Agricultural Specialization Threatens Sustainable Mental Health: Implications for Chinese Farmers’ Subjective Well-Being

Abstract

China’s agriculture is increasingly becoming more specialized. However, specialized production has disrupted traditional farming culture and may threaten sustainable mental health. This study takes Chinese farmers’ subjective happiness and agricultural production outsourcing as the research object, in an attempt to reveal the possible unhappy impacts of Chinese-style agricultural specialization represented by agricultural production outsourcing. First, we construct a theoretical framework of the relationship between agricultural production outsourcing and farmers’ subjective well-being. Secondly, based on more than 3800 household survey data collected by the Chinese Academy of Social Sciences in 2020, we use the classical econometrics and psychological analysis methods such as the Ordered Probit model and the instrumental variable estimation to conduct a rigorous impact assessment. The results show that for every doubling of agricultural outsourcing expenditure, the probability that farmers think they are very happy decreases by about 21%, and the probability that they think they are relatively happy decreases by about 9%. The groups affected by the negative psychological impact mainly include farmers growing rice and corn, farmers in hills and mountains, and farmers with small-scale operations. Further analysis shows that outsourcing risks, the weakening of farmers’ professional autonomy, and family split caused by agricultural outsourcing bring unhappiness, and the increase in income cannot offset the negative psychological effect of outsourcing. The findings of this study may bring inspiration to other countries with agricultural outsourcing markets and programs to improve the national subjective well-being.

1. Introduction

The classical theory of division of labor expounds three reasons why the division of labor improves productivity. First, the division of labor improves the work proficiency of workers; second, the division of labor saves the time for workers to switch jobs; third, the division of labor can easily lead to the invention of related machinery [1,2,3]. For example, if 10 people make needles individually, each person can only make 20 needles per day. If 10 people work together and each person is responsible for one process, then each person can produce 4800 needles per day. However, it seems difficult to deepen the division of labor in the agricultural sector, as the growth of crops is governed by natural rules [1,3,4]. People can only work in a certain production link during a specific season, such as land preparation, sowing, irrigation, plant protection, and harvest. It is impossible to keep the labor force engaged in a certain process uninterruptedly.

In the 20th century, large-scale production patterns and large agricultural machinery applications began to prevail in Western countries. The discussion on the division of labor in agriculture gradually shifted to Asia [5,6,7]. There is a growing view that it is unnecessary to overemphasize the particularity of agriculture. Technological progress and capital substitution for labor can also reduce the production cost per unit area and realize the division of labor economy in the agricultural sector in a roundabout way [8,9,10,11]. In reality, advances in agricultural technology have led to the separation of production processes, as the farm scale is often smaller than the optimal production scale of advanced technologies [12,13].

In countries where small-scale farming is common, there is a role for agricultural production outsourcing. Farmers can purchase outsourcing services provided by certain organizations with specialized production conditions to cope with the endowment defects in some or all links of family agricultural production. When there is a great demand for agricultural outsourcing in a certain region, economies of scale driven by outsourcing services can be generated [4,5]. This development model is theoretically tenable, and its economic performance in practice has been affirmed by most scholars and policy makers [14,15]. Based on this, China is vigorously promoting agricultural outsourcing services [11]. According to data from the Price Department of the National Development and Reform Commission, the average production outsourcing service cost per mu of China’s three major grain crops in 2021 was CNY 223.82, accounting for 24.88% of the total production cost. Agricultural outsourcing has become an essential business decision for 220 million small farmers in China.

Everything has its advantages and disadvantages. Since ancient times, land has provided people with a stable annual output of food. Therefore, land is regarded by Chinese farmers as the “lifeblood”, the most basic means of production and the foundation for their livelihood [16]. The farming culture of working at sunrise and resting at sunset has given Chinese people a unique feeling about the land [17,18]. Farmers are prone to worry when they hand over their land to others for farming, because others do not cherish the land as much as they do [16,19]. Furthermore, outsourcing can destroy farmers’ sense of accomplishment in providing food for their families by farming personally [20]. Over time, farmers’ sustainable mental health may be compromised. Following this logic, this study intends to reveal and quantitatively analyze the adverse effects of agricultural outsourcing on the subjective well-being of Chinese farmers for the first time.

The 2030 Agenda for Sustainable Development and the Sustainable Development Goals (SDGs) highlights “reduce by one third premature mortality from non-communicable diseases through prevention and treatment and promote mental health and wellbeing”. However, different cultural backgrounds and different disciplines give different concepts of “sustainable mental health”. The Lancet Commission on global mental health and sustainable development has listed key terms related to human mental health. Happiness and subjective life satisfaction rank first among them [21]. The reason is that obtaining what individuals really need brings happiness and positive psychological effects. That is, a state of well-being is inseparable from sustainable mental health.

As economic welfare, such as income and consumption, is increasingly regarded as an inadequate indicator of social civilization progress, individual subjective well-being has become a research hotspot in economics and psychology [22,23,24]. Happiness is the ultimate goal of all actions [25]. Governments around the world have gradually taken the promotion of national subjective well-being as a new pursuit “beyond GDP” [26,27]. In 2022, China had a rural population of 491 million, accounting for 34.78 percent of China’s total population (https://data.stats.gov.cn/easyquery.htm?cn=C01 (accessed on 15 August 2023)). The spiritual life of the large number of farmers needs attention [28]. Based on this, it is of great theoretical value and practical significance to study the changes in farmers’ subjective well-being during the development of agricultural outsourcing in China.

The marginal contribution of this paper may be as follows. Firstly, the research perspective focuses on the causal relationship between agricultural outsourcing and Chinese farmers’ subjective well-being, improving the relevant theories of the influencing factors of farmers’ happiness. Secondly, based on the household data from the nationally representative China Rural Revitalization Survey (CRRS), the Ordered Probit model and the Extended Regression Model are employed to empirically test the impact of agricultural outsourcing on farmers’ subjective well-being. Thirdly, this paper analyzes how crop type, terrain, and scale factors regulate the unhappy effect of agricultural outsourcing, which helps to reveal the well-being effect of agricultural outsourcing among different groups. Fourthly, this paper summarizes the mechanisms by which agricultural outsourcing directly and indirectly affects farmers’ subjective well-being, and uses data to verify them.

2. Background and Theory

2.1. Development and Welfare Effects of Agricultural Outsourcing in China

In 1978, China’s per capita arable land area was 0.1 hectares, only 3.7 percent of the world average (https://data.worldbank.org.cn/indicator/AG.LND.ARBL.HA.PC (accessed on 15 August 2023)). The situation of many people and little land determines that China needs to increase labor input to improve per unit yield. In the same year, the Household Responsibility System (HRS) was introduced in China, aimed at fully mobilizing farmers’ enthusiasm in production [29,30]. For a long period of time since then, China has presented an agricultural management pattern of “big country and small farmers”. However, over the past 45 years, China’s economy has undergone dramatic changes. The labor force has gradually concentrated from rural to urban areas and from agriculture to non-agricultural industries, resulting in a serious shortage of agricultural labor force. Consequently, the phenomenon of aging and part-time employment of the rural labor force has become increasingly prominent [8,9,31].

