Land Property Rights, Social Trust, and Non-Agricultural Employment: An Interactive Study of Formal and Informal Institutions in China

Abstract

Based on China’s structural transformation and the resulting rural social institutional changes, this paper clarifies how land certification affects non-agricultural employment, introducing trust culture as an informal institution and constructing a theoretical framework for their interactive effect on non-agricultural employment. Using data from the China Rural Household Panel Survey, the empirical research finds that land certification increases the likelihood of rural labor engaging in non-agricultural employment by reducing the risks of land loss, promoting land transfers, and facilitating credit financing. The interaction between land certification and social trust shows that increased social trust weakens the positive impact of land certification on non-agricultural employment.

1. Introduction

The transfer of rural labor from agriculture to non-agriculture is an essential path for developing countries to progress toward developed nations, and it is also an important subject in development economics. Few countries have a rural population as large as China’s (about half of the total population), even when the agricultural GDP has dropped to about 10%. This theoretical research examines the migration of rural labor starting from the models proposed by Lewis (1954), Ranis & Fei (1961), and Todaro (1969) [1,2,3]. However, economic theory views labor migration as a result of rational decisions made by individuals (or families) constrained by different institutional factors. In the case of China, traditional farmers are deeply rooted in agriculture, and land plays a significant role in their livelihoods, which creates a “sticky” relationship between rural labor and agricultural activities. Barriers to labor mobility due to land system constraints are a major cause of the misallocation of labor across sectors.

China’s land system comprises three key components: land property rights, land market, and land planning and management. Among these, the property rights system serves as the fundamental cornerstone, as a well-functioning land market mechanism can only operate effectively when property rights are clearly defined. Consequently, this study focuses on the impact of rural land property rights on non-agricultural employment—an issue that also has garnered extensive attention in the academic literature. Property rights determine the ownership of resources and the distribution of associated benefits, typically comprising a bundle of rights, including ownership, usage, and income rights. According to China’s Constitution and Land Administration Law, ownership of agricultural land is collectively held, while usage rights are allocated to rural households through contractual arrangements. The contracted management right is thus regarded as the usage right1 Land certification is a critical component of rural land property rights reform and a prerequisite for the marketization of land. At its core, land certification involves legally recognizing farmland rights through land surveys, registration, and issuing official documentation (such as land contract certificates) to property rights holders. This process clarifies land possession, usage, income distribution, and transferability rules, ultimately strengthening property rights [4].

As an institutional framework to regulate the recognition and use of farmland, land certification represents a significant achievement in China’s rural reform. It aligns with the global trend toward the marketization of rural property rights [5]. In fact, since the 1990s, the formalization of property rights has become a core reform strategy for promoting rural economic development in many developing countries [6]. For instance, Ethiopia launched a nationwide large-scale land certification reform in the late 1990s, issuing over 25 million rural land certificates by 2012 [7]. Similarly, Vietnam’s Land Law of 1993 granted farmers long-term, transferable land use rights, while Mexico’s 1992 land certification reform provided ejidos members with land ownership certificates, lifted restrictions on the transaction of communal land, and allowed land to be integrated into the market. These land certification reforms offer valuable empirical cases for examining the economic and social effects of property rights institutions, encouraging scholars to conduct empirical analyses on the specific mechanisms and policy implications of land certification under diverse socio-economic conditions.

Regarding how land certification affects non-agricultural employment, existing literature mainly focuses on land rights expectations, agricultural investment, and land endowment [8,9], suggesting that stable land rights can release surplus labor from the agricultural sector and promote rural labor’s transfer to non-agriculture, for example, De Janvry et al. (2015) found that the introduction of Mexico’s land certification program from 1993 to 2006 led to an increase in household out-migration [10]. However, it can also discourage labor migration by reducing the risk of land loss and encouraging long-term investment in agriculture [11]. While existing research has produced substantial findings, it may have overlooked the interaction between formal land property rights and informal rural institutions, which complicates understanding how and when these institutions affect rural labor’s socioeconomic behavior.

Institutions determine behavior and, in turn, economic performance, which has been widely recognized in academic circles. However, the relationship between formal and informal institutions remains contentious [12,13,14]. Broadly, two representative perspectives exist. The first perspective posits that formal institutions are embedded within the cultural foundations of informal institutions, making them complementary and mutually reinforcing [14,15,16]. When informal institutions align with formal institutions, they can reduce transaction costs and enhance institutional efficiency. For instance, Lambert-Mogiliansky et al. (2007) found that the impact of Russia’s bankruptcy law varied across regions, depending on the popularity of regional governors and the quality of their relationships with the federal government [17]. The second perspective suggests that formal and informal institutions may have a competitive relationship and can serve as substitutes [18,19]. For example, Méon & Sekkat (2015), using data from 46 countries between 1984 and 2009, examined the effects of social trust and the formal legal system on capital accumulation. Their findings confirmed that both social trust and the quality of formal legal institutions positively contribute to capital accumulation and indicated that these factors function as substitutes [20]. Despite these insights, research on the interaction between formal and informal institutions remains insufficient, warranting further exploration.

As traditional agrarian societies transition into modern ones, industrialization, marketization, and the resulting transformations in economic structures and social institutions have reshaped urban landscapes and, more fundamentally, profoundly impacted rural areas [21]. For a long time, the traditional agricultural civilization in China has led to a norm of immobility among rural laborers; the bond between people and land is robust, with a particularized trust2 mechanism based on kinship and locality [23]. With rural land reforms, land rights have undergone a historical transformation, and rural labor began to flow outwards and accumulate material and social capital. The once predominantly agrarian village economies have evolved into diversified rural economies integrating agriculture, commerce, and services. This shift has led to the differentiation of the rural population, with the ‘’second-generation rural migrants’’ (born in the 1980s and beyond) driving an intergenerational revolution, gradually loosening the traditional attachment between farmers and land. Rural land ownership’s internal differentiation and institutional transformation have progressively dismantled rigidly structured rural social relations and interpersonal interaction patterns. Social trust has become a crucial institutional force as migration alters rural social networks.

Thus, with the deepening of rural formal institutional reforms and the transformation of informal institutions, “Rural China, once profoundly rooted in agrarian traditions, has now transitioned into an integrated urban-rural China”. The increasingly fragmented smallholder economy, the declining attachment between farmers and land, and the rapid restructuring of rural social relations have significantly altered the institutional environment for rural non-agricultural employment [24,25].

Against this backdrop, this paper focuses on two key questions based on the historical institutional transformation from rural to urban China: (1) How does land certification affect non-agricultural employment? (2) How does land certification, as a formal institution, interact with social trust, an informal institution, to influence non-agricultural employment? Therefore, this paper constructs a theoretical analytical framework based on the theory of land property rights and land function theory and further conducts empirical research. It argues that the strength of land property rights is a crucial factor influencing rural labor’s non-agricultural employment while emphasizing the interactive role of social trust in this process.

