The impact of green subsidies on farmers’ low-carbon production transition

Achieving carbon neutrality is an important goal in the global response to climate change and sustainable development, and reducing greenhouse gas emissions is fundamental to carbon neutrality. Livestock farming, as an important part of agriculture, is also an important source of greenhouse gas emissions (Crippa et al., 2021; Jiang et al., 2023). Agricultural data released by the Food and Agriculture Organization of the United Nations (FAO) in 2019 shows that, only from the perspective of the farming link, global livestock farming emits more than 51% of the total carbon emissions in terms of CO₂ equivalent ^[1]. The warming effects of CH₄ and N₂O produced during livestock farming are 25 and 298 times higher than those of CO₂, respectively, and coupled with the implied carbon emissions from feed and factor inputs during the farming process, it is an absolute high-carbon industry (Rosenzweig et al., 2020). Therefore, the development of low-carbon industries to cope with global climate change and promote stable economic growth cannot be separated from the deep involvement of livestock and poultry farming.

Many studies have explored the impacts on individual low-carbon production transitions in terms of individual characteristics, household business characteristics, market prices, degree of organization, technological progress, access to information and contract farming dimensions (Hou and Hou, 2019; Jiang et al., 2023; Lu and Qiao, 2024; Song and Dou, 2024). First, from the perspective of individual characteristics, most scholars believe that farmers with higher education, longer years of farming experience and higher levels of ecological cognition have a stronger willingness to transition to low-carbon production (Huang et al., 2022; Jiang et al., 2022; Zheng et al., 2022). Second, from the perspective of household characteristics, scholars generally agree that labor endowment, farming income and farming scale have a significant impact on farmers’ low-carbon production transition (Du et al., 2023; Li and Huang, 2023; Ling et al., 2023). Finally, from the external environment construction factors, incentive and constraint policy factors are also included in the analysis framework, with the advancement of the research, scholars gradually began to conduct independent research on special variables, such as environmental regulation and production trusteeship (He et al., 2022; Chen and Li, 2024; Tang et al., 2024). Some scholars have also explored the impact of low-carbon production from a financial perspective. Low-carbon agriculture-related financial products can effectively absorb a large amount of idle capital in society, thus solving the financial problems required for the development of low-carbon agriculture (Jiang, 2021; Jiang et al., 2023). In addition, digital finance, as a fusion of modern technology and financial products, plays a role in low-carbon production transformation (Li et al., 2023; Hong et al., 2024). Therefore, exploring the dynamics that drive farmers’ low-carbon production transition is the key to promoting the green and low-carbon transition of animal husbandry.

However, the low-carbon production transition with intertemporal positive spillover externalities requires substantial capital and technology investment, and the uncertainty of the market and the risk of unrecoverable costs are relatively large. For farmers aiming to maximize private utility, they have insufficient internal motivation to actively participate in low-carbon production transition, which seriously impedes the pace of green and low-carbon transition of livestock husbandry.

According to the theory of environmental pollution externality, the implementation of environmental regulatory policies through government intervention can internalize the externality of environmental resources, so as to achieve the synergistic effect of livestock and poultry farming to reduce pollution and carbon emissions. Green subsidies are regarded as an effective means to balance the contradiction between environmental management costs and private benefits.

Currently, studies related to subsidy policies and low-carbon production have not yet reached a consistent conclusion. On the one hand, most studies believe that government subsidies have a positive effect on farmers’ low-carbon production, which can make up for farmers’ “cost of compliance” and thus stimulate the adoption and updating of low-carbon production technologies (Li et al., 2019; Wu et al., 2020). However, on the other hand, some studies have pointed out that although the subsidy policy can, to a certain extent, improve the incentive of low-carbon production of the main body of farming, the effect of relying on government support to generate income is limited due to financial constraints and the cost of low-carbon production (He and Zhang, 2024). In addition, there are few empirical studies on the aspect of green subsidies on low-carbon production transition of farmers, and existing studies mainly focus on business and energy (Wang et al., 2022; Tang et al., 2024). For the livestock sector, some studies have confirmed that green subsidies have a significant effect on farmers’ low-carbon production behavior (Ke and Huang, 2024). The provision of green subsidies by the government can incentivize economic actors to innovate technologically, which, in turn, will influence farmers’ decisions to adopt low-carbon production practices (Zhou et al., 2023). Existing studies focus on the impact of subsidy policies on farmers’ low-carbon production behavior, but it is scarce to directly explore the conversion mechanism and impact effect of green subsidies on farmers’ low-carbon production transition and there is a lack of in-depth exploration of the impact of green subsidies on the heterogeneity of low-carbon production transitions of farmers of different sizes.