In response to it, the government called for the concentration of rural land to capable people and the development of large-scale land management. Over the past 20 years, the central government has continued to deepen the reform of the rural land property rights system, attempting to revitalize the rural land transfer market and promote large-scale agricultural management by stabilizing the farmland use right, but the effect is not as expected [29,32,33]. The reasonable explanation is that the effectiveness of this model is always limited by the incomplete land property rights [34]. According to the data of the Ministry of Agriculture, 66 percent of China’s arable land in 2020 was still managed by 220 million traditional small farmers, with an average arable area of 0.5 hectares and an average land ownership of 5.72 plots per household (http://zdscxx.moa.gov.cn:8080/misportal/public/publicationRedStyle.jsp (accessed on 15 August 2023)). Hence, the problem of fragmentation and decentralization of farmland management is serious [35].

The theory and practice of international agricultural development show that the transformation of agricultural production from traditional management mode to modern management mode such as scale, intensification, and mechanization is an inevitable way [5,10,31,36]. Since the idea of driving agricultural large-scale operation by land transferring is not feasible, the second development path of agricultural large-scale operation driven by production outsourcing has increasingly attracted the attention of academia and policy makers, in order to ensure the effective supply of agricultural products for more than 1.4 billion people [11,12].

The development of agricultural outsourcing in China can be traced back to the No.1 Central Document in 1983, which mentioned that “pre-natal and post-natal outsourcing services have gradually become an urgent need for the majority of agricultural producers”. After 25 years of exploration, the Third Plenary Session of the 17th CPC Central Committee emphasized the focus on the development of agricultural outsourcing in 2008. In 2017, the report of the 19th National Congress of the Communist Party of China clearly proposed “improving the agricultural outsourcing service system and realizing the organic connection between small farmers and modern agriculture development”. In 2018, the CPC Central Committee and the State Council issued the Strategic Plan for Rural Revitalization (2018–2022), pointing out that “it is necessary to improve the agricultural outsourcing service system, vigorously cultivate new service entities, strengthen the leading and supporting role of agricultural productive services in the modern agricultural industry chain, and build a new agricultural outsourcing service system with full coverage, regional integration and complete supporting facilities”. The No.1 Central Document in 2023 emphasized “implementing actions of promoting agricultural outsourcing services, vigorously developing outsourcing services such as alternative farming and planting, management and collection, and whole-process trusteeship, and encouraging the construction of regional comprehensive service platforms”. It can be seen that China regards agricultural outsourcing as a key content of rural revitalization and places high hopes on it.

The data released by the National Bureau of Statistics are collated in Figure 1. China’s gross agricultural outsourcing product rose from CNY 90 billion in 2003 to CNY 390 billion in 2021 (the values in Figure 1 have been reduced to comparable prices based on 2003), with a rapid growth rate. By the end of 2020, various types of outsourcing service providers in China exceeded 900,000, covering a service area of more than 1.6 billion mu. Among them, more than 900 million mu of grain crops were served, and more than 70 million small farmers were driven by services (http://www.gov.cn/zhengce/2021-07/16/content_5625385.htm (accessed on 15 August 2023)).

The existing theoretical studies and empirical experience tell us that agricultural outsourcing can help reduce the cost of farmers adopting agricultural technologies [15], improve the efficiency of resource utilization [36], and increase the yield and unit price of agricultural products [14,36]. Based on this, agricultural outsourcing improves household income and consumption expenditure [13], enhances farmers’ ability to develop independently and prevent poverty [37], and thus has considerable economic well-being effects. However, as far as we know, no studies have focused on some potential negative welfare impacts of agricultural outsourcing in China, so we plan to conduct some exploratory research on farmers’ subjective well-being.

2.2. Agricultural Outsourcing and Farmers’ Subjective Well-Being

Agricultural outsourcing may directly or indirectly affect farmers’ subjective well-being. On the one hand, the transaction risk of agricultural outsourcing increases their psychological pressure. On the other hand, agricultural outsourcing may also be associated with farmers’ subjective well-being through reduced employment autonomy, increased family income, and family split. The theoretical framework of this paper is depicted in Figure 2.

2.2.1. Direct Impact of Agricultural Outsourcing

Although outsourcing makes agricultural production economical and easy [13], the stability of outsourcing services usually faces challenges. Agricultural production outsourcing is actually a principal–agent relationship between supply and demand. Under the condition of information asymmetry, there will be a conflict of interests and priorities between the farmers who purchase outsourcing services and the service providers when they take actions [38,39]. First, farmers face the risk of uncertain working hours. The time of farming in the same region is roughly similar, which easily leads to the short supply of outsourcing services during the busy farming season. Therefore, farmers with outsourcing demand need to bear the risk of delaying farming time due to failing to search for services or queuing for services [12,40].

Secondly, agricultural output is the result of the synergistic effect between humans and nature, and people cannot accurately judge which part of factors lead to the change of agricultural output. Therefore, it is easy for service providers to adopt extensive production mode from the perspective of maximizing their own interests, without ensuring crop yield and quality, let alone considering problems such as soil fertility decline and environmental pollution [14,41]. The fuzzy correspondence between agricultural input and output makes it difficult to supervise the operation quality of employees and mechanical services. Thus, farmers are concerned about the risk of uncertain operational quality. The above two transaction risks bring uncertain expectations to farmers, which can be detrimental to happiness. If farmers lack the means to punish service providers, the problem will worsen [42,43].

Based on this, the first hypothesis is proposed:

H1. 

Agricultural outsourcing is not conducive to farmers’ subjective well-being due to its risks.

2.2.2. Indirect Impact of Agricultural Outsourcing

In practice, many aspects of agricultural production require a large number of working hours and corresponding production skills to match. Without the introduction of outsourcing services, farmers need to invest a lot of time and energy to complete the entire production process independently [7,11]. The underlying logic of the division of labor economy is the combination and allocation of different production factors. Farmers can outsource part or all of the production and operation links to the third-party agricultural outsourcing service organization with advanced technology, practiced skills, and price negotiation advantages through agricultural outsourcing services [5,15]. In other words, agricultural outsourcing can replace rural labor supply with higher efficiency, increase the opportunity cost of labor in the agricultural sector compared with outsourcing services, and help labor release from the land at a reasonable cost [8,9].