2. Theoretical Analysis and Research Hypotheses

2.1. Land Certification and Non-Agricultural Employment

2.1.1. The Land Certification Policy

Land is fundamental to rural livelihoods. Land property rights determine the distribution of resources and benefits in rural areas. The instability of land rights has significantly impacted rural labor’s employment choices. Given China’s vast territory and complex land property systems, rural land institutions have undergone a series of reforms. “The farmer shall have his land” was the core principle of the early land system, followed by a stage of collective land certification during the cooperative and commune periods. In 1978, China introduced a collective land system, allocating land to households based on population. Land ownership and use rights would be separated, with ownership belonging to the rural collective and the farmers owning the use rights. However, its obvious disadvantage is that when there are changes in the rural population, there is controversy over whether the land should be adjusted. The central government continually issued policies to extend land contract terms and limit land adjustments to stabilize the relationship between people and land and promote farmers’ agricultural investment. For example, in 1984, land contracts were extended to 15 years, and in 1993, the second round of contracts was extended to 30 years, with a nationwide policy of “no increase of land with population increase, no decrease of land with population decrease”. However, due to the lack of specific property rights system details at this stage, coupled with factors such as inadequate certification, unclear land boundaries, and land expropriation, the actual property rights of the land are not clear.

In 2013, China launched a nationwide land rights certification program, marking the most profound institutional reform in rural land tenure since the decentralization of land rights. Land certification is the right to confirm and register farmers’ contracted land management rights. The government utilized modern technologies like GPS to measure and define land boundaries, create a registration system, and issue land certificates. This reform aimed to stabilize farmers’ land use rights and protect their legal land claims. By the end of 2018, over 200 million households had received land contract certificates, covering 1.48 billion acres of farmland. The land rights certification process varied in its progress across regions, according to CRHPS data in 2017, with some areas like Gansu, Qinghai, Sichuan, Yunnan, and Chongqing reaching high certification rates of over 80%. Other areas, such as Hebei and Guangdong, had lower rates under 45%.

In addition, it is worth noting that in implementing the land certification policy in some developing countries, households must actively apply for and pay for land certification. Then, the government gradually invests more funds to promote the policy, such as in Albania and Nicaragua. Empirical research related to these programs makes it difficult to avoid the endogenous problem of why some households receive certificates and others do not. In contrast, under China’s institutional arrangements, local governments proactively issue certificates to all farmers free of charge, thereby alleviating the above concerns.

2.1.2. The Relationship Between Land Certification and Non-Agricultural Employment

As a formal expression of property rights formed during the process of clarifying land rights information and certifying land certification [4], land certification primarily influences rural labor’s non-agricultural employment through channels such as reducing protection costs, incentivizing investment, facilitating market transactions, and promoting credit financing [8]. The mechanism of this influence can be summarized in three aspects.

First, the effect of reduced land loss risk. Improving land rights stability can release excessive agricultural labor caused by the risk of land loss, enabling it to be allocated to non-agricultural activities [26]. When land rights are unstable, rural laborers have to spend time and resources to defend their land rights, which inhibits their migration to non-agricultural industries. If land rights are stable, contrarily, rural laborers do not need to maintain their land rights by “occupying and farming”, which is conducive to making migration decisions based on labor market signals.

Second, the agricultural investment effect. While the improvement in land rights stability may directly incentivize rural labor to invest long-term in agrarian activities by enhancing agricultural investment returns [27], the mechanization and scale development brought about by land certification may have a crowding-out effect on agricultural employment. For instance, it increases the likelihood of surplus family labor being shifted to non-agricultural sectors [28], and rising agricultural input prices have reduced the profits of small-scale farmers. Individuals with lower agrarian productivity will likely migrate for work as some farmers shift to large-scale production. In summary, we argue that land certification has a limited impact on suppressing rural labor migration to non-agricultural sectors by promoting agricultural investment.

Third, the land financing effect. Land certification clarifies farmland’s physical boundaries and property rights, reducing uncertainty in transactions and helping promote land leasing by rural labor [29]. At the same time, stable land rights enhance the collateral and mortgage value of land, improving rural households’ access to credit. The transfer and mortgage of land can directly promote rural non-agricultural employment, and rental income and financing returns can supplement the living expenses of non-agricultural employment, reducing costs such as travel expenses for farming during busy agricultural periods and labor costs, thus lowering the transition cost to non-agricultural employment and continuously advancing the transition to non-agriculture.

In summary, this paper proposes the following hypothesis:

Hypothesis 1.

Land certification promotes non-agricultural employment among rural laborers.

2.2. The Interactive Effect of Land Certification and Social Trust on Non-Agricultural Employment

Trust, a crucial component of social capital, is the foundation for the effectiveness of formal institutions and government policies [30]. Trust carries distinct meanings across disciplines such as sociology, psychology, and economics. Sociologists focus on how social networks generate or inhibit trust relationships, viewing trust as a mechanism to reduce the complexity of social interactions. Psychologists examine trust from the perspectives of emotions, attitudes, and beliefs, defining it as an expectation of others’ behavior based on social interactions and personal experiences, influencing individual decision-making [31]. Economists, on the other hand, emphasize the role of trust in lowering transaction costs.

Theoretically, social trust and land certification interact when influencing non-agricultural employment [32]. First, trust is deeply embedded in social structures and cultural contexts. With the profound transformation of rural social structures worldwide, increasing social mobility, and the widespread adoption of the internet, farmers’ awareness of the rule of law and market mechanisms has gradually strengthened. Consequently, their choices regarding social interactions and land transactions have become more rational and diversified. As a result, traditional particularized trust is slowly evolving into generalized social trust. However, due to variations in natural environments, cultural customs, and patterns of population mobility, different social structures and trust models have emerged within rural China, which can be broadly categorized into two types: (1) kinship-based familiar societies, where clan-based social relationships dominate and (2) anonymous societies, where social trust prevails and interactions are increasingly market-oriented and institutionalized. According to the “Groupthink Theory”, families’ reactions and behaviors vary across social structures [33]. For instance, in societies with particularized trust, villages with strong clan solidarity have high collective cohesion, where family behavior tends to follow collective will, and attitudes toward land transactions and non-agricultural employment are significantly influenced by neighbors, respected individuals, or clan leaders. In contrast, in more dispersed village structures, individual or family decisions are less influenced by external environmental factors [34].

Second, from the perspective of transaction costs, the high transaction costs in land markets arise from insecure land property rights and low levels of trust. Well-defined land property rights are conducive to reducing transaction costs and making rural labor feel more secure when conducting land transactions. Trust can enhance formal institutions’ effectiveness and supplement property rights’ stability in a well-established land property rights system. Conversely, in unstable land property rights systems, high trust can reduce the costs of information acquisition, negotiation, and contract enforcement, partially replacing weak land property rights systems. Based on the above analysis, this paper proposes two competing hypotheses on the interactive relationship between land certification and social trust in influencing non-agricultural employment: one where they substitute for each other, and the other where they complement each other.

2.2.1. Complementary Relationship

Land certification provides a legal safety net, enhancing rural labor’s sense of security in non-agricultural employment, while social trust further reduces their perceived risks through informal risk-sharing mechanisms, jointly promoting the transfer of rural labor to non-agricultural sectors. For example, in communities with high trust, farmers are more likely to utilize social networks to share the risks of external employment. Additionally, individuals in high-trust communities tend to hold more positive attitudes toward the effectiveness of land certification, believing that land certification policies reduce land loss risks. In contrast, due to higher subjective perceptions of land loss risk, individuals with low social trust tend to hold pessimistic views on the stability of land property rights, making them less willing to engage in non-agricultural employment. Therefore, social trust can enhance the effectiveness of land certification in non-agricultural jobs.

