Abstract
In the context of the global challenge of sustainable resource management and rapid urbanization in China, the scarcity of forestland resources presents a critical barrier to improving forestland management efficiency (FME) for rural households. The present study aims to explore the constraints of the small-size agricultural sector by assessing the relationship between forestland management size (FMS) and efficiency. The survey data is applied from the 505 rural households in Jiangxi Province, China from 2017 to 2018. This research employs a nonlinear relationship model and threshold effect analysis to assess how FMS influences FME. Furthermore, the moderating effect analysis conducted to assess the role of socialized forestry services in shaping this relationship. The key findings revealed that there exists positive and favorable correlation between FMS and FME, indicating that larger FMS tend to improve the efficiency of rural households. Socialized forestry services are enhance FME by progressively upregulating the marginal effects, indicating their critical role in optimizing resource allocation and promoting efficient forestland management. The threshold effect observed the impact of FMS on FME underscores the nonlinear nature of this relationship. It suggested that optimal efficiency can achieved at specific FMS thresholds. This study provides valuable insights into the dynamics of forestland management in China, and highlighting the requirement for policies and interventions that promote appropriate scaling of forestland management while leveraging socialized forestry services to enhance efficiency and sustainability for socioeconomic development in years to come.
Keywords: Forestland management, Farm size, Socialized forestry services, Management efficiency, Policy implications, Rural development
Subject terms: Environmental social sciences, Psychology and behaviour, Socioeconomic scenarios, Sustainability
Introduction
The key difference between collective and private forestland lies in the ownership of property rights, the collective forestland is owned by the community rather than individuals, which complicates its management and operation. The new round of collective forest tenure reforms (CFTR) in China has gained remarkable achievements1. The stock of collective forestland has increased by nearly 3.9 billion cubic meters, representing an increase of nearly 85%, compared with CFTR before. The extent of state-owned forestland has expanded nearly 400 million mu. The CFTR aims to stimulate the enthusiasm of forest farmers by redistributing and redefining the rights to use and operate forestland, thereby enhancing the utilization efficiency and ecological benefits of forest resources. However, “Allocation of forestland to households” policy, collective forestland is managed by family contracts, resulting in problems, i.e., forestland fragmentation, forestland tenure disputes and low FME. These problems make it difficult for farmers to break through the status of low-level and small-area operations2.
Improving the FME of rural households has emerged as a complex issue. In order to address the difficulties of smallholder economies, expanding the scale of forestland management to appropriate size and developing economies of scale. The Central Committee of the Communist Party of China (CPC) and the State Council has announced the ‘Deepening Reform Program of Collective Forest Tenure Reforms’, which emphasizes the necessity of developing a reasonable size. Forestland management is transitioning towards collectivization and reasonable size under the direction of central policy3. The farmers face challenges due to the inherent difficulties of operating a small size during the transformation process, including the absence of advanced technology4,5 and the substantial expense of labor6. Outsourced machinery service-based socialized service can substitute expensive labors’ with capital input by sharing machinery, hence facilitating the adoption of advance technologies7,8. It can effectively address the inherent inefficiency problems associated with the current low-level and small-area management by forest farmers.
Considering the size, extensive studies have been demonstrated on the correlation between land size and efficiency. As early as 1985, Feder proposes that the absence of performance monitoring limits the correlation between land size and efficiency9. With the continuous increase of world population, the availability of land is becoming restricted. Consequently, there is a growing interest among experts to study the relationship between the size of land and efficiency. Some scholars believe that there is a positive association between land size and efficiency. The small farms in Asia’s smallholder economy are more productive than large farms10. The research data analysis in China, it noticed that the mechanization is a major contributing factor in the positive correlation between land size and efficiency11. This positive relationship is influenced by the capital investment of farmers12. Some research scholars put forward the contrasting perspective that there is a negative relationship between land size and efficiency.
Ghose13 suggested that the advantage of small-size output than large-size production accounts for the negative association between land size and efficiency. Other researchers demonstrated that this negative correlation is caused by measurement errors and diminishes as time period14,15. Currently, there is no evident correlation between land size and FME. Some scholars observed that there is no clear optimal and appropriate size for different national regions and topographical conditions. It found that the direct relationship between land size and efficiency linked to socioeconomic development16. In order to establish policies that can effectively modify the management size in various regions and focusing on the reasonable size of land management17–19. Compared to the size and efficiency of agriculture management studied by research scholars, the size of forestry management is often larger, while the investment return cycle of forestry management is longer, and input–output ratio may not change significantly in short duration. The impact of size of forestry management on efficiency can supplement the existing research gap.
In-view of socialized forestry services, nowadays research scholars have assessing the prevailing issues and impacts associated with it. The current states of the socialized forestry services have some difficulties, such as outdated service equipments, insufficient people, and poor service quality20,21. The impact of socialized services on agriculture management can be observed through four dimensions, i.e., socialized services can expand land size and assist land size management22,23. The provision of socialized services can encourage the adoption of enviro-friendly production behaviors in rural households8,24. The socialized services had beneficial effects on farmers’ utilization of organic fertilizer25. However, these socialized services can counteract the adverse effects of farmers’ risk perception on their behavior in the use of organic fertilizer. Socialized services have the potential to downregulate the use of fertilizer by effectively distributing resources and promoting scientific and technological advancements26. It contributes to the enhancement of technical efficiency, and eco-efficiency. Moreover, this positive relationship is a spatial correlation27–29.
The existing literature primarily emphasizes the importance of scaling up operations to enhance FME and reduce costs. Given that the socialized forestry services can mitigate the limitations associated with expanding forestland, reduce the expenses of implementing technology and machinery in farmers’ management, and expedite the achievement of rural mechanization. This study aims to incorporate socialized forestry services as a moderating variable to supplement the research in this direction. At present, the current academic community has yet to establish a uniform method for measuring FME. Currently, most studies only consider single-factor productivity of individual crops, like rice and maize.
Therefore, this article applied total factor productivity to measure the FME of rural households. It also utilized the SFA and DEA applications to conduct a comparative analysis of FME, thereby enhancing the accuracy and comprehensiveness of the FME of rural households. It analyzed the effect of FMS on farmers’ FME and the contribution of socialized forestry services in moderating effects. The analysis was based on CFTR data from Jiangxi Province, China (2016–2018). These findings aim to provide theoretical and practical knowledge for accelerating the transformation of the smallholder economy and improving the FME of rural households in China.
However, this study proposes measuring the FME of farmers through total factor productivity, which provides more comprehensive and reflective assessment of its changes. In the point of research view, this study investigated how FMS affects the FME of farmers. Additionally, the study included an analysis of the role of socialized forestry services in order to determine the moderating effects. This approach offers a advance perspective for future research. In terms of heterogeneity analysis, conducted a complete assessment of heterogeneity into account the influence of gender variations and household types on the FME of rural households. To analyze the varying behavior of distinct groups, these parameters were employed as criteria for grouping in heterogeneity analysis. In terms of data selection, to enhance the robustness of the findings, dual data from the identical sample were select throughout the process of data selection.