Agricultural outsourcing not only changes the mode of agricultural production, but also reconfigures the household labor force. After receiving agricultural outsourcing services, farmers usually shift the labor released by their families to non-agricultural industries [31]. Existing literature points out that the main form of off-farm transfer of rural labor force in China is to work in cities [8]. At this point, the change of farmers’ occupation type will have a conductive effect on the happiness effect of agricultural outsourcing. It is concluded that the change of farmers’ occupation type and the resulting family split and income change will have an impact on happiness.

Self-determination theory emphasizes the freedom of individuals to act in accordance with their own values [44]. Frey et al. (2004) further proposed the economic concept of “procedural utility”, believing that people’s well-being is not only influenced by the outcome of a certain event, but also gains procedural utility in practice [45]. Although engaging in both agricultural and non-agricultural activities can bring benefits, different types of employment may lead to different procedural utility. Studies on developed countries have almost consistently found that workers (the employed) are less happy than self-employed people, because the employed have less autonomy than self-employed people [46,47,48]. But this conclusion seems to be situationally limited in developing countries. If the lack of employment in a country leads to people being forced to engage in self-employment, there will be no significant difference in the self-experience and well-being of the employed and the self-employed [49,50]. In fact, China does not have this constraint. Since the reform of opening up in 1978, China’s rapid economic development has been accompanied by a huge demand for rural labor in the urban sector, and even the phenomenon of “shortage of migrant workers” in which there are more employed jobs than rural labor [51,52]. Therefore, considering that part of the driving force for farmers to work in cities is agricultural outsourcing, we speculate the following:

H2a. 

Farmers shift from the self-employed to the employed due to agricultural outsourcing, which reduces the subjective well-being of farmers.

The subjective welfare impact of farmers withdrawing from agricultural production is not limited to this. After the outsourcing of agricultural production, the middle-aged laborers in households often go out to work for a living, and the resulting family split may further worsen the happiness of farmers. According to the Migrant Workers Monitoring Survey Report released by the National Bureau of Statistics, from 2002 to 2014, the number of migrant workers increased from 104.7 million to 168.21 million year by year, of which the proportion of migrant workers with their families was only about 20%. China’s household registration system excludes the floating population from regional social security and welfare, and migrant workers cannot guarantee the quality of family life [53].

In fact, in order to alleviate the economic pressure, most farmers choose to go to the city alone and leave their families behind in the countryside [8,31]. They regularly send money to their families and only return home during major festivals, such as the Spring Festival. The separation from family members brings a strong sense of loneliness to farmers, making it difficult for them to integrate into society and reducing their happiness [54,55]. Therefore, given that part of the driving force behind the split of farmers’ families is agricultural outsourcing, we speculate the following:

H2b. 

Farmers are separated from family members due to agricultural outsourcing and career change, which reduces their subjective well-being.

The above contents are the potential negative effects of agricultural outsourcing on subjective well-being. Every coin has two sides. We need to acknowledge the benefits of agricultural outsourcing to China’s agricultural economy and farmers’ welfare at the present stage. It is worth discussing that the income-increasing effect of agricultural outsourcing may be beneficial to farmers’ happiness.

Agricultural outsourcing is a manifestation of the division of labor economy. Outsourcing services can involve farmers in the external division of labor and reduce the transaction costs of this process through organization, thereby exerting the economies of scale of services [15]. Agricultural outsourcing services covering pre-production, mid-production, and post-production effectively improve farmers’ agricultural production efficiency and utilization efficiency of agricultural materials with external efficiency advantages [36], which is conducive to the growth of family agricultural income. More importantly, agricultural outsourcing removes the “lock-in” of land to the household labor and pushes the household labor into relatively higher-paying off-farm work. Farmers can obtain considerable non-agricultural income while ensuring agricultural income. Mi et al. (2020) also found evidence of the income-increasing effect of agricultural outsourcing [13].

If agricultural outsourcing has a good effect on household income growth, then it leads to an important topic in the field of psychology, namely the relationship between income and subjective well-being. Neoclassical economics assumes that utility comes from commodity consumption, so increasing income can better satisfy individual preferences and make people happy [23]. The famous “Easterlin paradox” holds that when economic growth exceeds a certain threshold, national happiness will not rise with the continuous growth of income. However, this view is based on the relative utility theory that common progress equals relative invariance [56]. Actually, the income level of Chinese farmers is generally low. The per capita income of farmers in 2022 was CNY 20,133 (about USD 2993), which is only 41 percent of the income level of urban residents. They belong to the group of material deprivation. In a certain period of time, if the outsourcing services bring significant income increase to the farmers who receive services, it will effectively improve their quality of life and improve their psychological utility and happiness compared with those who do not outsource agriculture [57]. Based on this, we speculate the following:

H2c. 

The increase in income alleviates the negative psychological effect of outsourcing and related factors on farmers to a certain extent, and enhances their subjective well-being.

3. Data, Variables and Modeling

3.1. Data Sources

The data used in this paper come from the CRRS, a large-scale national rural tracking survey initiated by the Chinese Academy of Social Sciences. The survey involves issues such as farmers’ status and subjective attitudes, family economic status, and agricultural production, which provides good data support for this study.

The CRRS randomly selects sample provinces from the eastern, central, western, and northeastern regions of China. The sample counties are selected by equidistant random sampling method in the sample provinces, and sample towns and villages are randomly selected in the sample counties. Finally, sample farmers are randomly chosen according to the roster provided by the village committees. The first round of CRRS large-scale survey of farmers and villages was carried out in 10 provinces of Guangdong, Zhejiang, Shandong, Anhui, Henan, Heilongjiang, Guizhou, Sichuan, Shanxi, and Ningxia, from August to September in 2020. The survey data covered 50 counties and 156 towns across the country. A total of 300 village survey data and more than 3800 farmer household questionnaires were obtained, collecting information about more than 15,000 family members. It should be pointed out that the questionnaire survey was jointly completed by faculty members and graduate students of the Chinese Academy of Social Sciences, mainly through in-household visits.

CRRS plans to conduct a follow-up survey every two years. This study uses the first round of survey data published in 2020. In terms of data collation, samples of farmers who are not engaged in agricultural production are excluded. Next, we match the village data with the farmer data. Finally, cross-sectional data with a sample size of 2100 were obtained. It should be pointed out that due to the missing of some variables, the actual sample size entering the model may be reduced, as reported in the regression results.

3.2. Variable Selection

(1)

Farmers’ subjective well-being.

