The Role of Procedural Fairness: Land Titling Programs and Agricultural Investment in China

Abstract

This study examines the moderating role of procedural fairness between land titling programs and agricultural investment. We constructed a theoretical model that introduces perceived security of land tenure and procedural fairness into the traditional “property rights-investment incentives” analytical framework. Moreover, we empirically analyze the impact of land titling and its procedural fairness on agricultural investment using data from the “Thousands of People, Hundreds of Villages” survey held in 2018 among 9596 households in China. The empirical analyses were conducted by using the ordinary least squares (OLS), probit, and instrumental variable methods. Our analysis showed that land titling in China significantly promotes agricultural investment by farm households and that procedural fairness has a significant positive moderating role in the investment incentive effect of land titling as well as significantly improving the institutional credibility of land titling and enhancing farmers’ perceived land tenure security.

1. Introduction

Clear and secure land tenure is recognized as the key to encouraging investment in agriculture and reducing poverty [1]. The relationship between land tenure security and agricultural investment and its mechanism of action has also attracted considerable attention in academic and policy circles. Theoretically, land tenure security can promote agricultural investment through three mechanisms: (1) enhancing the perception of land tenure security and increasing incentives for farmers to make long-term investments; (2) improving the traceability of agricultural land, facilitating the transfer of agricultural land, and increasing the rate of return on agricultural investment; and (3) making agricultural land eligible as collateral, easing credit constraints, and providing funds for agricultural investment [2,3,4].

When poverty eradication is the goal, new investment in agriculture is a major means to achieve it, for which security of tenure is a necessary condition and where land titling programs provide security of tenure [5]. Many land titling programs have been implemented in developing countries to improve land tenure security. However, the impact of land titling on investment in agriculture remains to be determined, with mixed results from different countries [6,7]. Early studies such as those by Alston et al. [8] in Brazil, Smith [9] in Zambia, Deininger et al. [10] and Holden et al. [11] in Ethiopia, and Saint-Macary et al. [12] found that land titling promotes land-related investments, including the planting of trees and permanent crops, soil conservation measures such as crop rotation, fertilization, and minimum tillage, and the construction of infrastructure such as terraces and bunds. More recently, using a geographic discontinuity design with spatial fixed effects, Ali et al. [13] found that land titling had a substantial impact on investment in Rwanda. By contrast, a randomized controlled trial in Zambia found that land titling had no impact on investment [14]. Using a simultaneous equation model, Navarro-Catañeda et al. [15] found that land titling had little effect on land investments in Peru. Therefore, the impact of land titling on agricultural investment may vary depending on social context. As Bromley [16] pointed out, land titling may be the wrong prescription for developing countries, and land deeds may only be symbolic abstractions that undermine the original social network and informal security rules. Property rights evolve from complex patterns of power relations and local circumstances, and the assumption that a small part (land titling) of this complex cultural and legal system can be transferred to a new network of complex relationships and operations elsewhere is naive.

The current land system in rural China is a household-responsibility system. The land is owned by the collective, and farmers receive contracts and management rights (the right to use) through their identity as village members [17]. Although this type of institutional arrangement provides farmers with greater incentives, frequent land redistribution still occurs, which has become a major threat to land tenure security in China [18,19]. Therefore, the Chinese government has strengthened land tenure security by prohibiting land reallocation, extending the contract period, and issuing land certificates [20]. The Rural Land Contract Law (RLCL) of 2002 provides farmers with legal rights to manage their land and requires local governments at or above the county level to issue and register land certificates to confirm contractually managed land rights. Subsequently, in 2008, the Central Committee of the Communist Party of China officially proposed a new round of land titling. In 2013, a pilot project was launched for all provinces and the new land titling program was planned to be completed in five years. By 2018, China had virtually completed this new round of land titling. Several studies have focused on land titling programs in China. Hong et al. [21] found that China’s land titling had a positive investment incentive effect on farmers who had no experience of land reallocation but was detrimental to farmers who had experienced large-scale land reallocation. Zhou et al. [22] found that the rural land titling programs positively affected households’ land-related investments (measured by organic fertilizer application). Several studies have also shown that land titling in China has helped to promote environmentally friendly investment behavior among farmers [23,24]. Other studies have focused on the impact of China’s land titling program on land reallocation, the land rental market, credit constraints, and agricultural productivity [25,26,27,28,29].

Some scholars have also questioned large-scale land titling policies. First, the issuance of land certificates may not imply increased security of land tenure [30,31], with certain scholars contending that the traditional village order can provide the basic land rights needed to stimulate small-scale investment, whereas reforms to formalize property rights have eroded pre-existing social networks and informal rules to provide security [32,33]. Furthermore, it has been claimed that due to dysfunctional governments in developing countries, land title certificates lack an appropriate legal system backing and authoritative structures, making such certificates more nominally symbolic rather than of practical use [16]. Second, land certificates need to function through the perceived security of land tenure. Even with formal land certificates, widespread distrust among farmers of central and local institutions in developing countries, coupled with legal provisions that can legitimize land expropriation, can undermine perceptions of land tenure security [34]. Past experiences of insecure tenure can also affect people’s perceptions of current tenure security. For example, past land adjustments in China may weaken farmers’ current perceptions of tenure security, thereby affecting the incentive effect of land certificates for investment [21]. Finally, the reasons for the difficulties in making land titling work are related to inequity and the possibility of elite capture, which can affect the implementation and enforcement of land titling policies [35,36]. Local institutions are known to implement titling procedures unfairly, with abuse of power and corruption being common, and where the benefits of land titling programs are captured by a few elites. Individuals with powerful positions in the local political hierarchy have more secure land tenure, so they invest more in land fertility and produce higher yields, while poor families and women are often excluded [7,37,38,39]. It is thus clear that land titling is not only a property rights issue but also a governance issue [40,41], where the lack of credible government commitment and unfair implementation procedures can seriously hamper the impact of land titling programs. Property rights are not only a social reality but also involve social consensus, which needs to be agreed upon and widely shared to be effective [16].