Land certification makes land tradable and financeable, while social trust expands financing options through informal financial networks. As an essential form of social capital, social trust establishes the basic moral norms in market economies [35]. Higher social trust not only makes people more willing to engage in transactions and reduces transaction costs but also broadens rural labor’s social networks, increasing access to market information and resources. Thus, rural labor with higher social trust is more likely to participate in land transfer and financial markets, improving financing efficiency and employment options. The joint effect of land certification and social trust drives rural labor’s transition to non-agricultural sectors.

In general, regarding risk-sharing, land certification provides legal protection, while social trust further disperses risk through informal mutual support networks, boosting rural labor’s confidence in non-agricultural employment. Regarding financing, land certification makes formal finance more feasible, while social trust compensates for the limitations of formal finance through informal financing channels, further supporting farmers’ transition to non-agricultural employment.

2.2.2. Substitutive Relationship

However, the interaction between land certification and social trust may also have adverse effects, with formal and informal institutions potentially competing for influence, creating a substitution relationship.

Land certification provides stable expectations for agricultural investment, which helps rural labor make long-term investments. However, the investment incentives brought by land certification may compete with the collective investment models facilitated by social trust. In agricultural investment, the individual property rights protection provided by land certification may weaken the necessity of collective action and social trust. In communities with high social trust, rural labor is more inclined to invest collectively through cooperative or communal forms. However, after land certification, as land certification grants individuals more decision-making autonomy and profit distribution rights, individual farmers may prefer to invest independently. In this case, land certification may reduce the demand for collective investment based on social trust, partially substituting the role of social trust in promoting non-agricultural employment.

At the same time, higher social trust reflects farmers’ subjective perception that their land use rights are protected by traditional morality, which in turn indicates a higher recognition of the legitimacy and desirability of land property rights. Individuals with higher social trust tend to have higher recognition of the legitimacy and desirability of their land property rights [36]. Implementing land certification may further enhance rural labor’s confidence and recognition of land, making them more inclined to invest in agriculture. Furthermore, when they possess land property rights and trust their community, they may receive more social support, such as agricultural technology and financial assistance, thus increasing agricultural production efficiency and investment returns and reducing the demand for non-agricultural employment.

Based on the above analysis, we propose the following hypothesis:

Hypothesis 2.

The level of social trust influences the effectiveness of land certification. Its combined effect depends on the relative strength of social trust in influencing perceptions of land loss risk, agricultural investment, and land financing.

Figure 1 presents the theoretical framework illustrating Hypothesis 1 and Hypothesis 2.

3. Materials and Methods

3.1. Data Source

The data used in this paper comes from the 2017 sample of the China Rural Household Panel Survey (CRHPS), which is synthesized from Zhejiang University’s Chinese Family Database (CFD) and Southwestern University of Finance and Economics’ China Household Finance Survey (CHFS). These data combine stratified, three-stage, and proportional sampling methods. It covers 29 provinces in China, excluding Hong Kong, Macau, Taiwan, Xinjiang, and Tibet, providing comprehensive data representation at multiple levels, including rural, urban, provincial, and national levels. The comprehensive dataset includes rural households’ basic demographic characteristics, household information, land system arrangements, and village land utilization, making it ideal for this study.

To capture the effects of land certification on non-agricultural employment, the study restricts the sample to individuals aged 15–64 according to the international labor age standard. This paper examines rural labor through three criteria: household registration status, rural residency, and possession of land contractual management rights. Rural laborers often migrate to cities for non-agricultural employment, leading existing literature to use rural household registration as a defining factor. Given that the land rights confirmation system is fundamental to rural management, the study focuses on rural areas. Finally, only individuals or families with land contractual management rights are influenced by land system changes; thus, the sample is restricted to rural registered laborers possessing these rights. After data cleaning, the study includes 10,703 rural labor samples from 4946 households in 359 villages.

3.2. Empirical Models

3.2.1. Baseline Model

To test the theoretical hypotheses proposed earlier, and considering the discrete nature of the non-agricultural employment variable, this paper uses the following Probit baseline model to examine the impact of land certification on rural labor’s non-agricultural employment3:

$${Employment}_i={\alpha }_0+{\alpha }_1{Land}_i+{\alpha }_2{Controls}_i+{\epsilon }_i$$

(1)

Among them, the subscript $i$ represents the $i$th individual and ${Employment}_i$ is the non-agricultural employment behavior of the$i$th individual. ${Land}_i$ is the key explanatory variable that stands for land certification, including whether the household has received the land contract management certificate. ${Controls}_i$ is a set of control variables, including the characteristics of individual, household, village, and province-fixed effect. ${\epsilon }_i$ is a random error term.

3.2.2. Interaction Relationship Model

To further analyze the interaction between land certification and social trust, we conducted an empirical analysis by constructing an interaction effect regression model

$${Employment}_i={\alpha }_0+{\alpha }_1{Land}_i+{\alpha }_2{Trust}_i+{\alpha }_3{Land}_i\times {Trust}_i+{\alpha }_4{Controls}_i+{\epsilon }_i$$

(2)

In Equation (2), ${Trust}_i$ represents the degree of trust that rural laborers have in strangers. The other variables are the same as in Equation (1). The coefficient ${\alpha }_3$ captures the interaction effect between land certification and social trust on employment. Its sign determines how trust influences the impact of land certification, which means the total effect of land certification on non-agricultural employment is given by ${\alpha }_1$ + ${\alpha }_3\times {Trust}_i$.

If ${\alpha }_3$ > 0, higher trust strengthens the positive impact of land certification on employment, and social trust and land tenure security will be complementary.

If ${\alpha }_3$ < 0, this implies that higher trust weakens the effect of land certification on employment, and in extreme cases, the effect could even become harmful. Thus, the sign and magnitude of ${\alpha }_3$ are crucial in determining the moderating role of social trust in this relationship.

3.3. Variable Definitions and Descriptive Statistics

3.3.1. Non-Agricultural Employment (N-AE4)

Based on the CHFS database, if a rural laborer’s employer is “engaged in land contracting and management”, the “non-agricultural employment” variable is assigned a value of 0, and a value of 1 is assigned otherwise. Additionally, because some scholars use rural households as the unit of analysis when exploring rural institutional reforms and employment issues, this paper further constructs a household non-agricultural employment variable, defined by two aspects: “whether anyone in the household is engaged in non-agricultural employment” (where “yes” is assigned a value of 1 and “no” a value of 0) and “the proportion of non-agricultural employment among household members of working age”, with values ranging from 0 to 1, where higher values indicate a higher proportion of labor engaged in non-agricultural employment.

3.3.2. Land Certification (LC)

As an essential means of national empowerment, land legal documents play a role in legally protecting farmers’ land property rights and constraining public governance. The land certification variable in this study is measured by land certification registration and certification, primarily based on the CFD questionnaire’s question “Does your family have a rural land contract certificate?” Specifically, if a rural household answers affirmatively, it is assigned a value of 1, otherwise, it is assigned a value of 0.

3.3.3. Social Trust (ST)

This paper uses the CHFS questionnaire’s question “How much do you trust strangers?” to measure the level of social trust among rural laborers, with responses ranging from 1 (“very distrustful”) to 5 (“very trustworthy”). Considering the externality of social trust [37], social trust in the village or county also influences individuals’ trust perceptions. The environment of trust reflects the informal institutional characteristics of social trust. Thus, we calculate two types of social trust environment variables—village social trust and county social trust—together depicting the level of social trust. Later, an interaction term of “social trust × land certification” is constructed to test these two variables’ complementary or substitutive effect on non-agricultural employment.