Subjective variables are an indispensable part of social science research, which can accurately reflect the spiritual needs of farmers and have great research value [20,28,43,46,53]. The measurement of subjective happiness usually includes single-indicator measurement and multiple-indicator measurement. Considering that the satisfaction of individuals with life or marriage in different regions may vary greatly depending on culture, and it is difficult to maintain an objective weight setting for multiple indicators, we insist on using the single-indicator measurement [20,53]. The question “Do you feel happy in your current life?” is selected from the farmers’ questionnaire to determine the happiness of farmers; the answer level is from “very unhappy” to “very happy”, a total of five levels. At the same time, in order to ensure the reliability of the research results, life satisfaction is adopted as a substitute dependent variable for a robustness test. Life satisfaction is also a common indicator to measure individual subjective well-being [23].

(2)

Agricultural outsourcing.

Strictly speaking, all the exchange relationships in which farmers hand over part or all of the links of agricultural production to legally independent market entities according to the principle of equivalence are agricultural outsourcing. The CRRS focuses on the inputs and outputs of household agricultural production such as ploughing, sowing, pesticide application, fertilization, drainage and irrigation, harvest and transportation, basically covering the whole process of agricultural production. Following the practice of scholars [41,58], we use the sum of the rental cost of agricultural machinery and the cost of hired labor per unit area spent by households in all links to represent the level of household agricultural outsourcing. Figure 3 shows the average cost of agricultural outsourcing in each surveyed province. Zhejiang, Anhui, and Guangdong provinces have the highest level of outsourcing, while Heilongjiang, Sichuan, and Guizhou provinces have the lowest level of outsourcing.

(3)

Control variables.

In order to avoid the interference of other factors on the main research content, we need to control other variables that may affect farmers’ subjective well-being. Therefore, based on relevant studies, the individual characteristics, family characteristics, and village characteristics are controlled, respectively [20,28,46].

Individual characteristics include age, gender, marriage, education level, religious belief, disease, medical security, and interpersonal conflict [42]. With the increase in age, farmers’ happiness may be enhanced. Compared with men, women may have higher levels of happiness. Marital status also greatly affects subjective well-being. Education attainment enhances happiness. Individuals with religious beliefs tend to be accompanied by higher happiness. The social security for farmers with medical insurance helps to improve happiness.

Family characteristics include family size, family income, financial assets and liabilities [43]. The number of family members, family income, and financial assets may be positively correlated with happiness. Debt may reduce happiness.

Village characteristics include natural disasters, industrial pollution, and social unrest [58]. All three factors can be detrimental to happiness.

The descriptive statistics of variables are shown in

Table 1. The definitions and descriptive statistics of variables.

Variable	Definition	Mean	SD	Min	Max
Panel 1: Subjective well-being
Famers’ subjective well-being	1 = very unhappy; 2 = unhappy; 3 = neutral; 4 = happy; 5 = very happy	3.754	0.824	1	5
Panel 2: Agricultural outsourcing
Agricultural outsourcing	Ln (household expenditure on machinery rental and hired labor per unit area last year)	3.710	2.199	0	9.903
Panel 3: Individual characteristics
Age	Farmer’s age	54.449	11.821	13	89
Gender	Male = 1; female = 0	0.757	0.429	0	1
Marriage	Married = 1; otherwise = 0	0.912	0.283	0	1
Education	Number of years of schooling since primary school	14.134	5.544	0	19
Religious belief	Yes = 1; no = 0	0.159	0.366	0	1
Disease	Diagnosis of chronic disease = 1; no = 0	0.378	0.485	0	1
Medical security	Having commercial medical insurance = 1; not having = 0	0.162	0.369	0	1
Interpersonal conflict	Dispute with others in the past 5 years = 1; otherwise = 0	0.073	0.259	0	1
Panel 4: Family characteristics
Family size	Number of family members	4.061	1.577	1	10
Family income	Ln (per capita household income)	9.258	1.477	0	15.427
Financial assets	Ln (total deposits and financial products)	10.028	4.002	0	17.518
Liabilities	Ln (the total amount owed)	2.983	4.925	0	13.998
Panel 5: Village characteristics
Natural disasters	Suffered from natural disasters in the past 3 years = 1; otherwise = 0	0.541	0.498	0	1
Industrial pollution	Having industrial pollution in the last 5 years = 1; otherwise = 0	0.090	0.286	0	1
Social unrest	Number of criminal cases last year	0.294	3.389	0	60

.

3.3. Modeling

Since subjective well-being is an ordered multi-categorical variable, it is appropriate for us to employ the Ordered Probit model, which is a causal inference model commonly used for ordered discrete data in psychology and econometrics research [24,28,53]. It is set as follows:

$$Happines{s}_i=\beta Outsourcin{g}_i+\gamma X_i+{\vartheta }_i+{\epsilon }_i$$

(1)

where the individual farmer is denoted by $i$, the happiness of farmer $i$ is denoted by $Happines{s}_i$, and the agricultural outsourcing situation of the farmer $i$ is represented by $Outsourcin{g}_i$. The control variables are denoted by $X$. The provincial dummy variable is represented by $\vartheta$, which is selected to control the influence of provincial factors on farmers’ happiness. The random disturbance term is represented by $\epsilon$. The coefficient that this paper focuses on is denoted by $\beta$, which represents the impact of agricultural outsourcing on subjective well-being.

4. Analysis of Empirical Results

4.1. Benchmark Regression

In this part, we examine the overall impact of agricultural outsourcing on farmers’ subjective well-being, and the results are shown in

Table 2. Benchmark regression results.

	Farmers’ Subjective Well−Being (Five Categories)
	(1) Ordered Probit	(2) Ordered Probit	(3) Ordered Probit	(4) Ordered Probit	(5) Ordered Probit	(6) OLS
Agricultural outsourcing	−0.536 ***	−0.590 **	−0.592 **	−0.569 **	−0.633 **	−0.487 **
	(0.248)	(0.263)	(0.277)	(0.285)	(0.297)	(0.206)
Age		0.010 ***	0.009 ***	0.009 ***	0.009 ***	0.006 ***
		(0.003)	(0.003)	(0.003)	(0.003)	(0.002)
Gender		−0.013	−0.015	−0.009	−0.023	−0.014
		(0.067)	(0.065)	(0.065)	(0.073)	(0.051)
Marriage		0.084 *	0.035	0.043	−0.022	−0.017
		(0.049)	(0.043)	(0.044)	(0.049)	(0.026)
Education		0.006 **	0.004 **	0.004 **	0.004 *	0.004 **
		(0.003)	(0.002)	(0.002)	(0.003)	(0.002)
Religious belief		0.096 ***	0.057 ***	0.071 ***	0.107 **	0.080 **
		(0.035)	(0.017)	(0.022)	(0.042)	(0.037)
Disease		−0.168 ***	−0.139 **	−0.139 **	−0.165 ***	−0.115 ***
		(0.059)	(0.059)	(0.060)	(0.057)	(0.041)
Medical security		0.135 *	0.098	0.099	0.070	0.047
		(0.076)	(0.079)	(0.078)	(0.079)	(0.053)
Interpersonal conflict		−0.406 ***	−0.385 ***	−0.391 ***	−0.343 ***	−0.255 ***
		(0.092)	(0.086)	(0.086)	(0.091)	(0.070)
Family size			0.012	0.012	0.025	0.016
			(0.018)	(0.018)	(0.016)	(0.011)
Family income			0.055 ***	0.054 ***	0.051 ***	0.039 ***
			(0.018)	(0.018)	(0.017)	(0.013)
Financial assets			0.028 ***	0.028 ***	0.025 ***	0.018 ***
			(0.007)	(0.007)	(0.006)	(0.004)
Liabilities			−0.025 ***	−0.024 ***	−0.026 ***	−0.019 ***
			(0.006)	(0.006)	(0.006)	(0.004)
Natural disasters				−0.110 **	−0.080 *	−0.045 *
				(0.048)	(0.046)	(0.026)
Industrial pollution				−0.048	−0.011	−0.015
				(0.115)	(0.108)	(0.074)
Social unrest				−0.010 ***	−0.006 ***	−0.003 **
				(0.002)	(0.002)	(0.001)
Provincial fixed effect	No	No	No	No	Yes	Yes
Pseudo R2	0.040	0.081	0.125	0.126	0.147
R2						0.104
N	2099	2092	2081	2081	2081	2081

