Does Livelihood Determine Attitude? The Impact of Farmers’ Livelihood Capital on the Performance of Agricultural Non-Point Source Pollution Management: An Empirical Investigation in Yilong Lake Basin, China

Abstract

Agricultural non-point source pollution is intricately connected to the rural population’s production and lifestyle. The heterogeneous composition of livelihood capital results in varied livelihood types, influencing the farmers’ attitudes and perceptions of the treatment projects. This ultimately causes discrepancies in the farmers’ evaluation of agricultural non-point source pollution control. In this study, a participatory evaluation method was employed to evaluate the performance of agricultural non-point source pollution control projects in the Yilong Lake Watershed of Yunnan Province and analyze the underlying reasons for the differing performance evaluations. The findings revealed that pure farmers’ performance evaluation value for agricultural non-point source pollution control projects in Yilong Lake Basin was 0.4811 (with the full mark being 1), with a general evaluation grade. Part-time business households had a performance evaluation value of 0.5969, also with a general evaluation grade, while non-farmers had a performance evaluation value of 0.7057, with a good evaluation grade. The performance evaluation value ranked from highest to lowest is non-farmer > part-time farmer > pure farmer. The main factor affecting the variation in farmers’ performance evaluation is the key index of different livelihood capital. If pollution control projects can promote the adjustment of farmers’ livelihood capital types, it can enhance not only the performance evaluation degree of farmers, but also the sustainability of farmers’ livelihoods and increase their adaptability to livelihood risks. Therefore, pollution control projects should consider farmers’ livelihood capital types and be implemented accurately to improve farmers’ satisfaction and sustainability.

1. Introduction

At present, agricultural non-point source pollution has emerged as the principal contributor to water pollution in China [1]. As per the “Second National Pollution Source Census Report” in 2017, agricultural sources discharged 10.67 million tons of chemical oxygen demand (COD), accounting for 49.77% of the total COD emissions in the nation. Furthermore, 216,200 tons of ammonia nitrogen (NH3-N) were emitted, making up 22.44% of the nationwide NH3-N emissions. Additionally, agricultural sources released 1.4149 million tons of total nitrogen (TN), which accounts for 46.52% of the total TN emissions in water pollutants, and 212,000 tons of total phosphorus (TP), constituting 67.22% of the entire TP emissions in water pollutants. The agricultural non-point source pollution perilously threatens China’s food security, food safety, and the sustainable development of society [2]. The Chinese government holds great concern regarding agricultural non-point source pollution. In October 2022, the report of the 20th National Congress of the Communist Party of China emphasized “enhancing soil pollution control at its source and treating new pollutants” and subsequently initiated a series of measures and policies for governance. For instance, in the “14th Five-Year Plan” for National Economic and Social Development, specific implementation objectives and strategies have been proposed for the administration of agricultural non-point source pollution: “Profoundly execute the reduction of pesticides and fertilizers, curb agricultural film pollution, enhance the utilization rate of agricultural film recycling, encourage comprehensive utilization of straw, and promote the resource utilization of livestock and poultry manure”.

Research on the management of agricultural non-point source pollution was initiated in the 1970s in Europe and America, with a focus on the natural mechanisms of non-point source pollution. From the perspective of pollution generation mechanisms, various legal, economic, and technological measures have been implemented to prevent and control pollution, along with corresponding strategies [3,4]. However, this engineering-centric approach may lead to merely treating the symptoms instead of addressing the root causes. Therefore, it is essential to evaluate the effectiveness of agricultural non-point source pollution control to continuously improve the measures and policies. Currently, most evaluations concentrate on engineering performance indicators [5,6], with little consideration of multi-stakeholder perspectives. As a matter of fact, since agricultural non-point source pollution is closely linked to the production and livelihood behavior of farmers and other stakeholders, it is crucial to incorporate more multi-stakeholder evaluation perspectives to systematically assess the effectiveness of agricultural non-point source pollution control. The overall evaluation of farmers regarding the effectiveness of agricultural non-point source pollution control not only provides the most intuitive evaluation result but also reflects their recognition of pollution control, which is strongly related to their willingness to implement environmentally friendly behavior. Agricultural non-point source pollution control has the most significant impact on farmers, and government-led pollution control efforts cannot succeed without farmers’ participation, cooperation, and support. Therefore, strengthening the evaluation of the effectiveness of agricultural non-point source pollution control from the perspective of farmers and studying the influencing factors will help comprehensively evaluate pollution control effectiveness and understand the factors influencing farmers’ production, management, consumption behavior, and adoption of modern technology. This, in turn, will enable the implementation of a series of policy measures, such as taxation, fiscal compensation, and education incentives, to guide farmers’ behavior, encourage voluntary adoption of environmentally friendly technology, and achieve the goal of controlling agricultural non-point source pollution [7,8,9].

Currently, domestic and foreign scholars’ debates on farmers’ behavior mainly focus on whether farmers are fully rational economic actors and different understandings of farmers’ rationality will directly affect the analytical perspective and results of farmers’ behavior. Herbert Simon argued that “rationality refers to the behavior that achieves the given goals within the given conditions and constraints” [10]. Based on Simon’s bounded rationality hypothesis, this paper believes that although farmers subjectively base their behavioral decisions on profit or utility maximization, their rational behavior in decision-making exhibits certain differences from fully rational behavior due to the influence of multiple factors such as individual abilities, experience, knowledge structures, preferences, and resource endowments. Farmers’ decisions regarding production are intricately linked with their strategies for sustaining their livelihoods, and their livelihood capital plays a crucial role in shaping these strategies. Livelihood capital not only reflects farmers’ basic livelihood status but also has a significant impact on their investment decisions, production management, human-land relationships, and behavioral intentions [11,12,13]. The strategies that farmers adopt to sustain their livelihoods determine the nature and intensity of their impact on the local ecological environment, with profound implications for the ecological environment [14,15]. In recent years, research on farmers’ livelihoods has become a hot topic in academia, covering various aspects such as poverty alleviation, ecological migration, ecological construction, tourism, cultural change, and environmental change. The academic community’s research on the relationship between policy systems, environmental change, land use, and farmers’ livelihoods has been relatively in-depth. For example, Sumac Elisa Cárdenas Oleas (2015) analyzed the impact and role of the government’s support for the quinoa industry on the sustainable livelihood development of poor rural populations under the background of policy formulation and implementation [16]; Bhandari et al. (2013) used the sustainable livelihood approach to analyze the degree of influence of household and community assets from a micro perspective on farmers’ livelihood transformation strategies [17]; Milan et al. (2014) conducted empirical research on the relationship between climate change, environmental change, and mountain migration, studying the differences in livelihood and population mobility patterns in households at different altitudes [18]. However, research on pollution control from the perspective of farmers’ livelihoods remains limited. While some scholars have explored the relationship between pollution control and farmers’ livelihoods [19], they have only examined the relationship between these two factors [20,21]. Other scholars have focused on the current situation and challenges of environmental governance from the perspective of social capital, exploring the role of social capital in environmental governance and making suggestions and recommendations [22]. However, these studies only examine the external environment and the behavioral control imposed by policies and institutions, without establishing the internal mechanisms through which livelihood capital influences livelihood behavior and, consequently, farmers’ evaluations of pollution control effectiveness.