Based on this, this paper takes microfarmers as the research landing point, adopts the endogenous switching regression model and empirically explores the effect of green subsidies on farmers’ low-carbon production transition under the counterfactual framework. The expansion of this paper lies in the following: first, the current research on low-carbon production transition of farmers focuses on a single dimension of factor input reduction, harmless treatment of waste and resource utilization of manure and sewage. On the basis of relevant research, considering the multidimensional and comprehensive characteristics of behavioral integration in low-carbon production transition, this paper constructs a comprehensive indicator system from the above three dimensions to make up for the shortcomings of the qualitative analysis in the existing research. Second, the existing studies only take green subsidies as a control variable that affects the production decision-making behavior of farmers, and lack the research that directly explores the effect of green subsidies on the low-carbon production transition of farmers. Based on the practical problems of insufficient endogenous motivation and poor carbon reduction effect of farmers’ low-carbon production transition, this paper explores the mechanism and effect of green subsidies on farmers’ low-carbon production transition from both theoretical and empirical levels, so as to enrich the research content of green subsidies in animal husbandry. Finally, existing international experiences are mostly based on large-scale agricultural systems, while China has a high proportion of small-scale farmers and there are differences in policy suitability. In addition, existing policy research focuses on the direct incentives of subsidies, this paper systematically decomposes the driving mechanism of low-carbon production transition based on microdata and by quantifying the extent of the mechanism’s impact, it makes up for the shortcomings of the current research in the literature and provides empirical evidence for the country to formulate a precise policy to help low-carbon production transition.

Farmers tend to pursue a balance of multiple objectives such as minimizing the risk of their behavior, maximizing profits and optimizing resource allocation when making decisions and their low-carbon production transition is a series of behavioral choices made under the combined effect of multiple factors such as capital endowment constraints and the external environment characteristics in the pursuit of maximizing personal returns. Therefore, the core of whether green subsidies can effectively promote farmers’ low-carbon production transition lies in whether green subsidies based on the ecological compensation mechanism can directly alleviate the production constraints of the farming subjects, and what kind of behavioral decision-making it will adopt after the constraints have changed. Based on this, the conversion mechanism of the impact of green subsidies on farmers’ low-carbon production transition as a financial transfer payment method can be summarized as follows.

Low-carbon production transition is a production decision with high inputs and risks, the uncertainty of expected returns will hinder farmers’ low-carbon production transition. Green subsidies enhance farmers’ certainty expectations of future returns from low-carbon production transition by releasing policy stability signals. On the one hand, the annual increase in the amount and form of green subsidies not only brings stable psychological expectations to farmers, but also reduces the uncertainty of low-carbon production transition benefits, thus prompting them to allocate more resources to livestock and poultry production and forcing them to adjust their production decision-making and input structure, which, in turn, affects the low-carbon production transition of farmers. On the other hand, green subsidies can, to a certain extent, alleviate the negative incentives triggered by rising factor prices and change the factor input behavior of farmers. From the perspective of cost reduction, whether farmers receive income support subsidies or production factor subsidies, they will reduce the cost of production inputs, thus promoting the increase of farmers’ operating income and stabilizing the expected cost returns of farmers (Chen et al., 2020). Based on this, this paper proposes H1:

It has been shown that green subsidies are based on the size of the farm and only when a certain size of the farm is reached, the relevant subject will be eligible for the subsidy (Pan et al., 2021). Large-scale farms are more likely to receive green subsidies from the government due to the advantage of the number of farms, and this incentive system of “more pay for more work” gives farmers a higher incentive to engage in large-scale farming (Guo et al., 2021). In addition, green subsidies directly change farmers’ budget constraints and production decisions, promoting them to invest more in production factors to expand the scale of their operations (Han et al., 2023). Farmers relying on the advantages of economies of scale can promote more effective integration and optimization of factors such as capital, technology and labor in the production stage, improve the efficiency of factor inputs, reduce the unit cost of production, stimulate their productive investment and implement more low-carbon production, thus forcing farmers to transform their low-carbon production. Accordingly, this paper proposes H2:

