of the monetary attribute to be estimated. MWTP effectively reflects the marginal rate of substitution between the payments and attributes [41].

Moreover, the willingness-to-pay (WTP) measures relative to different environmental scenarios can be obtained with Equation (7):

$$WT{P}_i=-\frac{1}{{\beta }_T}\left(V_0-V_1\right)$$

(7)

where V₀ and V₁ represent the indirect utility of respondents before and after the implementation of the improvement scheme.

In this study, the economic meaning of WTP can be concretized as the economic value of ecological well-being loss caused by the effect of a 1-unit level deterioration in rural–urban land conversion. Then, the relative decline of various ecological well-being attributes can be ranked, providing a reference for the priority of compensation policy.

2.3. Experimental Design

2.3.1. Attribute Identification

Millennium Ecosystem Assessment (MA) is the first major project in the world to comprehensively assess ecosystems and reveal the link between ES and human well-being. One of the core issues of the MA is how ES changes affect human well-being and what options can be adopted to enhance the sustainable use of ecosystems. In the conceptual framework proposed by the MA, ES constituting of the provision of services, cultural services, regulating services, and supporting services is closely related to human well-being. Human well-being, which is composed of security, basic materials for a good life, health, good social relations, and freedom of choice and action, is the core of the assessment.

Combining with the specific content of the human well-being elements proposed by MA and the characteristics of the agro-ecosystem, this study constructs a set of ecological well-being attributes for residents (see

Table 2. Attribute and levels of resident ecological well-being.

Attribute	Indicators		Attribute Levels
	Rural Residents	Urban Residents
Security	Waste Recycling Capability 1	——	Unchanged	Improved
	Frequency of agroecosystem-related meteorological disasters (such as drought, floods, soil erosion, and desertification)
Basic materials for a good life 2	Obtain daily staple food Obtain daily vegetables Obtain daily meat	——	Unchanged	Improved
Health	Satisfaction with air/water quality Safety of food/vegetables/meat consumption Pollution-related diseases		Unchanged	Improved
Good social relations	Nostalgia for rural life Children’s rural experiences	Frequency of eco-tourism Satisfaction with natural landscape	Unchanged	Improved
Freedom of choice and action	Livelihood choices		Unchanged	Improved
Payment (CNY per household per year)	——		0 50 100 200

Notes: (1) Daily garbage in urban areas depends on urban waste disposal system rather than the agro-ecosystem, so waste recycling capability is not considered in the security attribute of urban residents. (2) Because urban residents do not directly access basic materials from agricultural land, this attribute for urban residents is not analyzed in this paper.

). For the sake of comparing the different impacts of rural–urban land conversion on urban and rural residents, the ecological well-being attributes for these two groups are selected, respectively. Their detailed definitions and selection references can be found in our previous research [7].

2.3.2. Attribute Levels

In this study, the attribute levels of human ecological well-being are set as either “unchanged” or “improved”. From 1978 to 2016, China’s urbanization rate had increased by around 1% annually, accompanied by roughly the same rate of rural–urban land conversion. Some scholars have calculated that there was about 20% of excessive ecological loss caused by rural–urban land conversion between 1989 and 2003 in China [16]. Synthetically considering the growth rate of urbanization and ecological loss proportion in the past, this study chooses 10% as the improved level in 10 years. From a pre-pilot survey that relied on an open-ended question (i.e., how much would you be willing to pay?), the bidding prices of respondents were mainly between 50 and 200 yuan. Referring to the design of the contingent valuation method (CVM) for cultivated land protection in existing studies [15], payment options were often divided into 3 or 4 levels. Ultimately, 0, 50, 100, and 200 yuan per household/year were set as the payment attribute levels (Table 2).

2.3.3. Experimental Schemes

Due to a full factorial design proved unrealistic and costly, an orthogonal design experiment is created using the software SPSS 19.0. Ultimately, there are 6 choice schemes for urban residents and 12 choice schemes for rural residents.