Impact of forestland scale on rural households’ management efficiency
According to the theory of economic scale, average expenses will go down and profits rise as an organization’s scale grows30. The rural households can enhance the size of their land holdings by forming cooperatives or leasing land. The benefits from economic scale, resulting in lower technology and productivity costs, and ultimately improving FME31–33. In forestry management, within a specific scale, as the FMS increases, fixed assets, technology and management resources can be fully utilized, and the costs can be reduced through centralized procurement, unified planning and standardized management34, thus achieving economic scale.
However, the FMS exceeds a certain limit, the complexity and difficulty of management increases significantly. Farmers or managers are challenged by a lack of specialized skills, insufficient labor and lack of time to monitor the forestland, leading to reduce the management efficiency35. Large-scale forestland management will face more difficulty36. FMS increases the complexity of managing forest ecosystems, including biodiversity conservation, pest and disease management, and soil profile. In turn, ordinary foresters may face management constraints, i.e., lack of specialized skills, labor shortages, and challenges in monitoring and maintaining forestlands, which can negatively affect management efficiency. Further, the adoption of advanced technologies and management practices may become more difficult as management size expand. Optimum implementation costs and the challenge of financing large-scale operations become barriers to maintaining high management efficiency.
Combined with the analytical analysis, it is reasonable to infer that the relationship between forestry management services and efficiency shows an inverted “U”-shaped relationship37–39. It indicates that before reaching specific scale, management efficiency enhances with the expansion of management scale. However, after exceeding this scale, management efficiency begins to decline. Small-scale forestland management may inefficient due to limited resources and technology. In the upward phase of the inverted U-shaped curve, resources are allocated and utilized more rationally, contributing to efficiency. However, in the downward phase, as the scale of management increases further, resource allocation and utilization may become challenges, leading to wasted resources and reduced efficiency. It presents hypothesis 1, as stated-
Hypothesis 1
The relationship between FMS and FME as inverted U-shaped pattern.
The moderating role of socialized forestry services
In the China’s smallholder economy, the expansion of FMS will help to form economic scale, which reduces the management costs of rural households40. However, with the transfer of labors’, the older individuals and women increasingly dominate the current structures of rural labors’. This demographic trend poses challenges in terms of paying the expensive implementation of advance technologies41. At the same time, farmers also face the limitations of restricted accessibility42, which hinders the smallholder economy from achieving transformation. In recent years, there have been rapid development in socialized forestry services. It service system provides professional knowledge, technology and support to rural households engaged in forestry activities through multiple channels and platforms43. As a new service model, socialized forestry services improve efficiency by providing professional knowledge, technical support, market information and financial resources, and by reallocating resources to mitigate the limitations of forestry management services22.
The specific approaches of the implementation of socialized forestry services has the potential to mitigate the inefficiency of labor organization and substitute human labor with automated machinery. At present, an increasing number of young rural households are engaging in labor migration for off-farm employment, resulting in a decline in the quantity and quality of rural labor force43. Socialized forestry services can offer mechanization as a substitute for inefficient labor, enhancing FME. As a result, farmers usually prefer to outsource labor-intensive stages of production44.
Moreover, forestry socialization services can mitigate technological constraints and enhance the resilience of foresters to risks. Socialized services provides rural households with access to updated forestry knowledge and technology, enabling them to adopt more efficient and sustainable management practices. Agricultural production marked by high unpredictable hazards and substantial expenses45, which inhibit farmers’ enthusiasm to expand management land. The machinery and technology brought by forestry socialization services can partially reduce risks, effectively mitigate the limitations of agricultural production, promote efficient production and improve the economic performance of farmers’ households. It can enhance yield, reduce costs and improve resource utilization efficiency. It helps to directly improve the risk management of foresters and mitigate the effects of natural disasters, market fluctuations and other uncertainties, thus increasing the stability and sustainability of forestry activities.
The socialized forestry services can reduce the obstacles in the process of FMS to improve the FME of rural households, the following hypotheses 2 are proposed:
Hypothesis 2
Socialized forestry services encourage FMS contribution to improving rural households FME.
Straightforward positive or negative interactions, the FMS shows an inverted U-shaped relationship with FME. While increasing FMS can reduce expenses, the cost-saving effects are gradually weakened23. Assess the differentiation of the effects of different sizes on FME and analyze the optimal size. It is necessary to explore whether or not exists a threshold value for FMS, so propose hypothesis 3:
Hypothesis 3
The role of FMS on FME exhibits significant nonlinear threshold characterization.
Material and methods
Data source
In-view of Jiangxi Province’s economic development level and the distribution of forest resources, the research group applied stratified random sampling to select sample counties. Ten representative sample counties (cities) were selected in the northern, central, and southern regions of Jiangxi Province, China and the combination of typical and random sampling was adopted, with 2 townships in each sampled county, for each township, a total of 2–3 sample villages were chosen. Within each sample village, 10 farm households were selected at random, with a total of 505 samples from farm households surveyed through questionnaires and case interviews. The research team conducted two surveys in 2018 and 2019 to investigate the annually situation of the sample area in 2017 and 2018, respectively. After excluding invalid data, a total of 1008 data were obtained for 2017 and 2018, with a combined data validity rate of 99.4%.
The survey data provided the following information, such as the size and type of forestland owned by farmers, farmers’ input of capital and labor in the process of forestry management, income level and structure of rural households, demand and adoption of socialized forestry services by rural households, individual characteristics and family characteristics of rural households, and the characteristics of village, where the rural households are located. In order to effectively reduce the challenges of hypothetical bias, the research team adopted different strategies to minimize hypothetical bias. First, before the formal survey initiate, the research team conducted a pre-survey in the target area and distributed 50 questionnaires. In-depth conversations with farmers, the content of the questionnaires were designed and iterated to ensure that the questions clearly and accurately worded to avoid hypothetical scenarios, thus improving the authenticity of the data collected. During the data collection process, the project team implemented a strict “three-tier review” system, i.e., initial review by the investigator, review by the project leader, and final review by the data analyst. This approach ensured the accuracy and completeness of the data. If any information was missing or out of order, the investigators promptly made a telephone call to the respondents and correct errors. In addition, in order to increase farmers’ participation and concentration during the survey, the research team provided appropriate incentives by giving each respondent in a cash allowance of 20 RMB each.
Variable selection
Dependent variable
The study measured FME by using total factor productivity according to Helfand and Taylor38. The specific indicators are shown in Table 1.The input factors are land, labor, and cash, and the three factors were combined to measure inputs. Income was used to measure outputs46. Farmers’ FME was measured by calculating their input–output ratios. The land input was quantified based on the extent of forestland managed by rural households. The labor input was determined by the number of individuals involved in forestry activities, as well as the monetary and temporal resources invested by the household in self-employed labor, employed labor, and forestland maintenance. The output was assessed by the forestry income generated.
Table 1.