Yunnan Province boasts abundant water resources and numerous natural lakes, serving both as a crucial ecological barrier and water source protection zone in China and as a relatively vulnerable area in terms of ecology and economic development. The nine basin areas of Yunnan Province’s major plateau lakes are the most densely populated, with the most frequent human activities and the most developed economy. Reinforcing the protection of these nine major plateau lake basins plays a pivotal role in Yunnan Province’s development [23]. As one of the nine major plateau lakes in Yunnan Province, Yilong Lake has long maintained its water quality between Class V and inferior Class V, severely affecting the lake’s ecological service function and the ecological safety of local residents [24]. The agricultural non-point source pollution control project is a critical undertaking for environmental improvement in the Yilong Lake basin, with its governance performance and effects on local farmers’ livelihoods under wide-ranging scrutiny from various sectors of society.

Therefore, in summary, this study focuses on evaluating the performance of governance projects based on different types of livelihood capital among farmers, analyzes the effects of farmers’ livelihood capital on governance performance evaluation, constructs an analytical framework for the relationship between livelihood capital and governance performance, and provides empirical evidence from the agricultural non-point source pollution control project in Yilong Lake, Yunnan Province to comprehend the behavior and willingness of local farmers to participate in governance. This study will furnish crucial reference value to the government to introduce targeted policies and counter the detrimental impacts of farmers’ production activities on the environment. Additionally, it will provide a novel research perspective for evaluating the project’s performance in pollution control, along with offering guidance for environmental pollution control and the sustainable development of the basin economy.

2. Materials and Methods

2.1. Overview of the Study Area

Yilong Lake is situated in Shiping County, Honghe Hani and Yi Autonomous Prefecture, Yunnan Province, and serves as the source of the Pearl River, the third longest river in China and the largest river system in southern China. Being the only lake in Honghe Prefecture with an area exceeding 30 square kilometers, it is revered as the “mother lake” of Honghe Prefecture. Furthermore, it is also the southernmost plateau freshwater lake in China, thereby underscoring the strategic importance of ensuring the ecological security and sustainable development of the economic society in the Yilong Lake basin, not just for Shiping County, but also for the southern Yunnan region.

The Yilong Lake basin serves as the hub of economic activity in Yunnan Province’s Shiping County, boasting a high intensity of agricultural development. However, agricultural non-point source pollution has become the primary cause of water pollution in the lake [25]. Farmers’ use of unreasonable farming methods has resulted in the daunting task of reducing pollution. The high-polluting agricultural cultivation techniques are widespread in lakeside, dam, and semi-mountainous areas. The cultivation of economically valuable crops, such as high-profit vegetables and fruits, which contribute to increased pollution, is continually expanding. Crop rotation and replanting indices are escalating, while the usage of pesticides and fertilizers is increasing each year. Rural living and the pollution generated by free-range livestock and poultry contribute to 51.4%, 45.1%, 49.0%, and 70.4% of the total pollution load into Yilong Lake as per statistics, thereby posing challenges to agricultural and rural pollution control [26].

Presently, the government of Shiping County has initiated the Agricultural Non-Point Source Pollution Control Project within the Yilong Lake Basin. This project aims to curb non-point source pollution in farmlands on the north bank of Yilong Lake, along the urban river, and facilitate comprehensive improvements in large-scale livestock and poultry farming along the lake and river. It also involves construction efforts that strengthen agricultural environmental protection measures, such as the restoration of ponds to the lake, soil testing, and formula fertilization, farmland land circulation, and reduced use and control of pesticides and fertilizers. The project will mainly affect Yilong Town and BaXin Town, covering 12 communities and village committees, including Dong Cheng Community, Dashui, Darui City, Dou Di Wan, Gaojiawan, Maohe Village Committee, and Haixi, Wangjiachong, Old Street, Bailang, Baxin, and Xinjie Village Committees. They have jurisdiction over a total of 80 natural villages. However, due to the wide distribution of agricultural non-point source pollution in the Yilong Lake Basin and the project being in its initial stages, the supervisory management mechanism remains imperfect. Moreover, since governance policies and project implementation significantly impact farmers’ livelihoods, which are tied to their fundamental interests, they have exhibited strong resistance to the implementation of agricultural non-point source pollution control projects. This has resulted in slow progress, long cycles, and significant difficulty in control. This article primarily investigates villages along the lake in Yilong Town and BaXin Town (see Figure 1).

2.2. Data Collection

Using random sampling and Participatory Rural Appraisal (PRA), the research team conducted a field investigation on the management of agricultural non-point source pollution in the Yilong Lake Basin from July to September 2019. To facilitate the collection and processing of research data and ensure data validity, prior to conducting the questionnaire and interview with farmers, the grassroots staff of each village committee assisted in classifying farmers in the target natural villages into pure farmers, part-time farmers, and non-farmers according to the classification criteria for farmer livelihoods in this article. After classification, a random sample of the three types of farmers was obtained for investigation.

The research team performed field research on agricultural non-point source pollution control in the Yilong Lake Basin from July to September 2019, utilizing a random sampling method and Participatory Rural Appraisal (PRA). To ensure data validity and simplify the collection and processing of research data, the team collaborated with grassroots staff from various village committees to classify households based on the criteria provided in this article. The households were categorized into three types: pure farming households, part-time farming households, and non-farming households. Following this, questionnaire surveys and interviews were conducted with random samples from each of these household types.

The research team initially focused on 11 natural villages along the north bank, southwest bank, and west bank of the Yilong Lake in the jurisdiction of Yilong Town, which are close to Shiping County. These villages included the City East Community, Da Shui He, Da Rui Cheng, Dou Di Wan, Gao Jia Wan, and Mao He Village Committee, with a total sample size of 99 households classified into pure farming households, part-time farming households, and non-farming households. Later, the research covered 11 natural villages along the east bank, south bank, and southeast bank of Yilong Lake in the jurisdiction of Ba Xin Town, including Hai Dong, Wang Jia Chong, Lao Jie, Bai Lang, Ba Xin, and Xin Jie Village Committee, with a sample size of 95 households classified into pure farming households, part-time farming households, and non-farming households. Overall, the research collected 194 random samples, with 75 pure farming households, 63 part-time farming households, and 56 non-farming households.

During the course of our research, we conducted semi-structured interviews with households randomly sampled from the population. Each interview lasted between 20 and 40 min and was designed to cover a range of topics. These included the purpose of the questionnaire, the rationale behind the design of the response options, basic information about the households, their livelihood capital situation, their attitudes towards environmental governance, their evaluation of specific indicators related to agricultural non-point source pollution control, the specific factors affecting their livelihoods with regard to agricultural non-point source pollution control, and their opinions and suggestions on the topic. The five-level Likert scale, i.e., “very dissatisfied, dissatisfied, average, satisfied, very satisfied”, was used to evaluate farmers’ satisfaction with pollution control.