A diverse combination of choice schemes can provide respondents with more options to find the most appropriate one that can truly reflect their preferences, but it can also bring a heavy burden to the respondents and make them confused in decision-making. This study thus adopts a binary approach—the status quo scheme (Option A, all remain unchanged) and the improved scheme (Option B, partially improved) to simulate residents’ possible preferences to the changes of ecological well-being in the rural–urban land conversion. The improved scheme means that after the implementation of the ecological compensation program for 10 years, the selected attributes will achieve the target state, i.e., 10% higher than the status quo, and the respondents are willing to pay for this result. For each choice set, individuals were asked to choose one from 3 options: Option A, Option B, or neither. A sample of choice cards for urban residents is presented in

Table 3. Example of a choice set.

Choice Sets	Security	Health	Good Social Relations	Freedom of Choice and Action	Payment (CNY/ Year·Household)
Option A	Unchanged	Unchanged	Unchanged	Unchanged	0
Option B	Unchanged	Improved	Unchanged	Improved	50
Your choice is	Option A(…)		Option B(…)		Neither (…)

.

2.3.4. Questionnaire Collection

The questionnaire structure includes three sections, created for rural and urban residents, respectively. The first part requests respondents to describe the changes in agricultural ES caused by rural–urban land conversion. The second part aimed to inquire respondents’ WTP of promoting the corresponding ecological well-being. Referring to the preceding list of elements of ecological well-being, respondents were asked, “would you like to pay an amount of money to the local public sector to improve these elements?” The respondents who answered “yes” were required to choose the specific ecological well-being improving schemes they most preferred from the selection sets. However, respondents who answered “no” were requested to give specific reasons. The third part collected information about respondents’ socio-demographic characteristics.

The questionnaire collection was conducted via face-to-face interviews. In April 2016, the exploratory interview was carried out. From the pre-pilot survey, we found it was difficult for respondents (especially rural respondents) to understand abstract concepts associated with agricultural ES and ecological well-being. Then, we converted the introduction using plain language. For instance, rather than introducing and defining the term “ecological well-being”, we referred to “satisfaction with the local environment”; “rural–urban land conversion” was translated to “agricultural land decline”. Further, we revised the payment interval from relatively wide (0, 100, 250, 500) to more reasonable (0, 50, 100, 200). Then, the large sample survey was conducted from May to October in 2016. Ultimately, a total of 1421 questionnaires were issued, and 1125 questionnaires were collected, for a recovery rate of 79.2%. Of these, 623 questionnaires were collected from urban residents and 502 from rural residents. For the CE, real zero payment and resistant zero payment should be distinguished when screening questionnaires. “Real zero payment” refers to respondents choosing “neither” due to their limited financial ability. In this context, respondents actually prefer Option A (although they are dissatisfied with the status quo). However, respondents who belong to “resistant zero payment” category often argue that “it is the government’s responsibility to protect the agricultural ES”. In this survey, the resistant zero payment questionnaires for urban residents and rural residents were 98 and 59, respectively, which were treated as invalid questionnaires and eliminated. Thus, a total of 968 valid questionnaires were obtained, including 525 questionnaires for urban residents and 443 questionnaires for rural residents.

2.4. Variable Definition

The dependent variable in this study is whether an improvement scheme is selected (CHOC). When the respondents choose A (status quo) or the scheme reflects as real zero payment, CHOC = 0. If they choose B, CHOC = 1. Based on the above, the independent variables of rural residents include 6 attributes and 6 socio-economic variables. The independent variables of urban residents include 5 attribute variables and 7 socio-economic variables. To eliminate the restriction on the logit model created by the assumption that random variables obey an independent distribution of the same type, an alternative specific variable (ASC) is set to distinguish whether residents are willing to participate in the improved policy. If a respondent chooses A in all selection sets, it means that he or she has no interest in improving the status quo, and thus ASC = 0. If he or she chooses at least one B, ASC = 1 (see

Table 4. Variable definition.

