Government subsidies and corporate environmental, social and governance performance: Evidence from companies of China
Abstract
Environmental, social, and governance (ESG) performance is crucial for companies to attain sustainable development, which is a key reference for assessing the value and growth potential of a company. Government subsidies can provide incentives for companies to prioritize environmental preservation, meet their social duties, and improve their governance performance. This paper empirically examines the effects and mechanisms of government subsidies on corporate ESG performance, using an Ologit multiple ordered regression model based on data from Chinese listed companies. We find that not only do total government subsidies significantly improve firms' ESG performance, but both environmental and non-environmental subsidies also have the similar effect, albeit with different impact mechanisms. The analysis of the mechanism suggests that government subsidies can enhance corporate ESG performance by promoting green innovation, alleviating financing constraints, increasing charitable donations, and attracting social attention. This paper holds significant practical value as it presents empirical findings as a basis for reforming China's subsidy policies, showcasing the actual impact of diverse subsidy policies and the heterogeneity.
1 INTRODUCTION
The Chinese economy is transitioning from a period of swift growth to a phase of high-quality development. In September 2020, President Xi Jinping unveiled China's goal to reach peak carbon dioxide emissions by 2030 and achieve carbon neutrality by 2060 during his address at the 75th United Nations General Assembly session. Enterprises, as microeconomic entities, are embracing a new development paradigm centered around innovation, coordination, environmental sustainability, openness, and sharing. They are also exploring sustainable development models, rather than solely prioritizing maximization of shareholder profits. In this context, the concept of environmental, social, and governance (ESG) has emerged as a significant sociopolitical concern. ESG integrates public interest and sustainable development goals into the corporate value framework, alongside companies' environmental, social, and governance practices. This aligns with China's new stage of expansion and development philosophy (Wang, Wang, et al., 2022; Xie & Lyv, 2022). Enterprises that perform well in terms of ESG are also more likely to receive noteworthy societal attention and recognition (Qiu & Yin, 2019), and efficiently integrate social resources to create long-term advantages (Drempetic et al., 2020). As the concept of ESG gains currency and becomes a commonplace in corporate value assessments, enhancing corporate ESG performance has become a crucial issue for governments, companies, and the public.
The Chinese government has taken a keen interest in ecological issues and corporate environmental practices for some time, even though ESG principles were only relatively recently introduced in China compared with other developed nations. In 2009, Ministry of Environmental Protection Order No. 36 was issued by the Communist Party of China's Central Committee and the State Council, implementing the “prize to promote governance” approach, which aimed to comprehensively enhance the rural environment. To improve energy efficiency and protect the environment, the Ministry of Finance introduced the Energy Saving and Emission Reduction Subsidy Funding Management Interim Measures in 2015. This policy aims to reduce greenhouse gas emissions and enhance environmental conditions. In the socialist market economy of China, the government plays a crucial role as the primary regulator. Government subsidies act as crucial regulatory instruments and policy signals that steer enterprise conduct in government macro-control policies. For example, BYD COMPANY LIMITED, a Chinese automobile manufacturer, has received government subsidies for new energy over many years. These subsidies have enabled BYD to upgrade their new energy technology and expand industrial chains, consequently advancing BYD's “carbon reduction and emission reduction” strategy. In addition, the alleviation of corporate financial constraint brought about by government subsidies is also beneficial to the protection of BYD's employees' rights and interests, which improves BYD's ESG performance.
Chinese governments at all levels provide numerous types of subsidies to businesses, such as environmental, R&D, and talent subsidies. Most existing studies focus on R&D subsidies1 but pay little attention to environmental subsidies. However, in recent times, the Chinese government has augmented its support for environmental causes, in response to China's emphasis on sustainable development. The General Administration of Taxation (GAT) website released tax and fee regulations favoring eco-friendly initiatives in June 2022. The guidelines present 56 tax and fee incentives across four main areas, comprising of endorsing environmental protection, advocating for environmental conservation, urging inclusive resource utilization, and promoting progression in low-carbon industries. Despite the Chinese government's increasing interest in this topic, there has yet to be empirical examination into the efficiency of their environmental subsidies. Thus, this study not only explores the impact of subsidies on the ESG performance of firms, along with the probable mechanisms involved, but also compares environmental and non-environmental subsidies.
The current research has mostly explored the elements impacting corporate ESG performance from two perspectives: the internal and external corporate environment. On the one hand, internal corporate governance and technology evolution have significant influences on business ESG performance. The academic study analyses the influence of board structure (Shive & Forster, 2020; De Villiers et al., 2011), executive team characteristics (Cordano et al., 2010), company scale (Drempetic et al., 2020), and corporate digital transformation (Fang et al., 2023; Wang, Wang, et al., 2022) on corporate ESG performance. The study illuminates the role of board structure, executive team traits and corporate digital transformation in influencing corporate ESG performance. On the other hand, the external environment also influences corporate ESG performance to a great extent, among which the influence of institutional, economic and cultural environment on corporate ESG performance has received much attention from scholars, such as the role of environmental protection tax (Wang, Wang, et al., 2022), environmental protection legislation (Wang, Wang, et al., 2022), government debt (Zhang & Deng, 2022), capital market (Liu et al., 2023), financial technology development (Mu et al., 2023) and socio-cultural factors (Cai et al., 2016; Gillan et al., 2021; Liang & Renneboog, 2017) on corporate ESG performance.
Although previous literatures have discussed the factors influencing corporate ESG from internal and external perspectives (Li & Xu, 2022). However, the effect of government macro-regulation on company strategy is more substantial in the Chinese setting, with government subsidies being one of the most regularly employed economic regulatory mechanisms. There is currently a scarcity of research on the effects of government subsidies on corporate ESG performance, and even less writing on the effects of environmental subsidies on the variety of ESG performance. Based on the theory of institutional economics, high-quality economic development is primarily determined by the logical system and its execution. Government subsidies are a vital industrial policy that is a tangible expression of the formal institution. They shape and modify corporate behavior and facilitate enterprise strategic planning to meet governmental expectations. Subsequently, government subsidies have a considerable impact on corporate ESG performance and are not only a crucial economic tool for the government to stimulate economic growth, but also function as policy signals to the market. Government subsidies play a dual role in providing material support and directing the focus of enterprises toward their development (Mao & Xu, 2015).
It can be argued that the study of subsidies on corporate ESG performance is of great significance from both a theoretical and practical viewpoint. Conducting further research on the impact of environmental and non-environmental subsidies can aid in better understanding the diverse effects of subsidies on corporate ESG. This, in turn, can serve as a more effective means of supplementing current research. The focus of this paper is, therefore, to shed light on this subject matter. Can government subsidies improve ESG performance? Do different types of funding have varying effects on ESG performance? Our study examines the impacts of various subsidies on the E (environmental), S (social), and G (governance), significantly enhancing our findings.
Using data from Chinese listed enterprises, this paper investigates the impact and process of government subsidies on corporate ESG success, utilizing Ologit multiple ordered regression models. The empirical findings reveal that government subsidies help companies enhance their ESG performance, with both environmental and non-environmental subsidies being advantageous. The mechanism analysis demonstrates that government subsidies and non-environmental subsidies improve corporate ESG performance by promoting green innovation, alleviating financing constraints, encouraging corporate charitable donation and attracting social attention, while environmental subsidies work only through the mechanism of alleviating financing constraints and promoting utility green innovation. The heterogeneity analysis indicates that government subsidies have a more pronounced impact on ESG performance of heavily polluting firms, non-high-tech firms, and non-state-owned firms, reflecting the fact that government subsidies can be more effective in supporting the weak. In our further research, it was discovered that government subsidies have a positive impact on the E, S, and G by encouraging corporate environmental investment, poverty alleviation, and strengthening board independence, respectively.