As the main body of livestock and poultry production, farmers’ breeding behavior is the production decision made by pursuing the maximization of economic benefits and economic benefits are the main incentive to drive farmers’ green and low-carbon production (Yang et al., 2022). As low-carbon production transition has positive externality attributes, relying only on market regulation will not realize the optimal allocation of resources. The mechanism of green subsidies is to internalize external benefits or compensate for external cost control, motivate the main body of farming to adopt low-carbon production technology or facilities and equipment in production and continue to play a positive environmental externality effect or inhibit the negative environmental externality effect, so as to achieve the unity of economic benefits and environmental benefits. Moreover, green subsidies can directly reduce the pressure of financial shortages in low-carbon production transition of farmers by using “income-oriented” cash subsidies to enhance the investment capacity of farmers, and promote the adoption of low-carbon technologies or the purchase of ancillary facilities (Liu et al., 2020), thus improving the level of their low-carbon production transformation. Therefore, this paper proposes H3:

The data used in this study originated from the field research on hog farming conducted in Sichuan Province and Shandong Province in 2021. To ensure the representativeness of the samples and the scientific nature of the survey, a combination of multistage stratified sampling and random sampling was used for data collection. The survey object is limited to family members who have a direct influence on the decision-making of hog farming, and the content of the survey mainly includes farmers’ access to green subsidies, low-carbon production status, basic characteristics, household characteristics and village environment characteristics. A total of 550 questionnaires were distributed during the data collection stage, and after data screening and eliminating questionnaires with illogical and outliers, 498 valid questionnaires were obtained, with a validity rate of 90.55%.

Since low-carbon production transition of farmers is a latent variable that cannot be directly observed, livestock and poultry farming behaviors that can achieve the effect of emission reduction are within the scope of this study. Excellent hog breeding breeds can shorten feeding cycle and increase the number of penning, which is conducive to improving production efficiency and reducing greenhouse gas emissions per unit of livestock products; scientific feeding methods and feed types can also reduce the feed-to-meat ratio and improve the feed conversion rate by improving the feed formula, thus realizing the reduction of the implied carbon emissions; epidemic prevention and inspection, as well as veterinary medicine, will reduce the probability of livestock and poultry diseases and deaths and thus reduce greenhouse gas emissions; the use of biogas, microbial fermentation and other facilities and equipment for the harmless treatment and resource utilization of livestock and poultry waste not only reduces the amount of greenhouse gases produced, but also substitutes the products of waste treatment for fertilizers, feeds and fuels, which further improves its carbon reduction effect. Combined with the above analysis, this paper follows the principles of source prevention, process control and terminal management and constructs the evaluation index system of low-carbon production transformation of farmers from the three dimensions of the degree of reduction of factor inputs, the degree of harmless treatment of waste and the degree of utilization of fecal sludge resources and the specific index system is shown in Table 1. Given the complexity and non-linear relationship between the indicators, this paper adopts the entropy weight index method to determine the weights of the indicators and calculates the weighted average to obtain the transformation degree of farmers’ low-carbon production to ensure the objectivity and accuracy of the evaluation.

China’s current green ecology-oriented agricultural subsidy policy system presents the characteristics of both “yield subsidies” and “green subsidies,” with the yield subsidy system fundamentally aiming to guarantee food security and farmers’ basic rights and interests. Green subsidies, by contrast, are financial incentives that the government uses as the goal of environmental protection, emphasize the realization of ecological sustainability of agricultural production and motivate economic agents to change their behavior (Wu et al., 2023). The green subsidies for the livestock and poultry industry studied in this paper are financial transfers implemented by the government in different forms of livestock and poultry farming practices to promote the transition to low-carbon production by farmers, drawing on the study of Zou et al. (2024), which restricts green subsidies to those that aim at the protection of the ecological environment as the fundamental purpose of subsidies, including subsidies for good seeds, subsidies for manure resource utilization subsidies for harmless treatment, subsidies for standardized breeding, subsidies for green equipment and subsidies for special funds for pollution control and if a farmer receives any of the above subsidies, it is considered to be supported by green subsidies and is assigned a value of 1. Otherwise, it is assigned a value of 0.

To ensure the identifiability of the model, this paper selects “the number of times the government promotes green subsidies in 2020” as an instrumental variable for empirical analysis. The reasons for using this instrumental variable are as follows: first, the more the government promotes the green subsidy policy, the more it helps farmers understand the conditions of the green subsidy policy and increases the possibility of obtaining the green subsidy, which meets the requirement of instrumental variable relevance; second, the number of times the government promotes the green subsidy policy will not directly affect the transformation of farmers’ low-carbon production, which meets the requirement of exogeneity of the instrumental variable.