Therefore, this study focuses on procedural fairness in land titling. Procedural fairness assumes that people react positively or otherwise to the outcome of a decision based on their perceptions of the fairness of the process from which the decision was made and that a fair process is necessary as an indirect control tool to protect their long-term interests [42]. Further, procedural justice is critical for building trust and enhancing the legitimacy of law enforcement authorities within communities [43], particularly when information about whether an authority can be trusted is unavailable. People will resolve the question of how to interpret the authority’s decisions by relying on perceived procedural fairness [44]. Several studies have focused on the issue of procedural justice in compulsory land acquisition [45,46,47]. Most notably, Cao and Zhang [46] empirically analyzed the impact of procedural justice on farmers’ satisfaction during land acquisition in China. However, few studies have focused on how procedural fairness influences the policy effects of land titling. This study fills this gap by drawing on data from the Renmin University of China’s “One Thousand People, One Hundred Villages” survey, which specifically examined the process of land titling. This includes not only whether farmers receive land certificates but also whether land titling has followed the normative procedures of field surveys, public announcements in the village, signature confirmations, and other requirements, to explore the mechanism of its impact on agricultural investment and how procedural fairness affects perceived security of land tenure. We find a moderating role for procedural fairness in the investment incentives of land titling, with the impact of land titling on agricultural investment increasing as procedural fairness increases. Moreover, land titling that lacks fair implementation procedures may have no investment incentive effect. We also verify the existence of a mechanism through which procedural fairness affects perceived land tenure security and that procedural fairness significantly improves farmers’ perceived land tenure security and institutional credibility, thereby enabling the institutional functional intentions of land titling to be fulfilled.

The remainder of this paper is organized as follows. The Section 2 describes a model of farmers’ investment decisions that incorporates procedural justice. Moreover, it describes the perceived security of land tenure to explain the mechanisms through which land titling affects agricultural investment and presents three hypotheses. Section 3 describes the sampling, data collection, and model construction. Section 4 empirically analyses the impact of land titling and procedural fairness on agricultural investment. Section 5 summarizes the article and discusses the policy implications of the empirical results, providing suggestions and guidance on ensuring farmers’ security of land tenure and sustainable agricultural growth.

2. Theoretical Analysis

In theory and policy, farmers’ land tenure security takes three forms: land tenure security as perceived by farmers, as a legal structure, and as de facto land tenure security [48]. Several studies have highlighted the positive impact of perceived land tenure security on investment incentives [49,50]. Therefore, this study builds on and extends the theoretical models of Feder and Feeny [3] and Besley [4] by introducing farmers’ perceived land tenure security by considering the following equation: $\theta \left(s,p,v\right)\in \left[\mathrm{0,1}\right]$, where perceived land tenure security $\theta$ can be measured by a set of farmers’ perception variables and is influenced by normative procedures such as field surveys $s$, village publicity $p$, and signature verification $v$. When ${\theta }_s^{\prime }$, ${\theta }_p^{\prime }$, and ${\theta }_v^{\prime }$ are all greater than zero, the perceived security of land tenure is enhanced by the implementation of procedures such as field surveys, village publicity, and signature verification. Procedural fairness of land titling depends on whether the land certificate is substantive. Standardized procedures give farmers a sense that their rights can be protected in the long term, thereby improving their perceived land tenure security. Next, we analyze the impact of land titling on agricultural investment at three levels: the perception of land tenure security, access to credit, and the land market.

2.1. Perception of Land Tenure Security

First, to establish our theoretical approach, we consider a farmer who invests $k_t$ in a given piece of farmland in period $t$. The expected return in period $t+1$ is $R(k_t,P_{t+1})$, where $P_{t+1}$ is the land title status in period $t+1$, and it is assumed that $R(k_t,P_{t+1})$ is increasing and concave to $k_t$, that is, $R_1^{\prime }>0$ and $R_{11}^{\prime }<0$ are satisfied. The cost is $\mathrm{C}(k_t,P_{t+1})$ and $C_1^{\prime }>0$, $C_{11}^{\prime }>0$, $C_2^{\prime }=0$. The farmer’s optimal investment decision then satisfies the conditions of the following equation:

$$\underset{k_t}{\max }\left\{W(k_t,P_{t+1})\right\}\equiv \mathrm{R}\left(k_t,P_{t+1}\right)-\mathrm{C}(k_t,P_{t+1})$$

(1)

Taking the derivative of Equation (1) with respect to $k_t$, the first-order condition can be obtained as follows:

$${\displaystyle \frac{\partial \mathrm{W}}{\partial k_t}}=R_1^{\prime }-C_1^{\prime }=0$$

(2)

According to the implicit function derivation theorem, the effect of land tenure status on agricultural investment can be obtained as follows:

$${\displaystyle \frac{\partial k_t}{\partial {\mathrm{P}}_{t+1}}}=-{\displaystyle \frac{W_{12}\left(k_t,P_{t+1}\right)}{W_{11}\left(k_t,P_{t+1}\right)}}=-{\displaystyle \frac{R_{12}^{\prime }}{R_{11}^{\prime }-C_{11}^{\prime }}}$$

(3)

It is assumed that the risk of land loss for farmers in period $t+1$ is $\mathsf{\tau }\left(P_{t+1}\right)\in \left[\mathrm{0,1}\right]$, which is a function of their land title status $P_{t+1}$, and has ${\mathsf{\tau }}^{\prime }\left(P_{t+1}\right)<0$. That is, the risk of land loss for farmers is reduced by implementing land titling (the main risks of land loss in China include land reallocation and expropriation). However, farmers do not perceive the real risk of land loss. The risk of land loss perceived by farmers is $\mathsf{\theta }\mathsf{\tau }\left(P_{t+1}\right)$, which needs to be adjusted by the perceived security of land tenure rights θ. Then, the farmers’ expected return is $\mathrm{R}\left(k_t,P_{t+1}\right)=\left[1-\mathsf{\theta }\mathsf{\tau }\left(P_{t+1}\right)\right]F\left(k_t\right)$, where $F\left(k_t\right)$ is the output of agricultural investment. Here, we assume that farmers receive no compensation for their agricultural investment after losing their land. Consequently, it is possible to calculate the following:

$$R_{12}^{\prime }=-\mathsf{\theta }{\mathsf{\tau }}^{\prime }\left(P_{t+1}\right)F^{\prime }\left(k_t\right)\ge 0$$

(4)

Substituting this into Equation (3), the following is obtained:

$${\displaystyle \frac{\partial k_t}{\partial {\mathrm{P}}_{t+1}}}={\displaystyle \frac{\mathsf{\theta }{\mathsf{\tau }}^{\prime }\left(P_{t+1}\right)}{R_{11}^{\prime }-C_{11}^{\prime }}}{F}^{\prime }(k_t)\ge 0$$

(5)

Evidently, the investment incentive effect of a land title certificate ${\displaystyle \frac{\partial k_t}{\partial {\mathrm{P}}_{t+1}}}$ is non-negative and is affected by the perceived security of land tenure $\theta$, which increases with the increase of $\theta$. Moreover, the perceived security of farmers’ property rights $\theta$ is also affected by normative procedures such as field surveys $s$, village publicity $p$, and signature verification $v$. Therefore, as the standardization of land titling procedures improves, the investment incentive effect of land titling will also increase. This model implies that although the land titling program objectively reduces the risk of land loss for farmers, only the perceived risk of land loss affects farmers’ investment decisions. In this perception process, nonstandard land titling procedures may attenuate the role of land titling programs.