Note: ***, **, and * denotes significant at the level of 1%, 5%, and 10%, respectively; robust standard errors clustered to county level are presented in parentheses.

. For the sake of robustness, the regression results without adding any control variables are presented in column 1. Column 2 controls individual characteristics, column 3 controls individual characteristics and family characteristics, and column 4 controls individual characteristics, family characteristics, and village characteristics. On this basis, column 5 further controls the provincial fixed effect. In addition, the Ordered Probit model is employed in columns 1–5. Column 6 is estimated using OLS to compare with column 5.

As can be seen from Table 2, when no control variables are added, the estimated coefficient of agricultural outsourcing in column 1 is significantly negative at the level of 1%, supporting the hypothesis of “outsourcing causes unhappiness” proposed in this study. After the control variables are added successively, agricultural outsourcing in columns 2–4 is still unfavorable to subjective well-being at the significant level of 5%. Even when the provincial fixed effect is controlled, the direction and size of the coefficient of agricultural outsourcing in column 5 are still robust. For reference, the results in column 6 based on OLS estimation also reveal the impact of agricultural outsourcing on farmers’ unhappiness.

In terms of individual characteristics, farmers’ increasing age, higher education, and religious beliefs all significantly increase happiness. Chronic diseases and interpersonal conflicts are not conducive to happiness. In terms of family characteristics, having higher income and more financial assets can improve farmers’ quality of life and psychological utility. More debt increases farmers’ psychological pressure, reducing their happiness. In terms of village characteristics, natural disasters and social unrest significantly reduce subjective well-being. In conclusion, the influence of control variables on subjective well-being is basically in line with expectations. It should be pointed out that, due to the endogeneity of some variables and the focus of this study on the effect of agricultural outsourcing on happiness, we do not interpret the regression results of control variables too much here.

Considering that the estimated coefficients of the nonlinear regression model are not directly comparable, the average marginal effect is further calculated to interpret the influence of agricultural outsourcing on farmers’ subjective well-being. As shown in

Table 3. Average marginal effect of agricultural outsourcing.

Farmers’ Subjective Well-Being	dy/dx	Delta-Method Std. Err.	z	p > |z|
Very unhappy	0.0438498	0.0207819	2.11	0.032
Unhappy	0.0833333	0.0391236	2.13	0.029
Neutral	0.1293276	0.0607172	2.13	0.029
Happy	−0.0936312	0.0441657	−2.12	0.030
Very happy	−0.2108789	0.0990042	−2.13	0.029

, with every doubling of household agricultural outsourcing expenditure, the probability that farmers think they are very happy decreases by about 21%, the probability that they think they are happy decreases by about 9%, the probability that they think they are neutral increases by about 13%, the probability that they think they are unhappy increases by about 8%, and the probability that they think they are very unhappy increases by about 4%. It can be seen that agricultural outsourcing significantly reduces farmers’ happiness and increases the incidence of unhappiness, and this effect is relatively obvious.

4.2. Robustness Test

4.2.1. Replace Dependent Variables

In order to enhance the credibility of regression results, the common indicator of subjective well-being, “life satisfaction”, is used as a substitute dependent variable for testing [46]. The corresponding question in the questionnaire is “Generally speaking, how satisfied are you with your current living conditions?”. Similar to happiness, an ordered five-category variable ranging from “very dissatisfied” to “very satisfied” is generated. The regression results are shown in column 1 of

Table 4. Robustness test.

	(1) Replace Dependent Variables	(2) Replace Independent Variable	(3) Change the Estimation Method
Agricultural outsourcing	−0.765 ***		−0.921 **
	(0.269)		(0.423)
Proportion of outsourcing links		−1.457 **
		(0.641)
Control variables	Yes	Yes	Yes
Provincial fixed effect	Yes	Yes	Yes
Pseudo R2	0.119	0.146	0.145
N	2081	2081	2081

Note: *** and ** denotes significant at the level of 1% and 5%, respectively; robust standard errors clustered to county level are presented in parentheses.

. Agricultural outsourcing significantly reduces farmers’ life satisfaction at the level of 1%, which is consistent with the above theoretical analysis logic and empirical results.

4.2.2. Replace Independent Variable

The existing research does not always measure the level of agricultural outsourcing by the expenditure per unit area. Another commonly used measurement is the proportion of the number of outsourcing links to the total number of links. Supporters of this approach believe that agricultural production has natural attributes, and each production link is indispensable and equally important [59]. Therefore, we re-regress with the proportion of outsourcing links as a replacement independent variable. It can be found from column 2 of Table 4 that the proportion of outsourcing links significantly reduce farmers’ happiness, which supports the results above.

4.2.3. Change the Estimation Method

The ordered probability model includes the Ordered Probit and the Ordered Logit. They differ in that the former assumes that the random error term conforms to a normal distribution, while the latter assumes that the random error term conforms to a logical distribution. In order to avoid coefficient bias caused by the selection of estimation methods, the regression results using the Ordered Logit model are reported, as shown in column 3 of Table 4. The estimated coefficient of agricultural outsourcing is still significantly negative.