Variable Types	Name	Definition	Value
Dependent Variable	CHOC	Choice variable	0 = section A and real zero payment, 1 = section B
Independent Variables	SEC	Security	0 = unchanged, 1 = improved
	HEA	Health	0 = unchanged, 1 = improved
	GSR	Good social relations	0 = unchanged, 1 = improved
	BML	Basic materials for a good life 1	0 = unchanged, 1 = improved
	FCA	Freedom of choice and action	0 = unchanged, 1 = improved
	PAY	Payment (CNY)	0, 50, 100, 200
	ASC	Alternative specific constants	0 = participated in no improved schemes, 1 = participated in at least one improved scheme
	GEN	Gender	0 = male, 1 = female
	AGE	Age	actual observed number
	EDU	Education level	1 = primary or below, 2 = junior high school, 3 = high school, 4 = specialized technical school, 5 = bachelor or above
	EXP	Rural experience	0 = no, 1 = yes
	FSZ	Family size	actual observed number
	LAB	Number of household labours	actual observed number
	INC	Annual household income (CNY, ×104)	1≤1 (CNY, ×104), 2 = 1–3 (CNY, ×104), 3 = 3–5 (CNY, ×104), 4 = 5–9 (CNY, ×104), 5 = 9–15 (CNY, ×104), 6 = 15–25 (CNY, ×104), 7≥25 (CNY, ×104)

Note: The independent variable of basic materials for a good life is only used in rural residents’models. Rural experience is only used in urban residents’ models.

).

3. Results

3.1. Estimation Results

As seen from the regression results (

Table 5. Logit model results.

	Urban Residents						Rural Residents
Research Areas	Eastern Region (Guangdong)		Central Region (Hubei)		Western Region (Guizhou)		Eastern Region (Guangdong)		Central Region (Hubei)		Western Region (Guizhou)
Variable	Estimate	z-Statistic	Estimate	z-Statistic	Estimate	z-Statistic	Estimate	z-Statistic	Estimate	z-Statistic	Estimate	z-Statistic
SEC	1.1591***	5.5052	0.6946**	2.5297	1.6943***	4.7566	1.4055***	4.0893	0.2499***	2.8527	0.1651**	2.0134
HEA	1.1076***	4.7485	0.8983**	2.2459	1.9192***	4.5389	1.5002***	5.2199	0.2869**	1.9786	0.6037***	3.4936
BML	——	——	——	——	——	——	0.6300**	2.6163	0.3035**	1.9076	0.4130***	2.6222
GSR	0.6209***	3.9598	0.3966***	4.5015	0.9641***	3.6416	0.6663**	2.0277	0.6183***	2.8130	0.2315**	1.7988
FCA	0.4837***	2.8171	0.3788**	1.9816	1.0162***	3.5242	0.4393**	1.9268	0.4701***	2.5915	0.5985***	2.8244
PAY	−0.0144***	−5.3333	−0.0086**	−2.3590	−0.0213***	−4.4294	−0.0189***	−4.8462	−0.0056***	−3.2941	−0.0076***	−3.4545
ASC	0.1815*	2.1920	0.1226**	3.2091	0.1283***	−3.5640	0.5569*	1.6726	0.1571***	2.6370	0.1389*	1.7553
GEN	−0.0747	−0.7814	0.6363***	5.1018	0.2715	1.1428	−0.7055	−1.1851	−0.5040	−1.5429	−0.2157**	−1.6491
AGE	−0.0050*	−1.9231	−0.0113**	−2.0222	−0.0259***	−3.7528	−0.0037*	−1.7282	−0.0162***	−2.3824	−0.0164***	−3.0370
EDU	0.3094***	7.2626	0.0540**	2.1752	0.2757***	3.5043	0.6478***	3.5633	0.2552**	2.1943	0.2296***	2.7696
FSZ	−0.2110***	−5.7027	−0.0539	−1.1801	−0.1292**	−1.7528	−0.0850	−0.9004	−0.0329	−0.6714	−0.1687***	−3.4289
LAB	0.2574***	5.3962	0.1655**	2.2074	−0.0469	−0.4658	0.0975**	1.9632	0.1243*	1.8415	0.1784***	2.4338
INC	0.0206*	−1.7913	0.0344*	1.9007	0.2351***	3.2383	0.2531***	2.1251	0.3528***	3.7692	0.2410***	3.3754
EXP	0.0015**	1.9841	0.1589**	2.6463	0.2491**	1.9807	——	——	——	——	——	——

