Whether and How ESG Impacts on Corporate Financial Performance in the Yangtze River Delta of China

Abstract

ESG (Environmental, Social and Governance) is not only a comprehensive manifestation of sustainable development but also an influencing factor of economic benefits. This research focuses on evaluating the impact of environment (E), society (S) and corporate governance (G) on the corporate financial performance in China. As China is currently in the exploration stage of ESG, a socially approved ESG evaluation system has not yet been formed. This paper deliberately selected variables and composite methods for E, S and G and integrated them into the ESG index through learning from the western experience and Chinese actuality. Then, whether E, S, G and ESG have a significant positive impact on financial performance is evaluated through panel regression analysis based on a sample of 191 listed companies in the Yangtze River Delta of China from 2015 to 2020. The results show that E has a significant negative impact on corporate financial performance, G has a significant positive impact, and S has no significant impact. ESG performance has a less significant impact on accounting-based financial performance and no significant impact on market-based financial performance. This research will help understand the performance of ESG and promote ESG practices in China.

1. Introduction

The concept of ESG (Environmental, Social and Governance) originated from ethical investment and responsible investment and represents the three major factors of enterprises in the environment, society and corporate governance [1]. Like ethics and social responsibility, ESG is a guide for company strategic management, risk management and non-financial performance. Broadly speaking, ESG covers various issues related to environment (e.g., climate change, energy and water use, carbon emissions), social responsibility (e.g., fair trade principles, human rights, product safety, gender equality, health and safety) and corporate governance (e.g., board independence, corruption and bribery, reporting and disclosure, shareholder protection). In Western countries, ESG highly concerns and is recognized by the government, the general public and investors, and it has a large market participation group as well as a mature evaluation system. The ESG concept is not only a comprehensive manifestation of sustainable development but also an influence factor of economic benefits.

In China, with the introduction and implementation of the development concept of “innovation, coordination, green, openness, and sharing” [2], ESG has gradually attracted the attention of governments, financial institutions and investors. Similar to the Global Reporting Initiative (GRI), which originally focused on the environmental performance reporting framework, the early ESG-related regulatory documents of China also mainly focused on environmental information disclosure. In June 2017, the Ministry of Environmental Protection of China and the China Securities Regulatory Commission jointly issued the “Cooperation Agreement on Jointly Carrying Out the Environmental Information Disclosure of Listed Companies”, establishing and improving the mandatory environmental information disclosure system for listed companies. In September 2018, the “Guidelines for the Governance of Listed Companies” revised by the China Securities Regulatory Commission added the content of environmental protection and social responsibility, which clarified the responsibilities of listed companies to stakeholders, employees and the social environment. Meanwhile, it also highlighted listed companies’ environmental protection as well as the guiding role of social responsibility and established the basic framework of ESG information disclosure.

In the field of ESG practice, there are many precedents for serious consequences caused by companies’ violations of ESG concepts. High leverage, diversified cross-border operations and ignoring the governance elements of the ESG system led to Lehman Brothers’ applications for bankruptcy protection [3]. The Deepwater Horizon oil spill in the Gulf of Mexico in 2010 by BP caused an economic and environmental tragedy [4]. In recent years, with the growing voice of ESG, organizations and enterprises have been actively practicing the ESG concept. In July 2019, South Korea’s Posco Iron and Steel Co., Ltd. issued a five-year ESG bond worth US $500 million to the world and was committed to building an iron and steel industry ecosystem.

With the development of ESG practice, there has been a wealth of ESG academic research. Among them, the relationship between ESG and corporate financial performance is always a focus of debate among researchers [1]. For developing China, since the practice of ESG is still in the initial stage of understanding and exploration, what is the performance of ESG, and how does ESG impact corporate financial performance? The existing literature is scarce, which is not conducive to the practice of ESG. This paper studies the relationship between ESG and the financial performance of Chinese listed companies based on deliberately selected variables and composite methods for the environment (E), society (S), corporate governance (G) and ESG indices. This research will help improve the understanding of ESG and its economic benefits in China, as well as promote the high-quality, sustainable development of ESG in China.