2.2. Access to Credit

We also introduce the possibility of investment failure and the need for loans to obtain funds. The farmers’ initial endowment is y. If the investor invests $k_t$ in period t, the probability $q$ that the return in period $t+1$ is $F\left(k_t\right)$, or there is a $1-q$ probability that the investment fails and the return is zero. It is assumed that $F\left(k_t\right)$ is increasing and concave for $k_t$, i.e., that ${\mathrm{F}}^{\prime }\left(k_t\right)>0$ and ${\mathrm{F}}^{\prime \prime }\left(k_t\right)<0$ are satisfied. Due to information asymmetry, farmers can only take mortgage loans of loan amount $b_t$ and interest rate $r_t$. If a farmer’s investment fails, the bank sells the land. The cost is $\mathsf{\Phi }\left(\mathsf{\theta }{P}_{t+1}\right)$, and the price of disposing of the land is $D$, where ${\mathsf{\Phi }}^{\prime }<0$. This indicates that holding a land certificate can help reduce disposal costs. Moreover, the higher the perceived security of land tenure, the lower the disposal costs. Thus, irregular land titling procedures can increase the cost of disposing of land as collateral because the lack of the necessary field surveys, village publicity, and signature confirmation can lead to inaccurate four-dimensional area and ownership information in the land certificate, increasing additional social disputes and difficulties in disposing of collateral. In this case, the farmer’s optimal decision can be calculated as follows:

$$\underset{b_t,k_t}{\mathit{max}}\left\{u\left(y+{b_t-k}_t\right)+qu\left[F\left(k_t\right)-r_t{b}_t\right]+(1-q)\underset{\_}{u}\right\}$$

(6)

The first-order condition is obtained using the following:

$${\mathrm{F}}^{\prime }\left(k_t\right)=r_t$$

(7)

Under equilibrium conditions, the marginal product of agricultural investment should equal the interest rate (the price of funds). From Equation (7), we can obtain:

$${\displaystyle \frac{\mathrm{d}{k}_t}{\mathrm{d}{r}_t}}={\displaystyle \frac{1}{{\mathrm{F}}^{″}\left(k_t\right)}}<0$$

(8)

Thus, a farmer’s investment amount $k_t$ will decrease as the loan interest rate $r_t$ increases. From Equation (6), it is possible to determine the farmer’s loan demand $b_t=g\left(r_t\right)$ and satisfy $g^{\prime }\left(r_t\right)<0$; that is, the farmer’s loan demand will decrease as the interest rate increases. Then, the lender’s income can be calculated as follows:

$$\Pi \left(r_t,P_{t+1}\right)=q\left(r_t-\rho \right)g\left(r_t\right)+\left(1-q\right)[D-\Phi \left(\theta P_{t+1}\right)-\rho g\left(r_t\right)]$$

(9)

where $\mathsf{\rho }$ is the opportunity cost of the lender’s funds. Equation (9) includes two items: (1) the interest income earned by the lender if the farmer’s investment is successful, and (2) the case if the farmer’s investment fails, and the lender compensates for the loss by disposing of land as collateral. It is assumed that entering the market for agricultural land mortgage loans requires a fixed cost $S$ and $\mathsf{\Pi }\left(\mathsf{\rho },P_{t+1}\right)<$ S, i.e., financial institutions will not break even if they lend at the opportunity cost of funds. There must be a minimum interest rate level ${\mathrm{r}}_t^{*}>\mathsf{\rho }$ that satisfies $\mathrm{m}\mathrm{i}\mathrm{n}\left\{r_t\right|\mathsf{\Pi }\left(r_t,P_{t+1}\right)=\mathrm{S}\}$ for financial institutions to just break even. At this point, the following equation applies:

$${\mathsf{\Pi }}_{r_t}^{\prime }\left(r_t,P_{t+1}\right)>0$$

(10)

Assuming that the financial market is perfectly competitive at this interest rate level, a financial institution’s income from lending is equal to its cost of lending, calculated as follows:

$${\mathsf{\Pi }}_{P_{t+1}}^{\prime }\left(r_t,P_{t+1}\right)=\left(1-\mathrm{q}\right)\mathsf{\theta }\left[-{\mathsf{\Phi }}^{\prime }\right]\ge 0$$

(11)

From Equations (10) and (11), it is possible to obtain the following:

$${\displaystyle \frac{\mathrm{d}{r}_t}{\mathrm{d}{P}_{t+1}}}=-{\displaystyle \frac{{\mathsf{\Pi }}_{P_{t+1}}^{\prime }\left(r_t,P_{t+1}\right)}{{\mathsf{\Pi }}_{r_t}^{\prime }\left(r_t,P_{t+1}\right)}}=-{\displaystyle \frac{\left(1-\mathrm{q}\right)\mathsf{\theta }\left[-{\mathsf{\Phi }}^{\prime }\right]}{{\mathsf{\Pi }}_{r_t}^{\prime }\left(r_t,P_{t+1}\right)}}\le 0$$

(12)

where the loan interest rate decreases with land titling, and the degree of reduction is regulated by the perceived security of land tenure $\mathsf{\theta }$. The higher the perceived security of land tenure, the more obvious the reduction in the loan interest rate caused by the land title confirmation certificate. Further, from Equations (8) and (12), we can obtain:

$${\displaystyle \frac{\partial k_t}{\partial {\mathrm{P}}_{t+1}}}=-{\displaystyle \frac{\left(1-\mathrm{q}\right)\mathsf{\theta }\left[-{\mathsf{\Phi }}^{\prime }\right]}{{\mathrm{F}}^{″}\left(k_t\right){\mathsf{\Pi }}_{r_t}^{\prime }\left(r_t,P_{t+1}\right)}}\ge 0$$

(13)

Thus, land titling can enable financial institutions to lower lending rates by reducing the cost of disposing of agricultural land when it is used as collateral, allowing farmers to obtain more credit and increase agricultural investment. However, the premise is that land title certifications must be widely recognized and that details concerning boundaries, area, and ownership must be accurate and clear to improve the perceived security of land tenure. Otherwise, difficulties in disposing of land will not be reduced.