***, **, and * indicate that the estimated results are significant at the 1%, 5%, and 10% level respectively.

), all attributes are significant at the 5% level, and in the eastern and western regions, all attributes are significant at the 1% level. All ecological well-being attributes are positively related to the probability of choosing an option, indicating that ecological improvement programs can increase the expected utility of residents. The coefficient of payment is negative, meaning that with an increase in the payment for ecological improvement, residents’ utility decreases.

The estimation values of parameters reflect the preference intensity of interviewees. In other words, the greater the absolute value is, the stronger the respondents’ preference for this attribute is, and the higher the marginal utility of improving this attribute. Overall, in the western region, the conversion of rural land to urban construction has produced a remarkable feeling of loss among urban residents, with parameter values ranging from 0.9641 to 1.9192. From the attribute’s perspective, for urban residents, people in the eastern region have the strongest preference intensity for security (1.1591), while people in the central and western regions have the strongest preference intensity for health (0.8983 and 1.9192). Comparatively, western urban residents rank the livelihood issue as more important after rural–urban land conversion, with a higher parameter value (1.0162) of freedom of choice and action. For rural residents, the eastern region generally has the strongest intention to improve the ecological well-being attributes, with parameter values ranging from 0.4393 to 1.5002. Rural residents in different regions have different preferences. The strongest preferences of rural residents in the eastern and the western region are health (1.5002 and 0.6037), while in the central region, it turns to good social relations (0.6183).

3.2. Loss of Residents’Ecological Well-being

According to CE, the marginal WTP for a certain element of ecological well-being is the ratio of the no-monetary attribute coefficient and the coefficient for the payment attribute [40]. These implicit prices can also be ranked accordingly (

Table 6. Willingness-to-pay (WTP) and implied ranking of ecological well-being elements in urban residents.

Ecological Well-Being Constituents	Eastern Region (Guangdong)		Central Region (Hubei)		Western Region (Guizhou)
	MWTP	Rank	MWTP	Rank	MWTP	Rank
Security	80.49	1	80.77	2	79.54	2
Health	76.92	2	104.45	1	90.1	1
Good social relations	43.12	3	46.12	3	45.26	4
Freedom of choice and action	33.59	4	44.05	4	47.71	3
WTP (CNY per household per year)	234.12	—	275.39	—	262.21	—

and

Table 7. WTP and implied ranking of ecological well-being elements in rural residents.

Ecological Well-Being Constituents	Eastern Region (Guangdong)		Central Region (Hubei)		Western Region (Guizhou)
	MWTP	Rank	MWTP	Rank	MWTP	Rank
Security	74.37	2	44.63	5	21.72	5
Health	79.38	1	51.23	4	79.43	1
Basic materials for a good life	33.33	4	54.20	3	54.34	3
Good social relations	35.25	3	110.41	1	30.46	4
Freedom of choice and action	23.24	5	83.95	2	78.75	2
WTP (CNY per household per year)	245.57	—	344.42	—	264.71	—

), which can be used by policymakers to identify the magnitude of the decline of ecological well-being caused by the rural–urban land conversion.

Generally speaking, urban residents tend to be willing to pay for improvements in health and security (ranking 1 or 2), followed by good social relations and freedom of action and choice (ranking 3 or 4). The loss of economic value (MWTP) of the first two attributes is approximately 2–3 times that of the last two attributes. Thus, it can be inferred that the decline in health and security has a more serious impact on urban residents’ ecological well-being during rural–urban land conversion. The loss of the economic value of security and good social relations in these three regions are very similar. Health has the most obvious regional difference, with the central region exceeding the eastern region by CNY 27.53 per household every year. Considering the relative decline of each attribute, the rank of economic value loss in different regions is consistent overall. When all attributes are at the preferred level, the yearly WTP of a typical household for ecological payment programs can be calculated by adding the WTP of each attribute, in which the central region is the highest (CNY 275.39), followed by the western region (CNY 262.21) and the eastern region (CNY 234.12).

