How Can Forestry Carbon Sink Projects Increase Farmers’ Willingness to Produce Forestry Carbon Sequestration?

Abstract

The development of a forestry carbon sink project is an important way to achieve carbon neutrality and carbon reduction, and the collective forest carbon sink project is an important part of China’s forestry carbon sink project. As the main management entity of collective forests, whether farmers are willing to produce forestry carbon sinks is directly related to the implementation effect of the project. In this paper, a partial equilibrium model of farmers’ forestry production behavior was established based on production function and utility function, and the path to enhance farmers’ willingness to produce forestry carbon sink through forestry carbon sink projects was analyzed in combination with forest ecological management theory. In terms of empirical analysis, the PSM-DID econometric model was established based on the survey data of LY in Zhejiang Province, China, and the following conclusions were drawn: (1) With the receipt of revenues from forestry carbon sequestration projects and partial cost-sharing by the government, farmers’ participation in forestry carbon sink projects can save investment in forest land management. (2) The saved forestry production costs and forestry carbon sink project subsidies can make up for the loss of farmers’ timber income, so that the net income of forestry will not be significantly reduced. (3) The forestry production factors saved by farmers can be transferred to non-agricultural sectors and increase non-agricultural net income, so that the net income of rural households participating in forestry carbon sink projects will increase. The forestry carbon sink project can improve the utility level of farmers and increase the willingness of farmers to produce forestry carbon sinks by delivering income to farmers and saving forestry production factors. This study demonstrates that a well-designed forestry carbon sink compensation mechanism, combined with an optimized allocation of production factors, can effectively enhance farmers’ willingness to participate. This insight is also applicable to countries or regions that rely on small-scale forestry operations.

1. Introduction

1.1. Research Background

Global climate warming has become an increasingly severe issue, and forestry carbon sinks, as an important carbon reduction measure, play a crucial role in addressing climate change and promoting sustainable development. Based on prior assessment reports from the IPCC, forestry carbon sink projects represent a crucial strategy for mitigating carbon emissions due to their cost-effectiveness, economic viability, and numerous ancillary benefits (IPCC, 2023: Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change) To promote the development of forestry carbon sink projects, the Chinese government has not only actively participated in international forest carbon initiatives but also explored the establishment of a domestic voluntary emission reduction mechanism—namely, the China Certified Emission Reduction (CCER) system. This mechanism aims to develop a national carbon market and trading system to incentivize low-cost emissions reduction. However, due to issues such as a lack of transparency in project registration procedures and deficiencies in the verification system, China suspended the registration and issuance of new CCER projects at the end of 2017, effectively placing forestry carbon sink projects in a state of regulatory freeze. On 24 October 2023, the CCER project application channel was officially reopened, providing an important opportunity for the further development of China’s forestry carbon sink project (Data from the Central People’s Government of the People’s Republic of China: https://www.gov.cn/zhengce/zhengceku/202310/content_6912025.htm, accessed on 21 September 2024).

According to different pathways for enhancing forest carbon sinks, current international forestry carbon projects can be broadly categorized into three types: Afforestation, Reforestation, and Revegetation (ARR) projects; Improved Forest Management (IFM) projects; and Reducing Emissions from Deforestation and Forest Degradation (REDD⁺) projects. ARR projects aim to increase forest carbon stocks by converting non-forested land into forest through afforestation and reforestation. In contrast, both IFM and REDD⁺ projects are essentially forest management-based approaches that seek to improve carbon sequestration or reduce carbon losses by enhancing forest governance structures, reducing deforestation pressures, and promoting sustainable forest management through positive incentive mechanisms. Given the limited potential for further afforestation in China and the greater economic feasibility and scalability of IFM approaches, IFM is expected to become a primary pathway for future forestry carbon projects under China’s Certified Emission Reduction (CCER) system. Notably, approximately 60% of China’s forests are under collective ownership. Therefore, the development of IFM-based carbon sink projects in collective forests holds strategic significance for achieving China’s dual carbon goals and advancing rural revitalization. And whether the collective forest operators are willing to participate in the forestry carbon sink project mainly depends on the operating cost of forestry carbon sink and the positive externality internalization of carbon sink income. By the end of 2010, the contract-to-household rate of collective forest land in China was 88.6% (Data from the Central People’s Government of the People’s Republic of China net: https://www.gov.cn/lianbo/bumen/202408/content_6968739.htm, accessed on 21 September 2024), and by 2023, the transfer rate of collective forest rights in China was 10% (Data from the national bureau of forestry and grassland https://www.forestry.gov.cn/c/www/lcdt/528772.jhtml, accessed on 21 September 2024). In other words, farmers are the main operators of collective forest in China. Exploring how the forestry carbon sink project affects the income of farmers and the willingness of farmers to produce forestry carbon sink is of great significance to the long-term development of forestry carbon sink projects. From an economic rationality perspective, farmers’ willingness to participate in forestry carbon sink projects depends on whether their income is affected before and after participation, which is closely related to changes in their forest management practices. Therefore, this study examines how farmers’ shifts in production behavior after engaging in forestry carbon sink projects influence their income and utility, providing important insights into the long-term sustainability of such projects.

1.2. Development Background of the CCER Mechanism

To advance its dual carbon goals of peaking carbon emissions and achieving carbon neutrality, China began exploring the establishment of a domestic carbon trading market and a voluntary emission reduction mechanism—namely, the China Certified Emission Reduction (CCER) program—drawing upon the experiences of international carbon markets. In 2011, the National Development and Reform Commission (NDRC) issued the Notice on Launching Pilot Carbon Emissions Trading, initiating pilot carbon trading schemes in seven provinces and municipalities, including Beijing and Guangdong, with the traded product being voluntary greenhouse gas emission reductions. In 2012, the government released the Interim Measures for the Administration of Voluntary Greenhouse Gas Emission Reduction Transactions, which clarified that CCER projects could participate in carbon markets upon approval and verification. In September 2015, the Central Committee of the Communist Party of China and the State Council issued the Overall Plan for the Reform of the Ecological Civilization System, which called for the establishment of effective mechanisms to increase carbon sinks in forests, grasslands, wetlands, and oceans, thereby accelerating the development of carbon markets. Due to concerns over market standardization and regulatory capacity, China suspended approvals for new CCER projects in 2017. However, on 24 October 2023, the application channel for CCER projects was officially reopened, marking a renewed phase in the development of the national voluntary carbon offset market.

To address the challenges that led to the suspension of the China Certified Emission Reduction (CCER) program in 2017—such as insufficient transparency, inconsistent verification standards, lack of unified methodologies, and limited market participation—the Chinese government implemented a comprehensive reform of the CCER system prior to its relaunch in 2023. First, a clear and standardized “verification-registration-trading” process has been established, with a mandatory 20-day public consultation period for project disclosures. Second, a series of regulatory documents, including the Administrative Measures for Voluntary Greenhouse Gas Emission Reduction Trading (Trial), have been issued to provide legal and procedural guidance. The reforms introduce a dual third-party verification mechanism, requiring project developers to engage two independent and certified Validation and Verification Bodies (VVBs), thereby enhancing the credibility and impartiality of project assessments. Third, a unified national CCER registry and trading platform, operated by the Beijing Green Exchange, has been established to improve information transparency and standardize trading procedures. In addition, stricter requirements for methodology development and approval have been implemented to ensure methodological rigor and environmental integrity. Collectively, these reforms have addressed several of the institutional weaknesses of the previous CCER system and laid a more robust foundation for the development of a transparent, standardized, and credible voluntary carbon market in China.

The CCER program primarily complements China’s national carbon emissions trading scheme, which represents the mandatory carbon market. Forestry carbon sink products are generated in the form of projects, where project owners either manage the forests themselves or guide forest managers to implement scientifically based forest operations to produce carbon sequestration. These forestry carbon credits can then be purchased by enterprises that exceed their emissions quotas, allowing them to offset their surplus carbon dioxide emissions. In most cases, the project owner is also the project developer; however, in some instances, the project owner commissions a specialized carbon development company to carry out project development. As forestry carbon sink projects are implemented on state-owned or collectively owned forest land, the landowners may be state-owned forest farms, village collectives, or individual farmers. As of 2022, a total of 97 forestry carbon sink projects had been registered, including four main types: afforestation carbon sink projects, forest management carbon sink projects, Moso bamboo (Phyllostachys edulis) afforestation carbon sink projects, and Moso bamboo management carbon sink projects.