4.3. Endogeneity Discussion

The econometric model set in this paper may face the threat of the endogeneity problem, namely the problem of mutual causality and missing variables. Firstly, agricultural outsourcing and farmers’ subjective well-being may be mutually causal. Farmers’ subjective well-being can partially determine the decision of household agricultural outsourcing. For example, farmers who feel unhappy may be more reluctant to leave their land to others for farming [50]. Secondly, although the key factors that affect farmers’ subjective well-being are controlled as much as possible, it is still impossible to guarantee that there are no missing key variables in the model [58]. Thus, the Instrumental Variable (IV) method is needed to be employed for further processing. Specifically, the Extended Ordered Probit Regression (Eopribit) model in the Extended Regression Models (ERMs) is introduced in this paper, which can achieve IV estimation of the ordered model [60].

As for the selection of instrumental variables, this paper draws on the ideas of relevant research and selects two instrumental variables: (1) membership of production service cooperatives; (2) the distance between the village where farmers live and the township government. The rationality lies in the fact that farmers who join production service cooperatives are more likely to connect with and reach deals with outsourcing service providers, and farmers who are located closer to township centers are more likely to conduct agricultural outsourcing [12,61]. Moreover, no studies have found that membership of production service cooperatives and the exogenous geographical distance directly affect farmers’ happiness. The estimation results of IV are shown in

Table 5. Agricultural outsourcing and farmers’ happiness: IV method.

	Eoprobit	IV-Eoprobit (First Stage)	IV-Eoprobit (Second Stage)
Agricultural outsourcing	−0.633 **		−0.898 *** (0.294)
	(0.297)
Membership of production service cooperatives		0.188 ***
		(0.062)
Distance between village and town government		−0.018 ***
		(0.005)
Control variables	Yes	Yes	Yes
Provincial fixed effect	Yes	Yes	Yes
N	2081	2073
corr(e.AE,e.SWB)	-	0.933 *** (0.174)

Note: *** and ** denotes significant at the level of 1% and 5%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

For comparison, regression results of the Eoprobit model without the IV method and the IV-Eoprobit model using the IV method are respectively shown in Table 5. The regression results of the Eoprobit model without the IV method are consistent with the benchmark regression. After using the IV method, the two instrumental variables in the first-stage regression satisfy the correlation with the key independent variable at the level of 1%. The estimated coefficients of the instrumental variables are in line with expectations. After eliminating the endogeneity of agricultural outsourcing, the estimated coefficient of agricultural outsourcing in the second-stage regression is significantly negative and the influence effect increases, indicating that the “unhappy” effect of agricultural outsourcing would be underestimated if ignoring the endogeneity of the independent variable.

4.4. Heterogeneity Analysis

4.4.1. Crop Heterogeneity

The development of agricultural outsourcing in China focuses on the production of three major grain crops, including wheat, corn, and rice. However, there are differences in the development of service outsourcing among different crop types [42]. For example, in 2019 (sample year), the national average outsourcing investment of rice planting was the highest, while that of wheat and corn was lower. Considering the gap in development of outsourcing among different crops, a question arises: will crop heterogeneity change the relationship between agricultural outsourcing and subjective well-being?

Thereby, based on the main crops planted by the sample farmers, all samples are divided into three groups for separate regression, in order to reveal the impact of crop heterogeneity. The results are presented in

Table 6. Analysis of crop heterogeneity.

	(1) Wheat	(2) Corn	(3) Rice
Agricultural outsourcing	0.182 *	−0.779 ***	−0.871 ***
	(0.107)	(0.280)	(0.313)
Control variables	Yes	Yes	Yes
Provincial fixed effect	Yes	Yes	Yes
Pseudo R2	0.153	0.150	0.157
N	600	893	504

Note: *** and * denotes significant at the level of 1% and 10%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

It is not difficult to find that outsourcing makes corn and rice growers unhappy, among which rice growers suffer more serious psychological burden. On the other hand, outsourcing makes wheat growers happier, though the statistical significance of this effect is weak. A plausible explanation for these findings is offered as follows.

The mechanization of wheat planting in China started early, popularized widely, and had a strong ability to replace labor [10]. It can be seen from the “Compilation of Cost-Benefit Data of National Agricultural Products” that the average labor cost per mu of wheat planting in 2019 was the lowest among the three grain crops, only CNY 13.05, while the average labor costs per mu of corn and rice were CNY 22.17 and CNY 80.31, respectively. Generally speaking, outsourcing does not bring obvious psychological burden to wheat growers because of the high homogeneity and high operation quality of mechanical services. Nevertheless, the employed labor is more likely to produce the moral hazard of “lying down on the job”, which requires farmers to pay a certain supervision cost to ensure operation quality. This potential risk may aggravate the psychological pressure of corn growers and rice growers [42].

4.4.2. Terrain Heterogeneity

China has a vast geographical area, and the terrain varies across China, ranging from high mountains to flat rivers. Terrain is one of the key factors that may affect the efficiency of agricultural outsourcing [58]. In the plains, farmers can easily substitute labor for farming with large agricultural machinery. In the mountainous areas, farmers can only rely on small agricultural machinery or hired labor for production, because it is difficult for large-power agricultural machinery to move easily. Based on this, we attempt to analyze the subjective welfare effects of agricultural outsourcing in different terrains.

According to the terrain data in the CRRS village survey data, the sample farmers are divided into three groups, namely plain, hill, and mountain. The results of the group regression are shown in

Table 7. Analysis of terrain heterogeneity.

	(1) Plain	(2) Hill	(3) Mountain
Agricultural outsourcing	−0.043 *	−0.863 **	−1.039 ***
	(0.024)	(0.396)	(0.317)
Control variables	Yes	Yes	Yes
Provincial fixed effect	Yes	Yes	Yes
Pseudo R2	0.151	0.183	0.157
N	954	418	709

Note: ***, **, and * denotes significant at the level of 1%, 5%, and 10%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

From the results, it can be found that outsourcing mainly worsens the subjective well-being of farmers in the mountain group and hill group, but has little adverse effect on farmers in the plain group. The logic lies in the fact that the plain area is more suitable for cross-regional operation of high-power agricultural machinery services, which further causes high-horsepower agricultural machinery to exert the scale effect of service supply, improves the efficiency and economy of outsourcing operations, and reduces farmers’ expectations of outsourcing risks [9,11]. Farmers living in hilly and mountainous areas usually only rely on small hand-held agricultural machines or hired workers to complete agricultural production. At this time, the labor substitution efficiency of outsourcing services is poor and the operation risk is high [14,40]. The heterogeneous subjective welfare effect in Table 7 arises from this.

4.4.3. Scale Heterogeneity

China’s farmland management under the HRS has the characteristics of dispersion and fragmentation. Most of the farmland is operated by small farmers. Only 34 percent of the farmland is managed by 3.65 million large-scale households such as family farms. The expansion of operation scale can not only change the connectivity between plots and give full play to the economies of scale of agricultural outsourcing, but also influence household livelihood strategies [35].