The paper contributes to ESG research in several ways. First, it offers a new perspective by analyzing the factors influencing corporate ESG performance in detail. We expand on existing literature examining government subsidies and corporate ESG performance, identifying disparities in the functions of environmental versus non-environmental subsidies. Second, the paper elucidates the mechanism of government subsidies on corporate ESG performance. Not only does this paper examine the general impact mechanism of government subsidies on corporate ESG performance, but it also delves into the specific impact paths regarding corporate Environment, Social, and Governance performance. Finally, this paper holds significant practical value as it presents empirical findings as a basis for reforming China's subsidy policies, showcasing the actual impact of diverse subsidy policies and the heterogeneity.
The rest of this paper is structured as follows. Section 2 introduces the research hypothesis based on the literature review, explaining its derivation. Section 3 presents the research design, detailing the data, measurements, and research model. Section 4 presents the empirical results. Section 5 provides further analysis to enrich the outcomes. Finally, Section 6 concludes the article.
2 HYPOTHESIS DEVELOPMENT
Government subsidies provide motivation for companies to improve their ESG performance.2 New institutional economics points out that the behavior and development of an organization largely depend on the institutional environment in which it is located (North, 1990). On the one hand, according to Porter's assumption, reasonable environmental regulation will promote enterprise innovation, upgrade production technology and process, so enhancing firms' abilities to achieve both economic and environmental performance (Porter & Linde, 1995). Government subsidy is one of the important measures to regulate the market, which is an important institutional factor faced by enterprises. Therefore, government subsidies have great potential to enabling enterprises to realize economic, social, and environmental value. On the other hand, in terms of legitimacy theory, government subsidies usually incorporate authorized policy guidance and expectations, and companies receiving subsidies have to meet certain government requirements or risk losing legitimacy from the government (Zhang et al., 2013). Moreover, under the “spotlight” of media and the legitimacy pressure of the public, companies that receive government funding will be more cautious in their use of government subsidies and are likely to exhibit pro-social behavior to gain favorable evaluation. As a result, these companies will cultivate a positive public image and social performance, which will enhance their ESG performance.
Based on the criterion that the aim of environmental subsidies is whether to protect the environment or not, we specifically concentrate on subsidies concerning the environment and those that do not. Environmental subsidies aim to financially reimburse companies for any economic losses incurred due to emission reductions through cash transfers, tax reductions, and low-interest loans, among other incentives. Additionally, these subsidies seek to support and motivate firms to decrease pollution and switch to more environmentally friendly resources, while simultaneously enhancing their production technologies, to ensure sustainable development in the future. (He et al., 2022). We acknowledge that environmental subsidies may have impacted diverse firms due to their excessive emphasis on environmental issues that ignoring others (Bai et al., 2018). However, we think that its contribution to ESG is more prominent.
Non-environmental subsidies consist of a range of government subsidies that are oriented toward objectives other than environmental improvement. These subsidies comprise R&D subsidies, talent subsidies, industrial upgrading subsidies, employment subsidies, trade subsidies, and project operation subsidies (Nie et al., 2022). Non-environmental subsidies assist businesses in concentrating on their areas of expertise to cultivate products and services, obtain a competitive edge and enhance profit margins, establish a firmer material foundation for upholding and enhancing employee welfare, and empower businesses to fulfill their social obligations with greater efficacy. Furthermore, non-environmental subsidies offer tangible assistance to businesses in incorporating innovative skills, which supports the enhancement of their human resource structure and elevation of human capital. Therefore, non-environmental subsidies can contribute to the enhancement of the environmental, social, and governance performance of businesses.
Previous research has mainly focused on the impact of environmental subsidies on corporate green behavior (Acar & Yeldan, 2016; Bai et al., 2018; Georgakellos, 2011). However, there is a lack of understanding of the effects on corporate ESG performance. The literature on non-environmental subsidies focuses solely on R&D subsidies. Moreover, there is a dearth of literature that compares environmental subsidies with non-environmental subsidies. Based on the different usage of government subsidies, we further investigate the dissimilarity in impact mechanisms between environmental and non-environmental subsidies. Therefore, this paper proposes the following hypothesis:
H1..Government subsidies are conducive to improving corporate ESG performance.
H1a..Environmental subsidies are conducive to improving corporate ESG performance.
H1b..Non-environmental subsidies are conducive to improving corporate ESG performance.
Government subsidies firmly support companies to develop and upgrade green technology, product, and process (Liu et al., 2020), then enhancing green innovation capabilities which play a key role in improving ESG performance. The capital, technology, and human resources required by enterprises to carry out green innovation activities are substantial. External supportive funding, in the form of government subsidies, can act as a catalyst for green innovation (Li et al., 2021). Obtaining these subsidies means enterprises have additional funds to invest in the research and development of green innovation technologies and products, creating a snowball effect. Ultimately, companies can attain an environmentally sustainable production process and mitigate the negative effects of their actions on the environment. It, in turn, enhances their ESG performance. Furthermore, governmental subsidies facilitate firms to acquire green innovation experts with preferential policies and amass human capital related to green technology advancements (Fu et al., 2022). This can enhance companies' capabilities to innovate sustainably and boost their ESG performance. In particular, subsidies focused on both the environment and other areas can have a positive impact on firms' ESG performance by encouraging sustainable innovation.
Environmental subsidies are funds which are paid by the government to companies for emission reduction and environmental protection. As a result, firms will use part of the funds for green innovation research and development activities to meet the government's requirements and at the same time ensure their productivity (Bigerna et al., 2019). Green innovation not only facilitates firms in enhancing their resource utilization and mitigating pollutant emissions but also augments corporate green productivity, attains stakeholder recognition, and enhances corporate ESG performance (Li et al., 2022). Firms can broaden their green supply chains by utilizing environmental subsidies to their fullest extent (He et al., 2022). Although non-environmental subsidies are not specifically targeted toward environmental improvement, they enhance the financial flexibility of firms and increase the probability that these firms will utilize the funds for the development of green innovation skills, optimization of organizational structure and improvement of production management processes. In addition, R&D subsidies are directly beneficial for technological advancements. This paper argues that both environmental and non-environmental subsidies can improve corporate ESG performance. However, environmental subsidies may be less effective in promoting green innovation compared to non-environmental subsidies. This is because they focus more on environmental performance and do not have inherent requirements for technological progress. Therefore, we propose the following hypothesis:
H2..Government subsidies have a positive effect on ESG performance by promoting green innovation.
H2a..Environmental subsidies are conducive to improving corporate ESG performance by promoting green innovation.
H2b..Non-environmental subsidies are conducive to improving corporate ESG performance by promoting green innovation.
Government subsidies strengthen the ESG performance of companies by easing their financing constraints. Obtaining subsidies from the government is an effective way for most companies to increase their financial income (Yang et al., 2019). This is due to the fact that government subsidies are less expensive and quicker to acquire compared to bank loans and internally accumulated funds. Obtaining government subsidies boosts enterprises' confidence to invest in green production, improve social responsibility, employee welfare, and corporate governance. Signaling theory suggests that government subsidies signal to the market the government's support for the firms receiving them (Yan & Li, 2018). This indicator assists external investors in identifying government policy trends and enterprises with high potential, thereby providing more external financing and other resources to these businesses.
Environmental subsidies are predominantly earmarked funds, eligible for use by enterprises only toward supporting environmentally friendly production methods, thereby aiding them in mitigating the financial challenges associated with transitioning to “green” operations. Non-environmental subsidies have a broader range of applications, allowing businesses to utilize these funds to undertake initiatives such as reducing poverty, enhancing employee benefits, and optimizing governance structures. Therefore, we propose the following hypothesis:
H3..Government subsidies have a positive effect on ESG performance by alleviating corporate financing constraints.