In summary, the demand for forestry carbon trading is largely shaped by carbon quota regulations and offset ratio rules. At present, China’s carbon control policies primarily target key industrial sectors, where carbon allowances remain relatively generous. Moreover, the proportion of emissions that can be offset by forestry carbon credits is restricted to no more than 5% of the total quota obligation. As a result, Chinese enterprises currently purchase forestry carbon credits mainly to enhance their green image and gain additional reputational value rather than to fulfill compliance obligations. This leads to limited participation by buyers and insufficient market competition, resulting in a dysfunctional price formation mechanism. Given the incomplete nature of the forestry carbon market, the pricing of forestry carbon sink products is largely determined by direct government intervention—such as setting price ranges or floor prices—rather than through market-based competition. This administrative pricing mechanism lacks effective price discovery functions and thus fails to provide sustained incentives or constraints for forestry operators to supply carbon sink products. To ensure the long-term stability of forestry carbon trading, the Chinese government often offers supplementary incentives to producers through mechanisms such as cost-sharing.

1.3. Literature Review

The essence of IFM (Improved Forest Management) projects in China’s collective forest areas lies in incentivizing farmers to shift from traditional forest management practices to more sustainable forest operations. Existing studies indicate that farmers’ willingness to adopt improved management approaches largely depends on three key considerations: the additional management costs required for changing their forestry practices, the compensation they may receive, and the expected economic returns [1,2,3,4]. Therefore, it is necessary to examine the existing literature to understand in which specific aspects farmers’ forest management practices have changed under IFM projects and how such behavioral adjustments further affect their management costs and income levels.

First, when farmers shift from traditional forest management to sustainable forest management, the specific changes typically include delaying the rotation period; shifting from clear-cutting to selective logging; reducing harvesting intensity to optimize forest stand structure [5,6]; and increasing investments in thinning, replanting, and other silvicultural practices to improve stand quality. Additionally, the adoption of higher-cost organic fertilizers specifically designed for carbon sink forests is required to enhance carbon sequestration efficiency [7,8,9]. Research on Moso bamboo forest carbon management projects in China has found that intensive management of Moso bamboo forests can lead to improved stand structure and increased productivity. Such practices not only enhance carbon benefits but also generate higher economic returns from Moso bamboo timber and shoots [10].

Changes in farmers’ management practices further influence operating costs and income levels. Reduced harvesting intensity helps lower logging and transportation costs, thereby significantly decreasing overall production expenses [9]. However, increased investments in silvicultural activities and higher fertilizer costs may raise forest management expenses. Additionally, adjustments to stand structure may lead to short-term losses in traditional forestry income [11]. These changes in forestry production inputs directly impact the allocation of farmers’ production factors, potentially shifting labor and capital to other sectors. According to utility maximization theory and household production theory, farmers make rational choices by balancing profit and utility in their production decisions [12,13]. Consequently, the implementation of forestry carbon sink projects can expand employment opportunities, encourage non-agricultural employment, and facilitate rural livelihood transitions [14,15].

In summary, forestry carbon sink projects influence farmers’ income by altering their allocation of production factors. However, the existing literature presents divergent conclusions regarding whether these projects actually increase farmers’ income. The World Development Report suggests that forestry carbon sink projects not only mitigate global warming but also increase farmers’ income and alleviate rural poverty (World Bank, 2010: World Development Report 2010: Development and Climate Change. The International Bank for Reconstruction and Development/The World Bank, Washington, DC, USA, https://openknowledge.worldbank.org/handle/10986/4387, accessed on 21 September 2024). Some studies support this view, arguing that these projects reshape farmers’ income structures, enhance overall earnings, and improve welfare. However, the income-enhancing effects of such projects may be quite limited for smallholder farmers [16,17,18]. Studies by Nantongo and Vatn (2024), as well as Corbera et al. (2020), suggest that participation in forest-based carbon management projects does not necessarily lead to increased income for farmers, particularly in impoverished or marginalized regions [19,20]. The main reason is that transitioning from traditional forestry practices to carbon sequestration-oriented management reduces timber revenue, while the current carbon pricing mechanism fails to sufficiently compensate for the increased operating costs and opportunity costs. As a result, farmers may experience financial losses, and in some cases, forestry carbon sink projects could even exacerbate smallholder poverty.

The existing literature has revealed both the positive and negative impacts of forestry carbon sink projects on farmers’ income. However, most existing studies focus on the direct income effects of forestry carbon projects on farmers, with a strong emphasis on REDD⁺ initiatives, while research on IFM (Improved Forest Management) projects remains limited. Moreover, there is a lack of in-depth discussion on how ecological forest management practices influence farmers’ income and willingness to produce carbon credits through changes in the allocation of production factors.

This study conducts an empirical analysis using a sample of 233 survey responses from LY County, Zhejiang Province. Its marginal contribution lies in applying a partial equilibrium model based on production functions and utility functions to explore the optimal production decision-making process of farmers between public goods (carbon sequestration) and private goods (timber production). The study further examines how this decision-making process impacts farmers’ income and utility. By establishing a logical framework—“Forest Management Transition → Reallocation of Production Factors → Income Impact → Willingness to Participate”—this study systematically analyzes the mechanism through which forest ecological management transformations influence farmers’ production behavior. The findings offer important policy insights for regions dominated by smallholder forestry, providing valuable guidance on how to optimize the institutional design of forestry carbon sink projects. Additionally, this study helps identify potential constraints on the long-term sustainability of these projects, contributing to the development of more effective and equitable carbon sink policies.

2. Materials and Methods

2.1. Theoretical Framework and Research Hypotheses

Farmers’ forest management practices and production behaviors have evolved following their participation in forestry carbon sink projects. This study draws on the household production behavior framework proposed by de Janvry et al. (1991) and analyzes farmers’ decision-making in forestry carbon sink production based on the production function, utility function, and local equilibrium theory [21]. In addition, Becker’s (1965) theory of time allocation is employed to examine how households, under time and budget constraints, allocate resources among timber production, forestry carbon sequestration, off-farm employment, and leisure activities to maximize overall utility [22]. Although the theoretical framework of farmers’ participation in forestry carbon sink projects is primarily grounded in neoclassical assumptions of utility maximization and voluntary participation, it remains applicable in the Chinese context. In China, collective forestland is legally owned by village collectives; however, following the reforms of the collective forest tenure system, farmers have been granted household contract management rights, which are protected as property rights under the real rights system. As a result, farmers have become the principal agents responsible for management and production decisions within the household responsibility system. While wood and forest products are classified as private goods, forest ecological services—such as carbon sink—are quintessential public goods. Farmers can engage in the production of either wood or forest carbon sink through various forest management practices. Furthermore, participation in forestry carbon sink projects prompts farmers to alter their production decisions.

Farmers’ production decisions are influenced by the production possibility curve and the equal profit line [21]. Prior to their involvement in forestry carbon sink projects, farmers typically employ a traditional forest management approach centered on timber production. Let $P_W$ represent the price of wood, $K_w$ denote the capital input for wood production, $L_w$ signify the labor input for wood production, ${\omega }_k$ indicate the price of forestry capital inputs, and ${\omega }_L$ reflect the price of forestry labor. As a public good, forestry carbon sinks do not yield direct benefits for farmers; consequently, at this stage, their profit function related to forest management is represented as ${\pi }_1$.

$${\pi }_1=P_W{f}_w(K_w,L_w)-{(\omega }_k{K}_w+{\omega }_L{L}_w)$$

(1)

Recording the cost of forestry production $C_1={\omega }_k{K}_w+{\omega }_L{L}_w$; Wood yield ${Q_{w_1}=f}_w(K_w,L_w)$, then the isoprofit line $Y_1$ is:

$$Q_{W_1}={\displaystyle \frac{{\pi }_1+C_1}{P_W}}$$

(2)

At this time, the production decision of farmers is to produce wood as $Q_{W_1}$ and not to produce forestry carbon sink.

