Impacts of Tenure Security on Rural Households’ Forestland Investment: Evidence from Jiangxi, China

Abstract

This paper examines the distinct effects of actual and perceived security on forestland investment by rural households. To achieve this, we utilized Tobit and IV-Reg models to analyze repeated survey data from 500 households residing in 50 villages in Jiangxi Province during the years 2017 and 2018. We measured households’ investment in forest management by labor and cash inputs. The findings indicate that actual and perceived tenure security significantly influence forestland investment. Specifically, the possession of forestland certificates exhibiting a marked increase in labor and cash inputs. However, the logging quota system has a significant negative impact on cash input, but no significant effect on labor input. With regard to perceived tenure security, the evaluation and comprehension of the existing tenure policy by households contribute positively to both labor and cash inputs in forestland. From our analysis, it is recommended that the logging quota system be revised to incentivize farmers’ active participation in forest management, and the government should strive to raise awareness among rural households about forest tenure policy.

1. Introduction

Tenure security of land is widely recognized as a key factor in driving investment and fostering economic growth [1,2]. Land tenure reforms are often viewed as a primary approach to enhancing agricultural productivity and spurring investment in agriculture [3]. Since forestry is a long-term investment, forestland tenure and property right market rules are vital for affecting household investments [4].

Forestlands in China are classified under two categories of property rights: state-owned and collective-owned. The productivity and stock levels of collectively owned forests have been consistently low over time, primarily due to the failure of the tenure system and governance mechanisms, which have led to a lack of active forest management [5]. Following the reform and opening-up policies that began in the late 1970s, the Chinese government has implemented multiple rounds of collective forest tenure and governance system reforms [6], with a focus on improving individual tenure and promoting market liberalization [7,8]. The task of the most recent round of collective forest tenure reforms (CFTR) was to clarify forest property rights, establish clear forestland boundaries, and distribute legal certificates to families. These measures aimed to ensure that individual households possess the rights to utilize and manage their own forestland, ultimately improving forestland tenure security and fostering increased investment by families. By 2018, the Chinese government had allocated approximately 180 million hectares of collective forestland to households, accounting for 99% of the total collective forestland area [9]. With the CFTR in place, households have now assumed the primary role as management entities for collective forests. Consequently, the management and investment in forestland by farmers have become integral to forestry production, serving as a crucial avenue for enhancing the productivity of collective forests.

However, the CFTR did not achieve the desired and satisfactory results. The impact of forestland certification on farmers’ investments is temporary and unsustainable [10,11]. Although the reforms have greatly improved the tenure security of forestland, farmers still experience insecurity regarding their actual and perceived tenure [12]. Yin et al. (2013) discovered that while the forest tenure reform enhances the basic incentive structure, conflicting policies and inconsistencies can undermine any progress made [5]. For instance, China’s forestry sector implemented logging quotas in 1987 in response to the accelerated reduction in forest volumes. The State Forestry Administration prescribes these quotas for each province every five years. Subsequently, the provinces allocate these quotas to their corresponding counties, townships, and villages. According to the Forest Law, all logging activity necessitate prior approval from the appropriate government forestry agency. The logging permit system and the prohibition in natural forests have significantly limited the freedom of timber harvesting on family forestland [8,13]. Wang et al. (2013) also found that approximately 60% of the 2200 households surveyed lacked trust in the government’s initiatives to bolster tenure security [14]. Therefore, it is crucial to differentiate between actual and perceived tenure security and evaluate their possibly differing influences on households’ management behaviors, including their investments in forestland.

Numerous studies have explored the influence of land tenure security on investment. The majority of these studies have concluded that enhanced tenure security leads to increased investment and productivity [15,16]. However, a few studies have failed to identify any significant impact of land tenure security on investment [17,18,19]. The contradicting conclusions from land tenure research are largely derived from the variations in how to define and measure tenure security [20]. To address the issue of using various indicators to represent the concept of tenure security, Van Gelder (2010) has put forward a three-part conceptualization of tenure security that includes three key components: legal tenure security, actual tenure security, and perceived tenure security [21].

Empirical research on land tenure security and investment has primarily concentrated on either household perceptions of tenure security [22,23,24], or on current land tenure arrangements (actual tenure security) [25,26], such as the acknowledgement of land rights and land expropriation by local governments. However, current land tenure arrangements may also affect investment through channels other than tenure security perceptions. Perceived tenure security may not always align with the legal standing of a tenure scenario, and it may differ from the actual conditions [21,27].