H3a..Environmental subsidies have a positive effect on ESG performance by alleviating corporate financing constraints.
H3b..Non-environmental subsidies have a positive effect on ESG performance by alleviating corporate financing constraints.
Government subsidies make firms' cash flow more flexible, so promote firms to increase charitable giving (Meier, 2007), which is an effective way to improve ESG performance. Receiving government subsidies implies governmental interest in an enterprise (Zhang et al., 2013), and thus the enterprise ought to align its behavior with the government's expectations and policy orientation. Charitable donations represent an efficient means for enterprises to meet governmental expectations. After gaining more disposable cash flow, enterprises can make larger charitable donations (Wang et al., 2020), thus attracting the attention and recognition of the public, investors, and media. This, in turn, cultivates an image of social responsibility within the capital market and further advances their ESG ratings (Chen, 2021).
Environmental subsidies allow enterprises to collaborate and exchange with environmentally friendly public welfare organizations and make donations, which can increase the level of corporate donations and improve ESG. Non-environmental subsidies, which are widely applicable, can also reduce pressure on enterprise funds for research and development, infrastructure construction, and talent recruitment, enabling them to reserve more funds for charitable donations and gain favor in the capital market. Therefore, we propose the following hypothesis:
H4..Government subsidies improve corporate ESG performance by promoting charitable donation.
H4a..Environmental subsidies improve corporate ESG performance by promoting charitable donation.
H4b..Non-environmental subsidies improve corporate ESG performance by promoting charitable donation.
Enterprises that accept government subsidies will receive more attention from the capital market and the media. On the one hand, subsidized firms who have been highly concerned by society will acquire more resources from various stakeholders (Chai et al., 2023). In China, the government will support the key industries, and the relevant enterprises in the industry will receive government subsidies. According to signal theory, government subsidies send a signal to the market that the enterprise receiving the subsidy have been recognized and valued by the government (Kleer, 2010). This signal helps external investors identify government policy trends and high-potential enterprises, thus providing subsidized enterprises with more external financing and other innovative resources. As a result, companies will have more resources to upgrade green technologies, engage in social welfare and improve employee well-being to enhance ESG performance. On the other hand, subsidized firms who get much attention from the public will attach much importance on environmental protection, social issues, and corporate governance efficiency to maintain their good image and reputation (Wu, 2017). Subsidized companies mean they are at the top of the government's attention list, leading investors and the media to pay more attention to them. Under the “spotlight” of investors and media, enterprises are more likely to show pro-social behaviors to obtain positive evaluation, such as participating in charitable donations, launching public welfare activities, developing cooperative relations with nonprofit organizations, and carrying out corporate social entrepreneurship. If enterprise behavior is not accepted by the public, it could lose its legitimacy in the market and face further difficulties in development (Yan et al., 2021).
Both environmental and non-environmental subsidies assist organizations in attracting social attention, consequently enhancing their ESG ratings. The government has precisely delineated the funds aimed toward environmental protection subsidies, while publishing the list of organizations that utilize these funds. Such a move implies dual supervision from the government and the public over enterprises. If a corporation undergoes significant attention from the society, its unlawful expenditure of subsidies is more liable to be detected and publicly disclosed by the media or the citizens. Enterprises with a strong social focus generally use environmental subsidies with precision for corporate environmentalism and preservation. Furthermore, non-environmental subsidies have wider applicability and greater flexibility for the said firms to utilize such subsidies. When a company demonstrates a high level of social responsibility, it is more likely to receive public recognition and develop a positive reputation, which enhances the impact of non-environmental subsidies on ESG. Therefore, we propose the following hypothesis:
H5..Government subsidies improve corporate ESG performance by attracting social attention.
H5a..Environmental subsidies improve corporate ESG performance by attracting social attention.
H5b..Non-environmental subsidies improve corporate ESG performance by attracting social attention.
3 RESEARCH DESIGN
3.1 Sample selection and data description
Our data set comprises 3068 listed companies from 2009 to 2019, which we analyze to explore the impact of Chinese government subsidies on corporate ESG performance. We treat the data as follows: (1) exclude the companies listed after 2009; (2) exclude the samples with missing ESG data; (3) exclude the samples of ST companies; (4) exclude the financial companies. To reduce the impact of extreme values on the model, we apply a 1% tail shrinkage up and down to all continuous variables. The data of government subsidies are obtained from “government subsidies included in current profit and loss” under “nonoperating income” and “other income”in the annual reports of listed companies. ESG data originates from Shanghai Huazheng Index Information Service Co while CSMAR and WIND databases contain the source information for financial constraint, social donation, environmental performance, social performance, governance performance, and control variables. The number of green patents is matched from the national IPR patent database and the green list of international patent classification of WIPO. Due to the absence of certain data, an unbalanced data set ranging from 2009 to 2019 is utilized as the research sample for this paper.
The explained variable, firm ESG performance, is the composite ESG rating of China Securities in the benchmark regression and heterogeneity analysis. Following the method of Wang, Luan, et al. (2022), we provide values to nine categories, ranging from AAA, AA, A, BBB, BB, B, CCC, CC, C respectively from 9 to 1,3 where higher values represent better ESG performance of the corporation. This paper also incorporates Bloomberg ESG rating and HEXUN's CSR rating as alternative independent variables in the robustness test. Furthermore, we use E rating, S rating, and G rating from Bloomberg in the further analysis.
The independent variables in this article are government subsidies, which we obtained from the annual reports of publicly traded companies. To identify environmental subsidies, we followed the manual approach suggested by Zhao et al. (2015) that involved searching for terms such as “environment”, “environmental protection”, “waste gas”, “wastewater”, “energy”, and “green”. Environmental subsidies refer to government subsidies that are related to the environment. In contrast, non-environmental subsidies are government subsidies that have no connection with the environment.
There are four key mediating variables examined in this study—green innovation, financial constraint, charitable donation, and the investor attention index. In further analysis, the submechanisms of corporate environmental investment, poverty alleviation, and board independence are utilized.
The following are the control variables in this article: (1) Firm size, which is assessed by calculating the total assets at the end of the year. (2) Financial leverage, which is measured by the company's gearing ratio. (3) Return on total assets, which is figured out by dividing net profit by total assets. (4) Shareholding ratio of the largest shareholder, which is determined by the percentage of shares owned by the largest shareholder in the company. (5) Market value of the company, which is measured by Tobin's Q in this paper. (6) The profitability of the company is determined using dummy variables. A value of 1 is assigned if the company is profitable and 0 otherwise. The detailed variable definitions are shown in Table 1.