Upon farmers’ participation in the forestry carbon sink project, the forest management paradigm evolves into a multifunctional model that prioritizes the production of public goods, such as forestry carbon sinks. At this juncture, capital input is denoted as $K_C$ and labor input as $L_C$, with the output of forestry carbon sinks represented by $Q_c=f_C\left(K_C,L_C\right)$; the wood yield is expressed as $Q_{W_2}=f_w\left(K_C,L_C\right)$. By establishing a carbon market, the forestry carbon sink project assigns a monetary value to forestry carbon sinks and facilitates their trade. Consequently, these sinks transition from being public goods to commodities, allowing for potential financial returns for farmers. As a result, farmers’ income comprises both minimal timber revenue and earnings from the forestry carbon sequestration project. If we denote the price of forestry carbon sequestration as $P_C$, then the profit function for farmers engaged in forest management post-participation in this project can be articulated as ${\pi }_2$:

$${\pi }_2=P_W{f}_w\left(K_C,L_C\right)+P_C{f}_C\left(K_C,L_C\right)-{(\omega }_k{K}_C+{\omega }_L{L}_C)$$

(3)

It is posited that the equilibrium is optimal when farmers utilize $K_C*$ and $L_C*$, with the forestry production cost being $C_2$*, and the corresponding isoprofit line denoted as Y₂.

$$Q_{W_2}={\displaystyle \frac{{\pi }_2+C_2\mathrm{*}}{P_W}}-{\displaystyle \frac{P_{\mathrm{c}}{Q}_C\mathrm{*}}{P_W}}$$

(4)

The production decision of farmers is cut point A in Figure 1. After farmers participate in the forestry carbon sink project, the production of wood is $Q_{W_2}$*, and the production of forestry carbon sink is $Q_{c_2}$*.

The willingness of farmers to engage in forestry carbon sink initiatives is contingent upon the alteration in their utility level resulting from participation in such projects. Furthermore, the economic rationality of farmers dictates that their utility level is influenced by the changes in profit before and after involvement in forestry carbon sink initiatives. The change in profit for established farmers post-participation in these projects can be denoted as ∆π.

$$∆\pi =P_W\left[f_w\left(K_C,L_C\right)-f_W\left(K_W,L_W\right)\right]+P_C{f}_C\left(K_C,L_C\right)-[{\omega }_k\left(K_C-K_W\right)+{\omega }_L(L_C-L_W)]$$

(5)

Make:

$$∆Y_W=P_W\left[f_W\left(K_C,L_C\right)-f_W\left(K_W,L_W\right)\right]$$

(6)

$$Y_C=P_C{f}_C\left(K_C,L_C\right)$$

(7)

$$∆C={\omega }_k\left(K_C-K_W\right)+{\omega }_L(L_C-L_W)$$

(8)

∆π is further expressed as:

$$∆\pi =∆Y_W+Y_C-∆C$$

(9)

In Formula (9), $∆Y_W$ represents the reduction in timber income for farmers ($∆Y_W$ < 0), $Y_C$ denotes the income derived from forestry carbon sinks, and ∆C signifies the change in production costs associated with forestry following farmers’ participation in the carbon sink project. Consequently, an exploration of ∆π necessitates a comprehensive analysis of these three factors.

Initially, the alteration in forestry production costs (∆C) resulting from farmers’ participation in the forestry carbon sink project is examined. This change in production costs arises from a shift in forest management practices. Farmers will determine the optimal rotation period aimed at maximizing the net present-value income and engage in various other forest management activities throughout the production and management cycle.

Prior to engaging in the forestry carbon sink project, farmers seek to ascertain the optimal rotation period t that maximizes the net present value of timber revenue. The decision-making process regarding the rotation period aligns with the Faustmann model when solely considering timber revenue [23]. In this context, for a single crop rotation, the net present value of the anticipated forest land value, denoted as $NP{V}_w(t)$, can be articulated as follows:

$$NP{V}_w(t)=\left[\begin{array}{c}\left(P_w-C_v\right)\times \delta \times V(t)\end{array}\right]{\mathrm{e}}^{-rt}-C_e$$

(10)

In this context, $C_v$ represents the costs associated with timber harvesting and transportation, δ denotes the yield of timber, $V(t)$ signifies the volume of forest stock, and $r$ indicates the discount rate under continuous compound interest. Additionally, $C_e$ refers to operating costs; thus, when $NP{V}_w\prime (t)=0$—implying that farmers aim to maximize timber income—the rotation period t must satisfy:

$$r={\displaystyle \frac{\delta \left(P_w-C_v\right)V\prime (t)}{\delta \left(P_w-C_v\right)V(t)}}$$

(11)

Upon engaging in the forestry carbon sink project, farmers aim to ascertain the optimal rotation period t that maximizes the combined net present value of timber revenue and forestry carbon sink income [24]. In this context, for a single crop rotation, the net present value of forestry carbon sink income can be denoted as $NP{V}_c(t)$:

$$NP{V}_c(t)={\int }_0^t {\mathrm{e}}^{-rt}{P}_c\times m_c\prime (t)\mathrm{d}t-{\mathrm{e}}^{-rt}{P}_c\times m_c(t)$$

(12)

In this context, $m_c(t)$ denotes the quantity of carbon sequestered in the above-ground biomass of the forest. The integral ${\int }_0^t {\mathrm{e}}^{-rt}{P}_c\times m_c\prime (t)\mathrm{d}t$ signifies the cumulative net present value of the annual increments in carbon sequestration throughout the operational cycle, while ${\mathrm{e}}^{-rt}{P}_c\times m_c(t)$ reflects the loss of forest carbon sink during harvesting.

At this time, the total income of farmers is:

$$NP\mathrm{V}\left(t\right)=NP{V}_w\left(t\right)+NP{V}_c(t)$$

(13)

If $NPV\prime (t)=0$, indicating that the income from timber for farmers and the carbon sink revenue from forestry are both maximized, then the rotation period t must satisfy:

$$r={\displaystyle \frac{\delta \left(P_w-C_v\right)V\prime (t)}{\delta \left(P_w-C_v\right)V\left(t\right)-P_c\times m_c(t)}}$$

(14)

Given that $m_c(t)$ is an increasing function of t, the right-hand side of Equation (14) will consistently be less than that of Equation (11) for a fixed discount rate. Consequently, when accounting for forestry carbon sink income, the rotation period will exceed that determined solely by timber income; this extension in the business cycle will lead to a reduction in farmers’ felling frequency [5] and subsequently decrease labor associated with forestry harvesting. To ensure optimal output from forestry carbon sinks, it is imperative for farmers to maintain a healthy and stable forest ecosystem while cultivating multi-functional forests characterized by diverse ages, species mixtures, and structural layers. In contrast to traditional high-intensity cutting and weeding practices in forestry management, farmers should minimize human interference and engage in supplementary activities such as fertilization and selective planting or weeding as necessary [25]. This approach implies a reduction in both the frequency and intensity of mowing and weeding efforts by farmers, thereby decreasing employment related to these tasks. Overall labor input into forestry production can thus be diminished. On the other hand, in order to provide sustained incentives for forestry operators to continuously supply carbon sink products, project owners of forestry carbon projects in China typically return the full amount of carbon credit revenues to farmers and offer free inputs such as fertilizers. These measures serve as supplementary incentives from the government, aiming to lower the entry barriers for smallholders to participate in the carbon market. As a result, farmers are generally not required to make additional capital investments. Furthermore, project owners involved in forestry carbon sink initiatives typically design tailored forest management schemes for farmers while providing specialized fertilizers aimed at further reducing capital expenditures. If both labor inputs and capital investments decline concurrently within this framework, then overall production costs are expected to decrease (∆C < 0). Henceforth, this paper proposes Research Hypothesis 1:

H₁. 

The participation of farmers in forestry carbon sink projects can lead to a transformation in forest management practices, resulting in reduced production costs for forestry activities.

Given that $Y_C-∆C>0$, and $∆Y_W<0$, the sign of ∆π is contingent upon whether the subsidies from forestry carbon sink projects and the cost savings achieved by farmers can offset the decline in timber revenue. This necessitates a comparative analysis aligned with the actual conditions of collective forest development in China.

With the reform of the collective forest ownership system and the implementation of afforestation initiatives, China’s collective forest area has experienced substantial growth. In 2023, the total area of collective forest land in China is projected to reach 40.5 billion hectares, with 32.745 billion hectares designated as collective forests (data from the China people’s government of the net: https://www.gov.cn/lianbo/bumen/202408/content_6968614.htm, accessed on 22 September 2024). This indicates that there is limited potential for reforestation within China’s existing collective forest lands; moreover, compared to afforestation carbon sink projects, the costs associated with managing carbon sinks in forests are relatively low. Consequently, developing carbon sink projects through effective management of existing collective forests represents a crucial direction for enhancing China’s overall carbon sink strategy in these areas. The forestry development within these regions is influenced by policy frameworks and developmental objectives. As environmental improvement and ecological civilization construction become increasingly prioritized, policies promoting classified forest management continue to advance; this results in a reduction of commercially managed forest areas overseen by farmers. Additionally, a decline in the forestry labor force contributes to rising labor costs within this sector—these increases outpace those seen in wood prices—rendering forestry production more costly relative to income derived from timber sales [26].