Despite the significance of land tenure security in investment, the prior literature has given little attention to analyzing the impact of both actual and perceived land tenure security on farmers’ investment in forestland. The Chinese government has implemented collective forest rights system reforms aimed at stimulating farmers’ enthusiasm for production since 2003. As the execution of these policies varies significantly across different areas of China [12], this presents an important opportunity to empirically investigate the effect of actual and perceived land tenure security on forestland investment. Therefore, our study aims to contribute to the literature by empirically evaluating the distinct impacts of actual and perceived tenure security on farmers’ forestland investment.

Research indicates a potential reverse causal relationship between farmers’ perceived tenure security and their investment in forestland. Specifically, the act of investing in forestland, particularly in tree planting, might contribute to farmers’ perception of enhanced tenure security [17,28]. To achieve this objective, we employ Tobit and IV-Reg models in conjunction with household survey data collected from the southern collective forestry region of Jiangxi Province. Additionally, we intend to assist the government in formulating or refining pertinent policies to promote farmers’ investment in forestland by identifying the causes of forestland tenure insecurity.

2. Theoretical Mechanisms

Land tenure security is defined as the level of confidence that an individual’s land rights will be acknowledged and safeguarded in situations of particular disputes [29]. Vailable studies claim that LTS encourages investments through three channels: higher anticipated returns from investment (assurance effect), more efficient land markets that enable the transfer of land to more productive producers (transferability effect), and greater access to capital (collateralizability effect) [17,29].

Legal tenure security pertains to the legal status of land tenure and its safeguarding supported by state authority and a well-established set of legal rules [21]. The objective of the CFTR is to provide effective legal documentation for the protection of households’ forestland rights. Two significant measures taken in this regard are issuing land certificates to farmers and implementing the policy of separating the contracting right from the management right. Despite the uniformity of the laws and policies of the CFTR in the collective forest region, there are significant disparities between actual and perceived land tenure security. Hence, this study examines the impact of actual and perceived tenure security on forestland investment.

Forestland investment can be defined as any time or monetary expenditure in forest management activities, mainly including labor, cash, technology, and other production factor inputs [8]. Due to farmers having little technological investment in the actual production and operation process of forestry, forestland investment in this study mainly includes labor and cash inputs.

2.1. Impact of Actual Tenure Security on Forestland Investment

Actual tenure security refers to the real situation on the ground [21], which primarily depends on whether the established regulations and rules have been effectively implemented at the local level [12]. Apart from land titles, other factors such as local customs, norms, and duration of possession may also contribute to a high degree of tenure security [30]. In China, most empirical studies use indicators from households’ past experiences in land redistribution and possession of land certificates to measure actual land tenure security [31,32,33]. However, Zhong et al. (2009) have noted that land adjustment is not very frequent in most areas [18].

Two significant components of actual tenure security on forestland are possession of forestland certificates and the logging quota. First, due to difficulties in defining forest tenure and the complicated registration process of forestland tenure, forestland certificates are often not issued to farmers, which leads to deviations between the expected and actual results of policy implementation. Second, the newly amended Forest Law specifies that logging must be regulated, and any logging activity must gain approval from the public forestry agency. This implies that forestland users do not have complete rights to benefit from or dispose of their forestland, which reduces their anticipated income and negatively affects their investment behavior.

2.2. Impact of Perceived Land Tenure Security on Forestland Investment

Perceived tenure security refers to the perception of the likelihood of losing one’s land, which is distinct from the actual likelihood of that risk [21]. Personal characteristics, such as age, gender, or risk preferences, can be significant factors that contribute to household perceptions of tenure security. Hence, perceived tenure security may not necessarily align with the legal status of a tenure situation, and it may differ from the actual circumstances.