3.3.4. Control Variables

The control variables in this paper are categorized into individual, household, and village characteristics. Human capital in education and health plays a vital role in rural labor’s non-agricultural employment. Generally, rural labor with higher human capital investment and better health tends to have higher marginal returns from non-agricultural work, with a stronger ability to transition and better quality of transition. Household resource endowment and caregiving responsibilities also affect non-agricultural employment. Economic capital provides the material foundation for non-agricultural employment, while the proportion of elderly and young family members influences the push for local employment due to caregiving responsibilities. Additionally, village land endowment and economic development also affect non-agricultural employment. Per capita, cultivated land determines the tension between land and population in the village, while per capita income and economic crop cultivation area reflect the opportunity cost of rural labor migrating to cities. The proportion of out-migrating labor influences non-agricultural employment by affecting the village’s exposure to external market information. Moreover, regions with higher non-agricultural employment tend to have a stronger demand for land certification and are more likely to ensure the smooth implementation of land certification. Thus, controlling for this variable helps mitigate reverse causality issues in the land certification–employment relationship. Finally, this paper includes province-fixed effects as control variables to eliminate regional effects on the estimation results.

The specific variables include the following:

Individual characteristics: Gender (male = 1, female = 0); age and age squared; education (assigned values from 1 to 9 based on educational attainment from no schooling to PhD); marital status (married = 1, otherwise = 0); health (rated from 1 for “very poor” to 5 for “very good”). Household characteristics: Family size; the proportion of elderly (age 75+); the proportion of children (age 6 and below); family total assets (top and bottom 1% trimmed and natural logarithm taken). Village characteristics: Per capita income (natural logarithm of per capita disposable income); per capita cultivated land area (measured as cultivated land area per household); the proportion of economic crops; external labor force proportion (measured as the number of out-migrating workers divided by village population); the number of roads to the county center; the number of land requisition occurrences since 2000.

3.4. Data Description

Based on the previous data processing and modeling requirements, this study selects valid data from 10,703 rural laborers in 4946 rural households. Of these, 3439 households completed land certification in 2017, with a land certification rate of approximately 70%.

Table 1. Descriptive statistics of major variables.

Variable Type	Variable Name	Non-Agricultural Employment (N = 3875)		Agricultural Employment (N = 4803)
		Mean	S.D.	Mean	S.D.
Key independent variable	Land certification	0.6973	0.4595	0.6921	0.4617
Individual characteristics	Gender	0.6302	0.4828	0.4928	0.5000
	Age	37.2150	11.8999	50.5249	9.8936
	Age squared	1526.5250	941.6271	2650.6260	931.4842
	Marital status	0.7143	0.4518	0.9271	0.2600
	Education	3.3502	1.4063	2.3881	0.8990
	Health	3.7316	0.9380	3.1924	1.0320
Household characteristics	Family size	4.6919	1.6956	4.0729	1.7972
	Proportion of elderly (75+)	0.0316	0.0855	0.0337	0.0943
	Proportion of children (6−)	0.0592	0.1020	0.0506	0.0996
	Family total assets	12.3359	1.3973	11.8103	1.3837
Village characteristics	Per capita cultivated land area	1.6833	2.6537	2.2926	3.2423
	Per capita income	8.8227	0.7121	8.6700	0.7620
	Proportion of economic crops	34.3415	32.6719	35.7551	32.7432
	External labor force proportion	0.2194	0.1599	0.2117	0.1517
	Roads to the county center	1.8010	0.8696	1.6750	0.8262
	Number of land expropriation times	0.1272	0.4312	0.1094	0.4362

shows descriptive statistics of major variables for non-agricultural employment and agricultural employment laborers.

4. Results

4.1. Basic Model Regression

Table 2. Baseline estimation results.

	(1) N-AE	(2) N-AE	(3) N-AE	(4) N-AE	(5) Household N-AE	(6) Household N-AE Ratio
Land certification	0.0914 ** (0.0391)	0.0985 ** (0.0451)	0.0898 * (0.0483)	0.0227 * (0.0120)	0.1468 *** (0.0486)	0.0187 * (0.0114)
Gender		0.3480 *** (0.0320)	0.3386 *** (0.0349)	0.0971 *** (0.0084)
Age		−0.0387 *** (0.0135)	−0.0360 ** (0.0139)	−0.0147 *** (0.0034)
Age squared		−0.0002 (0.0002)	−0.0002 (0.0002)	−0.0000 (0.0000)
Marital status		−0.2075 *** (0.0661)	−0.1202 (0.0766)	−0.0353 ** (0.0173)
Education		0.2160 *** (0.0206)	0.2112 *** (0.0230)	0.0495 *** (0.0046)
Health		0.0816 *** (0.0186)	0.0825 *** (0.0206)	−0.0229 *** (0.0050)
Family size		0.0350 ** (0.0152)	−0.0019 (0.0158)	0.0114 *** (0.0040)	0.3282 *** (0.0186)	0.0373 *** (0.0032)
Proportion of elderly (75+)		−0.4681 ** (0.2049)	−0.5092 ** (0.2244)	−0.1504 *** (0.0564)	−1.0520 *** (0.2036)	−0.0916 * (0.0520)
Proportion of children (6−)		−0.7533 *** (0.2233)	−0.6336 *** (0.2371)	−0.1910 *** (0.0588)	−1.4866 *** (0.2611)	−0.2376 *** (0.0454)
Family total assets		0.0703 *** (0.0166)	0.0953 *** (0.0178)	0.0183 *** (0.0043)	0.1403 *** (0.0169)	0.0362 *** (0.0039)
Per capita cultivated land area		−0.0253 (0.0173)	−0.0229 (0.0171)	−0.0069 * (0.0043)	−0.0218 (0.0164)	−0.0058 * (0.0031)
Per capita income		0.1252 *** (0.0473)	0.1444 ** (0.0564)	0.0363 *** (0.0117)	0.1208 *** (0.0415)	0.0217 ** (0.0103)
Proportion of economic crops		−0.0032 *** (0.0010)	−0.0030 *** (0.0010)	−0.0009 *** (0.0003)	−0.0024 ** (0.0009)	−0.0007 *** (0.0002)
External labor force proportion		0.3797 * (0.2077)	0.3051 (0.2164)	0.0981 * (0.0542)	0.3021 (0.1919)	0.0341 (0.0480)
Roads to the county center		0.0519 (0.0377)	0.0740 * (0.0401)	0.0112 (0.0100)	0.0175 (0.0361)	0.0132 (0.0087)
Number of land expropriation times		0.0712 (0.0473)	0.0625 (0.0514)	0.0187 (0.0136)	−0.0660 (0.0545)	−0.0002 (0.0113)
Constant	0.7642 *** (0.2063)	0.1658 (0.5574)	−0.5168 (0.6010)	0.6638 *** (0.1412)	−2.7019 *** (0.4956)	−0.1537 (0.1103)
Fixed effect	Y	Y	Y	Y	Y	Y
Observations	8678	8588	7204	8588	4319	4319
R2	0.0421	0.3317	0.2954	0.3800	0.1594	0.1406

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively; Column (2) is the benchmark regression model; In the table, ‘‘Y’’ denotes “YES”; Regarding the explanatory power of R² in the regression model, we have provided a detailed discussion in Appendix A.2.

presents the regression results with clustered standard errors at the village level, examining the effect of land certification on rural labor’s non-agricultural employment. Column (1) controls for regional effects and shows a simple regression. The estimated coefficient for land certification is approximately 0.09 and is statistically significant at the 5% level. Column (2) adds individual-, family-, and village-level characteristic variables, and the results support the hypothesis that land certification increases the likelihood of rural labor transitioning from agriculture to non-agriculture.