Specifically, the average annually is 4.8 m³/ha (data from China’s National Development and Reform Commission’s official website: https://www.ndrc.gov.cn/xxgk/jd/jd/202111/t20211108_1303400.html, accessed on 22 September 2024), with each cubic meter increase in forest stock capable of sequestering approximately 1.6 tons of carbon dioxide (data from China’s People’s Daily online: http://env.people.com.cn/n1/2021/0114/c1010-31999463.html, accessed on 22 December 2024). This leads to an estimated average annual carbon sequestration rate for China’s forests of about 7.65 tons/ha. According to a long-term follow-up survey conducted by the Development Research Center of the National Forestry and Grassland Administration at fixed observation points, the average price of wood in China’s collective forest areas from 2002 to 2022 was recorded at CNY 116.07/m³ [26], while the current average price for forestry carbon sink trading stands at CNY 58/ton (data from Zhang Shougong, Chen Xingliang, Introduction to Carbon Sinks and Carbon Markets, accessed on 22 September 2024). Consequently, if farmers primarily focus on timber production, the average annual net present value (NPV) of forest land increases by approximately CNY 557.1/ha; conversely, if they concentrate on forestry carbon sinks, this NPV rises by around CNY 443.7/ha. Data from sample farmers’ forestry management practices indicate that since 2010, when focusing on timber production, their capital input has remained relatively stable at about CNY 198/ha [27]. Following participation in forestry carbon sequestration projects—wherein project owners provide necessary capital inputs—farmers are not required to make additional financial contributions; furthermore, reduced labor input associated with these projects will further lower overall production costs for farmers. Therefore, should farmers prioritize forestry carbon sinks over timber production, it is anticipated that their savings in production costs would exceed more than CNY 198/ha. In conclusion, both income derived from forestry carbon sinks and cost savings in production surpass those generated through traditional timber sales for farmers. Based on the above, this paper proposes Hypothesis 2:

H₂. 

Subsidies for forestry carbon sink projects and forestry production costs saved by farmers can make up for the loss of timber income, so that the average profit per hectare of forest land and net forestry income will not change significantly.

Changes in farmers’ production decisions also influence their forestry labor time choices, as illustrated in Figure 2. The horizontal axis represents the leisure time of farmers, under the assumption that profits derived from forestry production are allocated for consumption. The vertical axis indicates the increase in consumption that farmers can achieve due to these profits. Consumption and leisure are two critical components that farmers must choose between, with their preferences depicted by the indifference curve U. Line I denotes the budget constraint faced by farmers. Following participation in the forestry carbon sink project, farmers can reduce their forestry labor while maintaining the average profit per hectare of forest land; this implies that they can still realize a profit π even with increased leisure time. Consequently, the budget constraint shifts from I₁ to I₂, and the indifference curve transitions from U₁ to U₂, enabling farmers to attain higher utility.

Farmers will reconfigure the production factors saved by forest land management and the increased leisure time $L_2-L_1$. According to the rational smallholder school and the family production theory, since the marginal income of the non-agricultural sector is higher than that of the agricultural sector, farmers will allocate the saved production factors to the non-agricultural sector out of rational choice, so as to increase the non-agricultural net income. Since the marginal income of the non-agricultural sector is greater than that of the forestry sector, the increase in the non-agricultural net income of farmers can bring additional income and increase the net income of farmers. Therefore, this paper proposes research Hypothesis 3:

H₃. 

After participating in forestry carbon sink project, farmers’ non-agricultural net income and farmers’ net income will increase significantly.

Based on the above analysis, the forestry carbon sink project can increase the net income of farmers by granting project subsidies, saving farmers’ production factors, and increasing non-agricultural income, so that farmers can obtain higher utility in the production of forestry carbon sink and their willingness to produce forestry carbon sink can be increased.

2.2. Site Selection and Data Sources

This study selects ZYL Village, Yan Village, LF Village, and ZKK as research sites for the Moso bamboo forest carbon sink project in LY County, Zhejiang Province, based on several compelling reasons.

With the implementation of collective forest tenure reform and large-scale afforestation programs, the forest area within China’s collectively owned forestlands has increased significantly. As of 2023, the total area of collective forestland reached 171.4 million hectares (data from the China people’s government of the net: https://www.gov.cn/lianbo/bumen/202408/content_6968614.htm, accessed on 22 December 2024; data from China’s National Development and Reform Commission’s official website: https://www.ndrc.gov.cn/xxgk/jd/jd/202111/t20211108_1303400.html; data from China’s People’s Daily online: http://env.people.com.cn/n1/2021/0114/c1010-31999463.html), of which 145.6 million hectares were classified as forested (data from Zhang Shougong, Chen Xingliang, Introduction to Carbon Sinks and Carbon Markets https://www.gov.cn/lianbo/bumen/202408/content_6968614.htm, accessed on 22 December 2024). This indicates that the potential for further afforestation in collective forest areas is now very limited. Moreover, compared to afforestation-based carbon projects, forest management-based carbon projects entail lower implementation costs [28]. Therefore, promoting forest management carbon sink projects within existing collective forest areas represents a key direction for the development of China’s forest carbon market. In comparison to other carbon sink projects focused on forest management, Moso bamboo exhibits a remarkable capacity for carbon sequestration, characterized by its rapid growth cycle, quick returns on investment, and the relatively low costs associated with Moso bamboo forest management initiatives [9,29]. Consequently, the Moso bamboo forest management carbon sink project represents a significant component of China’s overall forest management carbon sink strategy. Currently, there are six such projects in China; aside from the one located in Tongshan County, Hubei Province, the remaining five are situated within Zhejiang Province [30], underscoring Zhejiang’s leadership in both the exploration of Moso bamboo forest resource reserves and the implementation of Moso bamboo forest management carbon sink projects. In 2022, Mx Township in LY County, Zhejiang Province, initiated a 30-year collective Moso bamboo forest management carbon sequestration project and established a preliminary transaction with Zhejiang Jinlong Renewable Resources Technology Co., Ltd. (Quzhou, China). Jinlong is set to acquire the carbon sequestration rights for 150,000 hectares of Moso bamboo forests located in Mx Township and Xikou Forest Farm over the next three decades, which is projected to yield approximately 120,000 tons of carbon credits. The parties involved have referenced the average price from the national carbon market in 2021 and adjusted it based on the annual carbon sequestration potential of Moso bamboo forest land within this project framework, thus determining that a subsidy of CNY 300 per hectare per year will be allocated to farmers during the first trading cycle. Throughout the duration of this project, farmers are mandated to manage their carbon-sequestering forests according to methodologies specified by the initiative and may receive dividends upon fulfilling designated carbon sequestration criteria. Should there be any instances of forest degradation or carbon leakage, farmers will be liable for compensation. This transaction represents the largest forestry carbon sink deal executed by enterprises in Zhejiang Province thus far regarding upfront payments for forestry-related carbon sinks (according to the Zhejiang province forestry bureau’s website http://lyj.zj.gov.cn/art/2022/10/24/art_1229001954_59039513.html, 22 December 2024), thereby inaugurating pre-collection storage and preliminary transactions related to forest-based carbon sinks. In May 2023, LY County was designated as part of the second batch of forestry carbon sink pilot projects in Zhejiang Province (According to the Zhejiang province forestry bureau’s website http://lyj.zj.gov.cn/art/2023/5/10/art_1276365_59051551.html, 22 December 2024). This designation underscores the significant impact that the Moso bamboo forest management carbon sink initiative in LY has had on farmers’ production and income, positioning it as a model for both the province and the nation.

It is important to note that the forestry development processes in the aforementioned four villages mirror those of most collective forest areas across China. Prior to the environmental rectification efforts initiated in 2014, Mx Township had developed a Moso bamboo product processing industry, which constituted the primary source of income for farmers in these four villages through both Moso bamboo harvesting and product processing. Following 2014, as a result of environmental rectification and an expansion of ecological non-commercial forest areas, restrictions were placed on the Moso bamboo product processing industry, prompting a gradual transformation within traditional forestry practices. This industrial transition first altered farmers’ income structures, with non-agricultural earnings emerging as their predominant source of revenue. Additionally, changes in employment patterns among farmers have led to a rapid increase in labor costs associated with forestry management; currently, average wages have risen to approximately CNY 250 per day per person—reflecting broader trends observed among small-scale farmers throughout China’s forestry sector. Consequently, the implementation and study of the LY carbon sink project are relevant across most collective forest regions in China and can provide valuable insights for future initiatives related to carbon sinks within China’s collective forests. For these reasons, ZYL Village, Yan Village, LF Village, and ZKK have been selected as research sites for this paper. A ten-day field investigation was conducted in these four villages in June 2024. The field investigation focused on, but was not limited to, the history of local forestry development, the transaction processes of forestry carbon sink projects, the evolution of project model selection, the benefits already generated, and future development prospects. During the survey period, face-to-face interviews were conducted with key stakeholders, including officials from the local forestry bureau, village cadres, project owners, cooperative leaders, and selected representative households. In addition to interviews, data were collected through farmer questionnaires, village cooperative questionnaires, and structured statistical forms. With the support of relevant government departments, first-hand materials such as policy documents at the municipal, county, and village levels, as well as household-level survey data, were also obtained.