Higher perceived land tenure security typically encourages land investments. Most empirical studies employ farmers’ perceived risk of losing their land as a measure of perceived tenure security [3,24]. In China, land ownership belongs to the state. Therefore, commonly used indicators of perceived land tenure security include understanding of the ownership of farmland contract rights [34,35] and the households’ expectation regarding land reallocations or land expropriations [12,36]. In 2014, the Chinese government separated rural land contract management rights into two distinct components: a contractual right (right of disposal) and a management right. This strategic move aimed to legalize land transfer and enhance the efficiency of land management. Under the institutional framework of “separation of three rights”, farmers’ understanding of the policy reflects their cognition of the security of forestland contract rights. It is the crucial factor in the investment decision of farmers. The different aspects of tenure security discussed above are summarized in Figure 1.

Based on the above analysis, the following hypothesis is proposed:

H1. 

Improved tenure security positively influences households’ forestland investment;

H2. 

Improved actual tenure security positively influences households’ forestland investment;

H3. 

Improved perception of tenure security positively influences households’ forestland investment.

3. Methods

3.1. Data Collection

The data used in this study were collected from a rural household survey that was carried out in the collective forest region of Jiangxi, which is known for its vast forest resources and is considered a typical forestry area. Jiangxi Province was chosen as the pilot province for the CFTR in 2004, given its status as the second-highest forest coverage province in China. By 2009, the province had successfully achieved the primary objectives of the reform.

The authors collected this data from 500 households in 10 counties in Jiangxi in 2017. And we conducted a repeated survey in 2018. Considering the forest resource status, socio-economic conditions and spatial distribution of counties within Jiangxi Province, a stratified random sampling method was used to select 10 sample counties. Then, 5 villages were selected from each county, and 10 households were randomly selected from the household registration list of each village. As a result, the sample encompassed 500 rural households from 50 villages in 10 counties. In addition, due to objective reasons, the sample of households actually surveyed in 2017 and 2018 were 502 and 503, respectively. To maintain data consistency, 500 samples were finally used after screening.

Each household questionnaire survey lasted approximately an hour, with face-to-face interviews conducted to ensure the validity of the responses. The questionnaire primarily encompassed household demographic characteristics, forestland resources, forest management, production expenditure, information related to CFTR, and farmers’ cognition to tenure security. Additionally, in each sampled village, a member of the village committee was interviewed to assess the socioeconomic and the implementation of forest tenure reform policies. The village survey mainly included basic information about the village (such as total population, distance from the village to the nearest town, and per capita income of the village), the proportion of forestland certificates issued, among other details.

3.2. Selection and Definition of the Model’s Variables

3.2.1. Dependent Variable

The dependent variable in this study is forestland investment, which includes labor and cash inputs. Labor input includes the total time invested by the family’s own labor and hired labor. Cash input includes seeds, fertilizers, pesticides, and other cash outlay in the process of farmers’ forestland management. Considering that the difference of man-day price in different regions and the incomplete substitutability of labor and cash inputs, the unit of measurement for labor input in forestland is “man days” instead of “yuan”.

3.2.2. Independent Variable

Actual forestland tenure security is evaluated using two main indicators in this study: possession of forestland certificates and the ease of obtaining logging quotas. Forestland certificates are an essential tool for verifying and protecting land ownership, but their issuance can be complicated due to challenges in forestland demarcation and policy implementation variations. Therefore, households’ ownership of forestland is a clear indicator that reflects the effectiveness of the forestland property rights system [33]. If a household does not possess a forest certificate, the assigned value for possession of forestland certificates is 0; if a household holds a certificate for a portion of forestland, the assigned value is 1; and if a household owns a certificate for all forestland, the assigned value is 2. Additionally, the ease of obtaining logging quotas is also an important factor in evaluating actual forestland tenure security since farmers are more likely to invest in forest management if they have an economic incentive to do so [37]. The ease of obtaining logging quota is assigned a value of 0 if it was easy for a household to obtain logging quotas; 2 if it was not easy, and 3 if it was difficult.

Perceived forestland tenure security is composed of two elements: farmers’ subjective comprehension of their contracting rights security and their satisfaction level with the collective forestland tenure policies. Farmers’ satisfaction with the policies reflects their trust in national empowerment, and a positive attitude will motivate them to participate in forestland management. In addition, because there may be endogenous relationship between perceived tenure security and farmers’ investment behavior, village-level perceived tenure security is selected as a tool variable, that is, the average quantity of other farmers’ perceived tenure security in the village where the farmer lives [22].

