The Impact of Government Subsidies on the Environmental Performance of Agricultural Enterprises

Abstract

Facing the pressure of green transformation, studying the relationship between government subsidies and the environmental performance of agricultural enterprises has significant theoretical value and practical significance for achieving sustainable agricultural development. Based on the micro data of 283 A-share listed agricultural enterprises in China from 2013 to 2023, this paper empirically analyzes the impact of government subsidies on the environmental performance of agricultural enterprises and its mechanism. The results show that there is an inverted U-shaped relationship between government subsidies and the environmental performance of agricultural enterprises, that is, when the government subsidies are within a certain scale, increasing government subsidies will have a positive impact on the environmental performance of agricultural enterprises. When the government subsidy reaches a certain scale, increasing the government subsidy will have a negative impact on the environmental performance of agricultural enterprises. External media attention (EMA) and internal control level (IC) play mediating roles in the impact of government subsidies on the environmental performance of agricultural enterprises. Heterogeneity analysis showed that for different types of subsidies, R&D subsidies and environmental protection subsidies had an inverted U-shaped impact on the environmental performance of agricultural enterprises. This study provides useful implications for improving methods of issuing government subsidies and enhancing the driving force of agricultural enterprises to carry out sustainable development actions.

1. Introduction

Global climate change and environmental pollution have had a profound impact on the economic and social development of humanity. Promoting green, low-carbon, and sustainable development has become a global consensus [1,2]. In recent years, the concept of sustainable development embodied by ESG has gained widespread attention. An increasing number of enterprises are integrating environmental protection and other sustainability factors into their investment decisions and business strategies [3,4]. Notably, agriculture, as a cornerstone of human survival, is increasingly recognized as a critical sector in global climate governance and biodiversity conservation, due to its heavy reliance on natural resources like land and water. Agricultural enterprises, as key actors in this sector, should undertake green transformation to reduce carbon emissions, improve waste and wastewater treatment, and promote sustainable agricultural practices [5]. However, agricultural enterprises typically face challenges such as small scales, limited risk resilience, and inadequate financial resources, which hinder their environmental performance and make it difficult to meet stringent environmental protection and market compliance requirements. In this context, government subsidy policies have emerged as a crucial driving force for improving the environmental performance of agricultural enterprises, particularly when financial constraints are tight. Subsidies can alleviate the pressure of environmental protection investments, thereby strengthening enterprises’ capacity for environmental governance. Consequently, a comprehensive analysis of the impact of government subsidies on the environmental performance of agricultural enterprises, and the mechanisms at play, is essential for achieving sustainable agricultural development.

As a major agricultural country, China has long relied on traditional agricultural production models, which have led to serious ecological issues such as soil degradation, water shortages, agricultural pollution, and carbon emissions. Particularly during the process of modernization, agriculture faces a series of challenges, including overcapacity, environmental pollution, and resource wastage. According to statistics, the intensity of chemical fertilizer use in current agricultural production in China has reached 400 kg per hectare, the effective utilization rate of pesticides is only about 38%, and the annual residue of agricultural films exceeds 450,000 tons. The ecological externalities of traditional agricultural practices have, to some extent, hindered the achievement of sustainable development goals. Although the Chinese government has set targets for “carbon peak and carbon neutrality” and has promoted the green transformation of agriculture through policy initiatives, agricultural enterprises continue to face numerous obstacles in their efforts to achieve this transformation. Due to the long production cycle and extended payback periods in agriculture, investments in green technologies and environmental protection facilities typically fail to deliver immediate returns. As a result, agricultural enterprises often lack the intrinsic motivation to prioritize environmental improvements and are heavily reliant on external resources and policy support.

To address these challenges, the Chinese government has been actively promoting the green transformation of agriculture through initiatives such as financial subsidies and support for green technology research and development. At the strategic level, ecological governance has been integrated into the broader agricultural development agenda through policies like the “Technical Guidelines for Green Development in Agriculture.” In 2023, the central government allocated over CNY 12 billion in subsidies for agricultural resource and ecological protection, focusing on projects such as crop rotation and fallow, organic fertilizer substitution for chemical fertilizers, and agricultural film recycling. However, according to resource dependence theory, the relationship between agricultural enterprises and their external environment is both dynamic and complex. Enterprises not only passively receive resources but also actively leverage external resources to maximize their own benefits. As government subsidies, a key external resource, continue to grow, agricultural enterprises may, driven by profit motives, reduce their investment in environmental protection. This could lead to a significant decrease in the effectiveness of environmental protection actions. So, can government subsidies enhance the environmental performance of agricultural enterprises? Is it true that more subsidies always lead to better outcomes? What exactly is the nature of the relationship between the two? Exploring this issue in depth not only helps optimize China’s agricultural green governance tools but also offers valuable insights for other developing countries facing pressures for industrial green upgrading and exhibiting characteristics of fiscal dominance. Particularly in balancing fiscal incentives with enterprise autonomy and preventing subsidies from being distorted into short-term profit-seeking behavior, China’s experience could provide important references for designing more sustainable agricultural environmental policies globally.

At present, research on the influencing factors of environmental performance of agricultural enterprises is relatively scarce. Most studies related to environmental performance take heavily polluting enterprises and industrial enterprises as the research objects, and few focus on this important and special group of agricultural enterprises. This is not only detrimental to the improvement of the sustainable development capacity of agricultural enterprises but also to the high-quality development of the entire agricultural industrial chain. Secondly, the current research conclusions on the impact of government subsidies on the environmental performance of enterprises are still inconsistent. Some scholars believe that government subsidies can improve environmental performance by alleviating financial pressure and stimulating the research and development of green technologies. Others believe that subsidies may reduce the intrinsic environmental protection motivation of enterprises, create a “dependency effect”, or even shift towards the minimum compliance standards, thereby suppressing environmental performance. The reason for the divergence in the above viewpoints might be that most studies hold that there is a linear relationship between government subsidies and the environmental performance of enterprises, while ignoring the possible nonlinear relationship between them.