The carbon sink project focused on Moso bamboo forest management in the study area commenced in 2022. Due to variations in the timing of farmers’ adherence to carbon sink forest requirements, some farmers had not yet aligned their operations with these standards by 2023. This paper randomly sampled two groups: those who complied with the carbon sink forest requirements and those who did not. The survey primarily utilized household questionnaires to gather essential information regarding household members from 2021 to 2023, agricultural land and forest management practices, participation in the forestry industry, and household income. A total of 233 valid questionnaires were collected after excluding invalid responses, as detailed in

Table 1. Statistical table of sample villages (unit: hectare, CNY).

Sample Village	Forest Area	Per Capita Forest Area	Per Capita Net Income	Proportion of Rural Households Mainly Engaged in Forestry Operations in Total Households	Sample Size
ZYL Village	563.87	2.33	15,000	38.0%	45
Yan Village	616.77	1.86	21,000	58.7%	73
Lf Village	347.02	0.84	21,500	50.8%	69
ZKK Village	340.33	1.32	26,000	27.0%	46
Total	1867.99	-	-	-	233

Note: CNY is the monetary unit of Chinese currency, and at the study period, CNY 1 was equal to USD 0.14.

. Among these, 205 farmers who adhered to the carbon sink forest requirements in 2023 constituted the experimental group, while 28 farmers who did not comply formed the control group. Given that the production cycle for local traditional Moso bamboo is two years, average values from 2021 and 2022 are employed to represent conditions prior to initiating the Moso bamboo forest carbon sink project; data from 2023 reflect outcomes following farmer participation in this initiative. Inputs and incomes associated with farmers’ engagement in carbon sink forests are adjusted according to its management cycle.

2.3. Empirical Model Specification

To estimate the effects of forestry carbon sink projects on farmers’ production and income, we adopt a Propensity Score Matching combined with Difference-in-Differences (PSM-DID) strategy, following the empirical framework developed by Heckman et al. [31]. This approach helps address observable selection bias and potential endogeneity, making it suitable for credible causal inference in non-experimental settings [32].

Referring to Heckman’s PSM-DID model [32], in this study, the participating farmers were categorized into an experimental group and a control group, with the treatment variable defined as $D_i=\left\{\mathrm{0,1}\right\}$, which indicates whether farmer i adhered to the carbon sequestration forest requirements in 2023. Specifically, if $D=1$, it signifies compliance; conversely, $D=0$ denotes non-compliance. The net forestry income for farmers is represented as $Y_i$, where $Y_{0i}$ refers to the net forestry income of those in the experimental group, and $Y_{1i}$ pertains to that of the control group. This paper aims to investigate the impact of $D_i$ on $Y_i$, as illustrated by the following formula:

$$ATT=E\left[Y_{1i}-Y_{0i}\right]=E\left[Y_{1i}-Y_{0i}|D=1\right]=E\left[Y_{1i}|D=1\right]-E\left[Y_{0i}|D=1\right]$$

(15)

The Average Treatment Effect on the Treated (ATT) represents the mean processing effect observed among participants. Following the fundamental principles of propensity score matching, individuals in the control group are paired with those in the experimental group based on their proximity across various characteristics. If rural residents are assigned a feature combination $X_i$, then $P(X_i)$ denotes the likelihood that they will adhere to carbon sink forest requirements in 2023. This implies that under condition X, the conditional probability of farmers being classified as part of the experimental group is articulated as follows:

$$P\left(X_i\right)=P(D_i=1|X=X_i)$$

(16)

Specifically, the propensity score is derived from binary logit regression. This paper employs the K-proximity matching method in conjunction with reference regression. Following data processing through propensity score matching, a difference-in-differences approach is utilized to assess the impact of farmers’ participation in forestry carbon sink projects on their income. The model is structured as follows:

$$Y_{it}={\beta }_0+{\beta }_1{policy}_{it}\times {year}_{it}+{\sum }_{j=1}^n{\alpha }_j{X}_{it}+{\epsilon }_{it}$$

(17)

In this context, $Y_{it}$ represents the explanatory variable value for the i-th farmer in year t. The variable ${policy}_{it}$ indicates whether the i-th farmer has adhered to the management requirements of carbon sequestration forests; a value of ‘1’ signifies compliance, while ‘0’ denotes non-compliance. The variable ${year}_{it}$ is utilized to determine if the local area falls within the accounting period for forestry carbon sink projects in year t, with 2023 designated as ‘1’ and all other years marked as ‘0’. $X_{it}$ serves as a control variable, ${\epsilon }_{it}$ represents the residual term, ${\beta }_0$ is defined as the constant term, and ${\beta }_1$ is identified as the parameter to be estimated.

2.4. Variable Selection and Descriptive Statistics

The specific variables selected for this study are as follows:

Explained variables: Given that farmers’ willingness to engage in forestry carbon sink projects is influenced by both costs and returns, rather than merely changes in income, it may be more pertinent to examine net income. Consequently, this paper utilizes net income as the explained variable. In the questionnaire, the definitions are as follows: Net forestry income = Total Forestry Income − Forestry Production Expenditures; Non-agricultural net income = Non-agricultural Income − Non-agricultural Expenditures; Net income of rural households = Total Rural Household Income − Household Operating Expenses (which include total expenditures on forestry, planting, and non-agricultural production) − Human Costs − Other Expenses.

Key explanatory variables: The primary explanatory variable in this study is whether farmers participate in forestry carbon sequestration projects, represented by the interaction term ${policy}_{it}\times {year}_{it}$. In the experimental group, data for 2023 are coded as 1, while all other years are coded as 0.

Control variables: Given that farmers’ production decisions, production functions, and non-agricultural employment are influenced by the natural characteristics of forest land, farmer attributes, and village economic development levels, control variables have been selected from these three dimensions.

This paper focuses on the characteristics of woodland plots with the largest area cultivated by farmers to represent natural features; specific variables include plot area (area of the largest plot), height of forest land, slope of forest land, and soil fertility. Due to economies of scale, a larger area of forest land may facilitate more efficient input utilization [33]. Both height and slope serve as critical indicators of forest quality; steeper slopes often correlate with greater heights and increased transportation challenges, which may deter farmers from investing additional production resources [34]. Higher soil fertility typically indicates more productive forest land.

The key variables influencing farmers’ decisions primarily include the age of the head of household and human favor expenditure. In rural households, decision-making is predominantly undertaken by the household head, whose characteristics significantly impact family production choices. The age of the household head may correlate with a lower likelihood of engagement in non-agricultural employment; conversely, human favor expenditure can serve as an indicator of social capital to some extent. A higher level of social capital typically results in more comprehensive information regarding forestry carbon sink projects, enhanced understanding of these initiatives, and increased willingness to participate [35]. The economic development level of a village is primarily assessed through village collective financial income, which reflects both economic progress and organizational capacity within the community. Greater organizational capability at the village level likely enhances farmers’ participation in forestry carbon sinks, thereby influencing their production decisions. Detailed descriptions and descriptive statistics for these variables are presented in

Table 2. Descriptive statistics of variables (unit: hectare, CNY).

Variable Type	Variable Name	Variable Declaration	Mean Value	Standard Deviation
Explained variable	Net forestry income	Forestry Net Income = Total Forestry Income − Forestry Production Expenditures	6785	15,277.34
	Non-agricultural net income	Non-agricultural net income = Non-agricultural − Non-agricultural Expenditures	48,737	55,025.17
	Net income of rural households	Rural Household Net Income = Total Rural Household Income − Household Operating Expenses − Human Costs-Other Expenses.	60,875	101,840.1
	Average input per hectare of forest land	Farmers’ input to forest land = (Total number of workers × labor price of different plots + capital input)/land area	23.22	338.76
	Average profit per hectare of forest land	Average profit per hectare of forest land = average income per hectare of forest land − average input per hectare of forest land	10.62	401.76
Key explanatory variables	Participation of farmers in forestry carbon sink projects	policyit × yearit interaction item, the 2023 data of the experimental group is 1, and the rest is 0	0.44	0.50
Control variables	Plot area	The area of the largest forest land plots	1.04	28.52
	Height of forest land	The height of the largest forest land: low = 0, medium = 1, high = 2	2.05	0.82
	Slope of forest land	Slope of the largest forest plots: flat = 0, gentle = 1, steep = 2	2.33	0.72
	Soil fertility	Soil fertility of the largest forest plots: poor = 0, medium = 1, good = 2	2.39	0.72
	Age of head of household	Age of head of household	66	9.42
	Human favor expenditure	Total household human expenditure	9336	7758.43
	Village collective financial income	Research village collective financial income	452,600	12.32

Note: CNY is the monetary unit of Chinese currency, and at the study period, CNY 1 was equal to USD 0.14.