2.3. Research Methods

2.3.1. Method for Classifying Livelihood Types of Farm Households

Drawing on relevant research findings on household types [27,28] and the specific context of the Yilong Lake basin, this investigation categorized farming households’ livelihood types based on the proportion of non-agricultural income in their total household income. Pure farming households were defined as those whose non-agricultural income accounted for less than 30% of their total household income; part-time farming households were those with non-agricultural income accounting for 30–90% of total household income; and non-farming households were those with non-agricultural income accounting for more than 90% of total household income. Among the sampled households, there were 75 pure farming households, representing 38.66% of the total surveyed, 63 part-time farming households, accounting for 32.47%, and 56 non-farming households, accounting for 28.87% of the total surveyed.

2.3.2. Indicators and Measures for Evaluating the Performance of Pollution Control Projects

Agricultural non-point source pollution has a strong connection with farm households, and its control effectiveness can be assessed through feedback from such households. Hence, this study focuses on evaluating and measuring performance indicators based on farm households’ perceptions. Taking into account the specific circumstances of agricultural non-point source pollution control projects in the Yilong Lake basin, the performance evaluation covers four aspects: “observable pollution control effects, the influence of control projects on farm household livelihoods, participation of farm households, and satisfaction of farm households” (as depicted in

Table 1. Performance evaluation indexes and weights of agricultural non-point source pollution control projects in the Yilong Lake Basin.

Level Indicators	The Weight	The Secondary Indicators	The Weight
Impacts of governance projects on farmers S1	0.4	impacts on actual labor T1	0.3
		Influence on cultivated land T2	0.3
		Influence on farmland fertility and crop yield T3	0.3
		Improvement of ecological awareness T4	0.1
Perceived governance effect S2	0.3	Improvement of human settlement environment T5	0.15
		Improvement of water quality of Yilong Lake T6	0.15
		Reduction in the application of chemical fertilizers and pesticides T7	0.4
		Improvement of farmland environment T8	0.3
Household participation S3	0.15	participation in decision-making T9	0.33
		Participation in planning T10	0.33
		Participate in the management and maintenance T11	0.34
Farmer satisfaction S4	0.15	Incentive policy and compensation T13	0.4
		Project construction planning and completion rate T14	0.2
		Post-project management and maintenance T15	0.2
		Environmental protection publicity, education, and training T16	0.2

).

• Specific indicators decomposition;

The impact of the pollution control project on farmers (S1) is assessed based on four dimensions: the effect on actual labor force (T1), cultivated land area (T2), soil fertility and crop yield (T3), and improvement of ecological awareness (T4).

The “perceptible effect of pollution control” (S2) is measured by four sub-indicators: improvement of the living environment (T5), the water quality of Yilong Lake (T6), reduction of fertilizer and pesticide use (T7), and improvement of the agricultural environment (T8). This aspect is a vital indicator for evaluating the environmental management performance of Yilong Lake, and the evaluation indicators in this paper are based on the farmers’ subjective perceptions.

Farmers’ participation (S3) in the project plays a crucial role in enhancing governance ability and environmental awareness among the public, as well as facilitating collaboration and communication among different stakeholders. The sub-indicators of farmers’ participation mainly include participation in decision-making (T9), planning (T10), and management and maintenance (T11).

Farmers’ satisfaction (S4) is an important indicator for measuring the project’s performance. As beneficiaries and participants in the governance process, farmers’ satisfaction is mainly decomposed into policy incentives and compensation (T12), project planning and completion rate (T13), post-construction management and maintenance of the project (T14), and environmental publicity, education, and training (T15).

2.

Determination of Weight Values;

This study adopts the Analytic Hierarchy Process (AHP) to assign weight values to the evaluation indicators through a combination of qualitative and quantitative methods based on expert ratings [29]. The AHP analysis method relies mainly on expert knowledge and experience to ensure a scientific and accurate reasoning process. It is a standardized decision-making method that has been widely used in various fields [30]. A panel of 10 experts in public management and agricultural pollution source control, as suggested by relevant research results [31], was invited to form an expert system. They were requested to fill out the judgment matrix for the first-level indicators in Table 1, following the scale instructions, to obtain the judgment matrix. To eliminate systematic errors, 10 farmers were randomly chosen, excluding those sampled in the survey. The judgment matrices for the relevant indicators were completed according to the scale table, as displayed in

Table 2. Scale description table.

Scale Value	Description
1	Both indicators are equally important
3	One indicator is more important than the other
5	One indicator is more important than the other
7	One indicator is more strongly important than the other
9	One indicator is more important than the other
2, 4, 6, 8	Judge the median value

.

Based on the aforementioned process, a judgment matrix table was obtained, which was inputted into the YAAHP (Yet Another AHP) software, resulting in the judgment matrices and weight values for each indicator. To eliminate the bias of individual perception of indicator weights, the average value of the judgment matrix weight from ten decision matrices was taken as the final weight. The calculation results are shown in Table 1.

3.

Calculation of farm household performance evaluation values.

The performance evaluation criteria were assessed based on the weighted average value of each indicator. The performance value of each evaluation indicator at each level was obtained by multiplying the weight value with the average value of the corresponding indicator. The performance evaluation values of the second-level indicators were then totaled and multiplied by their corresponding weight values of the first-level indicators. This resulted in the performance evaluation value of the first-level evaluation indicator. The total performance evaluation value of the farm household for a specific livelihood type was determined by summing up the performance evaluation values of each first-level indicator.

2.3.3. Livelihood Capital Evaluation Indicators and Measurement for Farmers

• Measurement indicators and weights of livelihood capital;

The study first designed measurement indicators suitable for assessing the livelihood capital of farmers in the Yilong Lake Basin, based on the indicator system of livelihood capital quantification research results [32,33]. The livelihood capital of farmers was evaluated based on the sustainable livelihood framework proposed by the UK Department for International Development (DFID), which includes human capital, natural capital, physical capital, financial capital, and social capital [34,35]. Five livelihood capital measurement indicators were adopted for this study, as shown in

Table 3. The index, value, and weight of livelihood capital of farmers in the Yilong Lake watershed.