Drawing on previous related studies (Si et al., 2022; Zeng et al., 2024), this paper selects variables that have a significant impact on green subsidy acquisition and low-carbon production transition as control variables from individual characteristics, household characteristics and external environment characteristics, respectively. In addition, considering the differences between regions, this paper introduces province dummy variables in the model.

The results of the parametric t-tests for the main characteristic indicators are shown in Table 2. Without controlling other variables, the difference in low-carbon production transition between farmers who received and did not receive green subsidies is significant at the 1% statistical level. In addition, the statistical results show that farmers who received green subsidies tend to exhibit characteristics such as higher education, lower risk appetite, higher share of breeding income, higher input of farming labor, participation in contract farming and use of the internet.

There are sample selection bias and endogeneity problems caused by omitted variables or reverse causality in the impact of green subsidies on the low-carbon production transition of farmers. Therefore, this paper uses the endogenous switching regression model proposed by Lokshin and Sajaia (2004) to assess the effect and extent of green subsidies on farmers’ low-carbon production transition.

To measure the impact of green subsidies on farmers’ low-carbon production transition, the following basic equation is constructed:

where Y_i is the farmers’ low-carbon production transition; X_i is control variables; green subsidies D_i is a dummy variable, D_i = 1 is that farmers received green subsidies, D_i = 0 is that farmers did not receive green subsidies; β_i, α is the parameter to be estimated; ε is the random error term.

Given that the variable D_i in equation (1) reflects the self-selection behavior of farmers based on the comparative production return analysis, there are some unknown factors that affect both farmers’ access to green subsidies D_i and low-carbon production transition Y_i. Therefore, the equation for farmers’ access to green subsidies is:

where $Di∗$ is the implied variable for receiving green subsidies D_i; Z_i is a set of exogenous explanatory variables affecting green subsidies, including age of the household head, education level; γ is the parameter to be estimated; and μ_i is the random error term.

Based on different endowment constraints, there are differences in farmers’ low-carbon production transition. When unobservable factors affect both green subsidies D_i and low-carbon production transition Y_i, it leads to a correlation between D_i and ε in equation (1), and thus direct estimation of equation (1) may lead to biased regression results. Therefore, the inverse Mills ratio λ_id, λ_in obtained by estimating equation (2) for green subsidy access is introduced into the equation for low-carbon production transition of farmers. For the whole sample, the potential low-carbon production transition with and without access to green subsidies can be expressed as follows:

where Y_id, Y_in are the low-carbon production transition of farmers with and without green subsidies, λ_id and λ_in are the inverse Mills ratios of farmers with and without green subsidies, respectively; $σμ2=var(μ)$⁠, σ_μd = cov(ε_d, μ); φ(·) and Φ(·) are the density function and distribution function following the normal distribution, respectively; σ_μn = cov(ε_n, μ), with $σμ2$ standardized to 1; and δ_id and δ_in satisfy the condition of zero mean. The endogenous switching regression (ESR) model is estimated by using the full-information maximum likelihood method to estimate the equations (2)–(4) in a joint manner. In addition, to ensure model identification, it is necessary to satisfy the condition that at least one variable in the vector Z of the green subsidy equation is not in the vector X of the low carbon production transition equation of the farmers and that the variable affects the access to green subsidy but does not directly affect the low carbon production transition of the farmers.

Based on equations (3) and (4) of the ESR model, the conditional expectations of low-carbon production transitions of farmers with and without access to green subsidies can be expressed as equations (5) and (6):

Meanwhile, their counterfactual scenarios for the conditional expectations of farmers who do not receive green subsidies if they do and the conditional expectations of farmers who do not receive green subsidies if they do transition to low-carbon production when they do can be expressed as equations (7) and (8):

The average treatment effect on the treated (ATT) of low-carbon production transitions for farmers who actually received green subsidies can be expressed as the difference between equations (5) and (7):

Similarly, the average treatment effect on the untreated (ATU) of low-carbon production transitions for farmers not receiving green subsidies can be expressed as the difference between equations (8) and (6):

The results of the estimation of the model linking the access to green subsidies and low-carbon production transition of farmers are shown in Table 3. In Table 3, Model (1) shows the estimation results of the influencing factors of farmers’ access to green subsidies, and Models (2) and (3) show the estimation results of the low-carbon production transition model for farmers who have access to green subsidies and those who have not access to green subsidies, respectively. The results of the likelihood ratio (LR) test of the independence of the two-stage equations indicate that the equation of obtaining green subsidies and the outcome equation are not independent of each other. The goodness-of-fit Wald test is significant at the 1% level, a result that justified the use of an endogenous transformation regression model. The error term correlation coefficients ρ₁ and ρ₀ are significant, indicating that there is a self-selection bias problem in the sample, i.e. whether farmers receive green subsidies or not is not randomly occurring and there are some unobservable factors.