Compared with the existing literature on ESG topics in China, the expected contributions of this paper are as follows. First, we conduct empirical research to analyze the impact of ESG on corporate financial performance in China so as to fill up the research gap. We not only provide evidence for the relationships among E, S and G performance and financial performance variables but also describe the mechanism of the relationships. Second, we comprehensively consider three aspects of ESG to build the ESG Index system learning from the Western experience and Chinese reality, while environmental, social responsibility and governance aspects are conducted separately in the existing studies. Third, our selected samples are the listed companies in the Yangtze River Delta. The Yangtze River Delta has a pivotal position and demonstration effect on the national economy, boasting a relatively developed market system and a mature capital market where the listed companies are expected to follow ESG practices and attain more advantages. Fourth, we provide recommendations and measures based on our empirical analysis, which will help promote the ESG practice and the implementation of high-quality integrated development in combination with the special national conditions of China.

The rest of this paper is presented as follows. The next section will introduce related literature and hypotheses. The third section describes research design, including the selection of samples, selection of variables, index synthesis method, data sources and empirical models. The fourth and fifth sections discuss the results of empirical research analysis, and the last section consists of conclusions and suggestions.

2. Literature Review and Research Hypothesis

2.1. Literature Review

2.1.1. ESG

In developed countries, ESG assessment data has originated from the earliest studies based on individual company annual report and company website data, then gradually developed to be available in the databases of some business information companies, and the ESG system is relatively mature [5]. For example, the KLD database and the Bloomberg database can provide queries of enterprise ESG performance data and ESG disclosure data, respectively.

The KLD database contains over 60 ESG performance indicators in seven categories for three dimensions of sustainability performance and uses a binary representation of ESG ratings. If a company meets the criteria established by the rating, the value of the corresponding variable is “1”. Otherwise, it is designated as “0”. The ratings in each ESG category are divided into two groups of indicators to measure the performance of best practices (strengths) and the most severe challenges (focus) [6].

The ESG disclosure information of the Bloomberg database was first released in 2009. Bloomberg currently compiles approximately 300 data points in about 11,000 companies in 63 countries. Bloomberg assesses the level of ESG activities of each company through the collection of public information. The raw data of Bloomberg’s ESG comes from corporate social responsibility reports, annual reports, corporate websites and other related company documents, which reflect the scope of publicly available information for investors. The data disclosed in the Bloomberg database have scores between 0.1 (lowest) and 100 (highest) [7].

Due to the lack of public data on enterprises, the current research on ESG evaluation in China mainly focuses on one aspect of ESG evaluation. For example, Shen and Ma [8] selected eight indicators from the three dimensions of pollution discharge, environmental management and social impact and used AHP to analyze the performance of environmental performance. Yang and Yang [9] used Runling Global Responsibility Rating Score as the evaluation of corporate social performance. He and Chen [10] selected key variables from the equity structure, the governance of the board of directors, the board of supervisors and the management of the management and then used principal component analysis to extract common factors for the key indicators and perform weighted summation to constitute the corporate governance performance evaluation. The literature on the comprehensive evaluation of ESG is very scarce and mainly focuses on the domestic references of international research. For example, the Industrial and Commercial Bank of China’s Green Finance Task Force [11] learned from the relevant practices of international rating agencies and selected 32 key indicators in 17 dimensions to build the ESG scores of the companies included in the SSE 180 Stock Index. The Green Finance International Research Institute of the Central University of Finance and Economics [12], China Securities Investment Fund Association, and Cao and Xu [3] summarized or constructed their respective ESG evaluations by learning from existing ESG indicators at home and in indicator systems abroad. The other type is related to empirical research. For instance, Qiu and Yin [13] focus on the relationship between ESG and corporate financing costs. Referring to these studies, the ESG index system in this study will be constituted through learning from the Western experience and Chinese actuality and will be applied in our empirical research.