Based on the panel data of China’s A-share agricultural listed enterprises from 2013 to 2023, this paper reveals the impact mechanism of government subsidies on the environmental performance of agricultural enterprises. Firstly, the dual fixed effects model was used to empirically test the nonlinear impact of government subsidies on the environmental performance of agricultural enterprises. Secondly, the transmission mechanism between government subsidies and the environmental performance of agricultural enterprises was comprehensively identified from the two core mechanisms of external media attention and internal control level. Finally, in order to further identify the impact of government subsidies on the environmental performance of agricultural enterprises under different levels of subsidies, the heterogeneity test was conducted.

The possible innovations of this paper are as follows: Firstly, this paper focuses the research object on agricultural enterprises and explores the influence mechanism of government subsidies on the environmental performance of agricultural enterprises based on the resource dependence theory. Secondly, based on the theory of diminishing marginal returns, this paper broadens the channels through which government subsidies have an inverted U-shaped impact on the environmental performance of agricultural enterprises from two paths: external media attention and internal control level. Thirdly, this paper classifies government subsidies into research and development subsidies and environmental protection subsidies, analyzes the differences in the impact of government subsidies on the environmental performance of agricultural enterprises under different types of subsidies, enriches the academic community’s understanding of the impact of government subsidies on the environmental performance of agricultural enterprises, and provides practical guidance for the precise formulation of subsidy policies and the efficient promotion of sustainable development.

2. Theoretical Analysis and Research Hypotheses

2.1. Research on the Environmental Performance of Enterprises

The concept of corporate environmental performance comes from externality theory. Pigou and Coase successively applied the externality theory to analyze environmental pollution and, respectively, proposed Pigouvian tax and Coase Theorem [6], that is, to maximize the overall social welfare by increasing the tax cost of enterprises and clarifying the property rights [7]. However, neoclassical economists believe that the primary goal of enterprises is to maximize shareholder value, and overemphasis on social responsibility will lead to low efficiency of resource allocation. There is no denying the fact that enterprises’ one-sided pursuit of profit will produce irreversible negative externalities of pollution to the environment. Some scholars have discussed the influencing factors of enterprise environmental performance from the perspective of internal environment and external environment. In terms of the internal environment of enterprises, the scale of enterprises [8], the composition of the board of directors [9,10], management characteristics [11,12,13], corporate innovation [14,15] and other factors play an important role in promoting corporate environmental performance. In terms of the external environment of enterprises, investors’ attention [16], environmental protection tax [17,18], environmental regulation [19,20], and other factors will have a significant impact on corporate environmental performance.

2.2. Research on the Economic Consequences of Government Subsidies

Government subsidy refers to the financial or resource support provided by the government to enterprises or industries in the form of financial transfer payment, tax reduction and exemption, aiming to guide resource allocation and promote the realization of public goals [21]. The existing research on the economic consequences of government subsidies mainly focuses on its impact on enterprise performance, investment, innovation, and other aspects. Some scholars hold the view of the “supporting hand”, where government subsidies not only promote the green technological innovation and transformation of enterprises by reducing the cost of green production and enhancing the green level of products but also help improve the digital transformation level of upstream and downstream enterprises through collaborative transmission among industrial chains [22]. There are also scholars who hold the view of the “predatory hand” that government subsidies lead to an increase in corporate debt ratio, the continuous decline of corporate operating efficiency, the failure to form long-term effective innovation power, and the weakening of independent innovation capacity [23]. Whether government subsidies can improve corporate environmental performance has not yet reached a consensus in the academic community. Some scholars believe that government subsidies will improve enterprises’ investment in environmental protection [24,25] and promote enterprises to fulfill their social responsibilities [21,26] and improve its social value [27], which can alleviate the problem of corporate financing constraints [18,28] to enhance their commitment to environmental responsibility. However, some scholars believe that government subsidies have no substantial help for enterprises to undertake environmental responsibility [29,30] and higher government subsidies will inhibit enterprises from fulfilling environmental responsibility [31]. Ren et al. [32] pointed out that government subsidies will lead to an increase in production factor costs and misallocation of resources, which is not conducive to enterprises’ green innovation.

2.3. Relationship Between Government Subsidies and Environmental Performance of Agricultural Enterprises

Resource dependence theory holds that each organization does not exist independently and must obtain necessary resource support through interaction with the external environment [9]. At the same time, depending on the importance and scarcity of resources, organizations have different degrees of dependence on the external environment [33]. Therefore, in a specific scenario, organizations need to communicate, interact, and negotiate with resource controllers in the external environment to shape a favorable environment for survival and development. Similarly, the healthy development of agricultural enterprises cannot be separated from the support of external resources. When carrying out environmental governance, agricultural enterprises are mainly faced with problems such as high sunk ecological input and insufficient endogenous power caused by the long cycle of market return. The Environmental Kuznets Curve (EKC) theory indicates that the relationship between environmental pollution and economic development is not a simple linear one but rather follows an inverted U-shaped trajectory of “deterioration first, then improvement” as the economy grows. The underlying mechanism lies in the fact that in the early stage of economic development, in pursuit of growth and profit, enterprises often sacrifice environmental quality. When the economy develops to a certain level, society’s attention to the quality of the ecological environment increases. Technological progress, policy supervision, and corporate responsibility awareness jointly drive the improvement of environmental quality. Therefore, for the initial stage of agricultural enterprises, moderate government subsidies can reduce the marginal cost of green investment for agricultural enterprises, promote them to strengthen pollution control, optimize resource allocation, and thereby enhance environmental performance.