Measurement details of input–output indicators of farm household business efficiency (all the following indicators are in logarithmic size).
| Type of indicator | Indicator | Definition of indicators | |
|---|---|---|---|
| Inputs | Land input | Forestland investment (acres) | Area of family forestland management (including family-contracted forestland and transferred forestland) |
| Labor input | Number of foresters (persons) | Total household labor minus number of permanent household migrant workers | |
| Household self-invested labor expenditures (in Yuan) | Opportunity cost of labor devoted to forestry management in households | ||
| Family self-referred labor time (days/person) | Labor time devoted to forestry in households | ||
| Expenditure on hired labor (in Yuan) | Total household expenditure on hired labor | ||
| Hired labor time (days/person) | Time spent working in the household‘s hired labor | ||
| Care time (days/person) | Time invested by households in current forestland management | ||
| Cash input | Total family forestry expenditure (In Yuan) | Total current household expenditures on forestry management | |
| Outputs | Total household income from forestry (In Yuan) | Gross income from current household forestry management | |
Independent variables
The FMS, which was calculated from the actual area operated by the farmers in the current year, serves as the present study independent variable. To make the data distribution more stable and prevent the extreme values from influencing the research findings, this study took a logarithmic treatment of the data. At the same time, to explore whether the effect of FMS on rural households’ FME is linear or nonlinear, this paper introduced the square term for forestland management size (SFMS) as another independent variable to be included in the regression model.
Mediating variables
Currently, there are two main ways the measure socialized services. Studies that use socialized services as core independent variables or core dependent variables mostly use the indicator construction method, such as socialized services are divided into different indicators according to the period function and assessed by the entropy value approaches47. Studies that use socialized services as mediating or moderating variables mostly use dichotomous variable accounting48. This study focused and the data’s accessibility led to the introduction of dichotomous variable to measure socialized forestry services, i.e., farmers received socialized forestry services = 1 and not received socialized forestry services = 0.
Control variables
To ensure the stability of the regression results and minimize the impact of omitted factors, it incorporated a set of control variables that could potentially influence the income of rural households. The study incorporated the individual characteristics of gender, age, and level of education. At the household level, the efficiency of different forestry business models varies to some extent. Therefore, this study introduced the management mode as a control variable. In-view of village characteristics, considering that the labor transfer situation will affects the human capital investment in forestry, labor transfer was included as a control variable.
Descriptive statistics
The average FME of rural households was 0.447, with a modest standard deviation, suggesting that the overall FME of rural households is often poor. The mean value of socialized forestry service was 0.218 and the standard deviation was small, which means that most of the farmer samples have not received forestry socialization service (Table 2). At the level of personal characteristics, the mean value of the gender dummy variable of the sample farmers was 0.943, indicating that males account for larger proportion of the sample data, which certain extent reflects that the current labor force in forestry operation is still dominated by males. However, at the same time, the mean value of the age of the surveyed farmers was 56.414, indicating that the current forestry operators shows an aging trend. The mean value of the dummy variable for education was 1.892, and the standard deviation was small, i.e., the education level of forest farmers was mainly in elementary school and junior high school, indicating the prevailing trend that the majority of contemporary forest farmers possess relatively low degree of formal education.
Table 2.
The list of variable description.
| Variables | Variable symbol | Variable definition | Mean | SD |
|---|---|---|---|---|
| Dependent variables | ||||
| Forestland management efficiency | FME | Total factor productivity | 0.447 | 0.258 |
| Independent variable | ||||
| Forestland management size | FMS | Logarithmic treatment of actual acreage operated during the year (acres) | 3.969 | 1.297 |
| Square term for forestland management size | SFMS | Logarithm of the square term of the actual area operated during the year (acres) | 7.938 | 2.594 |
| Commercial forest | CF | Logarithmic treatment of commercial forest area (acres) | 3.732 | 1.338 |
| Square term for commercial forest | SCF | Logarithmic treatment of commercial forest area square term (acres) | 7.463 | 2.677 |
| Mediating variable | ||||
| Socialized forestry service | FSS | Ever received socialized services? (yes = 1; no = 0) | 0.218 | 0.413 |
| Control variables | ||||
| Personal characteristics | ||||
| Gender | gen | Male = 1; Female = 0 | 0.943 | 0.231 |
| Age | old | Age of surveyed farmers | 56.414 | 10.253 |
| Education level | edu | Elementary school and below = 1; middle school = 2; middle or high school = 3; college or bachelor‘s degree or higher = 4 | 1.892 | 0.826 |
| Are engaged in forestry-related activities | FC | Yes = 1; No = 2 | 0.726 | 0.448 |
| Family characteristics | ||||
| Business model | MO | single-family = 1, joint-family = 2, family forestry = 3, corporate forestry = 4, joint-stock cooperative = 5 | 1.483 | 1.159 |
| Is forestry a major source of income | FMI | Yes = 1, No = 0 | 0.090 | 0.287 |
| Village characteristics | ||||
| Labor transfer | LT | Percentage of village labor migration | 0.540 | 0.238 |
The dummy variables mean value about the farmer business model was 1.483, indicating that the investigated farmers mainly operate in single-family and joint-family business models, such as the degree of organization of forest farmers’ business is still low. The mean value of the dummy variable representing farmers who rely on forestry income as their primary source of income was 0.090. It indicates that the majority of farmers not depends on forestry income as their main source of earnings. This could be attributed to the unpredictability and long-term nature of money generated by forestry activities, and farmers will seek for more sources of income to ensure the basic needs of their own and their family’s life. The mean value of the proportion of labor transfer in the village was 0.540, indicating that the current labor transfer phenomenon is frequent, which explains to the certain extent the aging trend of forest farmers in the previous section.
To explore the overall trend of rural households’ FME in the context of different FMS, this study presents preliminary trend analysis of FMS and FME (Table 3). The grouping results about the FME of farmers in the table shows that the farmers own FMS less than 100 acres. The mean value of efficiency was 0.432, when FMS greater than 100 acres but less than 200 acres, the mean value of FME was 0.466, when FMS greater than 200 acres but less than 300 acres. The mean value of FME was 0.476, when the FMS is larger than 300 acres but smaller than 400 acres, the mean value of FME was 0.556, when the FMS is larger than 400 acres but smaller than 500 acres, the mean value of FME was 0.426, and when the FMS is larger than 500 acres, the mean value of FME was 0.518. As the size continues to enhance, the FME of the farmers are upregulating.
Table 3.
Results of descriptive statistics of rural households’ management efficiency.
| FMS (acres) | Mean | SD |
|---|---|---|
| 0, 100 | 0.432 | 0.248 |
| 100, 200 | 0.466 | 0.270 |
| 200, 300 | 0.476 | 0.267 |
| 300, 400 | 0.556 | 0.266 |
| 400, 500 | 0.426 | 0.296 |
500,
|
0.518 | 0.314 |
Econometric model
Assessing the impact of FMS on FME is essential, so the benchmark model of the FMS on FME effect is given in equation as:
 |
1 |
Equation 1, 
denotes the FME, i.e. total factor productivity of household t in year i; 
denotes the FMS of household t in year I, 
it is a set of control variables that influence the FME of farmer’s household t in year i. The constant term is 
, and the coefficients to be estimated represented by 
and 
, and 
is the random error term. Observing the distribution pattern of the values of the explanatory variables in this study, and adopts the OLS regression model for analysis.