2.1.2. ESG and Corporate Financial Performance

In general, existing empirical studies have shown mixed and even contradictory results between ESG performance and corporate financial performance. Studies have found that there is a positive correlation between the two performances [14,15]. Velte (2017) selected 412 German companies from 2010 to 2014 to study the relationship between ESG and corporate financial performance. He found that ESG performance has a positive impact on corporate financial performance and that governance performance has the greatest impact on corporate financial performance. Margolis et al. (2009) summarized 251 individual empirical studies (214 manuscripts) and found that the average positive correlation coefficient between ESG performance and financial performance was 0.133. They also studied a sample of 106 pieces of research published since 1998 and found that the average positive correlation effect was only 0.090, indicating that the relationship between the two performances may actually weaken over time. However, some other studies claim that there is a negative correlation between them. Garcia et al. [4] considered the data of 365 listed companies in sensitive industries from Brazil, Russia, India, China and South Africa BRIC from 2010 to 2012. Their results showed that companies in sensitive industries had better environmental performance, and the companies with the best ESG performance tended to have lower profits. In addition, some studies have found that there is no significant correlation between ESG performance and corporate financial performance. Al-tuwaijri et al. [16] used the structural equation model to research and found that environmental performance has a positive impact on economic performance, but the impact of economic performance on environmental performance is insignificant. Kuo et al. [17] used a multilevel quadratic growth model to investigate the impact of airlines’ disclosure of ESG performance indicators; the results revealed that in the initial stages of implementation of ESG-based practices, airlines demonstrate a downward trend in return on assets. However, it gradually increases after a period of incorporation and implementation. Galletta et al. [18] conducted on 271 publications over the 1986–2021 period, introducing a variety of findings, including the top authors at the journal and institution levels, citations, keyword distribution, highly cited works, co-authorship and the most influential journals and authors.

Furthermore, we found that empirical research on ESG and financial performance focused more on developed countries such as Europe and the United States. In the past two years, we have discovered a small number of related studies on Asian countries. Alsayegh et al. [19] conducted empirical research on the impact of ESG information disclosure on EES sustainability performance in Asian companies from 2005 to 2017, showing that active ESG disclosure can enhance the company’s sustainability performance. Tuan-Hock et al. [20] studied the relationship between financial development and ESG using data from Asian countries/regions from 2013 to 2017, showing that financial development is directly proportional to the success of ESG. Zhou et al. [21] Used the ESG rating data of Chinese listed companies newly developed by SynTao Green Finance from 2014 to 2019, showing that the improvement of ESG performance of listed companies can improve the market value of the company, and the financial performance of the company presents an obvious mediating effect. Chen and Xie [22] used a sample of non-financial listed companies from 2000 to 2020 and applies the staggered difference-in-differences technique to eliminate the endogeneity problem, showing that ESG disclosure has a favorable effect on corporate financial performance.

2.2. Research Hypothesis

Stakeholder theory is the core theory of the ESG theme [4]. Freeman and McVea [23] believed that companies should make decision-making activities that conform to the interests of groups or individuals (i.e., stakeholders). In short, the stakeholder theory points out that, in principle, the sustainable value produced by a company is measured by meeting the specific social expectations of various stakeholders. In order to continuously achieve these expectations of stakeholders, sustainability management is required. Sustainability management activities are effective tools for stakeholder communication, reflecting the connections among stakeholder power, sustainability performance and sustainability reports [24]. As stakeholders are interested in the company’s sustainable development strategy, better ESG performances will lead to better sustainable performances, such as ESG ratings or reputation [25]. Eventually, as the trust of stakeholders increases, the company’s financial status will also be improved [14].

Kim et al. [26] believed that companies may engage in two types of ESG strategies. One strategy is that companies are committed to environmental and social ethical behaviors and consume a lot of resources to implement corporate ESG practices. Therefore, these companies may have beneficial and positive results on the long-term development and other related issues, thereby achieving better financial results and higher social legitimacy. Another strategy for participating in ESG practice activities involves participating in symbolic and opportunistic activities, and companies adopting this strategy are defined as “green washing” companies. They tried to symbolically improve corporate images, but basically, they did not substantively participate in the environmental and social responsibility governance mechanism. Therefore, these “green washing” companies might have a legitimacy gap due to their lower ESG practices and might not bring significant returns in the short and long terms [1].