As a form of policy-driven resource injection, government subsidies can enhance environmental performance in the early stages through three main channels. First, they help alleviate constraints on specific asset investments, enabling agricultural enterprises to overcome the initial barriers to clean technology R&D. Second, subsidies enhance the symbolic capital of environmental legitimacy by providing policy endorsement, thereby reducing frictional costs in interactions with environmental organizations. Third, subsidies facilitate the construction of resource exchange networks, strengthening enterprises’ bargaining power in collaborative innovation with research institutions. At this stage, government subsidies produce a resource enhancement effect, whereby environmental performance improves linearly with the intensity of policy support.

When the intensity of subsidies surpasses the critical threshold of an organization’s absorptive capacity, the negative effects of resource dependence begin to dominate. According to the power asymmetry principle of resource dependence theory [34], excessive subsidies lead to a restructuring of government–enterprise relations. The government, through its control over resources, intervenes in firms’ environmental decision-making, causing strategic attention to shift from long-term technological accumulation to short-term compliance investments. At this point, enterprises may suffer from three types of efficiency losses. First, policy resources substitute for firms’ independent R&D investment, creating a vicious cycle of “subsidy dependence–capability degradation” [35]. Second, managers tend to allocate subsidies to easily observable end-of-pipe governance projects, while avoiding the development of less visible emission-reduction technologies. Third, excessive subsidies reduce the opportunity cost of environmental non-compliance, inducing symbolic governance strategies. During this stage, the marginal cost of improving environmental performance exceeds the benefits of policy incentives, and the slope of the performance curve turns from positive to negative. Therefore, the following hypotheses can be put forward:

Hypothesis 1. 

There is an inverted U-shaped relationship between government subsidies and the environmental performance of agricultural enterprises.

2.4. Mediating Effect of EMA

External supervision exerted through media attention can compel agricultural enterprises to strengthen self-discipline. In order to preserve their reputation and public image, enterprises under high media scrutiny are more motivated to fulfill their environmental responsibilities [24,36]. As an effective complement to internal corporate governance, media oversight helps curb managerial opportunism, encourages management to respond to stakeholder concerns, and ultimately enhances governance quality [37]. The level of media attention directed toward agricultural enterprises is influenced by the amount of government subsidies they receive. As previously discussed, when subsidies remain within a moderate range, media attention may increase alongside subsidy intensity, thereby stimulating green transformation efforts [38]. First, government subsidies serve as a form of official endorsement, enhancing the credibility and market recognition of agricultural enterprises and attracting more favorable media coverage. Second, the positive signals conveyed by subsidies can improve access to external financing and increase the likelihood of sustained media interest. However, as the scale of government subsidies continues to expand, the level of media attention received by agricultural enterprises may begin to decline. First, excessive subsidies can foster over-reliance on external resources, undermining the firms’ endogenous capabilities and weakening their innovation momentum. Consequently, the media may perceive these enterprises as lacking novelty or developmental value, leading to reduced coverage. Second, the sustained increase in government subsidies may cause the public to experience “aesthetic fatigue” toward policy-driven narratives, diminishing interest in related news content and further reducing media attention. Based on this, this paper puts forward Hypothesis 2.

Hypothesis 2. 

EMA plays a mediating role in the impact of government subsidies on the environmental performance of agricultural enterprises.

2.5. Mediating Effect of Internal Control Level

Internal control ensures that agricultural enterprises comply with legal and regulatory requirements, maintain the accuracy and integrity of financial reports, and enhance operational efficiency. As a crucial risk management mechanism, effective internal control directly impacts the environmental performance of agricultural enterprises. It supports compliance management, reduces environmental pollution and resource waste, and helps fulfill corporate environmental responsibilities [39,40]. The impact of government subsidies on the internal control of agricultural enterprises depends on the subsidy size. When subsidies are within a certain range, they provide financial support that enhances the operational efficiency of agricultural enterprises, reduces operational risks, optimizes resource allocation, and strengthens internal control. Additionally, as an external institutional pressure, government subsidies encourage enterprises to improve their internal control systems and meet regulatory requirements by fostering self-discipline. However, as the scale of government subsidies expands, agricultural enterprises may become increasingly dependent on them, leading to inertia. Driven by short-term interests, enterprises may prioritize obtaining higher subsidies over improving environmental performance, undermining internal control mechanisms and contradicting the original intent of the subsidy policy. Moreover, higher compliance requirements may increase internal control costs, thereby diverting resources from green transformation investments. As a result, when subsidy levels become excessive, the effectiveness of internal control may decline, showing an inverted U-shaped relationship between subsidy scale and internal control. Based on this, this paper puts forward Hypothesis 3.

Hypothesis 3. 

The level of internal control plays a mediating role in the impact of government subsidies on the environmental performance of agricultural enterprises.

3. Methods and Data

3.1. Sample Selection and Data Sources

This paper takes China’s A-share agricultural listed enterprises from 2013 to 2023 as samples for research. Specifically, according to the 2012 industry classification standard of China Securities Regulatory Commission (CSRC), this paper selects “agriculture, forestry, animal husbandry and fishery” enterprises. And, in the manufacturing industry, “agricultural and sideline food processing industry”; “food manufacturing industry”; “wine, beverage and refined tea manufacturing industry”; “textile industry”; “leather, fur, feathers and their products and footwear industry”; and “wood processing and wood, bamboo, rattan, palm and grass products industry” are selected as the research objects. ESG rating data of China Securities are from the Wind database, government subsidies and control variables are from the CSMAR database, internal control level data are from the Dibo Internal Control and Risk Management database, and external media attention data are from the China Research Data Service Platform (CNRDS). The data samples used in this paper are processed as follows: (1) excluding the samples of enterprises with special treatment of stocks and delisting early warning stocks; (2) eliminating the samples with abnormal data; and (3) eliminating the samples with missing indicators of main variables. Finally, this paper obtains 2280 observed values of 283 listed agricultural enterprises. In addition, in order to eliminate the influence of outliers, the continuous variables are winsorized at the level of 1% and 99%.