2.3. Land Market

The above analysis does not consider the scale of farmers’ land management or markets for agricultural land. Thus, we now introduce these two factors into our model. It is assumed that farmers’ land management scale is ${\mathrm{A}}_t$, productivity is ${\mathsf{\alpha }}_t$, land endowment is $\overline{\mathrm{A}}$, and farmers’ agricultural production function is $f({\alpha }_t,k_t,A_t)$. To simplify the analysis, we consider only land input ${\mathrm{A}}_t$ and non-land inputs $k_t$. The farmer’s optimization problem can then be presented as:

$$\underset{k_t,A_t^{\mathit{out}},A_t^{\mathit{in}}}{\max }\left\{p_{t}f\left({\alpha }_t,k_t,A_t\right){-k}_t{r}_t+A_t^{out}\left(R_t-c^{out}\right)-A_t^{in}\left(R_t+c^{in}\right)\right\}$$

(14)

where $p_t$ is the price of agricultural products, $r_t$ is the capital interest rate, $R_t$ is the land rent, $c^{out}$ or $c^{in}$ are the unit transaction costs of transferring land out or in, respectively, and $c^i\left(i=in;out\right)=c^i\left(\mathsf{\theta }\mathrm{P}\right)$, and satisfies ${c^i}^{\prime }<0$. In other words, with the implementation of land titling, transaction costs decrease.

The first-order condition can be obtained as follows:

$$p_t{f}_{k_t}\left({\alpha }_t,k_t,A_t\right)=r_t$$

(15)

$$p_t{f}_{A_t^{in}}\left({\alpha }_t^{in},k_t,A_t\right)=R_t+c^{in}if{A}_t^{in}>0\&A_t^{out}=0$$

(16)

$$p_t{f}_{A_t^{out}}\left({\alpha }_t^{out},k_t,A_t\right)=R_t-c^{out}if{A}_t^{out}>0\&A_t^{in}=0$$

(17)

Farmers with agricultural productivity ${\alpha }_t\in ({\alpha }_t^{out},{\alpha }_t^{in})$ will not participate in the land market. It is assumed that the agricultural production function is a C-D production function, $f\left({\alpha }_t,k_t,A_t\right)={\alpha }_t{k_t}^{{\beta }_1}{A_t}^{{\beta }_2}$, and satisfies constant returns to scale, i.e., ${\beta }_1+{\beta }_2=1$, ${\beta }_1,{\beta }_1>0$. It is, therefore, possible to obtain the following:

$${\alpha }_t^{in}={\displaystyle \frac{R_t+c^{in}}{{\beta }_2{k_t^{*}}^{{\beta }_1}{A_t^{*}}^{{\beta }_2-1}}}$$

(18)

$${\alpha }_t^{out}={\displaystyle \frac{R_t-c^{out}}{{\beta }_2{k_t^{*}}^{{\beta }_1}{A_t^{*}}^{{\beta }_2-1}}}$$

(19)

$$k_t^{*}={\left({\displaystyle \frac{r_t}{{\alpha }_t{\beta }_1{A_t^{*}}^{{\beta }_2}}}\right)}^{{\displaystyle \frac{1}{{\beta }_1-1}}}$$

(20)

when taking the derivative of Equations (18) and (19) with respect to $c$, we obtain ${\displaystyle \frac{\partial {\alpha }_t^{in}}{\partial c^{in}}}>0$, ${\displaystyle \frac{\partial {\alpha }_t^{out}}{\partial c^{out}}}>0$. Thus, reducing transaction costs expands the scope of farmland transfer transactions, causing low-productivity farmers to withdraw from farmland management, thereby increasing overall agricultural productivity. Simultaneously, since ${\displaystyle \frac{\mathrm{d}\mathrm{c}}{\mathrm{d}\mathrm{P}}}=\mathsf{\theta }{c}^{\prime }<0$, land titling will reduce the transaction costs of agricultural land transfer; however, the degree of reduction will also be adjusted by the perceived security of land tenure $\theta$. The higher the perceived security of land tenure, the more obvious is its effect on reducing transaction costs. Therefore, land titling can promote the transfer of agricultural land. However, it must be based on the premise that it improves farmers’ perceptions of land tenure security. Failure to ensure procedural fairness in land titling programs inevitably reduces farmers’ perceived tenure security.

When taking the derivative of Equation (20) with respect to ${\alpha }_t$, we obtain ${\displaystyle \frac{\partial k_t^{*}}{\partial {\alpha }_t}}>0$. Agricultural investment also increases as agricultural productivity increases. Given this, land titling may allow farmers with low agricultural productivity to withdraw from land management and transfer their land to farmers with higher productivity levels. This resource allocation effect increases agricultural investment. The degree of standardization of the land titling process, such as field surveys, village publicity, and signature confirmation, can improve the security of land tenure as perceived by farmers, thereby improving the transfer of farmland and exerting investment effects.

Based on the above, we propose the following research hypotheses:

Hypothesis 1:

Land titling enhances the security of land rights and promotes agricultural investment.

Hypothesis 2:

Procedural fairness positively regulates the investment incentive effect of land titling.

Hypothesis 3:

The procedural fairness of land titling improves the perceived security of land tenure.

3. Data and Methods

3.1. China’s Land Titling Programs

China has conducted several rounds of land titling programs. In 1997, the “Notice of the General Office of the Central Committee of the Communist Party of China and the General Office of the State Council on Further Stabilizing and Improving Rural Land Contract Relations” stated that “the agricultural contract departments of the township (city) people’s governments shall promptly issue unified printed land contract management right certificates to farmers”. The first large-scale confirmation of rural contract land rights was made in Article 23 of the “Rural Land Contract Law”, promulgated in 2002, which stipulated that local governments at or above the county level should issue land certificates to contractors. The 2008 “Decision on Several Major Issues in Promoting Rural Reform” introduced a new phase of rural land titling programs on the agenda and initiated pilot work in villages and counties. From 2013, China promoted land titling programs throughout its provinces and proposed completing the confirmation, registration, and certification of rural land contract management rights within five years. By 2018, China had substantially completed the confirmation and certification of rural land rights.

Confirming, registering, and issuing land contract management rights involves coverage of various areas and requires a large amount of funding in relation to aerial photography, surveying and mapping, and other tasks mainly borne by locally-derived funding. Due to the different levels of financial support in different localities, land titling methods and progress have been uneven, especially in cadastral surveys and other aspects. Field surveys were not conducted in many locations, resulting in major errors. Public announcements have not been initiated, and farmers’ signatures have not yet been obtained in some places. These factors may affect the institutional credibility of land titling programs and farmers’ perceived land tenure security.