Livelihood Capital Measures	The Secondary Indicators	The Assignment	Pure Farmers	Part-Time Farmers	Non- Farmers	The Weight
human capital	The overall labor capacity of the family V1	the labor force of more than 6 people is 5, 5–6, 4, 3–4, 3, 1–2, 2, 0, 1.	2.8	3.8	2.3	0.568
	The education level of the adult labor force is V2	5 for junior college or above, 4 for senior high school or technical secondary school, 3 for junior high school, 2 for primary school, and 1 for illiteracy.	3.1	3.9	3.6	0.432
natural capital	Arable land area W1	household per capital arable land area	0.89 Mu/person	0.35 Mu/person	0.08 Mu/person	0.585
	The cultivated land quality W2	cultivated land quality is 5, 4, 3, 2 and 1	4.1	2.9	1.8	0.415
material capital	Per capital housing area X1	5 for more than 45 m2; 35–45 m2 is 4; 25 to 35 m2 for 3; 15 to 25 m2 is 2; Less than 15 m2 is 1.	3.4	3.6	3.5	0.252
	Housing structure X2	steel concrete building 5, brick concrete building 4; Brick and wood house 3; Civil Building 2; Vegetation house is 1 (per capital housing area and housing structure are two indicators quantified)	4.1	4.2	4.0	0.252
	Household fixed assets X3	The proportion of the number of consumer durable, household assets, types of agricultural facilities and equipment, livestock, and poultry owned by farmers in the total number of categories listed in the questionnaire (transportation, household appliances, furniture; Water pumps, weeding machines, harvesting machines, greenhouses, and other agricultural facilities and equipment)	67%	78%	61%	0.496
financial capital	Credit capacity Y1	The likelihood of obtaining a loan. It is 5, it is bigger than 4, it is usually 3, it is smaller than 2, and it is not 1.	3.2	3.9	4.6	0.348
	Cash income Y2	Per capital annual cash income	1.1 Ten thousand	1.4 Ten thousand	1.5 Ten thousand	0.416
	Government subsidy Y3	Government rent subsidy is 2500 yuan/mu per year	2225 yuan	875 yuan	200 yuan	0.236
Social capital	Risk assistance ability Z1	The possibility that relatives and friends will offer assistance to you when you are in trouble. 5 for sure, 4 for larger, 3 for general, 2 for smaller, and not 1.	2.4	3.5	4.6	0.275
	In social network Z2	family members, the number of people who had worked as township or village cadres, public servants, technicians, teachers, doctors, enterprise and public institution workers, soldiers, and other professions: 5 for more than 4, 4 for 3, 2 for 3, 1 for 2, and 1 for none	1.7	2.4	3.6	0.241
	Participation in community organizations Z3	5 represents every-time participation, 4 represents frequent participation, 3 represents average participation, 2 represents less frequent participation, and 1 represents little participation.	2.3	3.9	3.4	0.234
	Social trust Z4	The degree of trust of villagers, relatives and friends, and village cadres is 5, 4, 3, 2, and 1	4.1	3.4	2.7	0.250

.

Next, the weight of each indicator is determined. A panel of 13 experts, including grassroots personnel involved in the environmental management of Yilong Lake and those familiar with the livelihood status of farmers in the basin, was invited to form an expert system. The AHP method was used for data processing, and the weight calculation steps were the same as the weight method for evaluating the performance of agricultural pollution control projects (as shown in Table 2). The weight values for each indicator were obtained (as shown in Table 3). The weight determination results indicated that overall household labor capacity had the highest weight for human capital, at 0.568. For natural capital, the weight of per capital cultivated land area was higher, at 0.585. For physical capital, the weight of per capital fixed asset value of the household was higher, at 0.496. For financial capital, the weight of cash income was higher, at 0.416. For social capital, the weight of risk-assistance capacity was relatively higher at 0.275. Overall, the study found that in the Yilong Lake Basin, overall household labor capacity, cultivated land area, asset value, cash income, and risk-assistance capacity have a greater impact on farmers’ livelihood capital. Analyzing the composition of farmers’ livelihood capital in the context of environmental governance can provide insights into their resilience.

2.

Livelihood capital measurement.

The field survey data collected from Yilong Lake was standardized using the deviation standardization method, which involved applying the following formula to the data from different dimensions:

$${\mathrm{x}}^{*}=\frac{\mathrm{x}-{\mathrm{v}}_{\min }}{{\mathrm{v}}_{\max }-{\mathrm{v}}_{\min }}$$

(1)

where ${\mathrm{x}}^{*}$ represents the data after processing, x represents the data before processing, and the standardized value is obtained after calculation. Through the weight value of each index and the standardized value of data, the livelihood capital value of farmers can be obtained, and the formula is as follows:

$$\mathrm{W} = {\displaystyle \sum }_{\mathrm{i} = 0}^{\mathrm{n}}{\mathrm{P}}_{\mathrm{i}}{\mathrm{Q}}_{\mathrm{i}}$$

(2)

where P_i is the weight of the evaluation index and ${\mathrm{Q}}_{\mathrm{i}}$ is the standardized value of this index.

2.3.4. Analysis Model of the Impact of Livelihood Capital on Performance Evaluation

The least squares regression model is used to analyze the impact of explanatory variables on the dependent variable, which in this study is the relationship between the livelihood capital of farmers and the performance evaluation of different livelihood types. The method aims to minimize the sum of squared errors between predicted values and actual observations in the regression empirical model [36,37]. Therefore, this study uses the least squares method (OLS) to perform linear regression on the relationship between farmers’ livelihood capital and performance evaluation, in order to examine the impact of various components of livelihood capital on the performance evaluation of different types of livelihoods. The regression model is constructed using Eviews software, and the specific formula is as follows: The regression model is constructed using E-views software, and the specific formula is as follows:

$$\begin{array}{c}{\mathrm{Per}}_{\mathrm{n}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{V}}_1+{\mathsf{\beta }}_2{\mathrm{V}}_2+{\mathsf{\beta }}_3{\mathrm{W}}_1+{\mathsf{\beta }}_4{\mathrm{W}}_2+\\ {\mathsf{\beta }}_5{\mathrm{X}}_1+{\mathsf{\beta }}_6{\mathrm{X}}_2+{\mathsf{\beta }}_7{\mathrm{X}}_3+{\mathsf{\beta }}_8{\mathrm{Y}}_1+{\mathsf{\beta }}_9{\mathrm{Y}}_2+{\mathsf{\beta }}_{10}{\mathrm{Y}}_3+{\mathsf{\beta }}_{11}{\mathrm{Z}}_1+{\mathsf{\beta }}_{12}{\mathrm{Z}}_2+{\mathsf{\beta }}_{13}{\mathrm{Z}}_3+{\mathsf{\beta }}_{14}{\mathrm{Z}}_4+\mathrm{c}\end{array}$$

(3)

In Formula (3), subscript n represents different peasant households; C is a constant.

The specific research roadmap is shown below (see Figure 2):

3. Results

3.1. Farmers’ Performance Evaluation of Farmland Non-Point Source Pollution Control Projects

Based on the aforementioned method, the survey results were computed, and the performance evaluation values of farmland non-point source pollution control were derived for three types of households: pure farming households, part-time farming households, and non-farming households. The results are presented in

Table 4. Performance evaluation of non-point source pollution projects by farmers with different livelihood types.

Level 1 Indicators (Weight)	Secondary Indicators (Weight)	Pure Farmers	Part-Time Farmers	Non-Farmers
S1 (0.4)	T1 (0.3)	2.267	3.222	3.946
	T2 (0.3)	1.747	2.857	3.786
	T3 (0.3)	2.347	3.222	3.821
	T4 (0.1)	2.600	3.238	3.982
S2 (0.3)	T5 (0.15)	3.267	3.571	3.768
	T6 (0.15)	3.320	3.397	3.357
	T7 (0.4)	2.600	3.413	3.911
	T8 (0.3)	3.053	3.222	3.714
S3 (0.15)	T9 (0.33)	1.160	1.349	1.750
	T10 (0.33)	1.107	1.460	1.571
	T11 (0.34)	2.180	2.683	2.446
S4 (0.15)	T12 (0.4)	2.227	2.270	3.589
	T13 (0.2)	3.893	3.921	4.107
	T14 (0.2)	3.440	3.619	3.839
	T15 (0.2)	2.600	2.921	3.857

.