To further confirm the existence of relationship characterized by “inverted U-shaped” curve relationship between FMS and FME, an estimation model of SFMS on FME is established in this study, as shown in Eq. 2.
 |
2 |
Add the square term to the regression model, i.e., the SFMS, 
, and carry out quadratic regression; assess the significance of the coefficient 
of the quadratic term in the regression model based on the regression findings evaluate the importance of quadratic term coefficient. If 
is negative and 
is positive, then the relationship follows an inverted U-shape, otherwise, it does not exist.
It considered the moderating effect of socialized forestry services on farmers’ FME, so it sets the product term of socialized forestry services and FMS to included in the model.
 |
3 |
 |
4 |
In Eq. 3, and 4, 
is the household, t receives socialized forestry services in year i, 
it is the matrix composed of control variables, 
, 
are constant terms, The coefficients 
and 
are to be estimated after the including the interaction term, the error term 
follows a conventional normal distribution, and the rest of the variables are defined as follows. The remaining variables are adjusted to ensure consistency with Eq. 1 and 2.
Results
Impact of forestland size on rural households’ management efficiency
The impact of FMS on FME is shown in Table 4. Models 1 and 2 demonstrated the effect of FMS on FME of the rural households. Models 3 and 4 demonstrated the effect of SFMS on FME of the rural households. As shown in model 1, the expansion of FMS was found significantly contribute to FME through 1% significance level. Model 2 confirms that inclusion of control variables revealed significant relationship between FMS and FME, with significance level of 5%. Models 3 and 4 showed that after incorporating the squared term of SFMS, the coefficients of both FMS and SFMS are insignificant, i.e., there is no inverted U-shape. The surveyed sample of this study consistently shows a positive correlation between the FMS and FME.
Table 4.
Regression results of FMS on the management efficiency.
| Variables | Dependent variables: Management efficiency of rural households | |||
|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | |
| FMS | 0.066*** (0.015) | 0.041** (0.015) | 0.070 (0.057) | 0.083 (0.056) |
| SFMS | – | – | − 0.001 (0.018) | − 0.013 (0.018) |
| gen | – | − 0.015 (0.031) | – | − 0.015 (0.031) |
| old | – | − 0.000 (0.000) | – | − 0.000 (0.000) |
| edu | – | 0.006 (0.011) | – | 0.006 (0.011) |
| FC | – | 0.041** (0.018) | – | 0.041** (0.018) |
| MO | – | 0.016** (0.008) | – | 0.016** (0.008) |
| FMI | – | 0.106*** (0.033) | – | 0.106*** (0.033) |
| LT | – | − 0.096*** (0.036) | – | − 0.096*** (0.036) |
| cons | 0.362*** (0.026) | 0.396*** (0.071) | 0.362*** (0.026) | 0.396*** (0.071) |
*, **, and *** indicate significant at 10, 5, and 1% levels, parenthesis values are regression standard errors (SE).
In addition, this study conducted regressions on control variables, and the results are shown in Model 2 and 4. Models 2 and 4 shows that the gender, age, and education level of rural households not have significant impact on FME. However, it involved in forestry activities exhibit superior FME, and this finding is statistically significant at 5% level. It was discovered that the level of household characteristics, the rural household’s FME increased with the degree of organization of its business model. This relationship was significant at 5% level of significance. The FME of the rural households was higher when forestry income was the main source of income, and result was significant at 1% level of significance. Farmers will enhance their profit margin by finding more applications to enhance the efficiency of management when forestry is their primary source of income. Farm families’ FME decreased with increasing rural labors’ force transfer proportion at the village characteristics level. This relationship was analyzed at 1% significance level. In the countryside, ‘hollowing out’ will happen when the percentage of labor force transfer from rural areas rises49, and the shortage of labor will increase the possibility of abandonment, resulting in the decline of FME50,51. On the other hand, the transfer of rural labor is mostly for young adults, and the aging and feminization of rural surplus labor is increasing, which leads to relatively drops in labor quality and decrease in FME.
Mechanism analysis: socialized forestry services
The moderating influence of socialized forestry services is shown in Table 5. Model 5 reveals that socialized forestry services increase the FME of rural households and this result passes the 1% significance level. Model 6 indicates that the coefficient of the FMS on FME effect was improved by adding forestry socialized forestry services compared with model 1, and this result was significant at 1% significance level. The regression coefficients of FMS on FME was shown to significantly greater by the 5% significance level compared to Model 2 and 7 introduced control variables. In comparison to Model 3, model 8 demonstrated an increase in the regression coefficient of FMS on FME at 1% level of significance. This was achieved by substituting the squared term of FMS for the independent variable. The added control variables in Model 9 allowed it to exceed the 5% significance level, indicating a considerable improvement than Model 4 in the regression coefficients of FMS on FME.
Table 5.
Analysis of the moderating effect of socialized forestry services.
| Variables | Dependent variable: management efficiency of rural households | ||||
|---|---|---|---|---|---|
| Model 5 | Model 6 | Model 7 | Model 8 | Model 9 | |
| FMS | – | 0.020*** (0.006) | 0.029*** (0.007) | – | – |
| SFMS | – | – | – | 0.010*** (0.003) | 0.014*** (0.003) |
| FSS | 0.038*** (0.019) | 0.026 (0.019) | 0.178*** (0.064) | 0.026 (0.019) | 0.178*** (0.064) |
| fss#c.lnfms | – | – | 0.036*** (0.014) | – | – |
| fss#c.lnsfms | – | – | – | – | 0.018*** (0.007) |
| Control | – | – | Yes | – | Yes |
| cons | 0.438*** (0.009) | 0.359*** (0.026) | 0.325*** (0.029) | 0.359***(0.026) | 0.325*** (0.029) |
*, **, and *** indicate significant at the 10, 5, and 1% levels. Values in parentheses are regression SE.
The regression results showed that the coefficient value of the FMS was significantly positive. The coefficient value of forestry socialized services was significantly positive, and the coefficient value of the cross-multiplier term was significantly negative, which suggests that the marginal effect of the FMS on the FME is affected by forestry socialized services. Next, visualized the coefficient of the average marginal effect of FMS on the FME of rural households when they receive socialized forestry services or not. Figure 1 shows the average marginal effect of the FMS, and it can be seen that the adjustment coefficient of the FMS is 0.03 when farm households do not receive forestry socialized services, and the adjustment coefficient of the FMS is − 0.01 when they do not receive forestry socialized services. As seen in Fig. 1, with the increase of the FMS, the adjustment coefficient of the FMS on the FME by households that do not receive forestry socialized services is marginal effect gradually reduces, while the marginal effect of households receiving forestry social services on FME gradually increases. Therefore, the marginal effect steadily rises and socialized forestry services have beneficial moderating role in the process of FME by FMS.