In the process of rural–urban land conversion, rural residents in the eastern region experience the most serious harm to their health and security, with annual average losses of CNY 79.38 and CNY 74.37 per household, respectively. Freedom of choice and action is the least damaged, which probably related to the greater non-agricultural employment opportunities in eastern China. Rural residents in central China suffer the greatest decline in good social relations, for which they suffer an average annual loss of economic value of CNY 110.41 per household, followed by freedom of choice and action, and the loss of security sees the least decline. Health and freedom of choice and action are the most seriously damaged factors for rural residents in western China. The average loss of annual economic value is CNY 79.43 and CNY 78.75 per household, respectively, and the least damaged attribute is security. Generally, in the eastern and western regions, rural residents’ health is the most severely affected during rural–urban land conversion. The respondents’ WTP is very approximate, being CNY 79.38 and CNY 79.43, respectively. In the central and western regions, rural residents rank the basic materials for a good life in the middle with basically the same WTP of CNY 54.20 and CNY 54.34, respectively. The most significant spatial difference is good social relations. Rural residents in the central region have the strongest WTP for the improvement of this element, while there is little difference between the eastern and western regions. From the perspective of the total loss in economic value, the ranking for rural residents is consistent with that of urban residents; that is, central region (CNY 344.42 per year) > western region (CNY 264.71 per year) > eastern region (CNY 245.57 per year).

From the perspective of overall WTP, although urban residents have a greater economic ability than rural residents, their WTP is lower than that of rural residents in all regions. Among them, the WTP gap between urban and rural residents in central China is the largest (CNY -69.04;

Table 8. Comparison of urban residents and rural residents.

Ecological Well-Being Constituents	Eastern Region (Guangdong)		Central Region (Hubei)		Western Region (Guizhou)
	Urban Residents	Rural Residents	Urban Residents	Rural residents	Urban Residents	Rural Residents
	(1)	(2)	(1)	(2)	(1)	(2)
WTP	234.12	245.57	275.39	344.43	262.21	264.71
Balance of WTP (1)–(2)	−11.45		−69.04		−2.5

Notes: The measure unit is CNY per household per year.

).

4. Discussion

4.1. Differential Loss in Ecological Well-Being Attributes

Health and security. These two attributes are the most serious elements driving the loss of ecological well-being among urban residents, as well as eastern rural residents. Health and security can be recognized as the basic elements of human ecological well-being. As concluded by Leviston et al. [2], if the ability of ecosystems to provide adequate services for the most basic needs is compromised, the level of well-being in other domains becomes negligible. In other words, “basic requirements” for each individual must be satisfied first.

Basic materials for a good life. This attribute ranks in the middle or lower among rural residents’ losses, which is basically equal across all districts (ranking 3 or 4). However, the loss of basic materials for a good life in the central and western regions is CNY 20 higher than it is in the eastern region. This reveals that land conversion increases living expenses for western and central rural residents and makes their lives more difficult. Land is the foundation upon which rural households maintain their livelihood [41,42]. However, rural–urban land conversion, especially in the area of land expropriation, directly changes the way that rural residents get access to daily materials. Before rural–urban land conversion, they typically obtain staple food (rice, flour, and beans), vegetables, and meats from their own cultivated land. Only a small portion of supplementary vegetables and meat products were purchased from local market. After rural residents lose their agricultural land, the market becomes the only source of daily necessities, which greatly increases their living costs.

Good social relations. This attribute shows a sharp decrease for rural residents in the central region, where converges the conflicts between the most rapid urbanization and the largest scale of traditional agricultural lifestyles. The agro-ecosystem carries the inheritance of specific cultural types comprising agricultural activities and rural lifestyles [43]. Specially, rural residents in China have a strong psychological dependence on agricultural land, which is the basis of maintaining interpersonal relationships. In rural–urban land conversion, fragmented habitats are increasing, and patches are generally small and isolated [44,45]. Soga et al. [46]) described this progressive separation of humans from ecosystems as the “extinction of experience”, especially for children, and viewed it as a major public health issue.