3.2.3. Control Variables

This study also considers several control variables that may influence households’ investment decisions, including household characteristics, forestland characteristics, and village characteristics. To ensure consistency with previous studies [38,39], basic household indicators such as age, education level, and the number of household laborers were included. Forestland characteristics, such as forest area and forest types (i.e., bamboo and timber forests), were chosen due to their different values [40]. The village characteristics considered were population, distance from the village to the nearest town, and the per capita income of the village [41].

Table 1. The variables used in econometric modeling and their preliminary statistics.

Variables	Definition	Mean	SD
Dependent variables
Labor input	Labor input (including own labor input and hired labor) in the process of farmers’ forestland management (days)	70.522	227.77
Cash input	Cash input (including seeds, fertilizers, pesticides, and other cash outlay) in the process of farmers’ forestland management (yuan)	6316.45	64,097.93
Independent variables
Actual tenure security
Certificates	Possession of land certificates (0 = without certificates; 1 = with certificates for part of forestland; and 2 = with certificates for all forestland)	1.87	0.42
Quota logging	Ease of obtaining logging quotas (0 = easy; 1 = not very easy; and 2 = difficult;)	0.915	0.963
Perceived tenure security
Understanding	Whether households know the policy of separating contracting and management rights (0 = no; 1 = yes)	0.606	0.488
Evaluation	Personal evaluation of the forestland tenure reform policy (1 = bad; 2 = not so good; and 3 = good)	2.355	0.837
Satisfaction	Satisfaction with the forestland tenure reform policy (1 = not satisfied; 2 = partially satisfied; and 3 = satisfied)	2.538	0.683
Control variables
Age	Household head’s age (years)	56.44	10.21
Education	Education level of household head (1 = elementary school and below; 2 = junior school; 3 = high school; and 4 = college or bachelor’s degree or above)	1.895	0.818
Labor	The number of laborers in a household	3.95	1.82
Labor out	The number of out-migrant laborers in a household	1.89	1.31
Area	Total area of forestland operated by a household (mu)	118.79	276.60
Bamboo forests	Whether bamboo forest is dominant (0 = no; 1 = yes)	0.34	0.475
Timber forests	Whether timber forest is dominant (0 = no; 1 = yes)	0.42	0.494
Population	Village’s overall population	1622	63.25
Distance	Distance from the forestland to the nearest town (kilometers)	7.246	6.069
Village income	Average annual income of farm households (yuan)	5524.60	2034.55

Note: The unit of forest area is mu (1 mu = 667 m² or 0.067 ha), and that of income is Chinese yuan (USD 1 = CNY 6.89 as of March 2023). SD is Standard Deviation.

provides a detailed description of these variables. Table 1 presents a summary of the chosen dependent and independent variables in this study.

3.3. Econometric Model

Drawing on descriptive analysis of the collected data, this paper develops a model to examine the impact of tenure security on farmers’ forest investment, encompassing both labor and cash investments. To achieve this, we utilize research data from farmers and employ regression analysis. Specifically, we investigate the effects of both actual and perceived tenure security on farmers’ investment in forestland. Furthermore, the explained variables in our empirical model are farmers’ labor input and cash input. To specify the model, we begin with the following equation:

$$y_i={\beta }_0+{\beta }_{1}securit{y}_i+{\beta }_3{X}_i+{\epsilon }_i$$

(1)

In Equation (1), $y_i$, the dependent variable, denotes the labor input and cash input in farmers’ forestland investment; $securit{y}_i$ represents the actual and the perceived tenure security; and $X_i$ means a set of control variables that encompass household characteristics, forestland characteristics, village characteristics, and more. The error term is denoted by ${\epsilon }_i$.

In previous studies, to address endogeneity problems, scholars have employed instrumental variables using lagged terms of the endogenous variables. However, due to limited short panel data in this study, lagged variables are not available to address the endogeneity of the core independent variables. Instead, some studies have used the mean of perceived tenure security among other farmers in the same village as a reasonable instrumental variable [22,42]. Therefore, in this paper, the average value of perceived tenure security of other households in the same village is used as an instrumental variable to address the issue of endogeneity within the model’s original variables. This instrumental variable is considered strictly exogenous for two main reasons. First, within the same village, the perceived tenure security of individuals can be influenced by the perception of others. Second, other households’ perceived tenure security does not have an impact on the behavior of individual households. Since the dependent variable in this study is a continuous variable, we can use IV-Reg to eliminate the endogeneity of the perceived tenure security variable. This method takes into account the influence of both instrumental and control variables. The equations for the instrumental and control variables are as follows:

$$\begin{array}{l}I*=\beta Lw+\alpha Xi+\mu \\ Lw=\gamma 1Zi+\gamma 2Xi+vi\end{array}$$

(2)

where $I^{*}$ is an unobserved variable, it will be influenced by farmers’ perceptions of tenure security and control variables. $L_w$ represents a set of control variables, including factors of householders and households. When using instrumental variables, it is important to test their validity. In addition, to test for weak instrumental variables, we performed a linear regression of Equation (2).