Fig. 1.
Plot of average marginal effect coefficients.
Robustness test
Substitution of independent variables
Three robustness testing approaches were used in this work to assure the validity of the aforementioned conclusions. The amount of commercial forest was logarithmized to replace the measurement of the core independent variable, which was FMS in the current year. The regression results are presented in Table 6. Significance level (1%) test is tested by Model 10, demonstrating the beneficial relationship between the size of the commercial forest and the FME of rural households. The results of Model 11 with control variables indicated that there is a significant relationship, significant at the 5% level of significance, between the extent of commercial forests and the FME of rural households. By testing at the 1% level of significance, Model 12 demonstrated positive association between the logarithmic treatment of the squared measure of commercial forest size and rural household FME. Model 13 confirms that the squared term of commercial forest size was still positively related to rural households’ FME after adding control variables and this result was significant at 5% level of significance. Robustness is demonstrated by the conclusion, which in line with the previous regression results.
Table 6.
Regression results of commercial forest size on management efficiency.
| Variables | Dependent variable: Management efficiency of rural households | |||
|---|---|---|---|---|
| Model 10 | Model 11 | Model 12 | Model 13 | |
| CF | 0.025*** (0.007) | 0.016** (0.007) | – | – |
| SCF | – | – | 0.012*** (0.003) | 0.008** (0.003) |
| Control | – | Yes | – | Yes |
| cons | 0.356*** (0.026) | 0.471*** (0.077) | 0.356*** (0.026) | 0.471*** (0.077) |
*, **, and *** indicate significant at the 10, 5, and 1% levels. Parentheses values are regression SE.
Changing the independent variable measure
Subsequently, this paper replaces the measurement of the explanatory variables and continues to measure the FME by using the SFA model and regressing the FMS and SFMS. Table 7 shows the regression analysis. When testing at the 1% level of significance, Model 14 suggests FMS and FME are positively correlated in the absence of control variables. Even with the inclusion of control variables, Model 15 testifies at the 1% level of significance that there is still positive association between FMS and FME. Without the control variables, Model 16 reveals positive association between the squared term of FMS and FME. This result passes the test at the 1% level of significance. Model 17 shows that there was still positive relationship between FMS on FME after considering the control variables and this result was significant at 1% significance level. The conclusion after replacing the way efficiency is measured consistent with the above and confirms the robustness of the research findings.
Table 7.
Robustness tests: replacing the dependent variable measures.
| Variables | Dependent variable: management efficiency of rural households | |||
|---|---|---|---|---|
| Model 14 | Model 15 | Model 16 | Model 17 | |
| FMS | 0.021*** (0.006) | 0.013*** (0.006) | – | – |
| SFMS | – | – | 0.010*** (0.003) | 0.006*** (0.003) |
| Control | NO | YES | NO | YES |
| cons | 0.362*** (0.025) | 0.392*** (0.071) | 0.362*** (0.025) | 0.392*** (0.071) |
*, **, and *** indicate significant at the 10, 5, and 1 levels. Parentheses values are regression SE.
Replacement model
The present article replaces the model with the Tobit model, and Table 8 shows the regression results. Model 18 demonstrates that there was a positive correlation between the FMS on FME of the farmers and results verified at 1% significance level. After adding control variables in model 19, the results indicated that the expansion of the FMS led to an increase in the FME of rural households, and this result was significant at 5%. Using 1% significance level, Model 20 discovered positive relationship between the logarithmic treatment of the squared component of FMS and rural household FME. After controlling for other variables, Model 21 illustrates that the squared term of FMS is still positively correlated with FME, passing the significance test at 5% level.
Table 8.
Robustness tests: replacing the model.
| Variables | Dependent variable: Management efficiency of rural households | |||
|---|---|---|---|---|
| Model 18 | Model 19 | Model 20 | Model 21 | |
| FMS | 0.021*** (0.006) | 0.014** (0.006) | – | – |
| SFMS | – | – | 0.011*** (0.003) | 0.007** (0.003) |
| control | – | Yes | – | Yes |
| cons | 0.362*** (0.026) | 0.396*** (0.072) | 0.362*** (0.026) | 0.396*** (0.072) |
*, **, and *** indicate significant at the 10, 5, and 1% levels. Parentheses values are regression SE.
Analysis of heterogeneity
There are specific differences in individual and household characteristics among the sample households, which affects the FME of farm households to different degrees. In order to explore such individual differences, it is necessary to further refine the analysis of the sampling data.
Gender difference
Gender will affect the quality of the forestry labor force, which affects FME by affecting labor inputs. Gender serves as the foundation for group regression in the present study (Table 9). Model 22 illustrates that among the male sample, farmers’ FME increased with FMS, and this result was significant at 5% level. After substituting the squared term of FMS for the independent variable, Model 24 for the male sample still reveals positive correlation between FMS and FME, passing the 5% significance level test. In the female sample, Models 23 and 25 lack significant correlation between FMS and management efficiency.
Table 9.
Heterogeneity analysis based on gender.
| Model 22 | Model 23 | Model 24 | Model 25 | |
|---|---|---|---|---|
| Variables | Male | Female | Male | Female |
| FMS | 0.013* (0.007) | 0.004 (0.026) | – | – |
| SFMS | – | – | 0.007* (0.003) | 0.002 (0.013) |
| Control | yes | yes | yes | yes |
| cons | 0.346*** (0.066) | 0.945*** (0.321) | 0.346*** (0.066) | 0.945*** (0.321) |
*, **, and *** indicate significant at the 10, 5, and 1% levels. Parentheses values are regression SE.
Type of rural households
Income structure will affects the enthusiasm of farmers to engage in forestry labor, which in turn affects FME. In this study, income was mainly based on the forestry and used as the basis for group regression, and the questionnaire result was yes, then the dummy variable FMI was introduced and assigned a value of 1, and the questionnaire result was no, then the FMI was assigned a value of 0 (Table 10). Model 26 shows that there was no significant relationship between FMS and FME in rural households not dominated by forestry income. Model 27 observed a substantial positive correlation between FMS and FME for rural households who primarily rely on income from forestry. This correlation was statistically significant at 1% level of significance. Model 28 shows that after replacing the independent variable with SFMS, and SFMS still had no significant effect on FME among rural households not dominated by forestry income. Model 29 provides evidence, at significance level of 1%, that SFMS had positive relationship with the FME of rural households who rely on income from forestry. Therefore, this study finds the relationship between FMS and FME significant only among households that are dominated by forestry income.
Table 10.
Heterogeneity analysis based on type of rural households.
| Variables | Model 26 | Model 27 | Model 28 | Model 29 |
|---|---|---|---|---|
| Variables | FMI = 0 | FMI = 1 | FMI = 0 | FMI = 1 |
| FMS | − 0.043 (0.026) | 0.020*** (0.007) | – | – |
| SFMS | – | – | − 0.022 (0.013) | 0.010*** (0.003) |
| Control | Yes | Yes | Yes | Yes |
| cons | 0.852** (0.337) | 0.371*** (0.066) | 0.852** (0.337) | 0.371*** (0.066) |
*, **, and *** indicate significant at the 10, 5, and 1% levels. Parentheses values are regression SE.