3.2. Data Resources

The 2018 “Thousand People, Hundreds of Villages” social survey, organized by the Renmin University of China, involved conducting systematic and standardized social surveys in 300 administrative villages in 31 provinces, cities, and districts (excluding Hong Kong, Macau, and Taiwan). Questionnaires are divided into two levels: administrative villages and farm households. The respondent at the village level is the village cadre, while the respondent at the farm level is the head of household or his or her spouse. The sampling method uses a three-stage stratified sampling design in which counties are used as primary sampling units and divided into eight strata according to geographical area and poverty level. In the second stage, the administrative village is used as the sampling unit, and simple random sampling is used in each selected sample district. In the third stage, the sampling unit is the farmer’s household. We identified 30 farmers in each village through systematic or map sampling. We assume that the sampling accuracy would be similar to that of simple random sampling. Hence, the allowable sampling error of the data used in this study is 3% at the 95% confidence level, which is within the acceptable range. In total, 9596 valid farm household questionnaires were collected from 295 villages. The distribution of the samples is shown in Figure 1.

The main contents of the survey included details concerning the basic situation of villages and farmers, land rights, transfers, and management. The implementation procedures and processes of land titling are examined in detail to meet the needs of this study.

3.3. Empirical Strategies and Models

To examine the impact of land titling on farmers’ agricultural investment behavior, this study employs the following equation as our basic model:

$$Y_i={\beta }_0+{\beta }_1{Titling}_i+\sum _{j=2}^n{\beta }_j{X}_j+{\epsilon }_i$$

(21)

where $Y_i$ is the agricultural investment behavior of farmers, including the decision to invest and the investment amount. The decision to invest is estimated using a probit model as it is a dummy variable. The investment amount is then estimated using an ordinary least squares (OLS) model. ${Titling}_i$ refers to whether the farmer has a land certificate, $X_j$ refers to a set of control variables, and ${\epsilon }_i$ is a random error term. Considering the possible endogeneity problem of land certificates, omitted variables may simultaneously affect land certificates and investments, thus creating endogeneity. Therefore, this study supplements all the above regressions using the instrumental variable method. Two instrumental variables are selected in this study. Following Feng et al. [27], we use the land certificate-holding rate of n-1 other farmers in the same county as one instrumental variable for land titling. Whether a farmer’s land is titled or not is affected by the progress of land titling in that county. Meanwhile, the land certificate-holding rate of others in that county is unlikely to directly affect the farmer’s investment, thus providing a suitable instrumental variable. Further, the share of the most dominant surname in a village is used as a further instrumental variable. According to Qiu et al. [51], the informal institution of ancestral land traditionally held in villages may hinder the implementation of land titling. The more homogeneous the surnames in a village, the more likely it is that a clan culture and an ancestral land institution will develop, hindering the implementation of land titling. However, the share of the most dominant surname in a village itself is exogenous and does not directly affect the investment behavior of farm households.

To further examine how procedural fairness affects the investment incentive effects of land titling, cross-terms for both land titling and procedural fairness are added to the model as follows:

$$Y_i={\beta }_0+{\beta }_1{Titling}_i+\gamma {Fair}_i+\delta {Titling}_i\times {Fair}_i+\sum _{j=2}^n{\beta }_j{X}_j+{\epsilon }_i$$

(22)

where ${Fair}_i$ refers to procedural fairness and ${Titling}_i\times {Fair}_i$ is the interaction of the two.

Finally, to evaluate how procedural fairness affects farmers’ perceived land tenure security, we obtain estimates using the following equation:

$${Security}_i={\beta }_0+{\beta }_1{Fair}_i+\sum _{j=2}^n{\beta }_j{X}_j+{\epsilon }_i$$

(23)

where ${Security}_i$ is the perceived security of land tenure, obtained by summing a set of farmer perception variables, and ${Fair}_i$ is procedural fairness, estimated using the ordered probit model because the perceived security of land tenure is an ordinal variable.

3.4. Variable Definitions and Descriptive Statistics

Dependent variables: The agricultural investment referred to in this study includes both the construction of field ditches for irrigation and drainage as well as land leveling, application of organic fertilizer, and adoption of organic farming practices, which are related to the plot, and investment in agricultural machinery, which is not related to the plot, to provide a comprehensive assessment in relation to the type of investment in agricultural production. Two variables are introduced in this study: investment (dummy variable) and investment amount (continuous variable). Descriptive statistics of the variables are shown in

Table 1. Variable definitions and descriptive statistics.

Variables	Definition	Mean	Std. Dev.
Investment	Long-term investment in agriculture, yes = 1, no = 0	0.52	0.50
Investment amount	Total investment amount (CNY)	7121.29	36,649.15
Land titling	Whether the farmer holds a land contract management right certificate, yes = 1, no = 0	0.56	0.50
Survey	Whether the land was measured in the field when land titling, yes = 1, no = 0	0.64	0.48
Publicity	Whether the results of land titling are published, yes = 1, no = 0	0.51	0.50
Signature	Whether the farmer signed to confirm the results of the land titling, yes = 1, no = 0	0.54	0.50
Procedural fairness	Survey + publicity + signature	1.77	1.29
Perceived security	Farmers’ perceived security of land tenure	1.22	1.17
Sex	Sex of the head of household, male = 1, female = 0	0.92	0.27
Age	Age of the head of household	55.45	12.53
Education	Years of education of the head of household	1.16	0.69
Farming	Farming at home = 1, other = 0	0.50	0.50
Part-time	Half-worker and half-farmer = 1, other = 0	0.17	0.38
Non-farm	Non-farm payrolls = 1, other = 0	0.18	0.38
Family size	Total household size	5.37	2.78
Ratio of elderly	Proportion of elderly in the household	0.23	0.26
Ratio of minors	Proportion of minors in the household	0.15	0.17
Income	Gross household income (CNY, logarithm)	10.76	0.76
Plot size	Amount of land contracted by the family (mu)	6.67	11.33
Plot attribute	Percentage of paddy land	0.37	0.42
Terrain = mountainous	Mountain = 1, other = 0	0.28	0.45
Terrain = plain	Plain = 1, other = 0	0.36	0.48
Terrain = hilly	Hill = 1, other = 0	0.32	0.47
Transport	Distance from the county seat (km)	31.41	25.12
Economic	Per capita of income in the village (CNY, logarithm)	8.98	0.70
Region = east	Eastern region in China	0.38	0.49
Region = central	Central Region in China	0.30	0.46
Region = west	Western Region in China	0.31	0.46

Note: 15 mu equals 1 hectare.

.

Core independent variable: Land titling was measured regarding whether farmers have land contract management rights certificates, as land contract-management rights certificates are the core outcome of land titling and represent legal evidence for farmers’ land tenure.

Table 2. The difference in long-term investment in agriculture.