4. Results

4.1. Descriptive Results Analysis

Table 1 presents descriptive statistics for the sample villages. The average labor input for farm households was 70.52 days, while the average cash input was CNY 6316.45. In terms of forestland tenure security, 14.59% of households had certificates for only a portion of their forestland, and 4.51% of households did not possess a forestland certificate at all, indicating that the actual state of forestland tenure security does not completely align with its legal version.

Furthermore, 43.34% of respondents found it not very easy to obtain logging quotas, and 14.19% reported experiencing difficulty obtaining them. These observations suggest that the actual tenure of forestland might not be stable or favorable. Regarding the policy of decoupling land contracting and management rights, approximately 40% of farmers admitted their lack of comprehension about the policy.

In terms of personal opinion on the policy, 16.78% expressed a negative viewpoint, and 26.27% considered it to be poor. In terms of satisfaction with the policy, 39.12% were partially satisfied, and 14.96% were not satisfied. These findings indicate that there were variations in farmers’ understanding and views of the policy.

4.2. Analysis of Regression Results

Table 2. Regression results of farmers’ labor input.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5
Actual tenure security
Certificate	0.188 * (0.131)	——	0.199 * (0.128)	——	0.326 ** (0.135)
Logging quota	0.016 (0.077)	——	0.038 (0.074)	——	0.151 (0.145)
Perceived tenure security
Understanding	——	0.045 (0.094)	0.009 (0.091)	0.727 *** (0.219)	0.792 *** (0.233)
Evaluation	——	0.676 *** (0.262)	0.685 ** (0.250)	0.320 ** (0.127)	0.310 ** (0.131)
Satisfaction	——	0.208 ** (0.079)	0.186 ** (0.087)	0.197 *** (0.073)	0.148 *** (0.128)
Control variables
Age	−1.811 *** (0.657)	−1.783 *** (0.656)	−1.808 *** (0.657)	−1.364 *** (0.252)	−1.393 *** (0.255)
Education	−0.075 (0.129)	−0.186 (0.368)	−0.151 (0.145)	−0.060 (0.105)	−0.026 (0.025)
Labor	0.227 *** (0.039)	0.227 *** (0.038)	0.293 *** (0.068)	0.166 *** (0.043)	0.168 *** (0.042)
Labor out	−1.121 *** (0.171)	−1.159 *** (0.168)	−0.991 ** (0.441)	−1.364 *** (0.252)	−1.393 *** (0.255)
Area	0.075 (0.129)	−0.186 (0.368)	−0.151 (0.145)	0.166 (0.043)	0.168 (0.042)
Bamboo forests	0.002 ** (0.014)	0.002 ** (0.010)	0.002 *** (0.011)	0.002 ** (0.000)	0.002 ** (0.000)
Timber forests	−0.005 (0.001)	−0.005 (0.001)	−0.047 (0.007)	−0.038 (0.309)	−0.040 (0.029)
Population	−0.001 ** (0.000)	−0.001 *** (0.000)	−0.001 *** (0.000)	−0.001 *** (0.000)	−0.001 *** (0.000)
Distance	−0.041 *** (0.009)	−0.039 *** (0.009)	−0.037 *** (0.009)	−0.039 *** (0.009)	−0.037 *** (0.009)
Village income (ln)	−0.008 (0.006)	−0.006 (0.005)	−0.007 (0.004)	0.166 (0.143)	0.168 (0.142)
Instrumental variables	No	No	No	Yes	Yes
R-squared	0.0027	0.0131	0.1257	0.161	0.167
Prob > F	0.0000	0.0237	0.0000	0.0000	0.0000

Note: Standard errors in parentheses; ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively; see Table 1 for variable definitions.

and

Table 3. Regression results of farmers’ cash input.