| Variable | Notation | Definition | |
|---|---|---|---|
| Dependent variables | Environmental, social and governance (ESG) | huazheng_esg | Huazheng ESG rating index |
| Independent variables | Government subsidies | dummy_subsidy | 1 for receipt of government subsidies, 0 otherwise |
| ln_subsidy | ln (government subsidies + 1) | ||
| Environmental subsidies | ln_environ | ln (environmental subsidies + 1) | |
| Non-environmental subsidies | ln_non_environ | ln (non-environmental subsidies + 1) | |
| Control variables | Company size | size | Total assets |
| Financial leverage | lev | Asset liability ratio | |
| Return on total assets | ROA | Net profit/total assets | |
| Shareholding by major shareholders | top1 | Shareholding ratio of the largest shareholder | |
| Tobin Q | tobingq | Market value/(total assets - net intangible assets - net goodwill) | |
| Profitable or not | profit | Positive profit is 1, 0 otherwise | |
| Mediating variables | Green innovation | ln_totalpatent | ln [(inventive green patent applications + utility green patent applications) + 1] |
| ln_inventpatent | ln (inventive green patent applications+1) | ||
| ln_utilitypatent | ln (utility green patent applications+1) | ||
| Financial constraint | SA | Drawing on Hadlock and Pierce (2010) to construct an index using firm size and firm age | |
| Charitable donation | donation | ln (total annual social donations of the company) | |
| Social attention | baidu | A baidu index compiled using stock names as keywords | |
| Other variables | ESG | pengbo_esg | Bloomberg ESG Rating Index |
| hexun_esg | Hexun ESG Rating Index | ||
| E | pengbo_environ_esg | Bloomberg E Rating Index | |
| S | pengbo_social_esg | Bloomberg S Rating Index | |
| G | pengbo_govern_esg | Bloomberg G Rating Index | |
| Environmental investment | environ_invest | 1 if the enterprise has environmental investments, 0 otherwise | |
| Social performance | povertyreduction | 1 if there is a follow-up poverty alleviation program, 0 otherwise | |
| Governance quality | independent_director | Number of independent directors | |
The descriptive statistical results of the primary variables in this paper are listed in Table 2. Enterprises' average ESG performance is 4.144, with a minimum value of 1 and a maximum value of 8. In truth, companies do not receive an AAA grading in the Huazheng ESG ratings, leading to a score range of 1–8 for firms.4 The average value of government subsidies is 0.989, suggesting that 98.9% of businesses have received government assistance. The maximum value of the logarithm of government subsidies is 20.39, while environmental subsidies are 17.715, which is generally less than the non-environmental subsidies of 20.333, and from the perspective of the mean, environmental subsidies are only 5.119, which is significantly less than 16.107 and 15.964. The values of the other variables are within a reasonable range.
| Variables | N | Mean | SD | Min. | Max. |
|---|---|---|---|---|---|
| huazheng_esg | 22,915 | 4.144 | 1.047 | 1 | 8 |
| dummy_subsidy | 22,915 | 0.989 | 0.104 | 0 | 1 |
| ln_subsidy | 22,915 | 16.107 | 2.366 | 0 | 20.39 |
| ln_environ | 22,915 | 5.119 | 6.707 | 0 | 17.715 |
| ln_non_environ | 22,915 | 15.964 | 2.554 | 0 | 20.333 |
| size | 22,915 | 1.146 | 2.849 | 0.029 | 21.076 |
| lev | 22,915 | 0.413 | 0.206 | 0.048 | 0.959 |
| ROA | 22,915 | 0.042 | 0.058 | −0.272 | 0.198 |
| top1 | 22,915 | 35.231 | 14.855 | 8.98 | 74.96 |
| tobingq | 22,915 | 2.219 | 1.416 | 0.917 | 10.358 |
| profit | 22,915 | 0.925 | 0.263 | 0 | 1 |
3.2 Model specifications
4 EMPIRICAL RESULTS
4.1 Baseline results
To investigate the effect of government subsidies on corporate ESG performance, we employed the empirical model provided above. The outcomes are demonstrated in Table 3, with Column (1) revealing the regression coefficient of the dummy variable for government subsidies (β = 0.784, p < 0.01). This shows that subsidies generate higher ESG performance for companies compared to those that do not receive government subsidies. The logarithm of the subsidy amount serves as the dependent variable in columns (2) and (3). No control variables are included in column (2), while control variables are included in column (3). In column (3), the coefficient of the independent variable is 0.065 (p < 0.01), indicating that government subsidies have a positive impact on firms' ESG performance. Table 4 displays the findings from the marginal effects analysis, indicating that increased government subsidies result in higher likelihood of a firm's ESG rating improvement. Specifically, with every 1% increase in subsidies, the probability of a firm attaining an ESG rating of level 8 increases by 0.001% and the probability of falling to level 1 decreases by 0.1%.
| (1) | (2) | (3) | (4) | (5) | |
|---|---|---|---|---|---|
| Variables | huazheng_esg | huazheng_esg | huazheng_esg | huazheng_esg | huazheng_esg |
| dummy_subsidy | 0.784*** | ||||
| (7.76) | |||||
| ln_subsidy | 0.111*** | 0.065*** | |||
| (14.31) | (7.96) | ||||
| ln_environ | 0.013*** | ||||
| (4.58) | |||||
| ln_non_environ | 0.055*** | ||||
| (7.61) | |||||
| Observations | 26,405 | 25,647 | 22,915 | 22,915 | 22,915 |
| Controls | No | No | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively. Robust standard errors are clustered at the firm level.
| ESG = 1 | ESG = 2 | ESG = 3 | ESG = 4 | ESG = 5 | ESG = 6 | ESG = 7 | ESG = 8 | |
|---|---|---|---|---|---|---|---|---|
| ln_subsidy | −0.001*** (−7.01) | −0.002*** (−7.7) | −0.007*** (−7.98) | −0.003*** (−6.99) | 0.009*** (8.06) | 0.004*** (7.44) | 0.0003*** (5.02) | 0.00001* (1.85) |
| ln_environ | −0.0002*** (−4.41) | −0.0005*** (−4.52) | −0.001*** (−4.59) | −0.001*** (−4.38) | 0.002*** (4.61) | 0.001*** (4.48) | 0.0001*** (3.73) | 0.000002* (1.75) |
| ln_non_environ | −0.001*** (−6.76) | −0.002*** (−7.39) | −0.006*** (−7.63) | −0.003*** (−6.73) | 0.008*** (7.67) | 0.003*** (7.19) | 0.0002*** (4.95) | 0.00001* (1.85) |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively.
- Abbreviation: ESG, environmental, social, and governance.
Furthermore, the study investigates the impact of environmental and non-environmental subsidies. Columns (4) and (5) of Table 3 reveal that both environmental subsidies (β = 0.013, p < 0.01) and non-environmental subsidies (β = 0.055, p < 0.01) have a positive effect on enhancing firms' ESG performance. Additionally, non-environmental subsidies have a greater influence on the ESG performance of firms. This lends support to the validity of H1a and H1b. Based on the marginal effects analysis in Table 4, we similarly find that both environmental and non-environmental subsidies are overall favorable to firms' ESG performance. As environmental subsidies increase by 1%, the likelihood of a firm's ESG rating rising to 8 increases by 0.0002%, while the likelihood of the same rating falling to 1 decrease by 0.02%. When non-environmental subsidies increase by 1%, the likelihood of a firm's ESG rating increasing to 8 rises by 0.001%, and the probability of a rating dropping to 1 decrease by 0.1%.
We further validate the results of the base regression model in Table 5 using the Gologit model by relaxing the parallel lines assumption for the subsidy variable, in other words, the coefficients of the impact of subsidies on ESG with different ratings (1–8) in the model setup can be different values. Table 5 indicates that with the increase of government subsidies, the probability of ESG performance improvement is higher. Nevertheless, subsidies' effect on ratings exceeding 7 (i.e., an ESG rating of 8) is not statistically substantial. It is challenging for companies to achieve the highest ratings directly. A general ordered logit model for environmental and non-environmental subsidies can be found in Tables A1 and A2.
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| ESG rating | >1 | >2 | >3 | >4 | >5 | >6 | >7 |
| ln_subsidy | 0.029 | 0.031** | 0.051*** | 0.074*** | 0.157*** | 0.253*** | 0.328 |
| (1.33) | (2.56) | (6.85) | (10.02) | (10.60) | (4.54) | (1.42) | |
| Constant | 3.976*** | 2.564*** | 0.559*** | −1.760*** | −5.341*** | −10.04*** | −14.32*** |
| (10.50) | (10.29) | (2.86) | (−9.07) | (−18.36) | (−10.20) | (−3.47) | |
| Observations | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively.