The coefficients of other control variables show that male and unmarried rural laborers, who are younger, healthier, and more educated, are more likely to engage in non-agricultural employment. Larger family size, lower caregiving burden, and richer family resources are associated with a higher probability of non-agricultural employment. A larger proportion of economic crop areas relative to arable land, less urban attraction, and tighter land-population relations are linked with higher agricultural employment in the village, although these effects are weak. Higher proportions of out-migrant labor in the village enhance peer effects, making it more likely for other rural laborers to migrate to urban areas for work.

Although land contract rights signify that the individual is accepted as part of the community, considering some rural labor migration behaviors, this paper further restricts the sample to individuals with rural household registration, residing in rural areas, having land contracting rights, and being local residents. The estimated results in column (3) confirm the robustness of the baseline conclusion. Column (4) switches the baseline model to OLS regression to mitigate potential estimation biases from model selection and confirms the robust relationship between land certification and rural labor’s non-agricultural employment. Columns (5) and (6) focus on rural households, measuring key dependent variables such as “whether the household engages in non-agricultural employment” and “the proportion of non-agricultural employment within the household”, respectively. The results show that land certification significantly increases the proportion of rural households’ labor force engaged in non-agricultural employment.

4.2. Robustness Test

To eliminate potential omitted variable bias, the study adds variables related to social security participation and considers additional factors like cultures and macroeconomic and policy factors that might affect non-agricultural employment. It also tests the causal relationship between land certification and non-agricultural employment by using instrumental variables.

4.2.1. Excluding Other Influences

(1) Agricultural “Pull” from Social Pension Insurance. Farmers were granted land use rights after implementing the Household Responsibility System, making land a unique substitute for rural social pension security. As the Chinese government gradually provided public goods for rural society, rural residents enjoyed equal access to social security rights. This social security system changes their psychological expectations regarding retirement options by providing stable pension income to rural residents. It not only reduces their reliance on land as a pension security tool (forming a “substitution effect”) but also affects their land transfer and agricultural investment behaviors [38]. In other words, when individuals participate in social pension insurance, it may weaken or even nullify the effect of land certification on rural labor’s non-agricultural employment.

This paper first adds individual social pension insurance participation as a control variable to account for this. Based on the survey question “Which of the following social pension insurances are you currently enrolled in?”, samples that answered “None” are assigned a value of 0, while those participating in any insurance type are assigned a value of 1. For family-level discussion, the paper uses “the proportion of family members participating in social pension insurance” to measure this variable. Additionally, an interaction term between social pension insurance participation and land certification is constructed to examine the substitution relationship between rural formal pension security and land-based pension security. It is expected that participation in social pension insurance will weaken the pension security function of land, thereby reducing the positive effect of land certification on family non-agricultural employment.

Table A1. Robustness test results, excluding the influence of other factors.

	N-AE		Household N-AE	Household N-AE Ratio
	(1)	(2)	(3)	(4)
Land certification	0.1004 ** (0.0457)	0.1646 ** (0.0745)	0.2612 *** (0.0756)	0.0347 ** (0.0171)
Social pension insurance participation	−0.1135 ** (0.0443)	−0.0496 (0.0768)	−0.3694 * (0.2230)	−0.0665 (0.0530)
Land certification × Social pension insurance participation		−0.0947 (0.0852)	−0.4702 * (0.2410)	−0.0660 (0.0589)
Control variables	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y
Observations	8371	8371	4319	4319
R2	0.3293	0.3295	0.1644	0.1439

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES”.

in Appendix A reports the impact of land certification on rural labor’s non-agricultural employment under different social pension insurance participation conditions. The results show that when rural labor has access to social pension insurance, the probability of non-agricultural employment decreases. In other words, rural laborers who do not participate in social pension insurance are more motivated to transition from agriculture to non-agriculture, seeking higher monetary income and saving for future pensions. Furthermore, the estimated coefficient of the “Land certification × Social pension insurance participation” variable is negative, and it is statistically significant when the dependent variable is “whether the household engages in non-agricultural employment”. This suggests that when the proportion of family members participating in social pension insurance increases, the substitution effect between social pension insurance and land-based pension security becomes more apparent.

(2) Considering Regional Disparities. Given China’s significant regional economic disparities, for example, land certification implementation speed and non-agricultural employment opportunities’ availability vary across regions, we conduct subgroup regressions based on economic development levels, classified by eastern, central, and western areas and per capita GDP. The results in Appendix A

Table A2. Robustness test results, considering regional development differences.

	Regional Classification			Economic Development (Per Capita GDP)
	(1) Eastern	(2) Central	(3) Western	(4) High	(5) Medium	(6) Low
Land certification	0.0668 (0.0937)	0.0565 (0.0642)	0.2394 ** (0.0935)	−0.0182 (0.0821)	−0.0992 (0.0811)	0.2954 *** (0.0680)
Control variables	Y	Y	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y	Y	Y
Observations	2275	3466	2630	2704	2273	3394
R2	0.3619	0.3609	0.2718	0.3676	0.3658	0.2914

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; ** and *** indicate significance at the 5% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES”.

confirm the robustness of our baseline findings and reveal the heterogeneous effects of institutional reforms. The employment-promoting impact of land certification is most pronounced in less developed western regions with lower per capita GDP. This may be because lower agricultural marginal returns make non-agricultural employment more attractive, and the property income from land certification reduces migration costs. Additionally, surplus labor in these regions facilitates smoother transitions to non-agricultural sectors once institutional constraints are lifted. In contrast, saturated non-agricultural labor markets in most developed areas limit new employment opportunities for surplus labor released by land certification. Moreover, land appreciation often exceeds non-agricultural wages, incentivizing farmers to retain land rather than seeking off-farm employment. In some cases, this leads to “land-holding for speculation”, further weakening the role of land certification in employment restructuring.

4.2.2. Adding Possible Omitted Variables

(1) Cultural Factors. ① Confucian Culture. As an essential part of Chinese traditional culture, “Confucianism” may play a significant role in affecting the relationship between land certification and non-agricultural employment. On the one hand, Confucian culture’s “collectivist view” and the governance concept of “not worrying about scarcity, but worrying about inequality” may accelerate the implementation of land property rights systems, and it could lead local governments to adjust land distribution policies in response to population changes. On the other hand, Confucianism has religious characteristics, and religious beliefs tend to make people trust each other more and be more willing to cooperate [39]. Thus, an emphasis on Confucian culture is more likely to foster friendly, mutually beneficial relationships, which can benefit rural laborers by providing them access to non-agricultural employment information and resources.

Referring to Gu (2015) [40], this paper uses the “number of Confucian academies” (including Confucian official schools, academies, and temples) at the provincial level to measure the strength of Confucian cultural tradition in different regions. Data for this comes from historical texts such as the Daming Yitongzhi, Da Qing Yitongzhi, and local chronicles from the Ming and Qing dynasties. Theoretically, Confucianism emphasizes guiding and improving human behavior through ethical education, ultimately achieving “moral alignment with common customs”. The spread of Confucian thought through education became a consensus among the ruling class and intellectuals after the Song Dynasty, transcending ethnic and national boundaries. Therefore, using the number of Confucian academies recorded in historical texts to measure Confucian tradition can objectively assess the influence of Confucian culture.