The maximum value of agricultural non-point source pollution project performance evaluation is 1, and the minimum value is 0. This evaluation index ranges from 0.0000 to 1.0000 and is divided into five levels, ranging from level 1 to level 5. Level 1 corresponds to an evaluation index range of 0.0000 to 0.2000, while level 5 corresponds to an evaluation index range of 0.8000 to 1.0000. According to this level division, the meaning of the evaluation index in different ranges are as follows: level 1 is very poor; level 2 is poor; level 3 is general; level 4 is good; and level 5 is very good. To simply and intuitively reflect the final results of environmental governance, the performance evaluation index grading standards of the Yilong Lake farmland non-point source pollution control project are shown in

Table 5. The grading standard of Yilong Lake environmental governance project performance.

Evaluation Index	0.0000–0.2000	0.2001–0.4000	0.4001–0.6000	0.6001–0.8000	0.8000–1.0000
Level	1	2	3	4	5
Meaning	very poor	poor	general	good	very good

:

Based on the evaluation index system presented in Table 5, the performance grade of farmers with different livelihood types has shown in

Table 6. The performance evaluation value and performance grade of farmers with different livelihood types.

Level 1 Index	Pure Farmers		Part-Time Farmers		Non-Farmers
	Value Rating	Performance Rating	Value Rating	Performance Rating	Value Rating	Performance Rating
S1	0.4337	general	0.6882	good	0.7728	good
S2	0.5888	general	0.6754	good	0.7495	good
S3	0.2979	poor	0.3678	poor	0.3855	poor
S4	0.5755	general	0.6000	good	0.7592	good
Total performance	0.4811	general	0.5969	general	0.7057	good

:

Table 6 indicates variations in the assessment of the agricultural non-point source pollution control performance among different types of households. Pure farming households view the total performance of agricultural non-point source pollution control as 0.4811, with a performance level of “general”; part-time farming households rate the total performance value as 0.5969, with a performance level of “general”; non-farming households assess the total performance value as 0.7057, with a performance level of “good”.

Livelihood type emerges as a vital determinant influencing farmers’ evaluation of non-point source pollution control performance. Pollution control projects that have a greater impact on farmers’ livelihoods receive significantly lower evaluation results than those with less impact on livelihoods, but all show relatively high recognition and satisfaction with the ecological environment improvement performance of pollution control. In terms of the S1 index, which reflects the effect of control measures on farmers, the average evaluation values of pure farming households for labor force impact, cultivated land area, farmland fertility, yield, and household income are all lower than those of part-time farming households and non-farming households. From the S2 index’s perspective, which reflects the perceivable effect of control measures, pure farming households receive the lowest evaluation value, followed by part-time farming households, while non-farming households have the highest evaluation value. As for the S4 index, which mirrors farmers’ satisfaction, the assessment values of the three types of livelihood households follow the order of pure farming households < part-time farming households < non-farming households.

The agricultural non-point source pollution control projects in the Yilong Lake Basin are closely intertwined with agricultural production activities. Pure farming households, due to their larger cultivated land areas, are more directly affected in the short term compared to part-time farming households and non-farming households. For instance, the implementation of wastewater recycling and soil testing, and formula fertilization projects, which are directly related to the amount of fertilizer used and crop yield, can result in increased costs of water-saving irrigation and impact the fertility of farmland. Such policies can have the greatest impact on the interests of pure farming households, thereby affecting their willingness to participate and subsequently influencing their evaluation of the project’s control performance. From the perspective of farmers’ participation, the evaluation values of the three types of farming households are in the order of pure farming households < part-time farming households < non-farming households. The participation degree of farmers depends on their education level and economic interests. Education can enhance their environmental awareness and willingness to participate in environmental governance, which ultimately depends on the benefits and losses to oneself.

Overall, indicators that are closely related to farmers’ livelihoods make them more concerned about short-term direct benefits and losses. If the benefits are high, their evaluation will be high, and if the losses are significant, their evaluation will be low. Therefore, it is essential to consider compensation mechanisms for control projects while also strengthening promotion and education on farmers’ long-term ecological benefits awareness.

3.2. Impact of Livelihood Capital on Performance Evaluation of Governance Projects of Different Livelihood Types of Peasant Households

This article employs the least squares method to establish a regression model for investigating the correlation between livelihood capital indicators and the performance evaluation of agricultural non-point source pollution control projects among the three types of households. The outcomes are presented in

Table 7. The influence of livelihood capital on the project performance evaluation of farmers with different livelihood types.

Livelihood Capital Index		Performance Evaluation Value
Level 1 indicators	Secondary indicators	Pure farmers	Part-time farmers	Non-farmers
Human capital	V1	−0.044 ***	−0.163 ***	−0.007
		(−2.71)	(−5.89)	(−0.63)
	V2	0.001	0.111 ***	0.014 *
		(0.07)	(3.28)	(2.03)
Natural capital	W1	−1.085 ***	−0.050 *	−0.097
		(−2.12)	(−1.00)	(−0.40)
	W2	−0.063 ***	0.002	0.005
		(−3.40)	(0.13)	(0.41)
Physical capital	X1	−0.006	0.002	−0.001
		(−0.81)	(0.12)	(−0.07)
	X2	−0.013	0.074	0.009
		(−1.87)	(2.47)	(0.67)
	X3	0.093 *	0.337 ***	−0.566
		(0.30)	(0.48)	(−1.21)
Financial capital	Y1	−0.006	−0.010	0.011 ***
		(−0.81)	(−0.78)	(0.51)
	Y2	0.437 **	0.166 *	0.480 ***
		(4.46)	(2.69)	(5.69)
	Y3	1.609 ***	−0.370	−0.647
		(3.01)	(−0.81)	(−1.06)
Social capital	Z1	−0.003	−0.004	0.070
		(−0.20)	(−0.30)	(3.06)
	Z2	0.016	0.004	0.096 ***
		(1.88)	(0.29)	(3.65)
	Z3	0.002	0.005 ***	0.009
		(0.17)	(0.14)	(0.77)
	Z4	0.072 ***	−0.003	−0.003
		(3.27)	(−0.23)	(−0.26)
	_cons	1.821 ***	3.226 ***	2.735 ***
		(6.05)	(5.69)	(9.99)

Note: ***, **, and * represent significance at the 1%, 5%, and 10% levels, respectively, and the brackets represent the T statistic.

. It is evident that households with distinct livelihood types exhibit dissimilar evaluations regarding the effectiveness of agricultural non-point source pollution control projects in the Yilong Lake basin. The marked disparities in their livelihood capital configuration are likely the primary cause. As the households are influenced by different livelihood capital indicator factors, their evaluations of project performance vary accordingly.