To address the endogenous issue whereby enterprises with higher ESG ratings may receive more subsidies, we introduced the instrumental variable approach in our model. Columns (1) and (2) present the IV calculation results of ordinary least square method and columns (3) and (4) show the IV-Oprobit results.6 To rectify the endogenous problem, we employed the studies of Zhang et al. (2015) and Xu et al. (2022) to use local government fiscal revenue as the instrumental variable as there is a strong connection between the amount of government subsidies granted to firms in the area and the local fiscal receipts. There is no direct correlation between the fiscal revenue capacity of the region and the ESG ratings of micro firms within the region. In general, local government fiscal revenue affects government subsidy expenditure, but it does not directly impact the internal behavior and performance of enterprises. The coefficient of local government fiscal revenue in the first stage is significantly positive (β = 0.088, p < 0.01), and the coefficient of government subsidies in the second stage is also significantly positive (β = 0.215, p < 0.1). As indicated in columns (3) and (4), the coefficient of instrumental variable exhibits a significant positive trend (β = 0.088, p < 0.01) in the first stage. Furthermore, in the IV-Oprobit model, the government subsidy coefficient also reflects a significant positive association (β = 0.202, p < 0.05). The weak instrumental variable test, which has an F-value of 28.18, significantly surpasses the critical value (16.38) under 10% bias, thus rejecting the weak instrumental variable hypothesis. Therefore, we can contend that the instrumental variable is not weak in nature. Consequently, even after taking into account endogeneity, the incentivizing influence of government subsidies on ESG performance of companies remains viable (Table 6).
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| OLS | IV-oprobit | |||
| Variables | ln_subsidy | huazheng_esg | ln_subsidy | huazheng_esg |
| finance_in | 0.088*** | 0.088*** | ||
| (4.07) | (5.31) | |||
| ln_subsidy | 0.215* | 0.202** | ||
| (1.86) | (2.57) | |||
| Observations | 22,915 | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively. Robust standard errors are clustered at the firm level.
4.2 Mechanism test
4.2.1 Green innovation
Drawing on Jiang (2022), we analyze the intermediate mechanisms by analyzing the effect of mechanism variables on independent variables. Table 7 presents empirical findings on using total green innovation as a mediating factor. Tables 7.1 investigate the dissimilar impacts of inventive green innovation and utility green innovation as mediating factors respectively. It is evident from the results portrayed in columns (1) of Table 7 that governmental subsidies positively affect green innovation considerably (β = 0.011, p < 0.01) It means that green innovation has mediating role between government subsidies and corporate ESG performance. The findings in column (2) suggest that there is no significant positive impact of environmental subsidies on green innovation. In column (3), the regression coefficient for non-environmental subsidies (β = 0.009, p < 0.01) highlights their significant positive influence on green innovation, thereby demonstrating that green innovation serves as the mechanism through which non-environmental subsidies impact the ESG performance of enterprises.
| (1) | (2) | (3) | |
|---|---|---|---|
| Variables | ln_totalpatent | ln_totalpatent | ln_totalpatent |
| ln_subsidy | 0.011*** | ||
| (6.78) | |||
| ln_environ | 0.001 | ||
| (1.59) | |||
| ln_non_environ | 0.009*** | ||
| (6.12) | |||
| Constant | −0.129* | 0.040 | −0.098 |
| (−1.69) | (0.56) | (−1.30) | |
| Observations | 22,852 | 22,852 | 22,852 |
| Controls | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
| (1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
| Variables | ln_inventpatent | ln_inventpatent | ln_inventpatent | ln_utilitypatent | ln_utilitypatent | ln_utilitypatent |
| ln_subsidy | 0.009*** | 0.008*** | ||||
| (6.09) | (6.71) | |||||
| ln_environ | 0.001 | 0.001* | ||||
| (1.26) | (1.75) | |||||
| ln_non_environ | 0.007*** | 0.007*** | ||||
| (5.44) | (6.10) | |||||
| Constant | −0.119** | 0.010 | −0.094 | −0.112* | 0.014 | -0.090 |
| (−2.05) | (0.18) | (−1.64) | (−1.94) | (0.25) | (-1.58) | |
| Observations | 22,852 | 22,852 | 22,852 | 22,852 | 22,852 | 22,852 |
| Controls | Yes | Yes | Yes | Yes | Yes | YES |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | YES |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | YES |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | YES |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
Further, we divide green innovation into inventive green innovation and utility green innovation. The analysis of Table 7.1 reveals that both total subsidies and non-environmental subsidies enhance firms' ESG levels through both utility green innovation and inventive green innovation. Additionally, despite having no significant impact on total green innovation and inventive green innovation, environmental subsidies have a significant positive effect on utility green innovation (β = 0.001, p < 0.1). Thus, total green innovation, inventive green innovation, and utility green innovation serve as the mediating mechanism for both total and non-environmental subsidies, whereas utility green innovation are the mediating influence mechanism for environmental subsidies. This indicates that the findings partly support hypotheses H2, H2a, and H2b. The investment cycle of inventive green innovation is long and the risk of R&D is high, so enterprises are often reluctant to use subsidies for the research and development of innovative innovation with higher risks. The difficulties of utility green innovation are lower than innovative green innovation. Therefore, more environmental and non-environmental subsidies are used for practical patent development. As a result, we conclude that non-environmental subsidies strengthen both inventive and utility green innovation while environmental subsidies solely promote utility green innovation, contributing to an improvement in ESG performance.
4.2.2 Financing constraint
Since the KZ and WW indices contain many financial variables that are endogenous, such as cash flow and leverage, the SA index is employed as a gauge of financing constraint, drawing upon Hadlock and Pierce (2010) and Ju et al. (2013). A higher SA index implies that financing constraints are effectively alleviated, and government subsidies could help firms boost their ESG performance by mitigating their financing constraints. In columns (1), (2), and (3) of Table 8, the coefficient of government subsidies (β = 0.012, p < 0.01), environmental subsidies (β = 0.001, p < 0.01) and non-environmental subsidies (β = 0.011, p < 0.01) are significantly positive. It means that government subsidies, including environmental and non-environmental subsidies, have indeed eased the cash shortage of companies. Government subsidies, environmental and non-environmental subsidies improve corporate ESG performance through alleviating financing constraints. According to the empirical results, we can know that when subsidies increase by 1%, the likelihood of a firm's SA index rises by 1.2%. However, the coefficient of environmental subsidies is much smaller than those of total and non-environmental subsidies. This may be attributed to the requirement for firms to utilize the subsidy funds to enhance the environment and decrease power consumption. Consequently, subsidized firms have less money to spend on production and less impact on easing their financing constraints. Hypothesis H3, H3a, and H3b are supported.
| (1) | (2) | (3) | |
|---|---|---|---|
| Variables | SA | SA | SA |
| ln_subsidy | 0.012*** | ||
| (14.23) | |||
| ln_environ | 0.001*** | ||
| (4.68) | |||
| ln_non_environ | 0.011*** | ||
| (14.28) | |||
| Constant | −3.423*** | −3.236*** | −3.397*** |
| (−55.55) | (−52.54) | (−55.24) | |
| Observations | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
4.2.3 Charitable donation
Table 9 shows the findings of the mediating variable analysis on charitable donation. The results in column (1) reveal a significantly positive coefficient for government subsidies (β = 0.088, p < 0.01), we can conclude that by increasing the amount of corporate charitable donations, government subsidies can improve corporate ESG performance. In addition, the other columns in Table 9 suggest that charitable donations do not mediate the relationship between environmental subsidies and ESG performance, but mediate the relationship between non-environmental subsidies and corporate ESG performance. Because environmental subsidies are used for special fields, while non-environmental subsidies have various usages. It allows firms to invest in philanthropy by using non-environmental subsidies instead of environmental subsidies. These findings confirm the validity of hypotheses H4 and H4b, while H4a is dismissed.