The results show that among the personal characteristics, the level of education has a significant positive effect on access to green subsidies. Generally speaking, the higher the education level of the household head, the better his ability to obtain information and understanding, which helps him to understand and make himself meet the conditions for obtaining green subsidies. Among the household characteristics, number of years of farming, the share of breeding income, farming labor input and contractual agricultural participation have significant positive effects on access to green subsidies. Farmers who have been farming for a longer period of time tend to have a better understanding of government policies, which helps them to better comply with the policy requirements and thus increase the likelihood of receiving green subsidies. Typically, the higher the share of breeding income and the higher the input of farming labor, the larger the scale of family farming and the livestock and poultry farming industry is now shifting to scale and the government has introduced a series of supportive policies to encourage the development of large-scale farmers, so that large-scale farmers will receive more green subsidies. Participation in contractual agricultural increases farmers’ access to information, while leading enterprises or cooperatives help farmers meet the conditions for obtaining green subsidies, which increases their chances of obtaining green subsidies. Among the external environmental characteristics, financial support has a significant positive impact on access to green subsidies. Financial support can reduce the financial risk of farmers in the production process, and this sense of financial security makes farmers more willing to invest in green production technology and thus meet the conditions for applying for green subsidies. In addition, the province dummy variable is more significant, which may be due to differences in policy conditions in different regions, resulting in different manifestations of access to green subsidies.

To test the instrumental variable validity, this paper uses two-stage least squares method for instrumental variable testing. The first stage of the two-stage regression is significantly positive at the 1% level and the F-value of the first stage is 23.26, which excludes the weak instrumental variable problem. The results of the second-stage regression show that the impact effect remains significant after mitigating the endogeneity problem, thus confirming the validity of the instrumental variables.

In terms of individual characteristics, the age of the household head variable has a significant negative effect on the low-carbon production transition of farmers. This suggests that as the household head grows older, they tend to be more conservative in their thinking and show less adaptability and motivation for low-carbon production transitions. Risk preference has a significant positive effect on the low-carbon production transition of farmers. The low-carbon production transition process exhibits significant high-risk characteristics, and farmers with a higher degree of risk appetite are willing to undergo low-carbon production transition to obtain higher transition returns. In terms of household characteristics, the share of breeding income has a significant negative effect on low-carbon production transition for farmers not receiving green subsidies, which may be due to the fact that the higher share of breeding income of farmers indicates that the family income is mainly derived from farming, while the transition to low-carbon production has to bear higher transition costs and the high-risk income uncertainty makes their willingness to transition low. Social network has a significant positive effect on low-carbon production transition of farmers who do not receive green subsidies. Farmers who frequently interact with each other can share the technology and experience of low-carbon production transition, which helps to stimulate the behavior of low-carbon production transition. In terms of village characteristics, environmental regulation has a significant positive effect on the low-carbon production transition of farmers receiving green subsidies. The higher the environmental regulation, the stricter the government constraints on farmers’ production behaviors, thus prompting them to make the transition to low-carbon production. In addition, compared to Sichuan Province, farmers in Shandong Province have a higher degree of low-carbon production transition and the differences in low-carbon transition in different provinces may be related to their own resource endowments and policies.

This part further measures the treatment effect of green subsidies on low-carbon production transition through equations (5)–(10), and the results are shown in Table 4. Overall, access to green subsidies has a positive average treatment effect on farmers’ low-carbon production transition, where the ATT effect value passes the 1% significance level test, indicating that for farmers who have access to green subsidies, the level of low-carbon production transition would be reduced by 0.640 if they do not have access to green subsidies, taking into account the counterfactual scenario. And the ATU results show that the level of low-carbon production transition of farmers who do not receive green subsidies will increase by 0.521 when they receive green subsidies. This result suggests that green subsidies have a positive effect in promoting low-carbon production transition of farmers. The mechanism of the impact of green subsidies on low-carbon production transition of farmers will be further discussed below.