Variables	Model 6	Model 7	Model 8	Model 9	Model 10
Actual tenure security
Certificates	0.778 ** (0.444)	——	0.953 ** (0.443)	——	0.894 *** (0.341)
Quota logging	−0.101 ** (0.059)	——	−0.085 * (0.059)	——	−0.229 ** (0.157)
Perceived tenure security
Understanding	——	−0.109 (0.110)	−0.035 (0.112)	0.687 * (0.368)	0.772 *** (0.221)
Evaluation	——	0.223 * (0.168)	0.183 * (0.169)	1.239 * (0.669)	0.123 * (0.154)
Satisfaction	——	0.474 * (0.267)	0.418 * (0.313)	0.459 ** (0.330)	0.418 ** (0.211)
Control variables
Age	−0.197 * (0.073)	−0.148 * (0.128)	−0.150 * (0.082)	−0.181 * (0.111)	−0.130 * (0.127)
Education	1.371 *** (0.249)	1.353 *** (0.257)	1.431 *** (0.262)	1.386 *** (0.252)	1.197 *** (0.731)
Labor	0.162 *** (0.065)	0.165 *** (0.065)	0.336 *** (0.126)	0.165 ** (0.065)	0.165 *** (0.064)
Labor out	−1.419 *** (0.297)	−1.398 *** (0.296)	−1.374 ** (0.219)	−1.401 *** (0.303)	−1.401 *** (0.292)
Area	0.015 (0.012)	0.001 (0.012)	0.001 (0.018)	0.001 (0.001)	0.001 (0.001)
Bamboo forests	0.033 (0.011)	0.037 (0.012)	0.021 (0.017)	0.021 (0.001)	0.020 (0.001)
Timber forests	0.073 (0.018)	0.075 (0.018)	0.064 (0.014)	0.005 (0.001)	0.005 (0.001)
Population	−0.033 (0.009)	−0.030 (0.009)	−0.029 (0.009)	−0.034 (0.009)	−0.027 (0.009)
Distance	−0.038 *** (0.009)	−0.035 *** (0.009)	−0.040 *** (0.009)	−0.051 *** (0.017)	−0.049 ** (0.023)
Village income (ln)	−0.030 (0.127)	0.010 (0.131)	−0.099 (0.137)	−0.123 (0.154)	−0.099 (0.137)
Instrumental variables	No	No	No	Yes	Yes
R-squared	0.035	0.036	0.013	0.038	0.155
Prob > F	0.000	0.000	0.000	0.000	0.000

Note: Standard errors are in parentheses; ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

show the results of the Tobit model regressions of forestland tenure security on labor and cash inputs, respectively. The following is a specific analysis of the impact on both labor input and cash input, respectively.

4.2.1. Impacts of Tenure Security on Farmers’ Labor Input in Forestland

Table 2 presents the regression models used in this study. Model 1 assesses the impact of actual tenure security on farmers’ labor input in forest management, while model 2 examines the impact of perceived tenure security on farmers’ labor input in forest management. Model 3 is a pooled model that considers the influence of both actual and perceived tenure security on farmers’ labor input in forestland. The coefficients in the regression models have consistent weights and signs, suggesting that the findings are robust and not affected by the specific model used. Overall, the results provide insights into the relationship between tenure security and farmers’ investment in forestland.

Since there is a reverse causal relationship between perceived tenure security and forestland investment, and a strong endogeneity problem between them. Therefore, this paper uses the mean value of perceived tenure security of forest land of other farmers in the same village as an indicator of perceived tenure security of farmers in this village as well as the instrumental variable. Model 4 and model 5 are the regression results after adding the instrumental variables of perceived tenure security.

In terms of actual tenure security of forestland, according to model 3 and model 5 in Table 2, the coefficient of the forestland certificates indicator is positive, and the coefficient of influence is increased by adding instrumental variables and passes the 5% significance test, which indicates that possession of forestland certificates had a significantly positive influence on labor input in forestland. Forest tenure certificates provide a basis for the legal protection of forestland against illegal land occupation and land conflicts. As a result, farmers holding forest tenure certificates usually perceive their forestland tenure as more secure and are more willing to invest in forestland. The difficulty of obtaining logging quota did not have a significant effect on labor input in forestland management, and although the coefficient was negative with the inclusion of instrumental variables, it did not pass the significance test.