| (1) | (2) | (3) | |
|---|---|---|---|
| Variables | donation | donation | donation |
| ln_subsidy | 0.088*** | ||
| (5.17) | |||
| ln_environ | 0.003 | ||
| (0.58) | |||
| ln_non_environ | 0.064*** | ||
| (3.91) | |||
| Constant | −0.717 | 0.629 | −0.345 |
| (−1.12) | (1.08) | (−0.54) | |
| Observations | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
4.2.4 Social attention
We use the Baidu index, utilizing stock name as a keyword, as a proxy variable for corporate social attention. The coefficient of the interaction term between social attention and government subsidies (β = 1.004, p < 0.01) in column (1) indicates that social attention is an effective mediating variable between government subsidies and firms' ESG performance. This implies that the more subsidies a firm receives, the more favorable it is in gaining higher social attention for itself. Furthermore, as illustrated in columns (2) and (3), the predicted value of the interaction coefficient between social attention and non-environmental subsidies is highly positive (β = 0.931, p < 0.01), whereas the interaction coefficient between social concern and environmental subsidies is considered statistically insignificant (β = −0.047, p > 0.1). The findings imply that social attention enhances the favorable impact of non-environmental subsidies on corporation ESG performance, although has no conspicuous effect on the correlation between environmental subsidies and ESG performance. One possible explanation for this disparity is that environmental subsidies, as a burgeoning class of designated funds, receive limited public notice, with greater attention instead directed toward non-environmental subsidies, for example, R&D subsidies. The findings suggest that hypotheses H5 and H5b are supported, while hypothesis H5a is not (Table 10).
| (1) | (2) | (3) | |
|---|---|---|---|
| Variables | baidu | baidu | baidu |
| ln_subsidy | 1.004*** | ||
| (7.01) | |||
| ln_environ | −0.047 | ||
| (−0.71) | |||
| ln_non_environ | 0.931*** | ||
| (7.12) | |||
| Constant | 17.34*** | 33.50*** | 18.54*** |
| (3.03) | (6.12) | (3.27) | |
| Observations | 20,057 | 20,057 | 20,057 |
| Controls | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes |
| Yearfixed effect | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
4.3 Heterogeneity test
In accordance with the “Management List of Listed Companies in Environmental Verification Industry Classification” (Environmental Protection Office Letter [2008] No. 373) formulated by the Ministry of Environmental Protection in 2008, we selected firms in our sample that belonged to heavy-polluting industries. The findings presented in Table 11.1 demonstrate that government subsidies have a pronounced influence on the ESG performance of companies that pollute heavily and those that do not. Heavy polluters' ESG performance is more sensitive to government subsidies (β = 0.082, p < 0.01). Similarly, environmental subsidies (β = 0.015, p < 0.01) and non-environmental (β = 0.058, p < 0.01) subsidies enhance ESG performance considerably more in heavily polluting firms than in nonheavily polluting firms.
| (1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
| Variables | pollute | non_pollute | pollute | non_pollute | pollute | non_pollute |
| ln_subsidy | 0.082*** | 0.061*** | ||||
| (4.65) | (6.70) | |||||
| ln_environ | 0.015*** | 0.012*** | ||||
| (2.77) | (3.46) | |||||
| ln_non_environ | 0.058*** | 0.056*** | ||||
| (4.09) | (6.74) | |||||
| Observations | 6182 | 16,733 | 6182 | 16,733 | 6182 | 16,733 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively. Robust standard errors are clustered at the firm level.
Table 11.2 demonstrates that government subsidies have a favorable impact on the ESG performance of high-tech and non-high-tech enterprises after distinguishing whether an enterprise is a high-tech enterprise or not. However, the effect on the ESG performance of non-high-tech enterprises is greater (β = 0.070, p < 0.01). Further examining the effects of environmental and non-environmental subsidies, we find that both environmental (β = 0.015, p < 0.01) and non-environmental subsidies (β = 0.058, p < 0.01) reflect a greater effect on the ESG performance of non-high-tech enterprises.
| (1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
| Variables | hightech | non_hightech | hightech | non_hightech | hightech | non_hightech |
| ln_subsidy | 0.055*** | 0.070*** | ||||
| (2.85) | (7.79) | |||||
| ln_environ | 0.011* | 0.015*** | ||||
| (1.94) | (4.60) | |||||
| ln_non_environ | 0.052*** | 0.058*** | ||||
| (2.98) | (7.37) | |||||
| Observations | 5817 | 17,098 | 5817 | 17,098 | 5817 | 17,098 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
The outcomes of Table 11.3 are noteworthy due to the fact that government subsidies have positively impact on the ESG performance of state-owned and non-state-owned enterprises. However, environmental and non-environmental subsidies have greater effects on non-state-owned enterprises than state-owned enterprises. Perhaps state-owned enterprises receive more subsidies and are more tightly regulated, so the marginal effect of subsidies on ESG performance is relatively weaker. Non-state-owned enterprises are faced with more serious resource constraints, and the bailout effect brought by government subsidies is stronger, so it is more conducive to private enterprises to improve their ESG performance by using government subsidies.
| (1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
| Variables | state | non_state | state | non_state | state | non_state |
| ln_subsidy | 0.060*** | 0.065*** | ||||
| (5.17) | (5.79) | |||||
| ln_environ | 0.010** | 0.011*** | ||||
| (2.08) | (3.22) | |||||
| ln_non_environ | 0.052*** | 0.056*** | ||||
| (5.28) | (5.27) | |||||
| Observations | 8263 | 14,652 | 8263 | 14,652 | 8263 | 14,652 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
4.4 Robustness test
4.4.1 Independent variables lagged by one period and inclusion of firm fixed effects
First, to mitigate the impact of endogeneity difficulties caused by reverse causality on the model, we substitute the independent variables with the logarithm of government subsidies from the previous period. The results presented in Table 12.1 confirm that lagged one-period government subsidies continue to have a significant positive impact on corporate ESG performance (β = 0.071, p < 0.01), with regression coefficients for both lagged one-period environmental subsidies (β = 0.014, p < 0.01) and non-environmental subsidies (β = 0.060, p < 0.01) also exhibiting statistically significant relationships. Columns (4)–(6) are the empirical results of firm fixed effects models, the coefficient of government, environmental and non-environmental subsides are also significantly positive. These results illustrate the reliability and validity of this study's conclusions.
| (1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
| Variables | huazheng_esg | huazheng_esg | huazheng_esg | huazheng_esg | huazheng_esg | huazheng_esg |
| L.ln_subsidy | 0.071*** | |||||
| (8.42) | ||||||
| L.ln_environ | 0.014*** | |||||
| (4.58) | ||||||
| L.ln_non_environ | 0.060*** | |||||
| (7.66) | ||||||
| ln_subsidy | 0.020** | |||||
| (2.51) | ||||||
| ln_environ | 0.008** | |||||
| (2.42) | ||||||
| ln_non_environ | 0.017** | |||||
| (2.31) | ||||||
| Observations | 19,885 | 19,885 | 19,885 | 24,024 | 24,024 | 24,024 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | No | No | No |
| Province fixed effect | Yes | Yes | Yes | No | No | No |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Firm fixed effect | No | No | No | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
4.4.2 Replacement of independent variables and expansion of the data sample
Second, we verify the robustness of the empirical results by replacing the independent variables. The explanatory variable in column (1) is the ratio of subsidy size to net profit while column (2) shows the size of environmental subsidies divided by operating profit. We find that the coefficient of government subsidies is also significantly positive in columns (1) and (2). Additionally, we broaden the sample in column (3) to incorporate firms listed after 2009 and observe that the outcomes remain resilient. To address potential endogeneity issues arising from sample selection factors, we use the Heckman two-stage approach to validate our findings. The results in columns (4) and (5) indicate that the impact of subsidies on ESG remains significant even after the inclusion of IMR. The coefficient on IMR is insignificant, indicating that there is no sample self-selection in the benchmark regression (Table 12.2).