Variables	Untitled		Titled		t-Test
	N	Mean	N	Mean
Investment	4176	0.452	5420	0.57	−0.118 ***
Investment amount (CNY)	4176	5499.383	5420	8370.943	−2871.56 ***
Construction of field ditches and roads	4176	0.056	5420	0.077	−0.021 ***
Land leveling	4176	0.122	5420	0.143	−0.021 ***
Purchase of agricultural machinery	4176	0.124	5420	0.169	−0.045 ***
Application of organic fertilizer (farmyard manure)	4176	0.237	5420	0.299	−0.062 ***
Adoption of organic farming practices	4176	0.017	5420	0.024	−0.007 **

Note: ** p < 0.05, *** p < 0.01.

shows a significant difference in long-term agricultural investment behavior between farmers with and without land certificates, with farmers with land certificates having significantly higher engagement in the construction of field ditch roads, land leveling, purchase of farm machinery, application of organic manure (farmyard manure), and adoption of organic farming practices.

Mechanical variables: Procedural fairness is considered in relation to land surveys, publication of results, and signature verification. Among these, the land survey variable has the highest degree of completion, with 64% of farmers carrying out land surveys, whereas the other variables had relatively poorer completion rates. The overall procedural fairness variable is obtained by adding the results of the three variables of land survey, publication of results, and signature verification together. These procedures are conducive to the full participation of farmers in the land titling process and in clarifying actual tenure rights to reduce land disputes. Farmers’ perceived security of land tenure consists of a set of variables as perceived by farmers, typically expressed in the following terms: “Farmers have the final say in land transfer”, “The village can no longer reallocate the land”, “More investment can be made in the land”, “It can be mortgaged for loans”, “The rent for renting or leasing the land will be higher”, “The land now belongs to the farmer”, and “Farmers do not have to return the land to the collective even if they move to the city”, which provide a comprehensive understanding of the degree of farmers’ perceived security of land tenure.

Control variables: Five levels of control variables, namely, household head, family, land plot, village, and region, are selected for this study. At the household-head level, we control for sex, age, education, and occupation. At the household level, we control for household size, proportion of older adults, proportion of minors, and annual household income. At the plot level, we control for the size and type of land contracted by the households. At the village level, we control for terrain, transport, and economic and regional characteristics.

4. Results and Discussion

4.1. Impact of Land Titling on Agricultural Investment

To test Hypothesis 1 (i.e., whether land titling can provide incentives for farmers to invest in agriculture), we estimated results using two equations. First, we estimated whether farm households make investments using probit and IV-probit models, and the results are shown in

Table 3. The impact of land titling on the investment.

Variables	Dependent Variable: Investment (1 = yes, 0 = no)
	Probit	Probit Marginal Effect	IV-Probit	IV-Probit Marginal Effect
Land titling	0.164 ***	0.059 ***	0.200 ***	0.072 ***
	(0.029)	(0.011)	(0.046)	(0.017)
Sex	0.049	0.018	0.073	0.027
	(0.057)	(0.021)	(0.060)	(0.022)
Age	−0.006 ***	−0.002 ***	−0.005 ***	−0.002 ***
	(0.002)	(0.001)	(0.002)	(0.001)
Education	−0.016	−0.006	−0.016	−0.006
	(0.023)	(0.008)	(0.024)	(0.009)
Farming	0.734 ***	0.264 ***	0.715 ***	0.258 ***
	(0.045)	(0.015)	(0.047)	(0.016)
Part-time	0.679 ***	0.244 ***	0.680 ***	0.246 ***
	(0.057)	(0.020)	(0.059)	(0.021)
Non-farm	−0.012	−0.004	−0.000	−0.001
	(0.060)	(0.021)	(0.062)	(0.022)
Family size	0.020 ***	0.007 ***	0.015 **	0.005 **
	(0.006)	(0.002)	(0.006)	(0.002)
Ratio of elderly	−0.142 **	−0.051 **	−0.162 **	−0.059 **
	(0.067)	(0.024)	(0.069)	(0.025)
Ratio of minors	0.216 **	0.078 **	0.230 **	0.082 **
	(0.091)	(0.033)	(0.095)	(0.034)
Income	0.015	0.005	0.001	0.001
	(0.020)	(0.007)	(0.021)	(0.008)
Plot size	0.005 ***	0.002 ***	0.006 ***	0.002 ***
	(0.001)	(0.001)	(0.002)	(0.001)
Plot attribute	−0.118 ***	−0.042 ***	−0.114 ***	−0.041 ***
	(0.035)	(0.013)	(0.036)	(0.013)
Terrain = mountainous	−0.149 *	−0.054 *	−0.025	−0.008
	(0.080)	(0.029)	(0.086)	(0.031)
Terrain = plain	−0.163 **	−0.059 **	−0.064	−0.023
	(0.081)	(0.029)	(0.086)	(0.031)
Terrain = hilly	−0.119	−0.043	0.000	−0.000
	(0.081)	(0.029)	(0.086)	(0.031)
Transport	0.001 *	0.000 *	0.001	0.000
	(0.001)	(0.000)	(0.001)	(0.000)
Economic	0.044 *	0.016 *	0.051 **	0.019 **
	(0.023)	(0.008)	(0.024)	(0.009)
Region = central	0.372 ***	0.137 ***	0.388 ***	0.144 ***
	(0.037)	(0.013)	(0.038)	(0.014)
Region = west	0.469 ***	0.172 ***	0.498 ***	0.183 ***
	(0.038)	(0.014)	(0.040)	(0.015)
Constant	−0.997 ***		−1.081 ***
	(0.323)		(0.335)
Pseudo R2	0.086
N	8512	8512	7921	7921

Note: Standard errors are in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01.

. Second, we estimated the amount of investment made by farm households using OLS and two-stage least squares (2SLS) regression, and the results are shown in

Table 4. The impact of land titling on the investment amount.

Variables	Dependent Variable: Investment Amount (CNY)
	OLS	IV-2SLS
Land titling	2583.158 ***	2852.854 ***
	(471.011)	(716.094)
Sex	1006.018	1022.501
	(782.444)	(813.713)
Age	−193.141 ***	−185.310 ***
	(31.271)	(32.041)
Education	783.694	480.292
	(592.885)	(502.865)
Farming	3171.069 ***	2817.277 ***
	(418.880)	(444.014)
Part-time	1160.476	1015.810
	(779.542)	(800.669)
Non-farm	−3381.624 ***	−3475.406 ***
	(672.242)	(698.717)
Family size	98.732	35.633
	(86.446)	(95.557)
Ratio of elderly	95.201	−112.465
	(1210.341)	(1166.596)
Ratio of minors	13.534	235.270
	(1557.609)	(1628.493)
Income	1789.166 ***	1801.969 ***
	(385.949)	(392.804)
Plot size	579.688 ***	574.298 ***
	(134.254)	(186.547)
Plot attribute	−160.621	159.748
	(795.932)	(869.017)
Terrain = mountainous	−676.253	−1692.376
	(1547.089)	(1862.548)
Terrain = plain	1164.248	35.520
	(1592.622)	(1807.112)
Terrain = hilly	−808.794	−1790.700
	(1531.642)	(1762.819)
Transport	−13.523	−18.951
	(13.241)	(14.580)
Economic	−198.709	−267.594
	(442.510)	(499.319)
Region = central	823.458	851.835
	(733.077)	(782.763)
Region = west	−273.131	83.747
	(745.956)	(693.171)
Constant	−9138.591	−7317.588
	(5725.347)	(6219.639)
R2	0.112	0.091
N	8512	7981

Note: Robust standard errors are in parentheses; *** p < 0.01.