If farmers can expand the scale of operation by leasing land, or even grow into family farms, they will be less likely to go out to work. This is because the income from a large-scale farming operation is sufficient to feed a family. These clues imply that scale heterogeneity will change the subjective welfare impact of agricultural outsourcing [17].

Hence, two methods are adopted to reveal the above effect. First, in order to obtain subsidies, self-employed households with an operation scale larger than 100 mu (6.67 hectares) usually apply for registration as a family farm in the local business department. Farmers are divided into “family farm” group and “non-family farm” group based on the questions set by the CRRS. Second, the interaction term of outsourcing and farmers’ self-reported operation scale is added into the benchmark model for regression, and the coefficient of the interaction term is what we care about. The results are shown in

Table 8. Analysis of scale heterogeneity.

	(1)	(2)	(3)
	“Family Farm”	“Non-Family Farm”	Subjective Well-Being
Agricultural outsourcing	0.242 ***	−0.650 **	−0.643 **
	(0.091)	(0.308)	(0.313)
Operation scale			0.049 **
			(0.021)
Agricultural outsourcing × operation scale			0.041 ***
			(0.010)
Control variables	Yes	Yes	Yes
Provincial fixed effect	Yes	Yes	Yes
Pseudo R2	0.225	0.148	0.152
N	77	1979	2081

Note: *** and ** denotes significant at the level of 1% and 5%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

The results in column 1 of Table 8 demonstrate that agricultural outsourcing significantly improves the subjective well-being of large-scale farmers. On the one hand, large-scale contiguous farmland is more economical for mechanical operations, and farmers do not need to worry about the destruction of the quality of the leased land. On the other hand, large-scale households do not suffer from the psychological distress of the change of occupational type, and they still work for themselves [20]. The results in column 2 are consistent with the results of benchmark regression. Outsourcing creates psychological pressure for a wider range of small-scale farmers, who may suffer from multiple psychological threats of outsourcing risk, occupation type, and family split.

The results in column 3 of Table 8 confirm our conjecture. The coefficient of the interaction term is positive at the significance level of 1%, indicating that the expansion of operation scale can alleviate the unhappiness of farmers brought by production outsourcing.

In summary, Section 4.4 confirms the universality of the unhappy effects of agricultural outsourcing from the perspective of heterogeneity.

4.5. Mechanism Analysis

4.5.1. Outsourcing Risk

In order to verify the validity of Hypothesis 1, the mechanism is tested by combining more variables set by CRRS. Considering the availability of data, we set the interaction terms of “mechanical operation speed”, “social trust”, and agricultural outsourcing in the model, and select the “expectation of changes in grain yield” to regress agricultural outsourcing. The results are presented in

Table 9. Mechanism test: outsourcing risk.

	(1)	(2)	(3)
	Subjective Well-Being	Subjective Well-Being	Expectation of Changes in Grain Yield
Agricultural outsourcing	−0.720 **	−0.675 **	−0.554 ***
	(0.315)	(0.331)	(0.184)
Mechanical operation speed	−0.212 **
	(0.093)
Agricultural outsourcing × mechanical operation speed	−0.092 ***
	(0.029)
Social trust		0.773 ***
		(0.158)
Agricultural outsourcing × social trust		−0.135 ***
		(0.036)
Control variables	Yes	Yes	Yes
Provincial fixed effect	Yes	Yes	Yes
Pseudo R2	0.158	0.164	0.239
N	1707	2081	2076

Note: *** and ** denotes significant at the level of 1% and 5%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

The results of column 1 in Table 9 show that the higher the speed of machinery operation, the more unhappy agricultural outsourcing makes farmers feel. Since the price of agricultural machinery services is generally fixed per unit area, farmers will inevitably worry about the quality of outsourcing work when the machinery is operating at a faster speed [24]. The results in column 2 indicate that social trust weakens farmers’ unhappiness caused by agricultural outsourcing. When the farmers’ trust in the outside of the family is higher, it is more difficult for them to worry about the quality of outsourcing operations.

When “expectation of changes in grain yield” is taken as the dependent variable for regression, the result in column 3 tells us that agricultural outsourcing reduces farmers’ expectation of grain yield. In short, the above clues confirm the existence of outsourcing risk mechanism, so the hypothesis H1 is verified.

4.5.2. Occupation Type

The second mechanism is tested in this part. That is, whether agricultural outsourcing will promote farmers’ occupation type to shift from self-employment to employment, thus worsening subjective well-being. Accordingly, the variable of “proportion of employed persons among migrant workers last year” is generated, and its interaction term with agricultural outsourcing is added into the benchmark regression model, the results of which are shown in

Table 10. Mechanism test: occupation type.

	Subjective Well-Being
Agricultural outsourcing	−0.527 **
	(0.240)
Proportion of employed persons among migrant workers	−0.425 ***
	(0.136)
Agricultural outsourcing × proportion of employed persons among migrant workers	−0.244 **
	(0.107)
Control variables	Yes
Provincial fixed effect	Yes
Pseudo R2	0.163
N	1147

Note: *** and ** denotes significant at the level of 1% and 5%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

It is not difficult to find from Table 10 that agricultural outsourcing exacerbates farmers’ unhappiness when the number of employees among household migrant workers increases. After the outsourcing of agricultural production, the farmers’ occupation changes from self-employment to employment, resulting in psychological gap and unhappiness. In other words, when the number of self-employed persons among the household migrant worker increases, agricultural outsourcing will alleviate farmers’ unhappiness. Thus, the influence mechanism of occupation type is verified, confirming the accuracy of hypothesis H2a.

4.5.3. Family Split

In order to verify the existence of the influence mechanism of family split, two variables are generated: “working time of migrant workers last year” and “number of times that migrant workers returned home last year”. Similarly, the interaction terms between the above two variables and agricultural outsourcing are added into the benchmark regression model for regression, the results of which are presented in

Table 11. Mechanism test: family split.

	Subjective Well-Being
	(1)	(2)
Agricultural outsourcing	−0.794 **	−0.601 **
	(0.353)	(0.282)
Working time of migrant workers	−0.035 **
	(0.014)
Agricultural outsourcing × working time of migrant workers	−0.027 **
	(0.011)
Number of times of that migrant workers returned home		0.042 *
		(0.023)
Agricultural outsourcing × number of times of that migrant workers returned home		0.033 ***
		(0.009)
Control variables	Yes	Yes
Provincial fixed effect	Yes	Yes
Pseudo R2	0.168	0.164
N	1007	1000

Note: ***, **, and * denotes significant at the level of 1%, 5%, and 10%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

The results in column 1 of Table 11 demonstrate that the longer the working time of migrant workers in the family, the more unhappy farmers feel about outsourcing. After agricultural outsourcing, when one or some family members choose to work outside the township, the longer they work, the less time they spend with their families. Ultimately, family split brings pain [53]. The results in column 2 indicate that the more migrant workers return home, the more outsourcing can alleviate farmers’ sense of unhappiness. In conclusion, the impact mechanism of family split, namely hypothesis H2b, is confirmed.