Freedom of choice and action. It is the second most damaged attribute of rural residents’ ecological well-being in central and western China, largely resulting from regional economic development and individual conditions. Studies have shown that household livelihood plays a vital role in addressing uncertain external events, especially in less developed areas [47,48]. Rural–urban land conversion directly reduces the natural capital of rural residents, which basically deprive their way of making a living by farming. In our survey, younger rural residents typically choose to work in provincial cities. While the older residents, limited by age, education level and vocational skills, either engage in basic labour in nearby factories for meagre wages, or idle at home to take care of their grandchildren. In general, the loss of their freedom of choice and action is permanent because the rural–urban land conversion is irreversible [7].

Similarly, Wang et al. [9] also established indicators to measure changes in human well-being. Their research results showed that most aspects of human well-being were improved, except for security. The main reason for the inconsistent with this study is that most of the well-being indicators they selected are economic sides. For example, the promotion of health depended on the popularization of medical insurance, and good social relations benefited from the public investment of education, which are little related to the provision of the capacity of agricultural ES. Finally, they also underlined that the well-being associated with natural ecological resources was dysfunctional or even too low [9].

4.2. Stakeholders’ Differences in Ecological Well-Being Loss

As the major interest groups of rural-urban land conversion, the economic value loss experienced by rural and urban residents provides a reference for determining ecological compensation standards. As previous studies indicated, stakeholders in rural–urban land conversion are experiencing different changes in ecological well-being [7,49]. In our research, rural residents’ ecological well-being declined more markedly than urban residents’, which can be probably explained by their relationships with agro-ecosystem. Regardless of the interaction with the ecosystem or the geographic distance from cultivated land, the linkage between urban residents and the agroecosystem is relatively indirect [50,51]. While rural residents, especially in developing regions, often directly obtain livelihood resources from agro-ecosystem, being commonly constrained in their ability to other sources of capital and particularly vulnerable to ES changes [52]. They are forced to leave their original residence, lose agricultural land, and experience direct ecological deterioration. This vulnerability towards external changes has also been observed by previous studies [53,54,55]. Generally, the interference intensity of rural–urban land conversion for rural residents is greater than urban residents.

There is another perspective to explain the lower WTP of urban residents. Since ES is characterized by non-excludability [56], urban residents highly benefit from agro-ecosystem protection. In the absence of institutional constraints, they have been accustomed to being “free riders” [56]. Resistant zero payment can provide some evidence for this inference. Urban residents, as a relatively wealthy group in China, have a higher resistance ratio (15.73%) than rural residents (11.75%).

4.3. Spatial Differences in Ecological Well-Being Loss

Comparing the WTP across these three regions, it is noted that residents in the central region have the highest WTP for reducing the risk of ecological well-being decline, followed by those in the western region and then those in the eastern region. These results can be interpreted from the conclusion of the MA; that is, the access to external changes is heavily mediated by resource conditions and socioeconomic circumstances [29].

In the central region, the historical close relationship with agricultural lands makes the residents attach more importance to the agro-ecosystem. Tremendous rural–urban land conversion triggered off by the on-going urbanization has been greatly changing their way of life, living environment and social structure. Taking Wuhan as an example, its urban area reached 628 km² in 2017, which was nearly 3 times larger than 2007 (222 km²). Massively land cover change has increased the frequency of pollution issues and urban waterlogging. The western region, as the important ES supplier defined by MFOZD in China, also has been disturbed by rural–urban land conversion. If we take into account the fact that the payment ability of residents in the western region is lower than that in the eastern and central regions, the actual loss of ecological well-being of residents in this region is higher than the estimated. In the eastern region, the volume of rural–urban land conversion has been relatively small in recent years, as agricultural land resources are almost exhausted in the early urbanization. The reason for less damage to residents’ ecological well-being in this area may be that the amount of ecological services provided by cultivated land is originally limited.