② Clan Culture. The interaction between political governance and the power of informal social norms constitutes an important context for implementing land property rights in clustered villages. As an informal institution, clans in villages make traditional relational governance play a significant role in village governance, weakening the enforcement of formal systems, such as issuing land legal documents. Moreover, the clan networks, which are based on blood ties, and the patronage networks formed through marriage ties, while effectively enhancing the probability of rural labor engaging in non-agricultural employment, may also hinder the non-agricultural transition of rural labor by increasing the provision of rural public goods [41]. To exclude this possibility, we add clan culture as an omitted variable in

Table 3. Robustness test results, adding possible omitted variables.

	N-AE			Household N-AE	Household N-AE Ratio
	(1)	(2)	(3)	(4)	(5)
Land certification	0.0985 ** (0.0451)	0.0997 ** (0.0453)	0.1014 ** (0.0458)	0.1476 *** (0.0493)	0.0191 * (0.0114)
Confucian culture	0.2167 *** (0.0645)		0.2079 *** (0.0673)	0.1819 ** (0.0727)	0.0411 *** (0.0153)
Clan culture		0.0287 (0.0717)	0.0350 (0.0730)	0.1242 * (0.0643)	0.0040 (0.0168)
Social pension insurance participation			−0.1124 ** (0.0444)	−0.7110 *** (0.1464)	−0.1140 *** (0.0337)
Control variables	Y	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y	Y
Observations	8588	8567	8350	4309	4309
R2	0.3317	0.3317	0.3293	0.1646	0.1432

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES”.

measured by “whether the village has a dominant surname”. The dominant surname in a village usually indicates a relatively strong clan network. The higher the proportion of dominant surnames, the stronger the clan culture.

Table 3 sequentially controls for the above two omitted variables based on the baseline model. From the estimation coefficients in columns (1) and (2), the promoting effect of land certification on non-agricultural employment remains significant. Columns (3) to (5) simultaneously control for Confucian culture, clan strength, and social pension insurance participation, and perform regressions at the family level. The estimated results confirm the robustness of the baseline conclusion.

(2) Macroeconomic and Policy Factors. First, we measure macroeconomic factors using the marketization index and population mobility ratio. The marketization index serves as a key indicator of the economic institutional environment, directly influencing the efficiency of land factor mobility [42]. The population mobility ratio reflects the openness of the labor market, regions with high mobility may experience “migration saturation”, whereas in areas with restricted mobility, high migration costs make institutional constraints a primary barrier to labor mobility. In such cases, land certification plays a more significant role by reducing exit barriers. Controlling for population mobility helps isolate the effects of spontaneous migration from those induced by land certification.

Second, we use provincial urbanization rates and the proportion of non-agricultural registered residents as proxies for policy factors. Higher urbanization rates are typically associated with more significant infrastructure investments facilitating non-agricultural employment. The proportion of non-agricultural registered residents reflects household registration (hukou) reform. In regions with strict hukou controls (low non-agricultural hukou proportion) and underdeveloped social security systems, farmers rely more on land as a safety net, making the effects of land certification more pronounced. In contrast, in areas with greater hukou flexibility (high non-agricultural hukou proportion), the government may provide more employment assistance or public goods, and the impact of land certification on employment may weaken. Controlling these variables allows us to isolate the potential effects of urban expansion and household registration policy reforms, preventing the misattribution of labor mobility changes to land certification.

Based on this framework, we incorporate the interprovincial marketization index using data from the 2016 China Market Index Database. We measure population mobility as the ratio of in-migrants from other provinces to total provincial population, and define urbanization rate as the ratio of urban to total population. The proportion of non-agricultural hukou holders is a proxy for hukou policy reform5. The results in Appendix A

Table A3. Robustness test results, considering macroeconomic and policy factors.

	Non-Agricultural Employment
	(1)	(2)	(3)	(4)	(5)
Land certification	0.1014 ** (0.0458)	0.1014 ** (0.0458)	0.1014 ** (0.0458)	0.1014 ** (0.0458)	0.1014 ** (0.0458)
Social pension insurance participation	−0.1124 ** (0.0444)	−0.1124 ** (0.0444)	−0.1124 ** (0.0444)	−0.1124 ** (0.0444)	−0.1124 ** (0.0444)
Confucian culture	0.2079 *** (0.0673)	−0.0357 (0.0692)	0.0188 (0.0433)	−0.0055 (0.0212)	−0.0185 (0.0189)
Clan culture	0.0350 (0.0730)	0.0350 (0.0730)	0.0350 (0.0730)	0.0350 (0.0730)	0.0350 (0.0730)
Marketization index		0.3151 * (0.1659)		0.0956 (0.4938)
Population mobility ratio			6.5703 * (3.4597)		0.4176 (0.4630)
Urbanization rate				2.6591 (7.8545)
Proportion of non-Agricultural registered residents					−0.0421 (0.1244)
Control variables	Y	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y	Y
Observations	8350	8350	8350	8350	8350
R2	0.3293	0.3293	0.3293	0.3293	0.3293

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES”.

show that after controlling for these variables, the land certification coefficient remains significant, confirming that its effect is independent of regional marketization, migration trends, and urbanization levels, thereby reinforcing the robustness of our baseline findings.

4.2.3. IV-Probit Estimation Results

Do the regression results in Table 2 still suffer from sample selection bias or estimation bias due to omitted unobservable variables? For example, in promoting land certification, local governments often follow a “first-easy-then-difficult” and “first-local-then-wider” approach, which may result in non-random variation across regions in whether rural households can obtain land contracting certificates. Moreover, land certification is a selective action by local governments, and factors such as the policy enforcement capacity of grassroots organizations, land disputes, and land adjustments could affect the implementation of land certification and, at the same time, influence local rural labor’s non-agricultural employment.

To address the potential endogeneity issues discussed above, this paper introduces the “village land certification rate” as an instrumental variable, defined as the average land certification rate of other rural households in the same village (excluding the individual). This variable is used in the IV-probit estimation method to re-examine the impact of land certification on rural labor’s non-agricultural employment decisions. The rationale for selecting this instrumental variable is as follows: the faster land certification proceeds in a rural household’s village, the higher the likelihood that the household will receive its land certificate. Moreover, the village land certification rate measures the overall land property rights environment and does not directly affect the individual’s non-agricultural employment decision.

Table 4. Robustness test results, introducing instrumental variables.

	Benchmark Regression	Correlation Test		Instrumental Variable Estimation
	(1) N-AE	(2) N-AE	(3) Household N-AE Ratio	(4) N-AE	(5) Household N-AE Ratio
Land certification	0.0985 ** (0.0451)			0.2916 ** (0.1295)	0.0740 * (0.0393)
Village land certification rate		3.0561 *** (0.0866)	2.1818 *** (0.0986)
Control variables	Y	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y	Y
Observations	8588	8349	4304	8349	4304
R2	0.3317	0.2658	0.1973	——	0.1372

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES.”

, column (1) reports the baseline regression results at the individual level. Columns (2) and (3) show the correlation test results at the individual and family levels, respectively, indicating a significant positive association between the village land certification rate and the household’s land certification status. Column (4) presents the instrumental variable estimation results under the IV-probit model. The estimated coefficient for land certification is statistically significant at the 5% confidence level and is higher than in the baseline regression results. This suggests that endogeneity issues such as omitted variables, which are difficult to measure, may have led to an underestimation of the positive effect of land certification on rural labor’s non-agricultural employment. In column (5), when the dependent variable is changed to the proportion of non-agricultural employment within the household, the estimated results confirm the robustness of the baseline conclusion.