In terms of farmers’ perceived tenure security, it can be seen from models 3 and 5 that after adding the instrumental variables, understanding changes from insignificant to significant at the 1% level, probably due to the endogeneity of this variable. In comparison, Evaluation and Satisfaction had a significant positive influence on farmers’ labor input in forest management, which was significant at the 5% and 1% levels, respectively (model 5), indicating that the better and more satisfied farmers are with the collective forestland ownership policy, the more they will increase their labor input in forestland management.

Regarding the effect of control variables on farm labor input, it can be seen from model 5 that Age and Labor out had significant negative impacts on labor input in forest management, both significant at the 1% statistical level, implying that the older the head of the household is, the lower the labor input of his or her household. Since the average age of the sample farmers was close to 60 years old, and in the collective forest areas studied, the location of the farmers’ contracted forestland was mostly located in forest areas far from their residence. Therefore, the increase in the age of the farmer will have a greater negative impact on labor input in forest management, especially self-invested labor. Similarly, Labor out had a significant negative influence on labor input in forest management; the higher the proportion of Labor out in the household, the more the forestland management input of the whole household is bound to be adversely affected. Thus, the off-farm employment behavior of laborers in farming households limits their labor input to forestry.

Regarding the characteristics of forestland, the predominance of the Bamboo forest had a significant positive influence on labor input. This indicates that farmers with a larger share of bamboo forest area had a positive effect on labor input. A possible reason for this is that the cutting and utilization of timber and economic forests are low due to the restriction of the quota logging policy, while the harvesting of moso bamboo is almost unrestricted. Therefore, farmers will increase the area of bamboo forest plantation and labor input.

In terms of village characteristics, Distance was negatively correlated with the labor input in forestland management, indicating that the farther the sample farmers’ villages are from the townships, the lower their labor input. Because the distance between villages and townships usually reflects the accessibility and road accessibility of the area where the forestland is located; the farther the distance from the township, the more unfavorable it is for its management.

4.2.2. Impacts of Tenure Security on Farmers’ Cash Input in Forestland

In Table 3, model 6 is a regression model considering the actual tenure security indicator alone, model 7 is a regression model considering the perceived tenure security indicator alone, and model 8 is a pooled model considering both actual and perceived tenure security variables. Model 9 and model 10 are regression results with the inclusion of the instrumental variable of perceived tenure security. The consistency of the coefficients in both types of models suggests that the results are robust.

According to model 8 and model 10 in Table 3, as shown in the actual tenure security, the inclusion of instrumental variables positively promotes the cash input in farmers’ forestland management with a coefficient of 0.894 at 1% level of significance for forestland certificates holdings, which suggests that farmers’ holding forestland certificates enhances their expectation of stability in obtaining income from forestland management, thus positively promoting farmers’ cash input behavior in forestland.

In Model 8 and model 10, the inclusion of instrumental variables increases the absolute value of the coefficient of the Logging quota indicators and negatively affects the cash input at the 5% significant level. This is because difficulties in obtaining logging quota will reduce farmers’ expected income from their forestland, which will lead them to reduce their cash input in forestland. The results of the Evaluation indicator of the forestland tenure reform policy are consistent with the results in Table 2, and the effect of the Understanding indicator changes from insignificant to significant after the inclusion of instrumental variables, further validating the endogeneity issue of this indicator. Satisfaction also has a significant positive contribution to the cash input in forestland. In conclusion, the more secure the farmers’ perceived tenure securities are, the more motivated they are to operate their forestland and the more cash input they make.

Model 10 reveals the impacts of control variables on cash input. Age was negatively associated with cash input in forestland, while Education had a positive correlation with cash input and is statistically significant at the 1% level. This suggests that older farmers tend to be more risk-averse and thus less likely to invest cash in forestland. Conversely, farmers with higher levels of education are more likely to earn non-farm income and may be more willing to invest financially in forestland. Moreover, the number of laborers in a household had a significant positive impact on cash investment, while the number of migrant laborers in a household and the distance between the village and township had a significant negative effect on cash input, significant at the 5% level. The above regressions of the sample data were all carried out in Stata 15.0 statistical software.