3.2. Variable Description

3.2.1. Explained Variable

Environmental performance of agricultural enterprises (EP). In recent years, the evaluation content of environmental performance includes not only clean air and water but also the environment, education, and other social fields in non-economic variables [41]. Measurement indicators are a key part of environmental governance, and the establishment of relevant and reasonable information base will help to make decisions [42]. ESG embodies the sustainable development concept of integrating corporate governance performance with environmental performance, which is consistent with China’s “dual carbon” goal and vision. Specifically, the index comprehensively considers the impact of enterprises on the environment, including how to control the emission of various pollutants and increase the use of various clean energy in the process of production and operation, as well as the treatment of waste generated after the process of production and operation. Therefore, the existing literature often uses corporate social responsibility score and China Securities’ ESG rating index to characterize the environmental performance of enterprises [43,44]. The ESG rating comprehensively reflects China’s environmental policy orientation, focusing on indicators such as carbon neutrality route, water consumption, waste discharge, and renewable energy and setting different index weights according to the characteristics of the enterprise industry. Due to its wide coverage, strong traceability and accurate calculation, ESG rating has become one of the most authoritative indicators to measure the environmental performance of enterprises. Therefore, in order to more accurately measure the environmental performance of enterprises, this paper selects the annual average of the environmental scores in the ESG rating of China Securities to represent the environmental performance of enterprises.

3.2.2. Explanatory Variable

Government subsidy (GS). The types of government subsidies obtained by agricultural listed enterprises mainly include agricultural support and protection subsidies, rural industrial development subsidies, agricultural science and technology development subsidies, agricultural green development subsidies, etc. According to the data of CSMAR, this paper uses the proportion of total government subsidies in the total assets of agricultural enterprises to measure the core explanatory variable. At the same time, in order to test the difference in the impact of different types of government subsidies on the environmental performance of agricultural enterprises, in the heterogeneity analysis part, this paper divides the government subsidies into R&D subsidies, non-R&D subsidies, environmental protection subsidies, and non-environmental protection subsidies, so as to explore the difference in the impact of different government subsidies on the environmental performance of agricultural enterprises.

3.2.3. Mechanism Variables

External media attention (EMA) and internal control level (IC). In order to further explore the mechanism of government subsidies influencing the environmental performance of agricultural enterprises, this paper uses external media coverage and internal control level as mediating variables. This paper selects online financial news and newspaper financial news to measure the external media supervision. The specific treatment method is to add 1 to the logarithm of the number of media coverage obtained by agricultural listed enterprises. Considering that investors, investment institutions and relevant stakeholders mainly rely on official financial news to obtain information, and the content of “We-media” tends to be entertaining, even if it reports relevant financial news, its credibility is low, so this paper does not include “We-media” into the consideration of external media attention. For the level of internal control, this paper uses the “internal control index” of Dibo Big Data Center to represent the internal control of enterprises [45]. In order to ensure that the coefficient value of regression results is appropriate, the values of external media attention and internal control level are reduced by 100 times.

3.2.4. Control Variables

This paper refers to relevant research [33,43,44], taking firm size (SIZE), firm age (AGE), leverage (LEV), net profit margin on total assets (ROA), growth rate of operating income (GROWTH), and ownership concentration (TOP5) as control variables, while controlling year and industry as fixed effects. The variable definition and descriptive statistics are shown in

Table 1. Variable definition and descriptive statistics.

Variables	Variable Symbols	Definition of Variables	Mean	SD	Min	Max
Environmental performance	EP	Annual average of environmental (E) scores in ESG ratings of China Securities	62.763	6.303	44.000	82.230
Government subsidies	GS	Ratio of total government subsidies received by enterprises to total assets of agricultural enterprises (%)	0.912	1.106	0.000	6.733
External media attention	EMA	The amount of media coverage a business gets	3.490	9.968	0.020	165.41
Internal control	IC	Enterprise Dib internal control index	7.339	1.208	0.000	9.978
Enterprise size	SIZE	Natural logarithm of total assets	21.943	1.265	19.098	25.962
Years on the market	AGE	The year of establishment of the business up to the current period	11.326	6.364	2.000	29.000
Leverage	LEV	Asset–liability ratio	0.382	0.199	0.061	0.880
Net profit margin on total assets	ROA	Net profit/average balance of total assets	0.048	0.065	−0.231	0.253
Concentration of ownership	TOP5	Shareholdings of the top five shareholders of the enterprise	0.441	0.149	0.174	0.886
Corporate growth	GROWTH	Operating income growth rate	0.147	0.374	−0.554	3.669

.

3.3. Model Setting

In order to test Hypothesis 1, this paper constructs a model to investigate the nonlinear relationship between government subsidies and the environmental performance of agricultural enterprises.

$${EP}_{i,t}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{GS}_{i,t-1}+{\mathsf{\alpha }}_2{GS}_{i,t-1}^2+{\mathsf{\alpha }}_n{X}_{i,t}+g_{\mathit{Industry}}+g_{Year}+{\mathsf{\epsilon }}_{i,t}$$

(1)

In the above formula, ${EP}_{i,t}$ represents the environmental performance of agricultural enterprise “i” during period “t” and ${GS}_{i,t}$ represents the government subsidy. Due to the lag effect of government subsidies, this paper uses the government subsidy lagging by one period for regression; ${GS}_{i,t-1}^2$ is the squared term of government subsidy. $X_{i,t}$ represents various control variables; $g_{\mathit{Industry}}$ and $g_{Year}$ are industry fixed effects and year fixed effects, respectively; and ${\mathsf{\epsilon }}_{i,t}$ represents the random disturbance term.