.

Weak instrumental variables and over-identification tests were performed to determine the validity of the instrumental variables. The first-stage regression results are shown in Table S1. Among them, the first-stage F-value of the 2SLS estimation was 4899.96, there were no weak instrumental variables, and the Hansen-J statistic of the over-identification test was 0.005, which meant that the null hypothesis could not be rejected, indicating that all the instrumental variables were exogenous and valid. Moreover, the Dubin–Wu–Hausman test result, with a value of 0.009, meant that the null hypothesis could not be rejected and that the explanatory variables were exogenous, indicating that land titling was not correlated with the error term after controlling for several factors at the household head, household, plot, village, and regional levels. Therefore, the results of our probit and OLS estimates were verified as practicable.

The regression results showed that land titling significantly increased the probability and amount of agricultural investment by farm households. Table 3 shows that land titling increased the probability of farm households investing 5.9–7.2% at the 1% confidence level. Similarly, using Table 4, it can be concluded that land titling increased the amount of investment made by farm households by approximately CNY 2583–2852 at the 1% confidence level. This finding is consistent with the findings of Hong et al. [21] and Zhou et al. [22], indicating that land titling has a significant investment incentive effect. Land titling can enhance the security of farmland property rights, protect farmers’ long-term investment returns, and promote long-term investment. Thus, Hypothesis 1 was confirmed.

Regarding the other control variables, the age of the household head and the proportion of older adults in the household had a significantly negative effect on long-term investment in agriculture, while household size and the proportion of minors had a significantly positive effect on long-term investment in agriculture. For older adults, the expected return on long-term investments had decreased, leading to a lack of investment demand. However, the larger the household size and proportion of minors, the higher the expected return from long-term investment, and the more likely they were to make long-term investments. The type of work of the household head also had a significant impact, with a higher dependence on agriculture leading to stronger incentives to make long-term investments in agriculture. Plot size and related attributes also significantly impacted long-term investment in agriculture: the larger the land scale and the lower the proportion of paddy land, the higher the probability of long-term investment in agriculture. Regional differences were also significant, with long-term investment in agriculture being higher in the Midwest than in the East.

4.2. The Moderating Role of Procedural Fairness

Furthermore, to test Hypothesis 2, this study introduced the interaction term of land titling and procedural fairness into the regression equation to test the moderating effect of procedural fairness on investment incentives for land titling. Specifically,

Table 5. The impact of land titling and procedural fairness on the investment.

Variables	Dependent Variable: Investment (1 = Yes, 0 = No)
	(1)	(2)	(3)	(4)
Land titling	−0.002	0.016	0.050	0.054
	(0.050)	(0.049)	(0.042)	(0.048)
Procedural fairness	−0.006
	(0.017)
Land titling × Procedural fairness	0.084 ***
	(0.023)
Survey		0.019
		(0.045)
Land titling × Survey		0.204 ***
		(0.061)
Publicity			−0.054
			(0.045)
Land titling × Publicity			0.209 ***
			(0.058)
Signature				−0.010
				(0.045)
Land titling × Signature				0.158 ***
				(0.061)
Control variables	Yes	Yes	Yes	Yes
Constant	−0.942 ***	−1.000 ***	−0.929 ***	−0.955 ***
	(0.324)	(0.324)	(0.324)	(0.324)
Pseudo R2	0.088	0.088	0.087	0.087
N	8512	8512	8512	8512

Note: Standard errors are in parentheses; *** p < 0.01.

shows the regression results where the dependent variable concerned whether an investment was made, estimated using the probit model, while

Table 6. The impact of land titling and procedural fairness on the investment amount.

Variables	Dependent Variable: Investment Amount (CNY)
	(1)	(2)	(3)	(4)
Land titling	−369.206	513.946	381.450	264.248
	(670.444)	(661.705)	(526.103)	(636.843)
Procedural fairness	−267.083
	(242.074)
Land titling × Procedural fairness	1521.690 ***
	(359.161)
Survey		−414.264
		(609.423)
Land titling × Survey		2975.227 ***
		(913.386)
Publicity			−1059.433
			(650.669)
Land titling × Publicity			4008.180 ***
			(893.512)
Signature				−449.925
				(617.536)
Land titling × Signature				3344.041 ***
				(900.040)
Control variables	Yes	Yes	Yes	Yes
Constant	−7990.218	−8755.481	−7875.945	−8171.846
	(5684.451)	(5712.357)	(5681.335)	(5706.075)
R2	0.114	0.113	0.114	0.114
N	8512.000	8512.000	8512.000	8512.000

Note: Robust standard errors are in parentheses; *** p < 0.01.

shows the regression results where the dependent variable concerned the amount of investment, estimated using the OLS method. The Dubin–Wu–Hausman test did not reject the null hypothesis, indicating that there was no endogeneity in relation to the explanatory variables and considering that the interaction term increased the complexity of the model. Thus, we used the above method to obtain more precise estimation results.

The regression results showed that the coefficients of the interaction terms of land titling and procedural fairness were all significantly positive at the 1% confidence level; that is, procedural fairness played a significant positive moderating role. From Table 5, it can be seen that with the improvement in procedural fairness and implementation of normative procedures (field surveys, result publication, and signature confirmation), the investment incentive effect of land titling gradually increased. In Table 6, we ascertained the magnitude of the moderating effect of procedural fairness. The results in Column (1) of Table 6 show that the investment incentive for land titling increased by approximately CNY 1522 for each additional normative procedure. Similarly, conducting land surveys, publishing results, and signing confirmations increased the investment incentive for land titling by CNY 2975, 4008, and 3344, respectively. This suggests that procedural fairness has become a key variable influencing the impact of land titling programs. A fair process helps farmers build trust in the authorities, makes them believe that their land rights will be protected in the long term, and allows them to invest with greater confidence. Furthermore, we find that the role of publicizing the results has the strongest moderating effect of all the procedures. This suggests that the openness and transparency of land titling programs are particularly important because they help reduce information asymmetry, making it easier for farmers to monitor and hold the government and relevant agencies accountable. This transparency enhances farmers’ sense of participation and belonging, significantly increases their trust in land titling policies, and thereby improves the overall effectiveness of the land titling program.