Analysis of threshold effects
Model setting
The impact of FMS on the growth of FME of rural households does not necessarily shows a simple linear relationship and may show a non-linear relationship with the change of FMS. To further explore whether it has different mechanisms of action on FME growth in the context of different FMS, this study conducted an empirical investigation employing the threshold model.
Assuming the panel dataset is 
, the single-threshold model can be expressed as:
 |
5 |
where, i denotes the sample rural households’ code, 
and 
are the explanatory and interpretive variables, 
and 
are the threshold variable and the threshold to be estimated, 
represents the indicator function, whose value is taken to be 1 or 0 according to the truth or falsity of the expression in the parentheses,
is the individual fixed effect, 
is the random perturbation term, 
and 
are the variable coefficients, 
, indicating the presence of threshold effects.
Unlike the exogenously given grouping criteria when using the grouping test method to examine the nonlinear effects between variables, in the threshold model, the threshold value 
is unknown and completely endogenously determined by the data characteristics of the sample. The grid search method is usually used to determine the optimal threshold value 
, when 
takes the value of 
, the threshold model of the residual sum of squares S(
) is the smallest, i.e., 
= argminS(
). After the optimal threshold is estimated, the threshold effect’s significance and the threshold estimate’s equality with its true value are further examined.
The above model assumes that only a single threshold exists, but in real problems, there may be two or more thresholds. Taking the existence of two thresholds as an example, the double threshold model is set as follows:
 |
6 |
In Eq. 2, 
, and other codes have the same meaning as Eq. 1. For the model of Eq. 2, the double threshold estimation is performed with reference to the optimization search method proposed by Hansen52. The assuming that there is a single threshold in the model, after searching for the first threshold value 
according to the single threshold model, fix 
, continue to search for the second threshold value 
, and if 
exists, it is necessary to fix 
; and then, search again for the first threshold value. Repeat the above process, and finally get the consistent estimator after the optimization of the two threshold values.
Threshold effect test
Before determining the specific form of the threshold model, the estimation of the threshold value is carried out first, and Table 11 shows the specific results. Figure 2 shows the F-value test of the threshold effect. Based on the information presented in the table. It is evident that the threshold estimation of the FMS was 22.899, p value was 0.000, LM-test for no threshold was 16.713. These estimations were located in the corresponding 95% confidence intervals, which means that our threshold estimation is in line with the actual situation.
Table 11.
Threshold effect test results.
| Threshold variables | Model | Threshold estimate | p value | LM-test for no threshold |
|---|---|---|---|---|
| FMS | Single threshold estimation | 22.899*** | 0.001 | 16.713 |
p values are obtained after 300 simulations using the “Bootstrap” method.
Fig. 2.
Threshold effect F-test plot.
Estimation of threshold parameters
After the threshold effect test, the results of the threshold model estimate with FMS as the threshold variable are presented in Table 12. The regression findings of Regime 1 and Regime 2 regarding the threshold value differed (Table 12). In Regime 1, when q ≤ 412 and the FMS was below the threshold value of 22.899, the coefficient of the impact of the FMS on FME was 0.446 and passed the significance test at 1% level. When agricultural science and technology innovation was between the first and second threshold values, the coefficient of the impact of the FMS on FME remained significantly positive (0.582) and the coefficient value increased less. From this, we can see that the FMS enhances, its impact on the FME of rural households is increasing, which may be mainly due to the following two aspects, such as, when the FMS is lower (less than the first threshold), the impact on the FME of rural households is smaller, in the stage of initial increase of the management model, for example, the production technology, labor, or socialized forestry services and other means of production, these factors of production may bring some efficiency improvement, but the allocation of the factors of production still exists a certain process of matching each other, so the FME has not reached a great value for the time being. The FMS is between the first and second thresholds, the impact on FME of rural households gradually increases, probably because rural households have already gained a series of management experience from the initial increase in FMS, and further increase in land production may lead to further marginal growth and thus arrive at the optimal and appropriate management size. In addition, if the FMS increases further, farmers may already face other production challenges, such as education, labor, and other limitations, which make it difficult to bring about the same magnitude of growth in the value of FME even if there is a further increase in the FMS. Overall, the role of forestland management on FME exhibits significant nonlinear threshold property.
Table 12.
Panel threshold model parameter estimates.
| Variables | Regime1 q < = 412 | Regime2 q > 412 |
|---|---|---|
| gen | − 0.013 | 0.137 |
| (0.032) | (0.111) | |
| old | − 0.0004 | 0.009 |
| (0.0008) | (0.005) | |
| edu | 0.009 | − 0.029 |
| (0.011) | (0.045) | |
| FC | 0.041 | 0.071 |
| (0.017) | (0.103) | |
| MO | 0.015 | 0.302 |
| (0.007) | (0.103) | |
| FMI | 0.155 | − 0.068 |
| (0.033) | (0.095) | |
| LT | − 0.093 | − 0.249 |
| (0.036) | (0.188) | |
| FMStI (FMStI < 22.899) | 0.446*** | − 0.352 |
| (0.069) | (0.479) | |
| FMStI (FMStI ≥ 22.899) | 0.582*** | − 1.292 |
| (0 0.310) | (0.587) | |
| Sum of Squared Errors | 58.498 | 3.227 |
| R-squared | 0.063 | 0.084 |
*, **, and *** indicate significant at the 10, 5, and 1% levels. Parentheses values are regression SE.
Discussion
The assessment in the relationship between forest management size and forest management efficiency yielded unexpected results and highlighted specific challenges. Contrary to the anticipated inverted U-shaped pattern, the empirical analysis revealed a consistent positive correlation between FMS and FME. This finding challenges the theoretical hypothesis, suggesting that the promotion effect of FMS on FME may diminish as the size expands due to diminishing marginal returns. The discrepancy between theory and practice underscores the complexity of China’s smallholder economy, where ideal scenarios of size expansion may not fully align with actual. Further exploration is required to ascertain the temporal dynamics of this relationship, particularly whether there exists a period of positive correlation before the onset of the inverted U-shaped pattern. Expanding the survey area and increasing sample size will be pivotal in verifying these conclusions, guiding future investigations.
While the utilization of 2017–2018 rural household data from Jiangxi Province, China ensures reliability through repeated sampling, it also introduces certain limitations. Geographical constraints restrict the generalize ability of conclusions beyond Jiangxi Province. Moreover, the temporal gap between data acquisition and the present day may result in lagged conclusions and overlook evolving trends. To address these challenges, future research will focus on acquiring longitudinal and nationwide data to capture long-term dynamics and mitigate geographical biases, thereby enhancing the robustness of findings.
Socialized forestry services as a moderating variable underscores their importance in the FMS-FME relationship. However, the study’s analysis only scratches the surface of this complex phenomenon. While it acknowledge the impact of various forms of socialized forestry services on FME, it fails to conduct a comprehensive comparative investigation into their differential effects.