3.2.1. The Influence of Livelihood Capital of Pure Farming Households on the Evaluation of Governance Project Performance

Within the human capital of households solely dedicated to farming, the V1 index measuring the adult labor force within the family has a substantial negative influence on the assessment of project performance. Specifically, for each incremental unit of family adult labor, the evaluation of governance project performance in pure farming households diminishes by 0.044. This denotes that the greater the labor force within pure farming households, the more significant the impact of agricultural source pollution control projects upon them. This observation is confirmed by the T1 index, revealing that the evaluation of governance projects’ impact on the actual labor force in pure farming households is the lowest among the three household types. The source pollution control projects enacted within the Yilong Lake Basin, including the reversion of farmland to forests, wetlands, and other endeavors, carry the greatest influence on the labor force of farming households. Given that pure farming households solely rely on income from farming, they are bound to experience income reduction in the short term if the cultivated land area significantly decreases and the labor force becomes idle. Thus, the more labor force present in pure farming households, the more probable they are to hold negative appraisals of governance projects and their effects.

Regarding natural capital, the W1 indicator measuring the cultivated land area has a pronounced negative impact on the evaluation of project performance for pure farming households. With every unit increase in the cultivated land area, the project performance evaluation declines by 1.085. The mean cultivated land area among pure farming households in the sample is 0.89 mu, the highest among the three household types. Agricultural source pollution control projects directly or indirectly affect cultivated land, with pure farming households being most significantly impacted. In the project performance evaluation system that pertains to cultivated land indicators, pure farming households are consistently given the lowest evaluations. The W2 indicator measuring the quality of cultivated land similarly has a negative impact on the evaluation of project performance for pure farming households. The soil in the Yilong Lake basin is relatively fertile, and it is well-suited for the cultivation of various crops such as vegetables, soybeans, wheat, and corn. The quality of cultivated land is directly proportional to the crop yield and economic benefits. Therefore, the better the quality of cultivated land, the more substantial the impact of the control project on the households, resulting in a lower evaluation. Pure farming households have the largest cultivated land area and the highest proportion of good-quality cultivated land. As a result, their evaluation is significantly and negatively influenced by cultivated land quality. This also explains why the evaluation of pure farming households is the lowest in indicators such as the impact of control projects on farmland fertility, the use of fertilizers and pesticides, and the improvement of the farmland environment.

In the realm of financial capital, subsidies from the government towards the Y3 index have a noteworthy and affirmative influence on the evaluation of the pure farmers’ performance in project governance. The term “government subsidies” denotes compensation payments made to farmers for various matters, including land expropriation and rent. With each unit increase in government compensation, the assessment of the pure farmers’ performance ascends by 1.609. Policies such as the restoration of farmland to forests and lakes, the restoration of wetlands, and the restoration of ponds to lakes offer a higher level of compensation and rent, resulting in greater satisfaction among farmers, and leading to enhanced performance evaluations. To pure farmers, farmland represents their livelihood capital and natural capital, yet it also connotes their significant reliance on farmland, limited ability to transition to other livelihoods, and therefore high expectations for compensation. The pure farmers demonstrate particularly low evaluations of incentive policies and compensation T12 indices, reflecting their dissatisfaction with compensation payments. Agricultural non-point source pollution control projects entail land recapture and leasing matters, and the fairness of compensation pricing and timely payment of compensation payments are closely linked to farmers’ economic interests. Presently, the compensation payment for the restoration of farmland to lakes is 5000 yuan per mu, payable as a lump sum, while the rent for the restoration of wetlands is 2500 yuan per year, payable every three years. This indicates that after the land is recaptured; if farmers do not expeditiously transition to other livelihoods, they will solely receive compensation payments and have no income. Projects such as the restoration of ponds to lakes and the restoration of farmland to forests do not offer any compensation payments. Of course, most of these farmlands were illegally occupied by farmers, and while farmers can understand the recapture of land, they have difficulty accepting the absence of compensation for crops grown on the land or fish raised in the ponds. Based on current market prices, farmers’ annual income from farming is around 10,000 yuan, and the compensation payments and rents for the control projects fall far below farmers’ expectations. For pure farmers, who are constrained by their livelihood capital, the unfavorable impact of government subsidies Y3 on performance evaluation explains why they have a low evaluation of incentive policies and compensation T12 indices.

The social trust level Z4 indicator within the social capital of pure rural households exhibits a significant positive impact on the evaluation of project governance performance in financial capital. With each one-unit increase in social trust level, the performance evaluation value of pure rural households increases by 0.072. The higher the social trust level of pure rural households, the more robust their willingness to engage in governance, and the better their evaluation of governance performance. Interviews with pure rural households reveal that they hold a relatively favorable view of the effectiveness of environmental governance and the government’s work in governance, believing that the implementation of environmental governance has greatly helped improve the local ecological environment. Despite their dissatisfaction with the amount and promptness of compensation, and their concerns about the duration of land requisition, most rural households still respond positively to policies. They dutifully submit their land within the prescribed time and express their support for pollution control projects.

3.2.2. Impact of Livelihood Capital of Part-Time Farmers on Governance Project Performance Evaluation

Within households engaged in part-time farming, the human capital indicator of V1, representing family adult labor, has a notable detrimental impact on the performance evaluation of governance projects. Specifically, each increase of one unit in family adult labor results in a decrease of 0.163 in performance evaluation. As part-time farmers lack an advantage in natural capital, family labor capacity is a crucial form of livelihood capital. Therefore, they must acquire not only agricultural production knowledge and technology but also other industry-specific livelihood skills. Governance projects place an increased urgency on the transition of family labor force and skills from non-agricultural to non-farming industries, which poses difficulties in adaptation and transition, testing the livelihood capacity of part-time farmers. Hence, the T1 indicator of labor impact on governance project performance evaluation is relatively low for this group. In contrast, the V2 indicator of adult labor education level has a significantly positive impact on governance project performance evaluation, increasing by 0.111 for each unit increase in the adult labor education level of households engaged in part-time farming. This finding indicates that the higher the education level of family labor, the more positively they participate in and evaluate governance work.

Concerning material capital, the X3 indicator of the quantity of family fixed assets has a substantial positive impact on governance project performance evaluation, increasing by 0.337 for each unit increase in family fixed asset quantity. Owning agricultural equipment and facilities, such as water pumps, weed cutters, harvesters, greenhouses, and more vehicles, enhances the efficiency of part-time farmers’ work, strengthens their ability to withstand risks, and enables them to carry out meticulous farming.

Concerning social capital, the Z3 indicator of participation in community organizations has a significantly positive impact on governance project performance evaluation, increasing by 0.005 for each unit increase in participation in community activities. Field surveys reveal that among the three types of farmers’ livelihoods, part-time farmers exhibit the highest proportion of participation in community activities, the strongest awareness of environmental protection, and the greatest support for collective work and public service. Part-time farmers achieve a higher evaluation than pure farmers and non-farmers on the T11 indicator of farmer participation in management and maintenance, indicating that more part-time farmers participate in the management and post-maintenance of environmental governance projects, and they evaluate their participation in this work more highly.

3.2.3. Impact of Non-Farm Livelihood Capital on Governance Project Performance Evaluation

The indicator of credit capacity Y1 in non-agricultural financial capital has a substantial positive impact on the evaluation of governance projects. Each increase in the credit capacity of non-agricultural households results in a performance evaluation value increase of 0.011. The livelihoods of non-agricultural households in the Yilong Lake Basin mainly involve tertiary industries such as accommodation, catering, and retail. Non-agricultural households with robust credit capacity possess stronger livelihood capabilities and are better equipped to tackle the impacts of governance projects.