5. Conclusions

This study constructs a theoretical framework of the impact of tenure security on farmers’ forestland input behavior, extends Van Gelder’s (2010) theoretical analysis of the three dimensions of tenure security using field research data from 500 farmers in Jiangxi Province from 2017 to 2018 [21], and empirically tests the impact of actual tenure security and perceived tenure security on farmers’ forestland input behavior. Theoretical analysis shows that tenure security promotes farmers’ forestland investment; actual tenure security and perceived tenure security have distinct effects on farmers’ forestland investment behavior. The empirical results show that, in terms of actual tenure security, farmers’ holding forestland certificates significantly promotes farmers’ labor input and cash input in forestland management; the logging quota indicator has a significant inhibitory effect on farmers’ cash input, while the effect on farmers’ labor input is not significant. Concerning perceived tenure security, the understanding and evaluation of tenure security policy have significant positive effects on both labor input and cash input in farmers’ forest management.

This study’s contributions can be summarized in the following three aspects. First, unlike most previous research, we investigated the distinct impacts of actual and perceived tenure security on rural households’ forestland investment. To measure actual tenure security, we examined the possession of forestland certificates and the ease of obtaining logging quotas, while perceived tenure security was mainly measured through farmers’ perceptions of the attribution of forestland property rights and their satisfaction with the policy or system of collective forestland property rights, drawing on the work of Hu et al. [35] and the characteristics of the forest land property rights system. Our findings indicate that overall, tenure security had a significantly positive effect on farmers’ forestland investment, both in terms of labor and cash investment. These findings align with previous research [33,43,44]. By exploring the distinct impacts of actual and perceived tenure security on forestland investment, this study provides a more nuanced understanding of the relationship between tenure security and forestland investment. The findings highlight the importance of both actual and perceived tenure security for promoting sustainable forest management practices and supporting rural households’ livelihoods.

Second, it is worth noting that the same variable may have different impacts on farmers’ cash input and labor input. The study revealed that the logging quota indicator had a negative impact on cash input in forest investment; however, it had no significant impact on farmers’ labor input. Inconsistent with earlier findings, they found that logging quota indicators had a significant negative impact on both farmers’ labor and cash investment in forestland [41]. The possible reason for this is that under the constraint of the logging quota system, although farmers will reduce labor input in timber forests, they will increase the afforestation area of moso bamboo and economic forests, as well as increase the input to forest planting and forest farming. Therefore, it shows that the impact of the logging quota on the overall labor input is not significant.

Third, to address the potential issue of endogeneity in our econometric analysis, we have used the average quantity of other households’ perceived tenure security in the same village as a reasonable instrumental variable. Our findings suggest that both the understanding and evaluation of the tenure security policy have significant positive impacts on farmers’ labor input in forestland management. Specifically, farmers’ satisfaction with the policy reflects their trust in state empowerment, and their attitude toward the reforms is an essential signal of trust and confidence [37]. A positive attitude towards policy will motivate farmers’ investment behavior in forestland [33,45].

From the above empirical research results, the following important policy implications can be drawn: Firstly, it ensures that all forestland certificates are issued to households, and boundary disputes in forestland are properly handled, in order to improve the actual tenure security. Secondly, the logging quota should be appropriately relaxed and farmers’ logging willingness should be included in the allocation process of logging indicators. At the same time, it is necessary to strengthen the supervision of the logging quota management department and reduce the “rent seeking” behavior caused by indicator allocation. Thirdly, local governments and forestry management departments should strengthen the promotion of policies related to forestland tenure reform, in order to enhance farmers’ perceived tenure security.

While this study has contributed to a better understanding of the relationship between tenure security and farmers’ forestland investment, there are still some limitations that require further investigation. First, this research has focused solely on collective forests in Jiangxi Province. Given the variations in economic development and human geography across China, our findings may not be generalizable to other regions. Therefore, it is important to conduct further research to examine the differences in forestland investment behavior among farmers in different regions. Second, forest management is a dynamic and evolving process. To fully comprehend the impact of tenure security on forest investment, it is necessary to study the dynamic processes over time. Future research should consider conducting longitudinal studies to examine the changes in forestland investment behavior over time and the effects of tenure security on these changes. In conclusion, while this study has advanced our understanding of the relationship between tenure security and forestland investment, additional research is needed to address the limitations and deepen our understanding of this important topic.