.

To supplement the econometric analysis and enhance the objectivity of the diagnostic results, we present a descriptive summary of key survey indicators in

Table 3. Summary of diagnostic survey results on farmer participation and behavioral changes.

Variable	Survey Question	Response Options	Percentage (%)
Participation	Through what channel did you learn about the forest carbon sink project?	(1) Promotion by village cadres; (2) Friends and relatives; (3) Internet, TV, etc.; (4) Enterprises operating carbon sink projects; (5) Others, please specify	(1) 100%; (2) 0%; (3) 0%; (4) 0%; (5) 0%
	What is your mode of participation in the carbon sink forest project?	(1) Self-management; (2) Leasing forestland to the project; (3) Working for the project; (4) Others, please specify	(1) 100%; (2) 0%; (3) 0%; (4) 0%
Changes in Management	Have you changed your forest management practices?	Yes/No	Yes: 87.98%/No: 12.02%
	Compared to non-carbon sink forests, how has the profitability of carbon sink forests changed?	(1) Profit increased; (2) Profit decreased; (3) No change	(1) 37%; (2) 36%; (3) 27%
Input Adjustments	Compared to non-carbon sink forests, how has the use of fertilizer changed?	(1) Increased; (2) Decreased	(1) 76%; (2) 24%
	Compared to non-carbon sink forests, how has labor input changed?	(1) Increased; (2) Decreased	(1) 17%; (2) 83%
Output Changes	Compared to non-carbon sink forests, how has the output (e.g., volume of timber) changed?	(1) Increased; (2) Decreased	(1) 43%; (2) 57%
Future Participation Willingness	What factors influence your willingness to participate in future forest carbon sink projects? Please rank the following.	(1) The forest designated as carbon sink is commercial forest; (2) Higher technical requirements and supervision than non-carbon forests; (3) High operating cost; (4) Long operating cycle; (5) Difficult to obtain carbon revenue; (6) Uncertain prospects of carbon sink forestry	The proportion of respondents who ranked each item as the most important factor: (1) 12%; (2) 17%; (3) 23%; (4) 27%; (5) 15%; (6) 6%

. The diagnostic survey provides important insights into farmers’ participation in the forest carbon sink project and their behavioral changes. As shown in Table 3, nearly all farmers (100%) became aware of the carbon sink project through promotion by village cadres, indicating that local governments and grassroots mobilization play a critical role in project participation. Regarding changes in forest management practices, there is variation in perceived profitability: 37% of farmers reported an increase in profits, 36% reported a decrease, and 27% perceived no change. According to local farmers, these differences may relate to factors such as forestland area, location, and previous management practices. In terms of input adjustments, 76% of farmers increased their use of fertilizers, which may reflect the technical requirements of the carbon sink methodology. Meanwhile, 83% of farmers reported a reduction in labor input, suggesting that carbon sink management may reduce labor-intensive activities such as logging. Nevertheless, 43% of forests were reported to have increased yields, while 57% experienced a decline. This indicates that under forest management practices oriented toward carbon sink production, production outcomes are heterogeneous, warranting further econometric analysis of farmers and plots with different characteristics.

Therefore, this paper first conducts a descriptive analysis of farmers’ costs and benefits at the plot level (

Table 4. Descriptive statistical table for the operation of sample plot (unit: hectare, CNY).

	Number of Plots	Average Plot Area	Average Labor Input per Hectare of Forest Land		Average Capital Input per Hectare of Forest Land		Average Input per Hectare of Forest Land		Average Income per Hectare of Forest Land		Average Profit per Hectare of Forest Land		Species of Trees
			Pre-project	Post-project	Pre-project	Post-project	Pre-project	Post-project	Pre-project	Post-project	Pre-project	Post-project	Moso bamboo
Experimental group	205	1.07	5641.65	2151.60	847.65	91.50	6489.30	2242.95	8299.95	4734	1810.50	2491.5
Control group	28	0.89	6748.35	6559.80	1130.10	1024.80	7878.60	7585.50	11,109.45	10,114.65	3231	2529.15
Grand mean	233	-	5774.70	3469.65	881.55	327	6656.25	3796.50	8637.60	6596.10	1981.50	2799	-

). A total of 233 woodland plots were sampled, of which 205 adhered to the guidelines set forth by forestry carbon sequestration projects, while 28 did not participate in such initiatives. The average input per hectare of forest land encompasses both labor input and capital input across various operational stages including seedling planting, irrigation and weeding, fertilization, Moso bamboo shoot harvesting, and logging activities. Given the significant variation in labor prices among different plots, labor input is adjusted based on the number of workers relative to each plot’s specific labor cost; capital input includes all expenses beyond labor—such as fertilizers, pesticides, transportation fees, and other associated costs. The average income per hectare of forest land comprises revenue from Moso bamboo products and subsidies related to forestry carbon sequestration projects. Descriptive statistics indicate that when farmers operate according to project requirements compared to traditional management practices, there is a reduction in both average labor input and capital input per hectare of forest land. Specifically, the average labor input per hectare of forest land decreases by CNY 3490.1, and the average capital input per hectare of forest land declines by CNY 756.2, thus resulting in an overall reduction in average input per hectare of forest land amounting to CNY 4246.3. Conversely, while the average income per hectare of forest land diminishes by CNY 3566.0 due to these changes in operation methods, notably, however, the average profit per hectare of forest land experiences an increase of CNY 681.0.

This study further performed descriptive statistics on the income of the sampled farmers (

Table 5. Descriptive statistical table for the income of sample farmers.

	Sample Household	Average Net Forestry Income		Average Non-Agricultural Net Income		Average Net Income of Rural Households
		Pre-Project	Post-Project	Pre-Project	Post-Project	Pre-Project	Post-Project
Experimental group	205	7249	6569	49,550	52,620	58,666	68,286
Control group	28	5818	5940	33,109	29,983	42,745	40,921
Grand mean	233	7077	6494	47,574	49,900	56,753	64,998

) and revealed that the non-agricultural net income constituted a significant portion of the total net income for these farmers. Notably, the average net forestry income within the experimental group declined following their participation in the forestry carbon sink project, whereas both the average non-agricultural net income and average net income of rural households experienced an increase. Building upon this descriptive analysis, this paper will employ an econometric model for subsequent empirical investigation.

3. Empirical Results

3.1. Propensity Score Matching

Propensity Score Matching (PSM) was initially employed to analyze the data, ensuring that the experimental and control groups were matched based on specified control variables. The co-supporting hypothesis and balance hypothesis served as foundational premises for the PSM methodology. This paper presents a bar chart illustrating the co-supporting field along with the results of balance tests for propensity scores (see Figure 3,

Table 6. Balance test results of control variables before and after matching.

Variable	Type	Mean Value		Standardization Deviation	Variation of Standardization Deviation	T-Value
		Experimental Group	Control Group
Plot area	Before matchmaking	16.0	13.3	12.3	93.1	0.47
	After matchmaking	12.6	12.8	−0.8		−0.14
Height of forest land	Before matchmaking	2.03	2.14	−12.5	28.0	−0.66
	After matchmaking	2.09	2.17	−9.0		−0.86
Slope of forest land	Before matchmaking	2.32	2.39	−9.8	17.8	−0.49
	After matchmaking	2.36	2.41	−8.0		−0.79
Soil fertility	Before matchmaking	2.38	2.43	−7.0	−21.2	−0.33
	After matchmaking	2.40	2.35	8.5		0.83
Age of head of household	Before matchmaking	65.2	66.5	−13.2	−40.3	−0.69
	After matchmaking	65.7	63.7	18.6		1.66 *
Human favor expenditure	Before matchmaking	8962.0	8137.5	15.5	96.6	0.62
	After matchmaking	8464.2	8436.2	0.5		0.05
Village collective financial income	Before matchmaking	45.7	41.2	38.7	90.2	1.78 *
	After matchmaking	43.8	43.4	3.8		0.37

Note: * is significant at the level of 10%.

and

Table 7. Overall goodness of fit statistics of the model before and after matching.