The cash income Y2 indicator also has a notable positive effect on the performance evaluation of governance projects. For every unit increase in non-agricultural households’ cash income, the performance evaluation value increases by 0.48. Non-agricultural households with higher cash income possess stronger livelihood capabilities and are less impacted by governance measures.

The case of non-agricultural households in Mao Ha Village serves as a prime example to demonstrate this relationship. In this village, a significant proportion of non-agricultural households were previously part-time farmers who shifted to non-farming livelihoods due to the implementation of policies such as the ban on fish farming in cages, the prohibition of “restaurants on boats”, and the ban on fishing during the non-fishing season. To aid these fishermen in transitioning to reasonable livelihood paths, the government established a new boat restaurant business in the ponds of the tourist center on the west bank of Yilong Lake. This fish pond is separated from Yilong Lake by a circumferential road, which not only preserves the tourist attraction of boat restaurants in Yilong Lake but also addresses the water pollution issue caused by traditional boat restaurants while also taking into account farmers’ livelihood concerns.

Several projects of Yilong Lake Tourist Fishing Boat Company are collectively invested in by Mao Ha Village residents. With government support and assistance, these villagers raised investment funds through borrowing, which are managed and operated by the company. Shareholders of the boat restaurant project receive dividends of approximately 40,000–50,000 yuan per year. Villagers who wish to become employees of the fishing boat company can also receive a monthly income. Thus, the villagers of Maoha Village are particularly content and supportive of the government’s environmental governance, and their performance evaluation of the source pollution control project is higher than the average evaluation of other villages along the lake.

In non-agricultural social capital, the social network Z2 indicator has a positive and significant effect on the performance evaluation value of governance projects. Each increase in non-agricultural households’ social network results in a performance evaluation increase of 0.096. Survey data reveals that the number of people employed in government agencies, public institutions, and village-level units in non-agricultural households is significantly higher than in part-time and pure farming households. The stronger the social network of non-agricultural households, the greater the promotional efforts of enterprise and institution family members in promoting environmental protection to farmers. This is particularly typical of the villagers in Xiaoshui, Dashui, Longjing, Shaqiao, and Renshou. In the policies of returning farmland to lakes and wetlands, even if the villagers are dissatisfied with the compensation, these socially networked farmers are the first to respond to the policy by surrendering their land. These non-agricultural households originally had very little cultivated land, and after surrendering their cultivated land, they became landless farmers. However, they responded to the governance policy in a timely manner under the persuasion of family members working in government and enterprise departments.

4. Discussion

This paper takes the agricultural non-point source pollution control project in the Yilong Lake basin as the background, and through a comparison and summary of the characteristics of different types of farmers’ livelihood capital, analyzes the effect of different types of farmers’ livelihood capital on the performance evaluation of agricultural non-point source pollution control projects.

Firstly, the agricultural non-point source pollution control project requires coordination among various stakeholders, with farmers being the most fundamental participating party. Their level of identification and participation is a crucial aspect of the long-term sustainability of pollution control [38]. Currently, the performance evaluation of pollution control focuses on natural engineering indicators, lacking subjective evaluation of the project’s performance by farmers [39,40,41]. Finding measurable indicators from the perspective of farmers and designing performance evaluation indicators based on the research objectives is an innovation and breakthrough from the previous focus only on environmental, social, and economic performance indicators. Comparing farmers’ livelihoods and environmental governance effects and analyzing the impact of farmers on environmental governance performance can solve the problem of academic emphasis on result performance in environmental governance project performance evaluation and provide a new perspective for agricultural non-point source pollution control and government performance evaluation. Surveying farmers’ subjective evaluation and analyzing influencing factors from the perspective of farmers’ livelihood capital framework can undoubtedly further propose policies for improving governance projects, promoting farmers’ recognition and participation in governance projects at the grassroots level, and greatly benefiting the successful implementation of projects [42]. However, this study uses cross-sectional data and lacks longitudinal comparative data, which fails to comprehensively reflect the influencing situation. The research project involves environmental protection issues, and many data are sensitive, which affects the comprehensiveness and accessibility of the data to a certain extent. There is still room for improvement in the scientific rationality and comprehensive synthesis of indicator selection and application methods.

Secondly, the livelihood capital of farmers in the Yilong Lake basin is generally low. The adaptation effect of farmers to changes in the living environment after non-point source pollution control is not ideal, and the sustainable development of their livelihoods is greatly restricted, mainly manifested in the insufficient buffering capacity of their livelihoods, poor stability, and low productivity [43]. Changing farmers’ livelihood strategies is an effective way to improve the performance evaluation of agricultural non-point source pollution control projects. In fact, in the performance evaluation of the Yilong Lake basin control project by the three types of farmers’ capital, pure farmers, dual farmers, and non-farmers, their dependence on the ecological environment quality is gradually decreasing [44,45], and their evaluation of the project’s performance is gradually increasing. Therefore, to improve the performance evaluation, it is necessary to start from a deeper perspective. This paper believes that changing livelihood types is an important way to improve performance evaluation, and the primary way is to increase the proportion of non-farmers. Since non-farmers have a relatively small dependence on and impact on the ecological environment, non-farm livelihoods reduce environmental damage and consumption while creating sufficient employment opportunities and sources of income [46,47]. Therefore, promoting farmers to shift from agriculture to non-farm activities and increasing the degree of diversification is a new opportunity chain for non-point source pollution control. However, this process is bound to be complex and slow. A more feasible method is to gradually shift pure farmers to dual farmers or non-farmers, encourage local residents to change their livelihood strategies or achieve livelihood diversification, and enable more pure farmers to achieve diversion through labor transfer or industrial upgrading, thereby promoting the overall improvement of project performance evaluation.

Thirdly, optimizing the structure of livelihood capital can improve the performance evaluation of agricultural non-point source pollution control projects. After the discussion above, it is worth focusing on how to effectively change farmers’ livelihood types. Since the differences in livelihood capital help explain environmental differences, non-farm activities promote the diversification of farmers’ livelihoods, enhance their autonomy, and reduce their dependence on the environment and resources [48]. Therefore, changing the combination of livelihood capital and increasing the proportion of non-farm livelihoods is an important way to fundamentally improve performance by changing livelihood types. Currently, the diversification of non-farm livelihoods is the main trend of China’s livelihood strategy transformation [49]. With the government’s promotion, it is easier to successfully transform the types of farmers’ income by increasing the proportion of non-farm capital. Therefore, it is necessary to adopt targeted strategies to adjust the composition of farmers’ livelihood capital to achieve livelihood diversification and transform their current situation.

(1)

For pure farmers, overall labor capacity, land leasing, family asset value, cash income, and social network are significantly positively correlated with the performance evaluation of agricultural non-point source pollution control projects. This indicates that the production and life of pure farmers have a direct dependence on the resources and environment in the basin, which indirectly reflects that the performance evaluation of pure farmers is largely determined by their existing abilities and capital. Therefore, to improve the performance evaluation of agricultural non-point source pollution control projects, it is necessary to first increase the cash subsidy for pure farmers. The government should provide various supports, carry out vocational skills training, education, and cultural training tailored to local conditions to improve the overall labor capacity of pure farmers; promote land leasing through measures such as providing small loans and government transaction platforms, strengthen social networks, and increase the proportion of human capital, financial capital, and social capital in the total livelihood capital, promote the transition from farmers to non-farmers, and fundamentally improve the ability to control and respond to agricultural non-point source pollution.