4.3. Mechanism Analysis

4.3.1. The Effect of Reduced Land Loss Risk

As a subjective perception, land loss risk depends on the extent to which farmers believe their land use rights may be unjustly or arbitrarily infringed upon. Considering this characteristic, this paper constructs a variable for perceived land loss risk based on the survey question: “How much benefit do you think land certification will bring farmers?” Responses such as “land rights are clearer, serving as the basis for land and agricultural subsidies, as a basis for land requisition compensation, and as a basis for protecting land rights in land disputes” are assigned a value of 1, indicating that rural labor can directly experience the substantial benefits of land certification without worrying about future expropriation. Other responses are assigned a value of 0, indicating that rural labor has not yet recognized the role of land certification in protecting land rights.

Table 5. Mechanism analysis results.

A	(1a) Perception of Risk of Losing Land	(2a) Land Transfer-out	(3a) The Time for Land Transfer-out	(4a) Land Financing
Land certification	0.3074 *** (0.0630)	−0.0303 (0.0685)	0.3026 * (0.1662)	0.2684 * (0.1527)
Control variables	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y
Observations	7889	8349	1191	6059
R2	0.0531	0.0816	0.1609	0.2129
B	(1b) Short-Term Agricultural Investment	(2b) Land Transfer-in	(3b) The Time for Land Transfer-in	(4b) Agricultural Machinery Value
Land certification	−0.0532 (0.0558)	−0.1617 *** (0.0562)	0.0761 (0.1491)	0.2440 (0.1851)
Control variables	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y
Observations	6440	8350	1470	7598
R2	0.3771	0.0676	0.1207	0.2077

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; * and *** indicate significance at the 10% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES.”

(1a) shows the regression results, indicating that land certification directly reduces rural labor’s perception of land loss risk, promoting non-agricultural employment.

4.3.2. Agricultural Investment Effect

To investigate whether the stability of property rights hinders non-agricultural employment by encouraging rural labor to invest in agriculture, this paper constructs relevant variables from both short-term and long-term perspectives. Specifically, the variable for short-term agricultural investment is “last year’s seed and fertilizer prices” (logged). For long-term investment, the paper uses the variables of “land transfer-in6” and “value of agricultural machinery” (logged) to measure it. Table 5 (1b) to (4b) presents the estimation results. The results show that land certification significantly suppresses land transfer behaviors. The estimated coefficients for agricultural input and self-purchased machinery value are statistically insignificant, indicating that rural labor does not consider long-term agricultural engagement in their planning. Land certification does not hinder non-agricultural employment through the promotion of agricultural investment.

4.3.3. Land Financing Effect

Land certification promotes the monetization of land by enhancing its value, thus improving its property value. This paper first measures land rental behavior and the duration of land transfer to capture land leasing situations. Considering that directly measuring “whether the household has borrowed” or “the number of household loans” may not adequately explain whether the loan is the result of land certification and may lead to bias, this paper uses “whether the household applied for a loan from the bank using land operation rights last year” to measure the mortgage value of farmland. Responses of “yes” are assigned a value of 1, while “no” is 0. Table 5 (2a) to (4a) reports the regression results, which show that land certification extends the duration of land transfer and promotes rural labor’s land-based financing behavior. This behavior is beneficial for human capital accumulation and entrepreneurial activities, and in turn, facilitates the non-agricultural migration of rural labor.

4.4. Interaction Between Formal and Informal Institutions: The Interaction Effect of Land Certification and Social Trust

Formal and informal institutions do not operate independently when shaping rural labor behavior; they may have “conflict” or “integration” effects. This section further examines the potential substitutive or complementary effect between land certification and social trust by constructing an interaction term between social trust and land certification to assess their combined impact on rural labor’s non-agricultural employment.

The estimated results in column (1) of

Table 6. Interaction effect results.

	(1) Individual Social Trust	(2) Village-Level Social Trust	(3) County-Level Social Trust
Land certification	0.2685 *** (0.0911)	0.5386 ** (0.2409)	0.7774 *** (0.2633)
Social trust	0.0781 ** (0.0361)	0.2427 ** (0.1213)	0.4436 *** (0.1374)
Land certification × Social trust	−0.0864 ** (0.0426)	−0.2282 * (0.1270)	−0.3531 ** (0.1369)
Control variables	Y	Y	Y
Fixed effect	Y	Y	Y
Observations	8308	8350	8350
R2	0.3294	0.3300	0.3310

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; *, ** and *** indicate significance at the 10%, 5% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES”.

show that land certification and social trust promote non-agricultural employment among rural laborers, utilizing the interaction relationship model in Section 3.2.2. However, there is a negative interaction effect between the two, indicating that strengthening social trust weakens the positive impact of land certification on non-agricultural employment, suggesting a substitutive relationship between them. Columns (2) and (3) in Table 6 further construct rural trust environment variables at the village and county levels. The estimation results confirm the robustness of the previous conclusion.

This finding aligns with the conclusions of studies such as Cui (2017) [43], which argue that social trust and formal institutions may have a substitutive relationship. The demand for formal institutions is lower in societies with higher social trust. For example, social trust can reduce the non-agricultural migration cost for rural labor by shifting their employment perceptions and transmitting labor market information, thus promoting non-agricultural employment. In other words, social trust can compensate for the shortcomings of formal institutions. At the same time, in environments with high social trust, land certification may enhance farmers’ protection and control over their land, encouraging them to engage more actively in agricultural activities and reducing their willingness and need to pursue non-agricultural employment.

Since social trust only measures the level of trust in strangers (i.e., generalized trust), it is difficult to fully capture the entire scope of trust. Additionally, individual social trust is relatively subjective and may directly influence rural labor’s non-agricultural employment behavior, potentially leading to endogeneity issues. Therefore, this paper differentiates between social trust and particularized trust, based on the agricultural production conditions of rural households, and divides the sample into “rice-growing areas” and “wheat-growing areas”. These two areas are used as proxy variables for the two types of trust, aiming to estimate the interaction between land certification and social trust more accurately.

The different production cooperation modes derived from the planting structures in these areas lead to distinct regional cultural characteristics. In rice-growing areas, the construction of irrigation systems and the increased demand for sowing and harvesting during busy agricultural seasons enhance interaction and cooperation among village members, strengthening interpersonal relationships within the village and clan, and forming a particularized trust model based on kinship and locality. In contrast, due to shorter planting cycles and labor shortages, wheat-growing areas developed labor employment networks with longer cooperative radii and a generalized trust model [44].

Theoretically, the rice-wheat planting structure formed due to geographic, climatic, and environmental factors should not directly affect the non-agricultural employment outcome variable. The survey question specifically characterizes this, “Does your family plant rice or wheat?”

Table 7. Interaction effect results; a re-examination based on planting structure.