With the increasing problems of environmental safety and social safety, the international community, including governments, the general public and investors, is paying more and more attention to the ESG system. China also attaches great importance to the development of ESG. On 30 September 2018, the “Code for Corporate Governance for Listed Companies” revised by the China Securities Regulatory Commission (CSRC) established the basic framework for ESG information disclosure in China. In recent years, major enterprises and financial institutions have gradually increased their ESG practical activities.

With the help of the ESG supervision, it is possible to better guide the value investment direction of investors and promote the environmentally friendly behavior, social responsibility behavior, the non-financial information disclosure and corporate governance of listed companies. As a result, a positive interaction between stakeholders is formed to promote the high-quality and sustainable development of listed companies, so as to improve corporate financial performance.

Therefore, we assume that ESG practices can increase the financial performance of Chinese companies and formulate the following hypothesis:

H1. 

ESG has significant positive impacts on corporate financial performance.

After decomposing ESG from three dimensions, H1 becomes three sub-hypotheses.

H1a. 

E has a significant positive impact on corporate financial performance.

H1b. 

S has a significant positive impact on corporate financial performance.

H1c. 

G has a significant positive impact on corporate financial performance.

3. Research Design

3.1. Selection of Samples and Variables

3.1.1. Selection of Samples

Regarding data collection, three criteria were employed for the selection of companies. First, the Outline of the Yangtze River Delta Regional Integration Development Plan (2019) issued by the State Council of the Communist Party of China states that the Yangtze River Delta region is one of the regions with the most active economic development, the highest degree of openness and the strongest innovation capacity in China. Thus, we considered the listed companies in the Yangtze River Delta region of China. Second, we did not consider financial companies because of the specificity of their operational activity. Third, we excluded stocks with ST (Special Treatment) and *ST during the sample period. After filtering, the total samples were composed of 1146 observations from 191 companies from 2015 to 2020. The data were collected from three different databases. The basic data of ESG performance were collected first from the China Stock Market & Accounting Research Database (CSMAR) and the Institute of Public and Environmental Affairs (IPE) website (http://www.ipe.org.cn/, which accessed on 6 January 2022). All financial data were collected from the Wind database (https://www.wind.com.cn/newsite/edb.html, which accessed on 6 January 2022). In order to eliminate the influence of outliers, Winsorize reduction processing was performed on all continuous variables at the 1% and 99% quantiles.

3.1.2. Selection of Dependent Variable

The dependent variables of this research adopt accounting-based and market-based financial performance measurement indicators. Specifically, return on assets (ROA) is used as the accounting-based financial performance indicator, and Tobin’s Q is used as the market-based financial performance indicator. ROA and Tobin’s Q have been widely used in business research to measure corporate financial performance [1]. The calculation method of ROA is to divide the income before special items by total assets, representing the company’s income after controlling total assets. Therefore, a higher ROA indicates better financial performance and comparability among companies with different operating sizes [27]. Similar to the research of Zhang Wang & Joseph Sarkis (2017), Tobin’s Q is used to measure a company’s operational efficiency and its ability to generate good financial performance. A Tobin’s Q value greater than “1” indicates that the market price of the company is higher than the replacement cost of the company and that the company has better financial performance than its accounting books [28]. Existing empirical literature shows that there are mixed and contradictory results between corporate ESG performance and accounting-based and market-based financial performance. In this paper, the ROA and Tobin’s Q of Chinese listed companies are used to re-research the relationship between them.

ESG Index System

(1) Environmental performance

At present, Chinese ESG is still in the exploratory stage. The Code of Corporate Governance for Listed Companies in China revised by the China Securities Regulatory Commission in September 2018 just established the basic framework for ESG information disclosure. Most existing ESG studies use the proxy indicators of environment, social responsibility and corporate governance. In terms of corporate environmental performance (E) in ESG, due to the lack of publicly available data, we learned from the research of Qiu and Yin [13]. The number of environmental penalties in the year disclosed by Institute of Public and Environmental Affairs (IPE) website is used as the basic measurements of corporate environmental performance. We sum up the number of environmental penalties suffered by listed companies and their subsidiaries during the accounting year. Since the data are reverse indicators, we use the maximum value in the data to subtract the current indicator value to convert the reverse indicator into a positive indicator. It can be seen that the higher the E index is, the less the environmental punishment is and the better the environmental performance is.