Moreover, when procedural fairness was introduced, the original land titling variable became insignificant, suggesting that land titling without procedural fairness and standardized procedures (land surveys, publication of results, and signature confirmation) is ineffective and has no investment incentive effects. This finding can be likely explained as being due to a gap between perceived land tenure security and legal land tenure security, with normative procedures such as land surveying, publicizing results, and signing confirmations closing this gap through fostering an increase in the farmers’ perceived land tenure security, thereby increasing the credibility and practical effectiveness of land titling. The next step was to validate this mechanism.

4.3. Impact of Procedural Fairness on Perceived Land Tenure Security

Finally, to test Hypothesis 3, we analyzed the effect of procedural fairness on the perceived security of land tenure using an ordered probit model. The results in

Table 7. The impact of procedural fairness on the perceived land tenure security.

Variables	Dependent Variable: Perceived Security
	(1)	(2)	(3)	(4)
Procedural fairness	0.162 ***
	(0.009)
Survey		0.374 ***
		(0.025)
Publicity			0.329 ***
			(0.024)
Signature				0.419 ***
				(0.025)
Control variables	Yes	Yes	Yes	Yes
Pseudo R2	0.024	0.021	0.020	0.024
N	8512	8512	8512	8512

Note: Standard errors are in parentheses; *** p < 0.01.

show that procedural fairness significantly improved the perceived security of land tenure at the 1% confidence level and that normative procedures such as land surveys, publication of results, and signature confirmation all improved farmers’ perceived security of land tenure. If the land is not measured in the field, the tenure boundaries cannot be clarified, leaving a possibility open for others to infringe on farmers’ land rights. Similarly, if the results of land titling are not published or signed by farmers, excluding farmers from the land titling process reduces their confidence in the government in securing their land rights. A fair process helps farmers feel that their rights will be secure in the long run, thereby increasing their confidence in land titling and reducing insecurity in relation to land tenure, which, in turn, reduces transaction costs and the cost of using farmland as collateral, meaning that land titling can truly serve as an investment incentive. This finding suggests that procedural fairness is the cornerstone of farmers’ institutional trust and that a decline in institutional credibility due to procedural unfairness will render land titling a mere formality without substantive benefit or utility.

4.4. Limitations

Although this paper analyzes in detail the impact of the land titling program on agricultural investment in China and assesses the moderating role of procedural fairness, it still has some limitations. First, this paper examines the transmission mechanisms of land titling and its procedural fairness affecting agricultural investment in terms of perception of land tenure security and access to credit and the land market, but it does not test all the transmission mechanisms in terms of empirical tests, such as the role of access to credit and the land market. Second, the data in this paper come from one year of survey data, not a long panel tracking survey. Although we control for a number of variables, such as household head, family, plot, village, and region, and use an instrumental variables approach to overcome endogeneity, we may still miss some heterogeneity factors. Future research should consider using panel data or more heterogeneity analyses to gain a more nuanced understanding of the role of land titling and its procedural fairness.

5. Conclusions and Implications

5.1. Conclusions

In this study, we constructed a theoretical model that introduced perceived land tenure security and procedural fairness into the traditional “property rights security—investment incentives” analytical framework to examine the moderating role of procedural fairness on the investment incentive effects of land titling. We analyze how procedural fairness affects the investment incentive effect of land titling through three channels: perceptions of land tenure security, access to credit, and the land market. We empirically analyzed the effects of land titling and its procedural fairness on agricultural investment using data from the 2018 “Thousands of People, Hundreds of Villages” survey of Renmin University of China. The dataset covered 9596 farm households in 31 Chinese provinces. The empirical analyses were conducted by using the ordinary least squares (OLS), probit, and instrumental variable methods. Our analysis showed that: (1) Land titling in China significantly promotes agricultural investment. The land titling programs increased the likelihood of farmers making long-term agricultural investments by 5.9% to 7.2%, and it increased their investment amounts by CNY 2583 to 2852. (2) Procedural fairness plays a significant positive moderating role in the investment incentive effect of land titling (as procedural fairness increases, the incentive effect of land titling on agricultural investment increases). For each additional standard procedure, the investment incentive for land titling increases by about CNY 1522. Conducting land surveys, publishing results, and signing confirmations increase the investment incentive for land titling by CNY 2975, 4008, and 3344, respectively. (3) Procedural fairness significantly improves the institutional credibility of land titling and enhances farmers’ perceptions of land tenure security.

Through empirical research, we found a positive relationship between land titling programs and agricultural investment, providing evidence to support ongoing efforts to improve land tenure security in developing countries. It is clear that key elements of land titling, such as clarifying land boundaries and issuing land certificates, improve household tenure security and thus increase incentives for farmers to make long-term investments. However, if the land titling process is not fair, this will weaken farmers’ confidence in the land titling program and their perceived security of land tenure, making it difficult for land titling to have an incentive effect on investment. The key to making land titling work effectively is credible government commitment and widespread acceptance by farmers.

5.2. Policy Suggestions

Our research not only has policy implications for China, but can also provide lessons for developing countries where a large number of land titling programs have been undertaken.

First, to promote sustainable agricultural investment, national authorities should establish a credible land tenure system, rather than simply issuing formal land certificates. Our research shows that land titling without fair procedures does not serve as an investment incentive. Countries should carry out land titling programs with the full participation of local farmers and ensure the fairness and transparency of the policy, so as to effectively play the policy effect, give farmers long-term and stable property rights expectations, and promote agricultural investment.

Second, in order to provide farmers with a credible commitment, it is necessary to further standardize the process of land titling programs. The authorities should establish a complete process of land investigation, ownership confirmation, result publication, and certificate issuance, pay attention to the resolution of land disputes and conflicts revealed in the land titling programs to protect the rights and interests of disadvantaged groups, and ensure procedural fairness in the land titling process to increase farmers’ confidence in the land titling program. However, more standardized procedures imply higher costs for land titling, and large-scale land titling may face a cost-benefit dilemma. In the future, more attention should be paid to measuring the costs and benefits of land titling programs to guide policy practices.

Finally, we recommend that the authorities enhance the development of credit and land markets by improving procedural fairness. When the land titling process is open and transparent and ownership is clear, the likelihood of land disputes is greatly reduced. Reducing disputes can not only reduce market uncertainty caused by information asymmetry and enhance the vitality of the land market but also improve the reliability of land as collateral, thereby promoting the development of credit markets. This improvement in the tradability of land and the availability of credit is particularly important for farmers and small businesses, as it makes it easier for them to obtain funds for agricultural investments.