Considering the possible differences in the impact of green subsidies on the low-carbon production transition of farmers with different endowment constraints, this paper groups farmers according to their number of years of breeding and farming scale. Given the limitations of the sample size, to ensure the accuracy of the analysis, this paper adopts the median as the criterion of division and divides the sample into two groups of “above the median” and “below the median” for analysis. Based on the endogenous switching regression model, the differences in low-carbon production transition among farmers with different endowment constraints are shown in Table 5.

The results in Table 5 show that green subsidies have a significant effect on low-carbon production transitions, both above and below the median, but the low-carbon production transitions of farmers with longer farming years and larger farm scale is higher than that of farmers with shorter farming years and smaller farm scale. With the growth of farming years, farmers have accumulated rich farming experience and have a deeper understanding of the environmental impact factors in the farming process and this accumulation of experience helps farmers to understand the importance of low-carbon production transition, while large-scale farmers are more likely to achieve economies of scale in low-carbon production transitions and have higher levels of low-carbon production transitions.

Based on the results of the main regression analysis in the first part of the article, there is a significant promotion effect of green subsidies on farmers’ low-carbon production transition, but there is still a need for a more in-depth test of how green subsidies affect low-carbon production transition. As the article’s theoretical analysis points out, green subsidies mainly affect farmers’ low-carbon production transition through three mechanisms: the expectation effect, the scale effect and the wealth effect. First, the expected effect mechanism. This paper adopts “household income (taking logarithm)” to characterize the expected benefit, the higher the expected benefit, the more farmers are willing to carry out low-carbon production transition. Second, the scale effect mechanism. Green subsidies stimulate farmers to expand the scale of farming, which generates the scale effect, forcing farmers to adjust the farming structure and prompting them to carry out the low-carbon production transition and this paper adopts the “2020 hog slaughter (head)” to characterize the scale of farming. Third, the mechanism of wealth effect. This paper adopts “bio-feed cost, epidemic prevention cost, technology payment amount and equipment purchase and maintenance cost (taking logarithm)” to characterize the transition investment capacity and the green subsidies can increase the investment capacity of the farmers’ low-carbon production transition and promote the low-carbon production transition of the farmers. In view of this, this paper adopts the new mediation effect test process to test the existence of the above three influencing mechanisms by applying the hierarchical regression method respectively.

The results of Model (1) in Table 6 show that the regression coefficient of green subsidies on farmers’ low-carbon production transition is 0.172, and it is significant at 1% statistical level. The estimation results of Model (2) in Table 6 show that green subsidies have a significant positive effect on farmers’ expected returns. The regression coefficients of green subsidies and expected return in Model (3) passed the significance test, indicating that expected returns play a mediating role in the relationship between green subsidies affecting farmers’ low-carbon production transition, and it is a partially mediating effect and its proportion in the total effect is 0.0675, i.e. green subsidies can influence farmers’ low-carbon production transition by increasing their expected returns, which verifies the existence of the mechanism of the expected effect and H1 is verified. Due to the limitation of length, only the mechanism of the expected effect is analyzed in detail. Similarly, H2 and H3 are also verified. In summary, green subsidies can act on farmers’ low-carbon production transition through the mediating effects of expected return, scale effect and wealth effect and the mediating effects of the three paths account for 6.75%, 1.80% and 7.85% of the total effect, respectively.

To test the robustness of the above results, this paper uses the propensity score matching (PSM) method to estimate the effect of obtaining green subsidies on farmers’ low-carbon production transition and the results are shown in Table 7. It can be found from the calculation results that the three matching methods of PSM are consistent in estimating the effect of farmers’ low-carbon production transition, and the ATT passes the test at the 1% level of significance. The effects obtained from different matching algorithms are averaged and their average effect is 0.168, which indicates that the receipt of green subsidies has a significant effect on the degree of low-carbon production transition of the farmers, and this is consistent with the results of the estimation of endogenous switching regression model.

In the previous main regression section, green subsidy is used as a dichotomous variable, to further explore the impact of green subsidy level on farmers’ low-carbon production transition, this paper uses “green subsidy amount” and “green subsidy intensity” to replace the core explanatory variables and green subsidy intensity is characterized by the proportion of green subsidy amount to total subsidy amount. As shown in Table 8, the coefficients of green subsidy amount and green subsidy intensity on farmers’ low-carbon production transition are both significantly positive, indicating that the increase in the level of green subsidy can help to promote farmers’ low-carbon production transition, which verifies the main conclusions of this paper again.