To test Hypothesis 2 and Hypothesis 3, this paper constructs the following model to investigate the effect mechanism of government subsidies on the environmental performance of agricultural enterprises:

$$M_{i,t}={\beta }_0+{\beta }_1{GS}_{i,t-1}+{\beta }_2{GS}_{i,t-1}^2+{\beta }_n{X}_{i,t}+g_{\mathit{Industry}}+g_{Year}+{\mathsf{\epsilon }}_{i,t}$$

(2)

$${EP}_{i,t}={\gamma }_0+{\gamma }_1{GS}_{i,t-1}+{\gamma }_2{GS}_{i,t-1}^2+{\gamma }_3{EMA}_{i,t}+{\gamma }_4{IC}_{i,t}+{\gamma }_n{X}_{i,t}+g_{\mathit{Industry}}+g_{Year}+{\mathsf{\epsilon }}_{i,t}$$

(3)

In Formula (2), $M_{i,t}$ is the level of external media attention and internal control; ${\beta }_0$ is the constant term; ${\beta }_1 ,{\beta }_2,{ \mathrm{and} \beta }_n$ are the parameters to be estimated; and ${\mathsf{\epsilon }}_{i,t}$ is the random disturbance term. If ${\beta }_2$ is negative, the slope of the curve is positive at the minimum value of ${ \mathit{GS}}_{i,t-1}$ and negative at the maximum value of ${ \mathit{GS}}_{i,t-1}$, and the inflection point of the curve is within the range of values of ${ \mathit{GS}}_{i,t-1}$. It shows that there is an inverted U-shaped relationship between government subsidies and external media attention and internal control level. In Equation (3), mediating variables are added on the basis of Equation (1). If ${\gamma }_1$, ${\gamma }_2,{ \gamma }_3$, and ${\gamma }_4$ are all significant, it indicates that external media attention and the level of internal control play partial mediating roles in the inverted U-shaped relationship between government subsidies and the environmental performance of agricultural enterprises.

4. Empirical Results

4.1. The Impact of Government Subsidies on the Environmental Performance of Agricultural Enterprises

To examine the impact of government subsidies on the environmental performance of agricultural enterprises, first of all, this paper selects the fixed effects model and the random effects model to estimate the samples. Secondly, a Hausmann test was conducted based on the estimation results of these two models, and the test results rejected the null hypothesis. Therefore, it is appropriate to select a fixed effects model to estimate the samples. The regression results of the double fixed effects model are shown in

Table 2. Government subsidies and environmental performance of agricultural enterprises.

Variables	EP	EP	EP
	(1)	(2)	(3)
GS		0.571 ***	0.574 ***
		(2.831)	(2.850)
GS2		−0.097 *** (−3.533)	−0.098 *** (−3.479)
SIZE	1.231 ***	1.031 ***	1.030 ***
	(3.846)	(3.710)	(3.721)
AGE	−0.015	−0.015	−0.158
	(−0.810)	(−0.810)	(−0.811)
LEV	−0.092 *	−0.090 *	−0.090 *
	(−1.720)	(−1.714)	(−1.718)
ROA	2.021 **	2.020 **	2.136 **
	(2.081)	(2.076)	(2.078)
TOP5	−0.001	−0.001	−0.001
	(−1.443)	(−1.440)	(−1.440)
GROWTH	−0.034 *	−0.026	−0.024
	(−1.846)	(−1.615)	(−1.611)
Industry-FE	NO	NO	YES
Year-FE	NO	NO	YES
Hausman test			χ2 = 28.38 p = 0.001
N	2128	1952	1952
R2	0.181	0.199	0.244

Note: ***, **, and * represent the significance level at 1%, 5%, and 10%, respectively; T-values in parentheses; column (2) uses the government subsidy data lagged by one period, and the number of observations decreases.

. Among them, column (1) shows the regression results with only the control variables added; column (2) shows the regression results with the control variables, core explanatory variables and their square terms added; and column (3) shows the regression results with further control for the fixed effects of the year and the industry. The regression results in column (3) of Table 2 show that the estimated coefficient of government subsidy is 0.574; the estimated coefficient of the square term of government subsidy is −0.098, both of which are significant at the level of 1%; and the explanatory power of the model is significantly enhanced. Therefore, government subsidies can promote agricultural enterprises’ environmental performance, and the influence of the government subsidies for agricultural enterprises environmental performance is inverted U-shaped. When the government subsidy is within a certain scale, increasing the government subsidy will have a positive impact on the environmental performance of agricultural enterprises. When the government subsidy exceeds a certain scale, increasing the government subsidy will have a negative impact on the environmental performance of agricultural enterprises. Hypothesis 1 is verified.

4.2. Robustness Test

This paper uses a variety of methods to conduct robustness tests, and the results show that the conclusion that there is an inverted U-shaped relationship between government subsidies and the environmental performance of agricultural enterprises is robust.

4.2.1. Instrumental Variables

Given that government subsidies are targeted, agricultural enterprises with better environmental performance may be more likely to receive such subsidies. Therefore, there may exist a bidirectional causal relationship between government subsidies and the environmental performance of agricultural enterprises. Although this study employs a fixed effects model as the primary estimation method to control for unobservable, time-invariant heterogeneity at the firm level, this approach cannot fully address endogeneity arising from time-varying factors. Consequently, we further adopt an instrumental variable approach as a robustness check. Specifically, referring to the practice of Xia et al. [46], this paper selects “the straight-line distance between the location of agricultural enterprises and Beijing” as the instrumental variable. The reasonableness lies in the following aspects: on the one hand, Beijing is the political, economic, and cultural center of China, and the straight-line distance between the location of agricultural enterprises and Beijing may affect the decision of policy makers to provide government subsidies to them [46]. On the other hand, the straight-line distance from the location of agricultural enterprises to Beijing is determined by geographical location, which is usually exogenous and meets the requirements of exogeneity of instrumental variables. Columns (1) to (3) of

Table 3. Robustness test.