(2) Social performance

For the evaluation of social performance (S) in ESG, most existing literature adopts the responsibility rating score of Hexun.com or Rankins CSR Ratings (RKS) as the social performance score of enterprises [9]. The indicators of Hexun.com and RKS are quite different from the social responsibility specified in the Code of Corporate Governance for Listed Companies in China (2018). We compare the criteria to select seven variables (all dummy variables. If “yes”, the value is 1. Otherwise, the value is 0) for principal component analysis. These variables are the company’s annual social donations, a dummy variable of whether to disclose the protection policy of shareholders’ rights and interests, a dummy variable of whether to disclose the protection policy of creditors’ rights and interests, a dummy variable of whether to disclose the protection policy of employees’ rights and interests, a dummy variable of whether to disclose the protection policy of suppliers’ rights and interests, a dummy variable of whether to disclose the protection policy of customers’ rights and interests and consumers’ rights and interests and a dummy variable of whether to disclose the safety production. Then, we use the first principal component as the corporate social responsibility performance score (S). The higher the S index value, the better the corporate social responsibility performance.

We use the principal component analysis method to extract the first principal component and perform Winsorize reduction processing at 1% and 99%, then define it as the corporate social performance (S) indicator. In the first principal component analysis, the loading factors of the seven variables are 0.0510, 0.4514, 0.2370, 0.4894, 0.1333, 0.3999 and 0.4677, respectively. The coefficients of all variables are consistent with our expectation. Among the seven variables, there are three important variables reflecting the S index. They are the dummy variable of whether to disclose the protection policy of employees’ rights and interests, the dummy variable of whether to disclose the protection policy of customers’ rights and interests and consumers’ rights and interests and the dummy variable of whether to disclose the protection policy of shareholders’ rights and interests.

(3) Governance performance

For the evaluation of governance performance (G) in ESG, most existing studies selected key variables from ownership structure, board of directors, supervisory board governance and management governance. These variables measure a company’s systems and processes that ensure that its board members and executives act in the best interests of its long-term shareholders [14]. Then, we use principal component analysis to extract common factors for key indicators and conduct weighted summation to form corporate governance performance scores [29,30,31]. This study uses their methods and selects the following nine governance variables: the proportion of shares held by the largest shareholder, the proportion of shares held by the top ten shareholders, the number of shareholders’ meetings, the proportion of tradable shares, the dummy variables of the integration of two positions (director and general manager concurrently) (if the president and general manager are the same person, the value is 1. Otherwise, it is 0), the management shareholding, the proportion of independent directors, the number of boards of directors, the dummy variable of whether it is state-owned or not (the value for state-owned enterprises is 1, and that for others is 0). Among them, the proportion of shares held by the largest shareholder, the proportion of shares held by the top ten shareholders, the number of shareholders’ meetings and the proportion of tradable shares reflect the equity structure mechanism. The dummy variable of whether it is state-owned or not reflects the corporate holding structure with Chinese characteristics. The dummy variables of the integration of two positions and the management shareholding reflect the governance mechanism of management. The proportion of independent directors and the number of boards of directors reflect other forms of governance such as directors.

We still use the principal component analysis method commonly used in statistics, to extract the first principal component and perform Winsorize reduction processing at 1% and 99%, then define it as the corporate governance performance score (G). In the first principal component analysis, the loading factors corresponding to the nine variables are −0.4514, −0.0926, 0.2813, 0.2998, 0.3816, 0.5076, 0.1011, 0.3029 and −0.5115, respectively. Except for the proportion of shares held by the top ten shareholders, the coefficients of the other eight variables are consistent with the results from Bai et al. [29] and Mao and Jin [31]. Among the nine variables, the dummy variable of whether it is state-owned or not, the management shareholding and the number of boards of directors are important variables reflecting the G indicator.