(2)

For part-time farmers, overall labor capacity, property value, and risk assistance capacity are significantly positively correlated with the performance evaluation of agricultural non-point source pollution control projects. Compared to pure farmers and non-farmers, part-time farmers’ work is more complex and they face more risks and difficulties. Therefore, the overall labor capacity requirements of part-time farmers are the highest among the three livelihood types. They not only require professional agricultural production capabilities similar to pure farmers, but also need high non-agricultural sensitivity similar to non-farmers. The complex livelihood type makes the risk assistance capacity requirements of part-time farmers fall between pure farmers and non-farmers. Therefore, it is very important to provide the necessary training for part-time farmers to improve the performance evaluation of the project. In addition to organizing experts to regularly train part-time farmers in agriculture, animal husbandry, and other knowledge, attention should also be paid to helping solve other problems encountered by part-time farmers in production and life. Both agricultural labor skills and non-agricultural development capabilities should be considered to meet both skill requirements. Furthermore, part-time farmers have a higher demand for risk avoidance, and the government should create conditions to play the role of agricultural insurance in protecting farmers, appropriately increasing premium subsidies so that planting and breeding industries are protected from natural disasters and accidents, and reducing economic losses caused by them.

(3)

For non-farmers, the indicators of the education level of the adult labor force, property value, credit capacity, and risk assistance capacity are significantly positively correlated with the performance evaluation of agricultural non-point source pollution control projects. Since the income of non-farmers does not depend on the use of agricultural resources, a feasible strategy is to reduce their compensation for agricultural non-point source pollution while providing convenient conditions for their production and operation, such as lowering their loan threshold, increasing their loan amount, reducing taxes and fees, providing technical support, and helping to reduce investment risks.

5. Conclusions

The livelihood type is an important factor affecting the performance of agricultural non-point source pollution control. Farmers of different livelihood types have different evaluations of the performance of control projects, which is closely related to the composition of their livelihood capital. Livelihood capital reflects the livelihood status of farmers, directly affecting their behavior willingness, in turn, affecting their livelihood type, thereby affecting their evaluation of the performance of agricultural non-point source pollution control. Therefore, this article constructs a framework for evaluating the performance of control projects based on livelihood capital and different livelihood types of farmers to understand their behavior and willingness to participate in control projects, the following conclusions can be drawn:

(1)

According to the actual livelihood status of farmers in the Yilong Lake basin, this article divides them into three types: pure farmers, part-time farmers, and non-farmers, to evaluate the performance of agricultural non-point source pollution control projects. The results show that the performance evaluation of pure farmers on the control project is 0.4811, and the evaluation level is general; the performance evaluation of part-time farmers on the control project is 0.5969, and the evaluation level is general; the performance evaluation of non-farmers on the control project is 0.7057, and the evaluation level is good. It can be seen that there are differences in the performance evaluation of farmers of different livelihood types, with the performance evaluation values ranked from high to low as non-agricultural households > part-time agricultural households > pure agricultural households. The performance results calculated by this method are consistent with our field research. This suggests that the performance evaluation system for environmental pollution control projects based on farmers’ livelihoods is scientific and reasonable.

(2)

Using livelihood capital as a framework, it was discovered that the indicator of household labor capacity in human capital has a significant negative impact on the performance evaluation of pure farmers and part-time farmers; the education level of adult labor has a significant positive impact on the performance evaluation of part-time farmers and non-farmers. In natural capital, the indicator of cultivated land area has a significant negative impact on the performance evaluation of pure farmers and a negative impact on the performance evaluation of part-time farmers; the cultivated land quality indicator has a significant negative impact on the performance evaluation of pure farmers. In material capital, the indicator of fixed assets of the household has a significant positive impact on the performance evaluation of part-time farmers. In financial capital, the credit capacity indicator has a significant positive impact on the performance evaluation of non-farmers; the cash income indicator has a significant positive impact on the performance evaluation of non-farmers and a positive impact on the performance evaluation of pure farmers and part-time farmers. The government subsidy indicator has a significant positive impact on the performance evaluation of pure farmers; the social network indicator in social capital has a significant positive impact on the performance evaluation of non-farmers; the participation in community organization activities indicator has a significant positive impact on the performance evaluation of part-time farmers; and social trust has a significant positive impact on the performance evaluation of pure farmers. Based on the above conclusions, it can be seen that the deep reasons for the differences in performance evaluation values lie in the different compositions of livelihood capital among farmers of different livelihood types. Key livelihood indicators that have a significant impact on performance evaluation values influence farmers’ attitudes and behaviors towards agricultural non-point source pollution control. Therefore, by adjusting the influential factors of significant livelihood capital indicators and targeting the changes in the livelihood capital capacity of farmers of different livelihood types, we can improve farmers’ sustainable livelihood and risk avoidance capacities, thereby changing their evaluation values of the performance of agricultural non-point source pollution control.

6. Policy Implications

Based on the discussion and conclusion, the following recommendations can be made for agricultural non-point source pollution control projects:

(1)

Strengthen the diversification of governance projects and compensation standards and methods from the perspective of farmers’ needs. Different compensation policies should be implemented based on the livelihood composition of farmers. For pure farmers, who have the largest cultivated land area and suffer the greatest impact from governance, the government should make corresponding adjustments to the actual situation and provide additional subsidies based on market demand. For part-time farmers, the government should establish a platform for capacity building, provide employment guidance and labor skills training, broaden the employment channels, and improve their labor mobility. For non-farmers, their compensation can be appropriately reduced, while facilitating their non-agricultural production and management, such as lowering their loan thresholds, increasing loan amounts, reducing taxes and fees, providing technical support, and helping to reduce investment risks to ensure their sustainable livelihoods.

(2)

Enhance policy transparency and information symmetry to improve farmers’ participation. Farmers’ participation in pollution control is evaluated as poor for all three livelihood types, which has a negative impact on their comprehensive performance evaluation. A new model that combines government control measures, market incentives, and farmer participation mechanisms is needed to fundamentally solve the problem. The transparency of governance policy formulation should be enhanced, and farmers’ participation in policy formulation and planning should be increased to construct a new model that combines government control and farmer participation mechanisms.

(3)

Promote farmers to change their livelihood strategies, optimize their livelihood capital structure, and improve their risk resistance. Encouraging farmers to transition from agriculture to non-farm activities is an opportunity to increase the degree of part-time farming and change the opportunity chain in non-point source pollution control. Gradually shifting from pure farmers to part-time farmers or non-farmers encourages local residents to change their livelihood strategies or achieve diversification, thereby enabling more pure farmers to transfer their labor or upgrade their industries to achieve diversion. Non-farm activities promote the diversification of farmers’ livelihoods, enhance their dominant position, reduce their dependence on the environment and resources, and form environmentally friendly behaviors to solve agricultural non-point source pollution problems at the source.