Variables	GS	GS2	EP	EP	EP	EP
	(1)	(2)	(3)	(4)	(5)	(6)
DIS	−2.366 ***	−12.531 ***
	(−3.610)	(−2.992)
DIS2	0.319 ***	14.749 ***
	(3.486)	(3.044)
GS			0.613 ** (2.317)	0.649 *** (3.247)	0.578 *** (2.956)	0.579 *** (2.956)
GS2			−0.116 *** (−3.103)	−0.010 *** (−3.481)	−0.098 *** (−3.480)	−0.010 *** (−3.482)
Controls	YES	YES	YES	YES	YES	YES
Industry/Year	YES	YES	YES	YES	YES	YES
Industry × Year	NO	NO	NO	NO	NO	YES
LM statistic	28.056 ***
Wald F statistic	21.706
N	1952	1952	1952	1560	1464	1896
R2	0.247	0.288	0.071	0.243	0.259	0.318

Note: ***, and ** represent the significance level at 1%, and 5%, respectively; T-values in parentheses.

show the regression results of the two-stage least squares estimation method. The regression results of the first stage are shown in columns (1) and (2). The regression coefficients of the instrumental variable with respect to the explanatory variable are both significant at the 1% level, indicating that the selection of the instrumental variable meets the correlation assumption. Column (3) shows the regression results of the second stage, and the impact of government subsidies on the environmental performance of agricultural enterprises remains significant. Finally, the Kleibergen–Paap rk LM statistical p value at the 1% significance level rejected the null hypothesis and passed the “insufficient identification of instrumental variables” test. The Cragg–Donald Wald F statistic of the weak instrumental variable test was 21.706, which was much higher than 16.38, indicating that the hypothesis of the weak instrumental variable was rejected. Therefore, Hypothesis 1 remains valid.

4.2.2. Propensity Score Matching (PSM) Test

In order to deal with the endogeneity problem caused by sample selection bias, the propensity score matching method is used in this paper. The specific approach is to select the median of government subsidies as the dividing line, with agricultural enterprises that receive lower government subsidies as the control group and those that receive higher government subsidies as the experimental group. Then, all the values of government subsidy data below the median are set to 0 and all the values above the median to 1. The control variable are selected as the matching variable and the nearest neighbor matching method is used for 1:4 nearest neighbor matching. The results of the balance test show that the standard deviations of the propensity scores of the control variables in the matching group after matching are all below 10%, indicating good balance and no significant differences. Then, Logit regression is performed on the matched data. The results are shown in column (4) in Table 3. The government subsidy is significant and the coefficient direction is positive, and the government subsidy square term is significant and the coefficient direction is negative. Therefore, after using PSM to deal with the endogeneity problem caused by sample selection bias, Hypothesis 1 is still valid.

4.2.3. Sample Change

Considering that municipalities directly under the central government are directly under the management and supervision of the central government and enjoy more financial resources and special policy support, this paper excludes the samples of municipalities for testing. The regression results show that Hypothesis 1 is still valid.

4.2.4. Model Transformation

In order to exclude the influence of the external factors that change with the industry year by year on the government subsidy and the environmental performance of agricultural enterprises, this paper replaces the industry fixed effect and year fixed effect in the original model with the interaction fixed effect of industry and year as a means of robustness testing. After replacing the interaction effect of industry and year with the interaction effect of industry and year in column (6) of Table 3, the government subsidy still significantly positively improves the environmental performance of enterprises. The results also show that the benchmark regression is robust to a certain extent.

4.3. Test of Influence Mechanism

4.3.1. Mediating Effect of EMA

To examine whether GS affects the environmental performance of agricultural enterprises through EMA, this study estimated models (2)–(3), and the regression results are shown in

Table 4. Mechanism test results.

Variables	EMA	IC	EP
	(1)	(2)	(3)
GS	2.532 ***	0.106 ***	0.320 **
	(3.651)	(4.301)	(2.025)
GS2	−0.512 ***	−0.017 ***	−0.052 ***
	(−4.333)	(−2.704)	(−2.731)
EMA			0.049 ***
			(2.868)
IC			1.213 ***
			(2.904)
Controls	YES	YES	YES
Industry-FE	YES	YES	YES
Year-FE	YES	YES	YES
N	1952	1952	1952
R2	0.076	0.257	0.292

Note: ***, and ** represent the significance level at 1%, and 5%, respectively; T-values in parentheses.

. The results in column (1) of Table 4 show that both GS and GS² are significant at the 1% level. The coefficient of GS is 2.532, and the coefficient of GS² is −0.512, indicating that the impact of GS on the external media attention received by agricultural enterprises is an inverted U shape. Meanwhile, the results of column (3) show that the coefficient of GS is significantly positive at the 5% level, the coefficient of GS² is significantly negative at the 1% level, and the coefficient of EMA is significantly positive at the 1% level, indicating that EMA plays a partial mediating role in the inverted U-shaped relationship between GS and the environmental performance of agricultural enterprises. Therefore, external media coverage is an important way for government subsidies to affect the environmental performance of agricultural enterprises, and Hypothesis 2 is established. The possible reasons are as follows: when the government subsidy is within a certain scale, with the increase in government subsidy received by agricultural enterprises, the media attention they receive also increases; However, when the government subsidy exceeds a certain scale, a continuous increase in the subsidy may lead to “aesthetic fatigue” of the external media, thus reducing the attention paid to the subsidy.