Conclusion and future research directions
The goal of this study is to assess the FMS and FME are correlated as an inverted U-shaped pattern. The analysis based on the sample data from Jiangxi Province during 2017–2018. Furthermore, the study intends to investigate how socialized forestry services may influence this relationship as moderating factor. The analysis of multiple linear regression and the robustness test revealed a positive correlation between FMS and FME. This is different from inverted U-shape hypothesis, which may due to the current basic national situation of the smallholder economy in China. In order to better understand, we incorporated socialized forestry services into the model. By analyzing the moderating effect, we discovered the impact of FMS on enhancing FME could be amplified by these services, with the effect becoming more pronounced over time. In the process of realizing economic of size, FMS faces more and more urgent challenges, and socialized forestry services can effectively mitigate these issues, thus continuously promoting the role of FMS in improving FME. However, the regression results for FMS on the FME of the existence of positive relationship, but does not understand whether this effect is non-linear. According to the findings of the threshold regression analysis, the FMS has a threshold value of 22.899. As the FMS increases, there is a corresponding increase in its impact on the FME of rural households. It suggests a significant nonlinear relationship between FMS and FME.
It concluded that there is a non-linear positive association between FMS and FME. Therefore, it is necessary to make full use of the impact of FMS on FME and formulate policies to expand the size to form the economic size effect according to local conditions. We should also effectively play the role of socialized forestry services, publicize and encourage farmers to use socialized forestry services, and accelerate the transformation of China’s smallholder economy.
Future research will delve in-depth into the nuanced roles of different types of socialized forestry services, exploring their distinct impacts on FME. Additionally, the mechanism through which socialized forestry services influence FME, particularly regarding labor substitution or complementarily, warrants thorough investigation in subsequent studies. By addressing these research gaps, a more comprehensive understanding of the interplay between socialized forestry services, FMS, and FME can be attained.
Acknowledgements
The authors appreciate the editors and anonymous reviewers for their valuable comments and suggestions.
Author contributions
Conceptualization, L.H. and X.L.; methodology, L.H. L.L. and X.L.; software,, L.H.L.L. and validation, L.H, L.L. and X.L.; formal analysis, L.H and L.L.; investigation, L.H, X.L. and X.L; resources, X.L.; data curation, L.H. L.L., and writing—original draft preparation, L.H.L.L., and X.L.; writing—review and editing, L.H. and L.L.; visualization, L.H. and supervision, X.L. and L.H.; project administration and funding acquisition, X.L. All authors have read and agreed to the published version of the manuscript.
Funding
The work was Supported by the Fundamental Research Funds for the Central Universities of Zhongnan University of Economics and Law (Project Name: Research on the influence mechanism, realization path and policy optimization of moderate scale management with Chinese characteristics; Grant Number: 202411011).
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Competing interests
The authors declare no competing interests.
Ethical approval
This study was conducted in accordance with the ethical guidelines for social science research established by the Chinese Ministry of Science and Technology. All data were collected through voluntary participation, and verbally informed consent was obtained from all surveyed households. To protect participant privacy, personally identifiable information was removed during data entry, and only aggregated data were used for research analysis. The study protocol was reviewed and approved by the Ethics Review Committee of the Jiangxi Province Rural Development Research Center at Jiangxi Agricultural University.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Xie, Y., Gong, P., Han, X. & Wen, Y. The effect of collective forestland tenure reform in China: Does land parcelization reduce forest management intensity?. J. For. Econ.20(2), 126–140 (2014). [Google Scholar]
- 2.Xiao, H., Hou, F., Li, X. et al. The impact of the forestland‐tenure disputes on the new type of forestry operating entity’s long‐term investment: The intermediary effects on credit constraints. In Natural Resources Forum (Blackwell Publishing Ltd., 2024).
- 3.Gao, X. P., Wang, L. & Liao, W. M. Policy guidance, poverty level and farmers’ forestry scale operation behavior. For. Econ. Issues2, 147–154 (2020). [Google Scholar]
- 4.Kong, F. B. et al. Impact of rural labor force transfer on farmers’ demand for socialized forestry services—A survey based on the production link of 1407 farmers. Forestry Sci.6, 132–142 (2018). [Google Scholar]
- 5.Kong, F. B., Ruan, H. & Liao, W. M. Analysis of the impact of socialized forestry services provision on poor farmers’ forest land input and output. For. Econ. Issues2, 129–137 (2020). [Google Scholar]
- 6.Wan, Q., Qin, T. & Pan, H. X. Analysis of influencing factors on farmers’ participation in policy forest insurance—an empirical study based on questionnaire survey of farmers in Fujian. Southeast Acad.3, 62–74 (2012). [Google Scholar]
- 7.Anang, B. T. & Asante, B. O. Farm household access to agricultural services in northern Ghana. Heliyon6(11) (2020). [DOI] [PMC free article] [PubMed]
- 8.Qing, C. et al. Impact of outsourced machinery services on farmers’ green production behavior: Evidence from Chinese rice farmers. J. Environ. Manag.327, 116843 (2023). [DOI] [PubMed] [Google Scholar]
- 9.Feder, G. The relation between farm size and farm productivity: The role of family labor, supervision and credit constraints. J. Develop. Econ.18(2–3), 297–313 (1985). [Google Scholar]
- 10.Fan, S. & Chan-Kang, C. Is small beautiful? Farm size, productivity, and poverty in Asian agriculture. Agric. Econ.32, 135–146 (2025). [Google Scholar]
- 11.Wang, J. et al. Is small still beautiful? A comparative study of rice farm size and productivity in China and India. China Agricult. Econ. Rev.7(3), 484–509 (2015). [Google Scholar]
- 12.Sheng, Y. & Chancellor, W. Exploring the relationship between farm size and productivity: Evidence from the Australian grains industry. Food Policy84, 196–204 (2019). [Google Scholar]
- 13.Ghose, A. K. Farm size and land productivity in Indian agriculture: A reappraisal. J. Develop. Stud.16(1), 27–49 (1979). [Google Scholar]
- 14.Desiere, S. & Jolliffe, D. Land productivity and plot size: Is measurement error driving the inverse relationship?. J. Develop. Econ.130, 84–98 (2018). [Google Scholar]
- 15.Gautam, M. & Ahmed, M. Too small to be beautiful? The farm size and productivity relationship in Bangladesh. Food Policy84, 165–175 (2019). [Google Scholar]
- 16.Rada, N. E. & Fuglie, K. O. New perspectives on farm size and productivity. Food Policy84, 147–152 (2019). [Google Scholar]
- 17.Xu, Y. et al. Exploring a moderate operation scale in China’s grain production: A perspective on the costs of machinery services. Sustainability11(8), 2213 (2019). [Google Scholar]
- 18.Li, M., Zhao, W., Tian, C., Li, Y. et al. Moderate operation scales of agricultural land under the greenhouse and open-field production modes based on DEA model in mountainous areas of southwest China. Heliyon9(11) (2023). [DOI] [PMC free article] [PubMed]
- 19.Ma, L. et al. Analysis of efficiency differences and research on moderate operational scale of new agricultural business entities in Northeast China. Sustainability15(12), 9746 (2023). [Google Scholar]
- 20.Wu, S., Guo, Y. & Ji, G. Development of socialized forestry service system in multiple governance perspective and change of government role. World For. Res.29(2), 54–59 (2016). [Google Scholar]
- 21.Liu, X.B. & Dai, L.Y. Analysis of the current situation and countermeasures of socialized forestry services at the grassroots level--Taking Zhenping County, Ankang City, Shaanxi Province as an example. Farmers’ Counselor(21), 72 (2018).