Type	Pseudo R2	LR Statistic	p-Value
Before matchmaking	0.033	5.70	0.575
After matchmaking	0.01	4.94	0.667

).

The results indicated that the majority of observations fell within the common value range, with only a limited number of samples lost—specifically, one sample from the control group and 27 samples from the experimental group. The balance test results reveal that prior to matching, there was a significant difference in collective financial income among villages at the 10% level; however, post-matching analysis showed that only the age of the household head remained significantly different at this level within an acceptable range. Furthermore, standardization deviations decreased after matching, both falling below 20%, which suggests that the matching process effectively balanced the distribution of control variables between the two sample groups. Based on these assessments, propensity score matching using K-proximity methods was employed to match participants in both groups. Following elimination of unmatched samples, a total of 205 samples were retained, comprising 22 from the control group and 183 from the experimental group.

3.2. Impact of Participation in Forestry Carbon Sink Projects on Plot Management

To examine changes in forestry production costs and profits following farmers’ participation in forestry carbon sink projects, this section conducts analysis at the plot-management level. Specifically, for each sample household, the largest forest plot in terms of area is selected as the plot-level observation. Based on PSM matching results, a regression model is constructed using average input per hectare of forest land and average profit per hectare of forest land as the dependent variables. The regression results are presented in

Table 8. Impact of participating in forestry carbon sink projects on land management.

	Average Input per Hectare of Forest Land		Average Profit per Hectare of Forest Land
	(1)	(2)	(3)	(4)
Participation of farmers in forestry carbon sink projects	−4083.6 *** (−5.11)	−3840 *** (−4.28)	1306.5 (1.23)	1298.7 (1.09)
Control variable	Out of control	Controlled	Out of control	Controlled
Constant	6843.2 ***	76,539.9 ***	2031.6	88,857.8
	(36.15)	(0.71)	(8.05)	(−0.62)
Individual effect	Controlled	Controlled	Controlled	Controlled
Time effect	Controlled	Controlled	Controlled	Controlled

Note: *** is significant at the level of 1%, respectively.

.

The study revealed that, in comparison to plots not adhering to the guidelines of forestry carbon sequestration projects, the average investment per hectare for carbon sequestration forest plots was significantly reduced by CNY 3840, while the average profit per hectare remained largely unchanged. Hypothesis 1 was confirmed, indicating that transforming farmers’ forest management practices can lead to a reduction in their forestry production costs.

3.3. Impact of Participation in Forestry Carbon Sink Projects on Net Forestry Income

Based on the analysis conducted at the plot level, this paper employs a differential method to further investigate the regression analysis of net forestry income, with the results presented in

Table 9. Impact of participating in forestry carbon sequestration projects on forestry net income.

Net Forestry Income	PSM-DID
	(1)	(2)
Whether farmers participate in forestry carbon sequestration projects	−939.07	−2856.87
	(−0.57)	(−1.58)
Control variables	Out of control	Controlled
Constant	6534.55 ***	−221,954.90
	(16.85)	(−1.02)
Individual effect	Controlled	Controlled
Time effect	Controlled	Controlled

Note: *** is significant at the level of 1%, respectively.

. The findings indicate that the coefficients are not statistically significant, suggesting that farmers’ participation in forestry carbon sink projects does not substantially diminish forestry net income. In other words, subsidies from these projects and savings on production costs can offset any potential losses in timber revenue. Empirical evidence gathered from interviews with farmers reveals that many expressed sentiments such as ‘hiring means losing money’, which reflects two underlying issues: first, current earnings from Moso bamboo harvesting in LY are relatively low; second, a shortage of local labor has driven up labor costs within the forestry sector. Typically, agricultural labor is provided by family members of farmers; when accounting for their labor costs, there is effectively no profit derived from logging activities. While subsidies associated with forestry carbon sink projects can mitigate some losses incurred during timber sales, they ensure that overall net income remains largely intact. Hypothesis 2 has been validated: following their involvement in these projects, farmers experience no significant detriment to their forestry income while also conserving inputs related to production factors. These conserved resources may be redirected towards leisure or alternative productive endeavors, thereby enhancing the utility levels for farmers.

3.4. Impact of Participation in Forestry Carbon Sink Projects on Non-Agricultural Net Income

Based on the analysis at the plot level, the production factors saved by farmers’ participation in forestry carbon sink projects make it possible for farmers’ non-agricultural net income to increase. Therefore, this section further explores whether participation in forestry carbon sink projects can bring about changes in non-agricultural net income. The regression results are shown in

Table 10. Impact of participating in forestry carbon sequestration projects on non-agricultural net income.

Non-Agricultural Net Income	PSM-DID
	(1)	(2)
Whether farmers participate in forestry carbon sequestration projects	6781.35 ***	6761.03 ***
	(4.01)	(2.17)
Control variables	Out of control	Controlled
Constant	45,500.03 ***	483,281.20 **
	(113.66)	(2.26)
Individual effect	Controlled	Controlled
Time effect	Controlled	Controlled

Note: ** and *** are significant at the level of 5% and 1%, respectively.

: farmers’ participation in the forestry carbon sink project will have a significant positive impact on non-agricultural net income. The non-agricultural net income of farmers participating in the forestry carbon sink project will significantly increase by CNY 6761.03, which is significant at the 1% significance level, and Hypothesis 2 is verified. Based on the employment information of family members in the survey, it is found that a certain proportion of permanent residents’ main source of income is casual jobs and temporary workers, which indicates that after the transformation of local industries, some non-agricultural jobs with lower technical requirements and flexible working hours can be provided. Farmers have basic non-agricultural employment skills and can adjust their employment channels at any time. Farmers transferred the leisure time increased by participating in the forestry carbon sink project to the non-agricultural sector, which promoted the increase in non-agricultural net income.

3.5. Impact of Participating in Carbon Sink Projects on Net Income of Rural Households

Based on the preceding analysis, participation in forestry carbon sink projects exerts a significant positive influence on non-agricultural net income. However, when compared to various income sources, farmers tend to prioritize changes in total income. Consequently, this section delves deeper into the alterations in farmers’ net income resulting from their involvement in forestry carbon sink initiatives. The dependent variable examined here is net income of rural households, with regression results presented in

Table 11. Impact of participating in forestry carbon sequestration projects on net income of farmers.

Net Income of Rural Households	PSM-DID
	(1)	(2)
Whether farmers participate in forestry carbon sequestration projects	4099.27 **	3849.34 **
	(2.17)	(1.82)
Control variables	Out of control	Controlled
Constant	53,692.29 ***	30,7164.3
	(119.85)	(1.21)
Individual effect	Controlled	Controlled
Time effect	Controlled	Controlled

Note: ** and *** are significant at the level of 5% and 1%, respectively.

: at the 5% significance level, farmer participation in forestry carbon sink projects demonstrates a substantial positive effect on their net income. Following engagement in these projects, the net incomes of rural households are projected to increase significantly by CNY 3849.34, thereby validating Hypothesis 3. In light of this analysis, it can be concluded that forestry carbon sink projects enhance farmers’ net incomes through project subsidies and the optimization of production factors while simultaneously boosting non-agricultural earnings, thus elevating the utility levels of participating farmers and enhancing their willingness to engage further in forestry carbon sink production.

3.6. Robustness Test

In order to make the regression results more robust, the robustness test of the above studies was carried out in this section, mainly including the replacement matching method and placebo test to verify the robustness of the results.

(1)

Change the matching method

Robustness tests for caliper matching and kernel matching were added in this paper.

Table 12. Caliper matching and nuclear matching test results.

	Caliper Matching		Nuclear Matching
Together Support the Test Results of the Hypothesis	Match Failure	Match Successfully	Match Failure	Match Successfully
Control group	1	27	1	27
Experimental group	29	176	25	180
Total	30	203	26	207
Balance test results	There were no significant variables after matching, and the absolute value of the standardization deviation of all variables after matching was less than 20%		None of the variables were significant before matching, and the absolute value of the standardization deviation of all variables was less than 20% after matching

shows the test results:

Based on the aforementioned test results, the double difference method was employed, with the regression outcomes presented in

Table 13. Robustness test regression results.