To avoid the interference of other subsidy policies leading to the estimation bias of the benchmark model, this paper further introduces two policy dummy variables, namely, the subsidy for breeding sows and the support subsidy for large-scale farms, on the basis of the benchmark model, with the values of ATT and ATU of 0.577 and 0.523, respectively, and through the test of significance, the result still holds true.

Using microsurvey data from Sichuan and Shandong provinces, the paper empirically analyzes the effect of green subsidies on farmers’ low-carbon production transition based on the theoretical framework of the impact of green subsidies on farmers’ low-carbon production transition by using an endogenous switching regression model. The study found that: first, green subsidies can promote farmers’ low-carbon production transition, which is manifested in the counterfactual hypothesis situation, if the farmers who obtain green subsidies do not obtain green subsidies, the low-carbon production transition will decline; if the farmers who do not obtain green subsidies obtain, their low-carbon production transition will increase. This study echoes the studies of Wu et al. (2020) and Ke and Huang (2024), both of which concluded that government subsidies have a positive impact on farmers’ low-carbon production behavior and can stimulate the adoption and updating of low-carbon production technologies. Second, there are group differences in the impact of green subsidies on farmers’ low-carbon production transition, and green subsidies have a higher impact on the low-carbon production transition of farmers with longer farming years and larger farm scale. This finding confirms previous research that there is significant heterogeneity in the factors influencing low-carbon production transitions among farmers (Zou et al., 2024). Third, green subsidies can act on farmers’ low-carbon production transition either directly, but also through the partial mediating effects of expectation affect, scale effect and wealth effect, the mediating effects of the three paths accounted for 6.75%, 1.80% and 7.85% of the total effect, respectively. This is consistent with previous studies that government-provided green subsidies can increase farmers’ incomes and induce farming restructuring, thereby influencing farmers’ decisions to adopt low-carbon production practices (Guo et al., 2021; Yi et al., 2024).

Based on the above conclusions, the following policy insights can be obtained: first, develop green subsidy policies for low-carbon farming, while ensuring the stability and continuity of subsidy policies to enhance farmers’ confidence in their expectations of future returns. In addition, encourage small-scale farmers to expand their production scale through cooperation and mergers to realize the scale effect, and use more of the subsidy funds to support the upgrading of farmers’ low-carbon production facilities and technologies to ensure that the scale expansion is in line with the goal of low-carbon production transition and provide additional incentives or subsidies to farmers who have achieved remarkable results in low-carbon production transformation. Second, to enhance the degree of low-carbon production transition of farmers. The government can raise farmers’ awareness of and support for low-carbon production transition through publicity and education activities, and at the same time organize professional organizations and research institutions to carry out technical training and consulting services for farmers to help them master the use of skills and promote low-carbon production transition in the aquaculture industry. Third, when formulating green subsidy standards, gradient subsidy standards can be set and in addition to considering the scale of farming, the environmental protection measures, farming efficiency, technological innovation and other factors of farmers should be taken into account to achieve a more equitable distribution of subsidies. In addition, to avoid conflicts of interest among farmers of different sizes, policy publicity and communication should be strengthened, policy transparency should be improved, feedback from farmers on policies should be collected regularly and subsidy policies should be adjusted and improved in a timely manner to ensure that policies can better serve farmers.

The endogenous transformation model used in this paper can well solve the problems of sample selection bias and endogeneity, but due to the limitations of the model, the core explanatory variables selected can only be binary variables and other models can be selected to use the amount of green subsidies or the percentage of green subsidies to conduct an in-depth investigation into the causal relationship between green subsidies on the transformation of low-carbon production of farmers, so that the results can be more representative. In addition, because the data used are micro research data, research data based on specific time period data may not be able to capture the dynamic changes and long-term trends of green subsidies on low-carbon production transition and subsequently, large databases can be used to analyze the impact of changes in subsidy policies on the dynamic effects of low-carbon production transition in different time periods, which will help to understand dynamic adjustments in the implementation process of the policy and the long-term effects.

Funding: This study is supported by the Youth Fund of the National Natural Science Foundation of China under the project “Research on low-carbon production transition of farmers under the incentive of green subsidy policy: conversion mechanism, carbon reduction effect and policy optimization” (No. 72303122).