4.3.2. Mediating Effect of IC

To examine whether GS affects the environmental performance of agricultural enterprises through enterprise IC, this study estimated models (2)–(3), and the regression results are shown in Table 4. The results in column (2) of Table 4 show that both GS and the quadratic term of GS are significant at the 1% level. The coefficient of GS is 0.106, and the coefficient of the quadratic term of GS is −0.017, indicating that the impact of GS on the IC of agricultural enterprises is in an inverted U shape. Meanwhile, the results of column (3) show that the coefficient of the primary term of GS is significantly positive at the 5% level, the coefficient of the secondary term of GS is significantly negative at the 1% level, and the coefficient of IC is significantly positive at the 1% level. This indicates that IC plays a partial mediating role in the inverted U-shaped relationship between GS and the environmental performance of agricultural enterprises. Therefore, IC is an important way for government subsidies to affect the environmental performance of agricultural enterprises, and Hypothesis 3 is established.

4.4. Heterogeneity of Government Subsidies

4.4.1. Heterogeneity Analysis of R&D Subsidies

Among the different types of government subsidies, R&D subsidies primarily aim to support enterprises in technological innovation and research activities, addressing the externalities faced by agricultural enterprises in green innovation and technological advancement. However, when R&D subsidies exceed a certain threshold, stakeholders may become concerned about the risks associated with high-risk R&D activities, which could, in turn, reduce investment in environmental performance. In contrast, non-R&D subsidies focus on supporting daily operations, infrastructure development, and human resource growth, and may have a less significant impact on the environmental performance of agricultural enterprises. To validate the above analysis, this paper first conducts a text search of government subsidy lists using keywords related to R&D subsidies, such as “research”, “development”, “technology”, “industry-academia-research collaboration”, and “science and technology”. If the text contains these keywords, the subsidy is classified as a government R&D subsidy. The total R&D subsidies are then summed and divided by the total assets of agricultural enterprises to serve as a measure of government R&D subsidies. Additionally, non-R&D subsidies are measured by subtracting R&D subsidies from total government subsidies and dividing the result by the total assets of agricultural enterprises. The regression results are shown in

Table 5. Regression results of heterogeneity of government subsidy types.

Variables	EP	EP	EP	EP
	(1)	(2)	(3)	(4)
GS	0.883 ***	0.201	0.975 ***	0.220
	(4.413)	(1.078)	(5.127)	(1.101)
GS2	−0.308 ***	−0.003	−0.371 ***	−0.005
	(−4.735)	(−1.148)	(−3.190)	(−1.269)
Controls	YES	YES	YES	YES
Industry-FE	YES	YES	YES	YES
Year-FE	YES	YES	YES	YES
N	1952	1952	1952	1952
R2	0.247	0.243	0.241	0.248

Note: *** represent the significance level at 1%, respectively; T-values in parentheses.

, columns (1) and (2). Neither non-R&D subsidies nor the quadratic term of non-R&D subsidies are significant. The coefficient value for R&D subsidies is 0.883, and the coefficient value for the quadratic term of R&D subsidies is −0.308, both of which are significant at the 1% level. Therefore, the impact of government R&D subsidies on the environmental performance of agricultural enterprises follows an inverted U-shaped pattern, while the impact of government non-R&D subsidies on agricultural enterprises is not significant.

4.4.2. Heterogeneity Analysis of Environmental Protection Subsidy

Government environmental protection subsidies have clear policy objectives and are specifically designed to support agricultural enterprises’ investments in environmental protection, thus having a direct and significant impact on their environmental performance. However, when such subsidies exceed a certain threshold, agricultural enterprises may become more focused on maximizing their immediate benefits from these funds, potentially neglecting environmental projects that require long-term investment and continuous improvement. On the other hand, non-environmental protection subsidies are not specifically designated for environmental initiatives, meaning they may not have a substantial effect on the environmental performance of agricultural enterprises. To validate these assumptions, this study conducted a text search of government subsidy lists using keywords related to environmental protection subsidies, such as “environmental protection”, “exhaust emissions”, “wastewater”, and “pollution”. When the text included entries containing these keywords, the subsidy was categorized as a government environmental protection subsidy. The total environmental subsidies were then summed and divided by the total assets of agricultural enterprises to measure government environmental subsidies. Additionally, non-environmental subsidies were calculated by subtracting environmental subsidies from total government subsidies, then dividing the result by the total assets of agricultural enterprises. The regression results are shown in columns (3) and (4) of Table 5. The coefficients for environmental subsidies and the squared term of environmental subsidies are 0.975 and −0.371, respectively, both of which are significant at the 1% level. Neither non-environmental subsidies nor the squared term of non-environmental subsidies is significant. Therefore, government non-environmental subsidies have no significant impact on the environmental performance of agricultural enterprises, while environmental subsidies have an inverted U-shaped effect on the environmental performance of agricultural enterprises.

5. Discussions and Conclusions

5.1. Discussions

Based on the data of agricultural listed enterprises on China’s A-share market from 2013 to 2023, this paper systematically explores the impact of government subsidies on the environmental performance of agricultural enterprises and its mechanism of action and finds that there is A significant inverted U-shaped relationship between government subsidies and the environmental performance of agricultural enterprises. This conclusion indicates that with moderate fiscal support, agricultural enterprises can utilize external resources to overcome the high sunk costs and long payback periods of green investment, effectively enhancing their environmental governance capabilities [25]; However, when the scale of subsidies is too large, it may induce enterprises to rely on policy resources, thereby weakening their enthusiasm for independent technological research and development and environmental protection investment. This finding is similar to Li [31]’s research on manufacturing enterprises, which states that moderate subsidies can help enterprises enhance their environmental responsibility, while excessive subsidies pose a risk of resource misallocation.