| (1) | (2) | (3) | (4) | (5) | |
|---|---|---|---|---|---|
| Variables | huazheng_esg | huazheng_esg | huazheng_esg | dummy_esg | huazheng_esg |
| ln_subsidy | 0.065*** | 0.027*** | 0.063*** | ||
| (9.24) | (2.60) | (7.31) | |||
| subsidy_ratio | 0.003** | ||||
| (2.27) | |||||
| environ_ratio | 0.004** | ||||
| (2.14) | |||||
| imr | 0.219 | ||||
| (0.65) | |||||
| Observations | 24,024 | 24,628 | 25,230 | 21,169 | 20,791 |
| Controls | YES | YES | YES | YES | YES |
| Industry fixed effect | NO | NO | YES | YES | YES |
| Province fixed effect | NO | NO | YES | YES | YES |
| Year fixed effect | YES | YES | YES | YES | YES |
| Firm fixed effect | YES | YES | NO | NO | NO |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
4.4.3 Replacement of the dependent variable's measurement method
Finally, we change the measurement of ESG performance. In the original model, the index of ESG performance is the Huazheng Index. We replace the measures of corporate ESG performance with the Hexun Index and the Bloomberg Index. The positive effects of government, environmental, and non-environmental subsidies on corporate ESG performance remain significant. These findings support the robustness of the paper's conclusion (Table 12.3).
| (1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
| Variables | pengbo_esg | pengbo_esg | pengbo_esg | hexun_esg | hexun_esg | hexun_esg |
| ln_subsidy | 0.548*** | 0.129*** | ||||
| (16.39) | (7.05) | |||||
| ln_environ | 0.099*** | 0.012** | ||||
| (10.13) | (2.24) | |||||
| ln_non_environ | 0.448*** | 0.108*** | ||||
| (15.37) | (6.09) | |||||
| Constant | −6.918*** | 1.060*** | −5.340*** | |||
| (−11.30) | (3.18) | (−9.59) | ||||
| Observations | 22,915 | 22,915 | 22,915 | 21,260 | 21,260 | 21,260 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
5 FURTHER ANALYSIS
To further investigate the influence of government subsidies on the subcategories of ESG, this study scrutinizes the effects of subsidies on E, S, and G aspects and their underlying mechanisms. The findings presented in Table 13 indicate a significantly positive impact of government, environmental, and non-environmental subsidies on E, S, and G. Furthermore, according to the coefficients, they have greater impact on G, followed by S, and finally the E.
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | |
|---|---|---|---|---|---|---|---|---|---|
| Variables | pengbo_environ_esg | pengbo_environ_esg | pengbo_environ_esg | pengbo_social_esg | pengbo_social_esg | pengbo_social_esg | pengbo_govern_esg | pengbo_govern_esg | pengbo_govern_esg |
| ln_subsidy | 0.227*** | 0.624*** | 1.209*** | ||||||
| (8.58) | (11.39) | (12.34) | |||||||
| ln_environ | 0.054*** | 0.112*** | 0.189*** | ||||||
| (4.72) | (5.18) | (5.22) | |||||||
| ln_non_environ | 0.174*** | 0.514*** | 1.008*** | ||||||
| (7.30) | (10.54) | (11.46) | |||||||
| Observations | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively. Robust standard errors are clustered at the firm level.
Subsequently, we further analyze the mechanisms by which total and non-environmental subsidies affect E. As Bloomberg's database provides distinct E, S, and G ratings for companies, we utilize these to examine the question. Column (1) presents the regression coefficient of government subsidies is significantly positive (β = 0.074, p < 0.1). suggesting that government subsidies have a beneficial effect on environmental investment. In columns (2) and (3), the results also mean that both environmental and non-environmental subsidies have positive impact on environmental performance. Existing literature has proved that corporate environmental investment is beneficial to green technology, product, and process innovation, reducing corporate energy consumption and environmental pollution, and thus improve corporate ESG performance. Therefore, we conclude that environmental investment play mediation roles between government subsidies (including environmental and non-environmental subsidies) and E (Table 14.1).
| (1) | (2) | (3) | |
|---|---|---|---|
| Variables | environ_invest | environ_invest | environ_invest |
| ln_subsidy | 0.074*** | ||
| (2.62) | |||
| ln_environ | 0.036*** | ||
| (6.34) | |||
| ln_non_environ | 0.035* | ||
| (1.70) | |||
| Constant | −5.592*** | −4.578*** | −4.907*** |
| (−6.16) | (−5.42) | (−5.57) | |
| Observations | 22,708 | 22,708 | 22,708 |
| Controls | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
Furthermore, we examine how subsidies impact companies' S (social) ratings and describe the mechanism for this effect in detail. Columns (1), (2), and (3) demonstrate that government subsidies, including environmental, and non-environmental subsidies have positive effects on social performance. Environmental subsidies including those focused on rural environmental improvement and forestry reform and development, encompass poverty alleviation, thereby allowing for the enhanced capacity of environmental subsidies to improve corporate social governance. Non-environmental subsidies can enrich the cash flow of companies to invest in poverty alleviation, which is a national strategy in China. Enterprises that invest in poverty alleviation have a higher probability of receiving praise from the government. Therefore, government subsidies, environmental and non-environmental subsidies can improve firms' S (Social) by boosting their social performance (Table 14.2).
| (1) | (2) | (3) | |
|---|---|---|---|
| Variables | povertyreduction | povertyreduction | povertyreduction |
| ln_subsidy | 0.040** | ||
| (2.43) | |||
| ln_environ | 0.018*** | ||
| (3.44) | |||
| ln_non_environ | 0.033** | ||
| (2.14) | |||
| Constant | −2.180*** | −1.624*** | −2.059*** |
| (−4.23) | (−3.69) | (−4.07) | |
| Observations | 10,950 | 10,950 | 10,950 |
| Controls fixed effect | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
Finally, we analyze the mechanism by which government subsidies affect G (governance). We use the number of independent directors of a firm to measure its governance performance. Table 14.3 indicates both government subsidies, environmental and non-environmental subsidies have positive impact on governance quality. Previous studies show that the improvement of corporate governance will help companies make scientific decisions, enhance the efficiency of internal resource allocation, protect the interests of minority shareholders and employees, and then improve firms' G (Governance). According to the empirical results and existing literature, government subsidies, environmental and non-environmental subsidies improve G (Governance) through promoting governance quality.
| (1) | (2) | (3) | |
|---|---|---|---|
| Variables | independent_director | independent_director | independent_director |
| ln_subsidy | 0.003*** | ||
| (4.66) | |||
| ln_environ | 0.0004* | ||
| (1.71) | |||
| ln_non_environ | 0.003*** | ||
| (4.43) | |||
| Constant | 1.365*** | 1.409*** | 1.371*** |
| (108.84) | (164.90) | (114.40) | |
| Observations | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes |
- Note: z-values in parentheses, ***, **, and * indicate significant at the 1%, 5%, and 10% levels, respectively. Robust standard errors are clustered at the firm level.