(4) ESG

The next step is the construction of ESG performance variables. ESG performance is to synthesize the performance of enterprises in E, S and G into a comparable value. Like many other studies, a structured expert scoring method is used to obtain the weights of E, S and G (Industrial & Commercial Bank of China Green Finance Research Group, 2017). Based on the scores of experts, the weights of E, S and G are 3-2-5 (i.e., E accounts for 30%, S accounts for 20%, and G accounts for 50%). Combining the single empirical research results of ESG in the previous part and the Research Report on ESG Evaluation System of Chinese Listed Companies issued by China Securities Investment Fund Association, it is mentioned that “among the three contents of ESG, governance factors are relatively significant, followed by environmental factors, and finally social factors”. Because the weight of 3-2-5 reflects the relative importance of stakeholders, we believe that it is reasonable. (Weight is an important indicator that affects the empirical results. In fact, empirical analyses with other weights are also made. The results can be found in the fifth section.) At the same time, we standardize the E, S and G indicators, and then weight them according to the weight of 3-2-5, so as to obtain the ESG index of the enterprise.

Control Variables

This research also introduces control variables related to enterprise scale, enterprise financial risk and enterprise profitability to eliminate the influence of irrelevant variables. Specifically, ① Use the natural logarithm of total assets (SIZE) to control the size of the company. ② Use the liquidity ratio (LR) and the quick ratio (QR) to control the impact of short-term solvency (financial risk). Intuitively, a higher leverage ratio may indicate a higher financial risk, so the financial performance is poor. ③ The three-year net profit growth rate (NETBENE) of business activities is used to reflect the profit growth of an enterprise. Rangan [32] believes that the company experiencing higher growth needs to allocate more working capital to its investment, therefore, it may affect their short-term profitability and ESG implementation. The use of asset turnover rate (ASSET_TURN) reflects the management quality and utilization efficiency of company. These variables usually adopted in the finance literature [1,13]. We perform Winsorize reduction processing on all control variables at 1% and 99%. In order to eliminate the impact of business cycle fluctuations, this study also uses the year as a control variable. The definition and description of variables are shown in

Table 1. Definition and Description of Variables.

Variable Name and Variable Symbol	Data Sources	Variable Definition	Theoretical Basis
Return on assets ROA (Dependent variable)	CSMAR	=Net profit/Total assets	Wang and Sarkis [1]
Tobin’s Q Tobin’s Q (dependent variable)	CSMAR	=Market value/Book value	Wang and Sarkis [1]
Environmental performance of ESG E (independent variable)	IPE website	=The maximum value of environmental punishment data of all listed enterprises and their subsidiaries in 2012–2017—the number of environmental penalties of listed enterprises in the current year.	Qiu and Yin [13] Wang and Sarkis [1]
Social performance of ESG S (independent variable)	CSMAR	First principal component: the company’s annual social donations, a dummy variable of whether to disclose the protection policy of shareholders’ rights and interests, a dummy variable of whether to disclose the protection policy of creditors’ rights and interests, a dummy variable of whether to disclose the protection policy of employees’ rights and interests, a dummy variable of whether to disclose the protection policy of suppliers’ rights and interests, a dummy variable of whether to disclose the protection policy of customers’ rights and interests and consumers’ rights and interests, a dummy variable of whether to disclose the safety production	Luo [33]; He and Chen [10] Qiu and Yin [13]
Governance performance of ESG G (independent variable)	CSMAR	First principal component: the proportion of shares held by the largest shareholder, the proportion of shares held by the top ten shareholders, the number of shareholders’ meetings, the proportion of tradable shares, the dummy variables of the integration of two positions (director and general manager concurrently), the management shareholding, the proportion of independent directors, the number of boards of directors, the dummy variable of whether it is state-owned or not	Mao and Jin [28]
ESG performance ESG (independent variable)	E, S, G	E, S and G indicators were standardized first, and then weighted according to the weight of 3-2-5	China Securities Investment Fund Association (2018)
Company size SIZE (control variable)	Wind	=Natural logarithm of total assets	Harjoto and Jo [34] Wang and Sarkis [1]
Liquidity ratio LR (control variable)	Wind	=Current assets/Current liabilities	Garcia et al. [4] Qiu and Yin [13]
Quick ratio QR (control variable)	Wind	=Quick assets/Current liabilities	Garcia et al. [4] Qiu and Yin [13]
Net profit margin growth NETBENE (control variable)	Wind	=3-year average growth rate of net profit	Garcia et al. [4] Qiu and Yin [13]
Asset turnover ASSET_TURN (control variable)	Wind	=Total turnover/Total assets	Garcia et al. [4] Qiu and Yin [13]

.