In terms of the mechanism of action, this paper further verifies the mediating role played by external media attention and the level of internal control in the process of government subsidies influencing the environmental performance of agricultural enterprises. Moderate subsidies attach policy recognition labels to enterprises, enhance their public trust and market reputation, attract more media attention, form reputation constraints, and prompt them to fulfill environmental protection responsibilities more proactively [38]. At the same time, subsidies have, to a certain extent, strengthened the internal control system of enterprises and improved the efficiency of resource allocation and environmental compliance. However, when the subsidy scale exceeds the upper limit of the enterprise’s absorption capacity, the enterprise may gradually shift its focus from the long-term accumulation of green technologies to short-term compliance. Media supervision also weakens due to “aesthetic fatigue”, leading to lax internal control and a decline in environmental protection investment. This result is consistent with the findings of Yang et al. [38] regarding the relationship between government subsidies and media supervision, and also echoes the crucial role of internal control proposed by Tao et al. [39] in promoting green behaviors of enterprises.

Furthermore, heterogeneity analysis indicates that both R&D subsidies and environmental protection subsidies have an inverted U-shaped impact on the environmental performance of agricultural enterprises, while non-R&D and non-environmental protection subsidies do not show significant effects. This indicates from the side that subsidies with clear innovation and environmental orientation can better stimulate the green transformation momentum of agricultural enterprises. However, when their scale is too large, they may also suppress green investment due to the over-reliance effect.

Overall, this paper makes three key contributions to the existing literature. First, it breaks through the previous research perspective that mostly focuses on high energy-consuming or heavily polluting industries and focuses the analysis object on agricultural enterprises, enriching the empirical evidence of the relationship between government subsidies and environmental performance in the context of agricultural green transformation. Second, based on the resource dependence theory and the law of diminishing marginal returns, this paper reveals the inverted U-shaped nonlinear relationship between government subsidies and the environmental performance of agricultural enterprises. It also systematically examines the mediating roles of external media attention and internal control, offering deeper insights into how fiscal tools influence environmental performance through both external oversight and internal governance. Finally, by differentiating between R&D subsidies and environmental protection subsidies, this study identifies the heterogeneous impacts of various subsidy types on the environmental performance of enterprises. This provides valuable practical insights for optimizing the structure of agricultural subsidies and informing the design of sustainable policies.

Although this article systematically explores the impact of government subsidies on the environmental performance of agricultural enterprises, several limitations remain. First, this study relies solely on data from agricultural enterprises listed on China’s A-share market, which may limit the generalizability of the findings to non-listed small and medium-sized enterprises or agricultural enterprises in other emerging economies with different institutional environments. Future research should expand the sample to include non-listed companies and conduct cross-national comparative analyses to verify the universality of the inverted U-shaped relationship. Second, while media attention and internal control have been identified as mechanisms, the specific green innovation capabilities of enterprises and supply chain pressures, which may also serve as key driving factors, have yet to be tested. Subsequent studies should further explore the mechanisms between government subsidies and the environmental performance of agricultural enterprises and clarify their interactions with government subsidies.

5.2. Conclusions

Based on the data of China’s A-share agricultural listed enterprises from 2013 to 2023, this paper empirically tested the impact of government subsidies on the environmental performance of agricultural enterprises, discussed the two major mechanisms of external media attention and internal control level, and further analyzed the difference of the impact of government subsidies on the environmental performance of agricultural enterprises under different types of subsidies. The results show that, first, there is an inverted U-shaped relationship between government subsidies and the environmental performance of agricultural enterprises. Second, government subsidies affect the environmental performance of agricultural enterprises through external media attention and internal control level. Third, in terms of different types of subsidies, R&D subsidies have an inverted U-shaped impact on the environmental performance of agricultural enterprises, while non-R&D subsidies have no significant impact on the environmental performance of agricultural enterprises. Environmental protection subsidies have an inverted U-shaped impact on the environmental performance of agricultural enterprises, while non-environmental protection subsidies have no significant impact on the environmental performance of agricultural enterprises. Based on the research conclusions, this paper obtains the following policy implications:

From the government’s perspective, first, it is necessary to establish a dynamic evaluation system for the intensity of subsidies. It is recommended that a negative list management approach be implemented, targeting non-R&D and non-environmental protection subsidies and reducing fiscal expenditure on ineffective subsidies. Additionally, the subsidy withdrawal mechanism should be enhanced. When the environmental performance of enterprises surpasses a critical threshold, their dependence on subsidies should be gradually reduced, and incentive transmission should be facilitated through market-based tools, such as the carbon trading market and green credit. Second, the government needs to make the quality of environmental information disclosure a prerequisite for subsidy applications. Enterprises should be required to disclose core indicators, including pollutant treatment technology paths and environmental protection equipment operation data, thus ensuring the optimization of their internal control systems. At the same time, the establishment of an environmental data open platform will provide information support for public opinion supervision, with provisions for subsidy recovery and credit downgrading for enterprises exposed by the media. Finally, based on the ecological characteristics of agricultural enterprises, the government can increase the proportion of R&D subsidies for entities that obtain green food and organic product certifications. This will help form a synergistic effect of policy tools.

From the perspective of enterprises, first, they need to move away from the inertia of scale-oriented subsidies and focus on strategically allocating subsidy funds. Priority should be given to targeted investments in R&D and environmental protection subsidies. Second, enterprises should restructure their internal control processes by establishing an independent environmental protection fund audit unit. This will help prevent subsidies from being diverted to non-environmental expenditures through strict budgetary constraints. For enterprises that are experiencing performance decline, a subsidy replacement strategy can be explored, wherein stock subsidies are converted into equity investments in ecological technologies. This would help overcome the technical barriers to environmental governance by leveraging external innovation resources.