Matching Method	Average Input per Hectare of Forest Land	Average Profit per Hectare of Forest Land	Net Forestry Income	Non-Agricultural Net Income	Net Income of Rural Households
Caliper matching	−3959.10 *** (−4.44)	1351.95 (0.26)	−2929.34 (−1.60)	6834.02 ** (3.77)	3960.60 ** (1.86)
Nuclear matching	−11,249.25 *** (−10.63)	168.15 (0.12)	−2701.24 (−1.51)	6828.04 *** (3.82)	4066.69 ** (1.95)
Individual effect	Controlled	Controlled	Controlled	Controlled	Controlled
Time effect	Controlled	Controlled	Controlled	Controlled	Controlled
Constant	Controlled	Controlled	Controlled	Controlled	Controlled

Note: ** and *** are significant at the level of 5% and 1%, respectively.

. The findings from both matching methods align with the baseline regression, indicating that farmers’ participation in forestry carbon sink projects can significantly enhance the non-agricultural net income and net income of rural households. Additionally, it notably reduces the average input per hectare of forest land while having no discernible impact on the net forestry income or average profit per hectare of forest land; furthermore, the baseline regression remains largely stable.

(2)

Placebo test

In order to further exclude the interference of other unknown factors on the experimental results and ensure that the results obtained in this paper are caused by the forestry carbon sink project, the placebo test was carried out in this study: new samples were randomly selected several times in the sample for regression. Specifically, this paper conducted 500 samples in the study sample as the virtual experimental group and control group. The kernel density distribution diagram of the test is shown in Figure 4 (from left to right are the kernel density distribution plots of the placebo test for Average input per hectare of forest land, Non-agricultural income, and Farmers’ net income), and most values are around 0, indicating that the significant correlation between independent variables and dependent variables is a small probability event, and the regression results are significant only when affected by policies. The changes in average input per hectare of forest land, non-agricultural income, and farmers’ net income are not random results but represent a robust causal effect caused by participation in forestry carbon sink projects.

4. Discussion

This paper employs economic theories, including the partial equilibrium model and utility function, to analyze farmers’ behavior in producing forestry carbon sinks. It further examines strategies to enhance farmers’ willingness to engage in forestry carbon sink projects through the lens of forest ecological theory. Utilizing 233 valid questionnaires collected from the LY Moso bamboo forest management carbon sink project site in Zhejiang Province, a PSM-DID model was established to assess the impact of forestry carbon sink projects on production factor inputs at the plot level, as well as their effects on various income streams and total income. The study suggests that after participating in forestry carbon sequestration projects, farmers can modify their forest management practices and reduce land management costs (

Table 14. A summary table of research hypothesis testing and conclusions.

Research Hypothesis (H)	Method (S)	Empirical Results (RC)	Conclusions (A)
H1: The participation of farmers in forestry carbon sink projects can lead to a transformation in forest management practices, resulting in reduced production costs for forestry activities.	PSM-DID	RC1: At the 99% confidence level, compared to plots that do not adhere to the guidelines of forestry carbon sequestration projects, the average investment per hectare for carbon sequestration forest plots decreased significantly by CNY 3840, while the average profit per hectare remained largely unchanged.	H1 passed the test. A1: after participating in forestry carbon sequestration projects, farmers can modify their forest management practices and reduce land management costs.
H2: Subsidies for forestry carbon sink projects and forestry production costs saved by farmers can make up for the loss of timber income, so that the average profit per hectare of forest land and net forestry income will not change significantly.		RC2: Participation in forestry carbon sequestration projects has no significant impact on forestry net income.	H2 passed the test. A2: the subsidies received from forestry carbon sequestration projects, combined with the savings in forestry production costs, can compensate for the loss of timber income, ensuring that farmers’ net forestry income remains unaffected.
H3: After participating in a forestry carbon sink project, farmers’ non-agricultural net income and farmers’ net income increased significantly.		RC3: Participation in forestry carbon sequestration projects has a significant positive impact on farmers’ non-agricultural net income and farmers’ net income, with confidence intervals of 99% and 95%, respectively.	H2 passed the test. A3: the conserved production resources can be redirected towards non-agricultural sectors, thereby enhancing non-agricultural income and ultimately increasing the net income of farmers involved in the forestry carbon sink initiative.

, A1). The subsidies received from forestry carbon sequestration projects, combined with the savings in forestry production costs, can compensate for the loss of timber income, ensuring that farmers’ net forestry income remains unaffected (Table 14, A2).

The findings of this study are generally consistent with existing research on farmers’ participation in forest management-based carbon sink projects. Consistent with the studies by Köthke and Dieter and Assmuth and Tahvonen [7,24], our results suggest that in order to enhance forest carbon sequestration, farmers are required to extend rotation periods and adopt selective logging practices. Furthermore, aligning with the findings of Roongtawanreongsri et al. and Gong [15,35], participation in forestry carbon projects contributes to livelihood transformation, facilitates an increase in off-farm income, and ultimately helps improve household income levels.

5. Conclusions

This study uses survey data from the Moso bamboo forest carbon sink project in LY County, Zhejiang Province, China, to examine how forestry carbon sink projects affect farmers’ production decisions and income levels, thereby sustaining farmers’ utility levels. Employing a partial equilibrium framework based on production and utility functions, combined with empirical analysis using the PSM-DID method, the results indicate that participation in forestry carbon sink projects leads to changes in farmers’ forest management practices. This study concludes that sustaining farmers’ utility levels and their willingness to engage in carbon sink production requires a multi-path approach involving both market mechanisms and government support. Under the combined effect of project dividends and government-supported cost-sharing mechanisms, forestry production inputs are reduced, and the income from carbon sink projects compensates for potential losses in timber revenue, thereby maintaining farmers’ net forestry income. The production resources saved by farmers can be reallocated to non-agricultural sectors, allowing farmers to obtain additional off-farm income and achieve overall income growth.

The main contributions of this study are as follows. First, unlike many existing studies that primarily focus on the direct income effects, this paper systematically analyzes the reallocation of production factors among forestry carbon sequestration, timber harvesting, and non-farm employment. It establishes a logical chain of “forest management transformation → production factor reallocation → income effects → participation willingness”. Second, by situating the analysis within the context of China’s collective forest tenure system and carbon market development, this study addresses the lack of research on how small-scale forestry farmers balance the provision of ecological public goods with the pursuit of private interests in an incomplete market environment. Notably, smallholder-based forestry management is also common in many developing countries. According to the FAO Global Forest Resources Assessment (F, 18%–29% of global forests are directly owned or managed by individuals, households, communities, or indigenous peoples, with large portions of forest resources in Latin America, Africa, and Southeast Asia being sustainably managed through community organizations and smallholder operations. (FAO, 2020: Global Forest Resources Assessment 2020: Main Report. Food and Agriculture Organization of the United Nations, Rome, Italy). Thus, the findings of this study offer not only policy relevance for China but also valuable insights for optimizing the institutional design of forestry carbon sink projects in smallholder-dominated regions worldwide.

Although this study provides important empirical support for the impact of forestry carbon sink projects on farmers’ production decisions and income, it still has certain limitations. The survey data used in this study may not fully capture the regional heterogeneity of forestry carbon sink projects across different areas. Future research could expand the scope of study to include samples from different provinces or countries, thereby improving the generalizability of the findings. Additionally, as the carbon trading market and government policies continue to evolve, this study does not fully account for future policy changes or carbon price fluctuations and their potential impact on farmers’ willingness to participate. Future studies could incorporate policy scenario analysis or economic simulation methods to explore the adaptability of forestry carbon sink projects under varying policy and market conditions.

Currently, forestry carbon sink projects have become an important breakthrough in China’s efforts to realize the economic value of ecological products; however, their institutional design still requires further refinement. Based on the findings of this study, the following policy recommendations are proposed to support the future development of forestry carbon sink projects.

First, in regions where the price of forestry carbon credits remains low, it is essential to improve the carbon trading system and pricing mechanisms, enhance price levels and stability, and thereby strengthen farmers’ confidence in participating in carbon sink projects. Second, a differentiated benefit-sharing mechanism should be established, with tiered compensation standards tailored to different types of forest operators, to ensure that carbon revenues sufficiently offset farmers’ opportunity costs. Third, the government can further assume part of the forest management costs by promoting technical support and providing production inputs such as fertilizers. In addition, effective training in carbon forest management—such as scientific tending and rational fertilization—should be offered to reduce farmers’ production costs and increase their returns. Furthermore, the siting of forest management carbon projects should prioritize collectively owned forest areas where harvesting is not the dominant activity and where marketization levels are relatively high. These regions typically face lower opportunity costs and have greater potential for labor mobility into non-forestry sectors. For farmers with a high dependence on forest income, targeted support for livelihood transition - through industrial subsidies, vocational training, and other policy tools - can facilitate their engagement in off-farm employment and reduce reliance on timber harvesting.