- 22.Yang, S. & Zhang, F. The impact of agricultural machinery socialization services on the scale of land operation: Evidence from rural China. Agriculture13(8), 1594 (2023). [Google Scholar]
- 23.Zhou, Z. et al. Land transfer or trusteeship: Can agricultural production socialization services promote grain scale management?. Land12(4), 797 (2023). [Google Scholar]
- 24.Shi, F. et al. Agricultural socialized services to stimulate the green production behavior of smallholder farmers: The case of fertilization of rice production in south China. Front. Environ. Sci.11, 1169753 (2023). [Google Scholar]
- 25.Ren, Z. Effects of risk perception and agricultural socialized services on farmers’ organic fertilizer application behavior: Evidence from Shandong Province, China. Front. Public Health11, 1056678 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Gao, E. K., Zhu, J. J. & Zheng, J. Impact of agricultural socialization services on fertilizer reduction—a double test based on panel data from 31 provinces and regions across China. Chinese J. Ecol. Agricult.4, 632–642 (2023). [Google Scholar]
- 27.Wang, L., Gao, X., Yuan, R. & Luo, M. Impact and spatial effect of socialized services on agricultural eco-efficiency in china: Evidence from Jiangxi Province. Sustainability16(1), 360 (2023). [Google Scholar]
- 28.Cai, B. et al. Agricultural socialized services empowering smallholder rice producers to achieve high technical efficiency: Empirical evidence from southern China. Front. Sust. Food Syst.8, 1329872 (2024). [Google Scholar]
- 29.Chen, T., Rizwan, M. & Abbas, A. Exploring the role of agricultural services in production efficiency in Chinese agriculture: A case of the socialized agricultural service system. Land11(3), 347 (2022). [Google Scholar]
- 30.Stigler, G. J. The economies of scale. J. Law Econ.1, 54–71 (1958). [Google Scholar]
- 31.Fei, R., Lin, Z. & Chunga, J. How land transfer affects agricultural land use efficiency: Evidence from China’s agricultural sector. Land Use Policy103, 105300 (2021). [Google Scholar]
- 32.Deininger, K. Collective agricultural production: A solution for transition economies?. World Develop.23(8), 1317–1334 (1995). [Google Scholar]
- 33.Hall, B. F. & LeVeen, E. P. Farm size and economic efficiency: The case of California. Am. J. Agricult. Econ.60(4), 589–600 (1978). [Google Scholar]
- 34.Kaarakka, L. et al. Improved forest management as a natural climate solution: A review. Ecol. Sol. Evid.2(3), e12090 (2021). [Google Scholar]
- 35.Song, H., Jiang, H., Zhang, S. & Luan, J. Land circulation, scale operation, and agricultural carbon reduction efficiency: Evidence from China. Discrete Dyn. Nat. Soc.2021, 1–12 (2021). [Google Scholar]
- 36.Juutinen, A., Tolvanen, A. & Koskela, T. Forest owners’ future intentions for forest management. Forest Policy Econ.118, 102220 (2020). [Google Scholar]
- 37.Muyanga, M. & Jayne, T. S. Revisiting the farm size-productivity relationship based on a relatively wide range of farm sizes: Evidence from Kenya. Amer. J. Agricult. Econ.101(4), 1140–1163 (2019). [Google Scholar]
- 38.Helfand, S. M. & Taylor, M. P. The inverse relationship between farm size and productivity: Refocusing the debate. Food Policy99, 101977 (2021). [Google Scholar]
- 39.Sheng, Y., Ding, J. & Huang, J. The relationship between farm size and productivity in agriculture: Evidence from maize production in Northern China. Am. J. Agric. Econ.101(3), 790–806 (2019). [Google Scholar]
- 40.Chen, Y. F., Sun, W. L. & Xue, G. X. Literature comment and theoretical study on the optimum management scale of farmland. China Land Sci.29(5), 8–15 (2015). [Google Scholar]
- 41.Mottaleb, K. A., Krupnik, T. J. & Erenstein, O. Factors associated with small-scale agricultural machinery adoption in Bangladesh: Census findings. J. Rural Stud.46, 155–168 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Kumar, C. S., Turvey, C. G. & Kropp, J. D. The impact of credit constraints on farm households: Survey results from India and China. Appl. Econ. Perspect. Policy35(3), 508–527 (2013). [Google Scholar]
- 43.Liao, W. et al. Influence of risk perception and policy support on the deviation of rural households’ demands and adoption behavior of the forestry socialized service. Front. Environ. Sci.11, 1211310 (2023). [Google Scholar]
- 44.Song, L. Negative impacts and solution countermeasures of excessive rural labor force transfer. J. Econ. Res.23, 45–46 (2015). [Google Scholar]
- 45.Deng, X., Xu, D., Zeng, M. & Qi, Y. Does outsourcing affect agricultural productivity of farmer households? Evidence from China. China Agricult. Econ. Rev.12(4), 673–688 (2020). [Google Scholar]
- 46.Jin, M., Chen, N., Wang, S. & Cao, F. Does forestry industry integration promote total factor productivity of forestry industry? Evidence from China. J. Clean. Prod.415, 137767 (2023). [Google Scholar]
- 47.Barry, P. J. & Fraser, D. R. Risk management in primary agricultural production: Methods, distribution, rewards, and structural implications. Am. J. Agric. Econ.58(2), 286–295 (1976). [Google Scholar]
- 48.Wu, A., Elahi, E., Cao, F., Yusuf, M. et al. Sustainable grain production growth of farmland–A role of agricultural socialized services. Heliyon (2024). [DOI] [PMC free article] [PubMed]
- 49.Li, Y. et al. The impact of labor migration on chemical fertilizer use of wheat smallholders in China-mediation analysis of socialized service. J. Clean. Prod.394, 136366 (2023). [Google Scholar]
- 50.Wen, Q., Li, J., Ding, J. & Wang, J. Evolutionary process and mechanism of population hollowing out in rural villages in the farming-pastoral ecotone of Northern China: A case study of Yanchi County, Ningxia. Land Use Policy125, 106506 (2023). [Google Scholar]
- 51.Liu, Z. Rural population decline, cultivated land expansion, and the role of land transfers in the farming-pastoral ecotone: A case study of Taibus, China. Land11(2), 256 (2022). [Google Scholar]
- 52.Hansen, B. Testing for linearity. J. Econ. Surv.13(5), 551–576 (1999). [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.