4.4. Policy Implications

In China, the absence of ecological compensation mechanism makes ecological well-being loss exist in the form of a negative externality, which offers a partial explanation for the excessive conversion of agricultural land during recent decades. There is a need to establish an effective ecological compensation mechanism for rural–urban land conversion from the perspective of ecological well-being. In this research, the WTP is interpreted as the economic value of ecological well-being attributes and reflects the loss in ecological well-being due to rural–urban land conversion. The public, including rural and urban residents, should be the beneficiaries of ecological compensation programs. Rural residents who have suffered from a direct ecological loss in rural–urban land conversion should be the major objects in ecological compensation. Besides, farmers who hold agricultural land, maintaining the supply of ES means waiving their development rights, should also receive subsidies from the ecological fund. New urban construction land users have gained huge profits at a lower land price without ecological cost. Following the “destroyer pays” principle, they should be the payers in this mechanism.

From the estimation results, there are spatial differences in the loss of ecological well-being between rural and urban residents, as well as in their specific attributes. These findings could provide reference for making spatial differentiated mandatory regulations or implementing incentive ecological compensation. A generalized ecological compensation mechanism paying for protection on land that requires no protective measures or underpaying for land that needs special protection [57] is not conducive to the sustainability of cultivated land protection and the efficiency use of scarce funds. The MFOZD of China provides a preliminary reference for identifying service providers (ecological service export area) and beneficiaries (ecological service import area). Given the non-exclusive and borderless characteristics of ES, most natural resources in developing regions are being directly or indirectly used by developed regions without full payment [26]. Therefore, the more developed and ecologically degraded regions (such as Guangdong Province), i.e., vested interests in rural–urban land conversion, should take greater ecological responsibilities, while the developing and less degraded regions (such as Guizhou Province) are supposed to have more net income. These implications could serve as a reference for developing and emerging economies like China to deal with the fierce competition between urban development and agricultural land conservation in rapid urbanization.

5. Conclusions

The sharp decline of agricultural land is an intuitive manifestation of rural–urban land conversion. Consequently, the original functions of the agro-ecosystem disappear or are limited, which reduce the ability of public to obtain various ES. As a result, public ecological well-being is declining. This paper attempts to quantify the ecological well-being loss of two major stakeholders (rural residents and urban residents) in rural–urban land conversion in six cities of three provinces in China using the CE method. In general, rural residents suffer more serious losses in rural–urban land conversion because of the more intimate and direct interaction with the agro-ecosystem. Additionally, the decline of their ecological well-being attributes shows great regional difference, while, there is little obvious spatial heterogeneity in the ecological well-being loss ranking of urban residents, and their most damaged attributes are mainly reflected on health and security. Comparing to overall loss of regions, residents in the central region (Hubei province) have experienced the most decline in ecological well-being due to intense urban expansion in recent years, followed by those in western region (Guizhou Province) and the eastern region (Guangdong Province). Eastern provinces, with the most agricultural land, has been occupied by urbanization and so should be the payers of cross-regional ecological compensation, while western regions as ecological functional zones should be compensated.

We acknowledge that the valuation accuracy of CE is subject to several limitations, including the payment levels, attribute identification, and number of choice sets. Especially in evaluating environmental values or loss, a trade-off has to be made between accuracy and complexity. For example, the binary approach adopted in this experimental design facilitates respondents in making choices among the multiple-choice sets, but it may prevent a few respondents’ preferences from being included in the choice sets displayed in the questionnaire, resulting in the respondents’decision to refuse to pay. It probably leads to an underestimation of the loss of residents’ well-being in rural–urban land conversion. However, the results in this study provide a quantitative reference to understand the impact of rural–urban land conversion on residents’ ecological well-being, urging land-use managers to pay sufficient attention to the ecological impact of land-use changes and introduce ecological compensation mechanism for rural–urban land conversion.