6 CONCLUSION AND RECOMMENDATIONS
We investigate the influence of government subsidies on ESG ratings of firms by employing data from listed Chinese companies spanning from 2009 to 2019. Based on benchmark regression results, we find that while obtaining the maximum ESG rating is challenging for enterprises, subsidies distributed by the government could tangibly enhance their overall ESG performance. We also note that similar outcomes are observed across different types of subsidies, including both environmental and non-environmental subsidies. As environmental concerns gain traction in the country, the government has shifted its focus toward environmental subsidy programs. Mechanism analyses indicate that government subsidies and non-environmental subsidies have a positive impact on firms' ESG performance by reinforcing firms' green innovation, mitigating financial constraints, increasing charitable donations and attracting social attention. Although environmental and non-environmental subsidies enhance corporate green innovation capacity, non-environmental subsidies strengthen both inventive and utility green innovation while environmental subsidies solely promote utility green innovation, leading to an improvement in ESG performance. The heterogeneity analysis indicates that government subsidies have a more pronounced impact on ESG performance in heavily polluting, non-high-tech, and non-state corporations.
In addition, this article provides practical policy suggestions for businesses to enhance their ESG assessments. The ESG performance of corporations is vital for attaining sustainable and superior economic growth, reliant on competent policy formulation and implementation (Wang, Wang, et al., 2022). A notable proportion of Chinese entrepreneurs express a sense of social responsibility. The government must incentivize companies to carry out more activities in support of environmental and social development and facilitate the further implementation of their ESG strategies. First, the government should heighten companies' understanding of the necessity of social responsibility to develop their ESG performance. Simultaneously, the government ought to offer financial incentives to businesses to guarantee their usual profitability while tackling social and environmental concerns. Second, in light of the heterogeneity analysis outcomes of this study, public authorities should channel their subsidies toward heavily polluting, non-high-tech, and private enterprises to bolster these corporate ESG performance. This can be achieved by guiding and incentivizing enterprises to hasten the adoption of advanced production technologies and processes, supporting the establishment of eco-friendly production capacity, and motivating enterprises to carry out research and development to innovate and showcase corporate social responsibility. Social media can be employed to steer public attention and motivate businesses to actively integrate ESG strategies. Finally, greater emphasis should be placed on enhancing the efficacy of environmental subsidies to prompt companies to enhance their environmental impact. These measures will enable the government to optimize the societal worth of subsidies.
Enterprises ought to develop and implement an ESG strategy and increase investment in prosocial initiatives, such as environmentally friendly innovation, safeguarding employee rights, and promoting social responsibility. National policymakers have heightened their expectations of enterprises to generate greater social and environmental value. For this reason, enterprises should comprehend the core requirements of national policy and amend their internal operations and governance models to enhance ESG performance. Furthermore, it is crucial for companies to manage their relationship with the government effectively to capture the policymakers' interest. Furthermore, it is crucial for companies to manage their relationship with the government effectively to capture the policymakers' interest. This approach can create a strong foundation for businesses to undertake ESG-related initiatives with government funding, establishing a reliable resource base.
AUTHOR CONTRIBUTIONS
Pei Peng: Conceptualization; data curation; formal analysis; funding acquisition; investigation; methodology; project administration; resources; software; supervision; writing—original draft; writing—review and editing. Mengzi Sun: Conceptualization; funding acquisition; investigation; methodology; project administration; writing—original draft; writing—review and editing.
ACKNOWLEDGMENTS
This paper was supported by the Graduate Student Innovation Fund of Shanghai University of Finance and Economics, “The Impact of US Industrial Subsidies on US-China Supply Chain Relations” (Project number: CXJJ-2022-406) and the Special Fund of Basic Scientific Research and Operation Expenses of Central Universities, “Research on the Construction of Digital Countryside and the Entrepreneurial Activity of Farmers“ (Project number: CXJJ-2022-405).
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.
ETHICS STATEMENT
Not applicable.
APPENDIX A
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| Variables | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| ln_environ | 0.008 | 0.008* | 0.016*** | 0.011*** | 0.017*** | 0.008 | −0.033 |
| (1.03) | (1.87) | (6.28) | (5.04) | (4.35) | (0.52) | (−0.44) | |
| Constant | 4.356*** | 2.977*** | 1.245*** | −0.663*** | −2.892*** | −5.896*** | −8.703*** |
| (25.74) | (18.64) | (7.96) | (−4.25) | (−18.24) | (−29.53) | (−16.56) | |
| Observations | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively.
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| Variables | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| ln_non_environ | 0.021 | 0.019* | 0.037*** | 0.069*** | 0.137*** | 0.23*** | 0.32 |
| (1.00) | (1.70) | (5.52) | (10.12) | (9.78) | (4.29) | (1.36) | |
| Constant | 4.102*** | 2.743*** | 0.77*** | −1.67*** | −4.99*** | −9.64*** | −14.19*** |
| (11.28) | (11.43) | (4.08) | (−8.87) | (−17.98) | (−10.21) | (−3.35) | |
| Observations | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 | 22,915 |
| Controls | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Industry fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Province fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Year fixed effect | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
- Note: The values in parentheses are z-values, and ***, **, and * represent significant values at the levels of 1%, 5%, and 10%, respectively.
REFERENCES
- 1 The literature on R&D subsidies includes works by Yuan et al. (2022); Czarnitzki and Toole (2007); Almus and Czarnitzki (2003); Goel and Ram (2001); Görg and Strobl (2007).
- 2 Solving social problems cannot rely on the government alone, but needs the market and the government to jointly seek solutions (Saka-Helmhout et al., 2022). Direct government investment in charity may not a sustainable way to address environmental and social issues. So the government should encourage enterprises to explore solutions of social problems through developing green innovation, engaging in industrial poverty alleviation and community building activities. Government subsidies not only enhance the cash flow flexibility of enterprises, but also boost the confidence of enterprises to participate in social construction and philanthropy, which is beneficial to enterprises to actively participating in ESG-related activities and improve ESG performance.
- 3 In reality, no companies receive an AAA rating in the Huazheng ESG ratings, resulting in a range of 1–8 for companies.
- 4 Through big data technology, the Huazheng ESG rating system selects more than 300 underlying data indicators, combining semantic analysis, natural language processing (NLP), and other intelligent algorithms, and selects 17 environmental key indicators, 13 social key indicators, and 14 governance key indicators as evaluation criteria for each theme. Compared with other ESG ratings, the Huazheng ESG Rating covers a wider time span, including data from 2009 to 2022, while the Hexun ESG Rating only covers 2010 to 2020 and the Bloomberg ESG Rating only covers 2011–2020. In addition, the Huazheng ESG ratings also give nine rating intervals, making the differences in ESG scores between companies more prominent. Therefore, we end up adopting the Huazheng ESG rating data as the primary dependent variable for the overall comparison.
- 5 Government subsidies exhibit bias toward specific industries. Furthermore, incorporating firm-fixed effects would pose complications in identifying industry-level characteristics and evaluating the impact of time-varying firm-level variables on independent variables. To alleviate potential issues with omitted variables, we incorporate both firm-level control variables and time-level fixed effects, in addition to controlling for industry and province fixed effects. Furthermore, we conduct empirical analysis in the heterogeneity test to examine the effect of firm variables that do not alter over time on companies' ESG ratings. Finally, the study accounts for firm-level fixed effects in the robustness tests and confirms that the results remain robust.
- 6 As there is currently no suitable method for adding IV to the Ologit model, the IV-Oprobit and IV-OLS methods have been employed here. Both the probit and logit models can be used to address discrete models. Nevertheless, empirical results from this paper suggest that the Oprobit model can also confirm the research outcomes. However, the logit model is more adaptable and easier to interpret, whereas the probit model is more suited for dichotomous variables. Therefore, the baseline model in this article continues to use the Ologit model.