4.5.4. Income Increase

This part tests the last influence mechanism, namely whether the expected income increase effect of agricultural outsourcing can partially offset farmers’ unhappiness. The farmers’ household planting income, wage income, and total income investigated by CRRS are used in the model. Also, three interaction terms between the above three variables and agricultural outsourcing are generated, respectively, which are added to the benchmark regression model. The results are shown in

Table 12. Mechanism test: income increase.

	Subjective Well-Being
	(1)	(2)	(3)
Agricultural outsourcing	−0.742 **	−0.627 **	−0.747 **
	(0.321)	(0.296)	(0.367)
Planting income	0.087 ***
	(0.016)
Agricultural outsourcing × planting income	0.055 **
	(0.023)
Wage income		0.068 *
		(0.036)
Agricultural outsourcing × wage income		0.045 ***
		(0.012)
Total income			0.049 **
			(0.019)
Agricultural outsourcing × total income			0.043 ***
			(0.008)
Control variables	Yes	Yes	Yes
Provincial fixed effect	Yes	Yes	Yes
Pseudo R2	0.150	0.148	0.159
N	1985	1976	2081

Note: ***, **, and * denotes significant at the level of 1%, 5%, and 10%, respectively; robust standard errors clustered to county level are presented in parentheses.

.

From the results, it is easy to find that the coefficients of the three interaction terms are significantly positive at least at the level of 5%. Agricultural outsourcing not only guarantees farmers’ income from grain planting, but also increases their working hours and obtains wage income, thus contributing to the increase in total household income, which has been confirmed by previous studies. Furthermore, the increase in various incomes alleviates the negative psychological impact caused by agricultural outsourcing, which is consistent with the findings of Howell et al. (2006) [57]. As Chinese farmers belong to a low-income group, the increase in income can bring positive psychological utility. Thereby, the influence mechanism of income increase is verified, confirming the accuracy of hypothesis H2c.

5. Discussion

This study focuses on farmers’ mental health issues during the development of agricultural specialization in China, and it establishes the logical connection between agricultural production outsourcing and farmers’ subjective well-being for the first time. What we want to appeal to is that each country has its own unique traditional culture. In a given cultural context, top-down economic policies can be a double-edged sword [16]. On the one hand, they may indeed help improve productivity. But on the other hand, they may have an impact on traditional culture and produce some unexpected negative effects [19].

While our view is unique, some earlier international studies have expressed the same concerns. For example, foreign direct investment in Estonia led to a serious problem of cultural conflict [62]. Another example is that unemployment has different negative psychological impacts on people in countries with different cultural backgrounds [63]. Data from 41 countries confirm that personal social relationships and sociocultural integration are highly correlated with happiness [64]. In the future, countries must take cultural and psychological factors as one of the key starting points of “happiness policies” in the pursuit of “happiness” [65].

Of course, this study inevitably has some limitations. First, subjective well-being is only one aspect of mental health. People feel that happiness does not necessarily mean that they are mentally healthy [21]. Second, due to data availability, we only use cross-section data from China. Future work could collect panel data from multiple countries to achieve international comparison and more rigorous causal inference. Third, the measurement of subjective well-being in this study is subjective happiness. Although it is feasible, future work can be better, such as introducing multiple-indicator measurement based on an entropy method or an expert scoring method [66].

6. Conclusions and Policy Implications

In reality, farmers’ mental health problems and low happiness lead to the continuous increase in farmers’ suicide rate and depression prevalence rate. Farmers’ sense of gain and happiness are the indicators that the academic community needs to pay attention to. The economic welfare effect of agricultural outsourcing has been confirmed by many research studies, but its impact on farmers’ subjective welfare is rarely addressed in the literature. Based on the theoretical analysis of the impact of agricultural outsourcing on farmers’ well-being, this paper empirically tests the impact of agricultural outsourcing on farmers’ well-being by using CRRS data in 2020 and employing the Ordered Probit model, extended regression model, and other econometric methods. The specific research conclusions include the following: (1) Agricultural outsourcing significantly reduces farmers’ well-being. The results of various forms of robustness tests such as substituting variables and the use of instrumental variable method to eliminate endogeneity still support this conclusion. (2) Agricultural outsourcing mainly damages the happiness of corn and rice growers, farmers in mountainous areas, and small-scale households. (3) The mechanism test results indicate that the principal–agent risk of agricultural outsourcing, the decrease in farmers’ occupational autonomy, and the split of families are all intermediate influence paths, and the income-increasing effect of agricultural outsourcing only partially alleviates farmers’ negative psychology.

According to the findings of this study, China can at least make improvements in the following aspects in the future:

Firstly, the government should actively build a local trading platform for agricultural outsourcing services, which can be used to publish and accurately match the supply and demand information of agricultural outsourcing in real time. It is important to establish an agricultural outsourcing database to trace the quality of outsourcing services. The formulation of relevant laws and regulations is also essential, and the behavior of agricultural outsourcing service providers should be regulated. Outsourcing service organizations need to strengthen the skill training of operators and strive to achieve the homogeneity of outsourcing services for the production of major grain crops. In addition, it would be excellent if the rural social security system could be improved, which can replace land as a guarantee for the survival of farmers [17].

Secondly, the development of agricultural outsourcing in non-plain areas should be paid enough attention. Research institutes and enterprises should actively develop and explore large-scale machinery and related production modes suitable for non-plain areas. Accordingly, it is necessary to continue to find ways to activate the land rental market and reduce the number of small-scale farms, including but not limited to supporting farmers with operating capabilities to grow into large-scale farmers [34].

Thirdly, local governments should vigorously develop non-agricultural industries within rural areas, and rely on the agricultural industry chain to cultivate a group of business entities engaged in agricultural product processing, e-commerce sales of agricultural products, rural tourism, etc., [67]. In addition to developing local industries, farmers’ relevant vocational skills training is particularly critical because we want farmers to work locally as much as possible. More importantly, there is evidence that rich family life can cope with the psychological distress of the employed compared with the self-employed. So enterprises need to strictly control the daily working time of the employed and try to provide them with decent living conditions [20].

Last but not least, we call on the international community to place mental health at the heart of sustainable development. Countries undergoing agricultural transformation need to invest more in rural communities, including but not limited to strengthening mental health education and opening clinics for mental disorders. All stakeholders need to work together for sustainable mental health. We call on rural communities around the world to build partnerships and engage farmers with mental disorders. If our recommendations can be fully implemented, this will contribute to the subjective well-being and sustainable development of the community.