	N-AE		Household N-AE		Household N-AE Ratio
	(1) Rice	(2) Wheat	(3) Rice	(4) Wheat	(5) Rice	(6) Wheat
Land certification	0.2017 ** (0.0866)	−0.0093 (0.0975)	0.2403 *** (0.0815)	−0.1167 (0.1143)	0.0476 *** (0.0177)	−0.0052 (0.0226)
Control variables	Y	Y	Y	Y	Y	Y
Fixed effect	Y	Y	Y	Y	Y	Y
Observations	2953	1966	1454	990	1456	994
R2	0.3416	0.3876	0.1889	0.2273	0.1635	0.1621

Notes: The values in brackets are the heteroskedasticity-robust standard errors at the village level; ** and *** indicate significance at the 5% and 1% statistical levels, respectively. In the table, ‘‘Y’’ denotes “YES”.

reports the estimation results, which show that, compared to wheat-growing areas, the effect of land certification on non-agricultural employment is more significant in rice-growing areas7. In other words, when traditional particularized trust dominates, the promotion effect of land certification on non-agricultural employment is more substantial, consistent with the previous research conclusions.

5. Discussion

In this section, we discuss this study’s potential contributions and limitations further and outline directions for future research.

First, this study contributes to filling gaps in the existing literature. Our findings indicate that rural land tenure systems positively influence non-agricultural employment, aligning with previous studies [4,26,45,46]. However, in contrast to these studies, our key innovation lies in examining how land certification (a formal institution) interacts with social trust (an informal institution) and theoretically elucidating the mechanisms through which this interaction shapes rural labor market transitions. This perspective deepens the literature on formal–informal institutional interactions and enhances the understanding of how institutional environments influence rural labor migration.

Existing research on the interaction between land certification and social trust primarily focuses on two aspects. On the one hand, some scholars examine how formal institutions shape social trust and explore their impact on macroeconomic indicators such as economic growth [43,47], capital accumulation [20], government governance [48], and firm performance [49,50]. On the other hand, other studies focus on how informal institutions influence land tenure systems, particularly the role of cultural values—such as egalitarianism, family traditions, and collectivism—in shaping property rights arrangements [51]. While these studies reveal the role of informal institutions in shaping social norms and economic behaviors, they largely overlook the perspective of social trust and fail to examine how land tenure systems are embedded in rural social networks, thereby influencing individual financial decisions.

The evolution of rural land tenure systems has reshaped the relationship between farmers and land and transformed the structure of social trust in rural areas. From the traditional “rural China” to an “urban-rural integrated China”, from a planned economy to a market economy, rural institutional changes in China have followed a trajectory of land tenure shifting from centralized control to decentralized empowerment. At the same time, these changes have fostered and reshaped the trust culture rooted in agricultural civilization and social interactions. In this context, how land tenure systems interact with social trust to influence labor market outcomes remains an underexplored topic.

Several studies are closely related to this paper, including Gongbuzeren & Li (2016), Dang et al. (2020), and Qiu et al. (2021) [32,52,53]. The first two studies highlight the complementary relationship between formal and informal institutions in rural land governance. For instance, Gongbuzeren & Li (2016) [32] examine the interaction between market mechanisms and customary institutions in rangeland management using case studies from the Qinghai–Tibet Plateau. Similarly, Dang et al. (2020) [52], drawing on household data from rural Vietnam, show that social trust can incentivize land investment when land management institutions are underdeveloped, demonstrating a complementary effect between the two. Qiu et al. (2021) [53] further extend this discussion by investigating the interplay between ancestral land and collective land tenure, revealing that ancestral land traditions promote a shift from interpersonal trust to contract-based trust, thereby increasing the marketization of land leasing. However, these studies do not systematically quantify how the interaction between land tenure institutions and social trust affects labor market dynamics, particularly in China’s rural development, leaving a significant research gap.

Using microdata from China, this study explores the impact of land certification and social trust on rural labor market transitions from theoretical and empirical perspectives. We find that both factors promote non-agricultural employment, but their interaction exhibits a negative effect, that is, an increase in social trust weakens the positive impact of land certification. Two possible mechanisms explain this finding: (1) in high-trust environments, farmers rely less on formal institutions, reducing the employment-promoting effect of land certification and (2) high social trust may strengthen farmers’ attachment to and control over their land, making them more inclined to remain in agricultural activities rather than seeking off-farm employment.

In summary, the main contributions of this study are as follows: (1) Developing an analytical framework for the interaction between formal and informal institutions, demonstrating how land tenure arrangements are embedded in social trust networks and influence farmers’ non-agricultural employment decisions, thereby expanding perspectives in institutional and development economics. (2) Highlighting the dynamic and heterogeneous nature of social trust and analyzing its moderating role in implementing rural land tenure reforms, deepening the understanding of the interplay between institutional change and socio-cultural transformation. (3) Providing policy implications for optimizing rural land tenure systems and promoting labor mobility, contributing to rural revitalization and the broader goal of common prosperity.

This study provides valuable insights into the relationship between land certification, social trust, and non-agricultural employment. However, there are some limitations. First, the analysis focuses primarily on the impact of land certification and social trust on non-agricultural employment without considering other potential influencing factors such as other informal institutions. Future research could adopt a more comprehensive analytical framework. Secondly, this study relies on cross-sectional data for empirical analysis, making it challenging to identify the long-term causal effects of land certification and social trust on non-agricultural employment. Future research could utilize panel data to address this limitation, allowing for examining information at different time points and providing dynamic analysis. In summary, while this study provides valuable insights into the role of land certification and social trust in shaping rural labor’s employment behavior, it is crucial to acknowledge these limitations and encourage further exploration in this field. Future research could incorporate panel data or experimental methods to adopt more rigorous identification strategies for validating the findings of this study. Additionally, further exploration of the dynamic evolution of social trust in response to institutional changes would provide a more comprehensive understanding of the interaction between formal and informal institutions. Ultimately, this would contribute to developing more effective policies and economic strategies in rural contexts.

6. Conclusions

Based on data from the 2017 CRHPS, this paper explores the issue of non-agricultural employment of rural labor in China from the perspective of land property rights and establishes and empirically tests a framework for understanding the mechanism through which land certification affects rural labor’s non-agricultural employment. The study reaches the following conclusions:

(1) Land certification mainly promotes non-agricultural employment among rural labor by reducing land loss risk, discouraging agricultural investment, and facilitating land financing. To address potential endogeneity issues, this paper excludes the agricultural “pull” effect of social pension insurance; considers regional development differences, introduces culture, macroeconomic and policy factors as possible omitted variables; and constructs the “village land certification rate” as an instrumental variable. The estimation results confirm the robustness of the positive effect of land certification on non-agricultural employment.

(2) From the historical perspective of China’s transition from rural to urban-rural China, a theoretical framework for the interactive effect of land property rights systems and trust culture on non-agricultural employment is established and further empirically tested. The findings show that land certification and social trust, as formal and informal institutions, respectively, have a substitutive relationship.

Based on the above conclusions, this paper offers the following policy implications: local governments should promote the differentiation of farmers based on their comparative advantages in agriculture and non-agriculture, optimizing labor resource allocation. For farmers with strong emotional ties to the land, who are skilled in farming or have advantages in agricultural production, efforts should be made to provide financial subsidies to reduce agricultural production costs, offer technical assistance, and support their development into new-type professional farmers. Farmers who are not skilled in farming should be encouraged to transfer land while strengthening training in non-agricultural employment skills to enhance their capacity for non-agricultural employment. The evolution of thousands of years of cultural tradition has deeply rooted informal institutions in China’s social structure, giving informal institutions a typical and powerful vitality in China. While maintaining the stability of rural land contracting relationships and strengthening the protection of farmers’ land contracting rights, it is essential to pay attention to the conflict and integration effects of trust culture rooted in planting structures on land certification and to tailor policies to local conditions.