3.2. Model Specification

This research evaluates whether E, S, G and ESG performance have impacts on corporate financial performance. Referring to the paper of Velte [14], financial performance measures are divided into accounting-based variable (ROA) and market-based variable (Tobin’s Q). Then, we establish the following models to evaluate the impact of ESG and their three components on the corporate financial performance including both accounting-based and market-based measures in China.

$$\begin{gathered} RO{A}_{it}={\beta }_0+{\beta }_1{E}_{it}+{\beta }_2{S}_{it}+{\beta }_3{G}_{it}+{\beta }_{4}SIZ{E}_{it}+{\beta }_{5}L{R}_{it}+{\beta }_{6}Q{R}_{it} \\ +{\beta }_{7}NETBEN{E}_{it}+{\beta }_{8}ASSET\_TUR{N}_{it}+{\epsilon }_{it} \end{gathered}$$

(1)

$$\begin{gathered} TOBIN{Q}_{it}={\beta }_0+{\beta }_1{E}_{it}+{\beta }_2{S}_{it}+{\beta }_3{G}_{it}+{\beta }_{4}SIZ{E}_{it}+{\beta }_{5}L{R}_{it}+{\beta }_{6}Q{R}_{it} \\ +{\beta }_{7}NETBEN{E}_{it}+{\beta }_{8}ASSET\_TUR{N}_{it}+{\epsilon }_{it} \end{gathered}$$

(2)

$$\begin{gathered} RO{A}_{it}={\alpha }_0+{\alpha }_{1}ES{G}_{it}+{\alpha }_{2}SIZ{E}_{it}+{\alpha }_{3}L{R}_{it}+{\alpha }_{4}Q{R}_{it}+{\alpha }_{5}NETBEN{E}_{it} \\ +{\alpha }_{6}ASSET\_TUR{N}_{it}+{\epsilon }_{it} \end{gathered}$$

(3)

$$\begin{gathered} TOBIN{Q}_{it}={\alpha }_0+{\alpha }_{1}ES{G}_{it}+{\alpha }_{2}SIZ{E}_{it}+{\alpha }_{3}L{R}_{it}+{\alpha }_{4}Q{R}_{it} \\ +{\alpha }_{5}NETBEN{E}_{it}+{\alpha }_{6}ASSET\_TUR{N}_{it}+{\epsilon }_{it} \end{gathered}$$

(4)

In the models, $RO{A}_{it}$ and $TOBIN{Q}_{it}$ are the data of the $ith (i=1,\cdots ,N)$ listed company in year $t (t=1,\cdots ,N)$. ${\epsilon }_{it}$ are independent and identically distributed random variables with 0 expectation and the same variance. The panel regression analysis model is used to test the three aspects of E, S, G and the relationship between ESG performance and corporate financial performance.

4. Empirical Analysis of E, S, G and Financial Performance

4.1. Descriptive Statistics

Table 2. Descriptive statistical analysis results.

Variable	Mean	Median	Standard Deviation	Maximum	Minimum
Panel A: E, S, G
E	12.54	13.00	0.49	13.00	0.00
S	14.62	6.51	40.97	250.87	0.00
G	5.76	4.78	2.13	25.35	0.48
Panel B: Financial performance
ROA	9.84	7.76	6.08	30.44	−9.25
TOBIN’S Q	2.33	2.21	1.99	9.01	0.12
Panel C: Control variable
SIZE	20.28	19.21	1.75	22.39	5.77
LR	2.87	2.23	12.01	9.97	0.33
QR	2.01	1.12	1.52	9.54	0.09
NETBENE (%)	112.24	50.77	401.51	2551.18	−901.37
ASSET_TURN	0.89	0.77	0.89	2.95	0.04

shows the descriptive statistical results of each variable. Panel A reflects the descriptive statistical results of independent variables, panel B reflects the descriptive statistical results of dependent variables, and panel C reflects the descriptive statistical results of control variables.

From Table 2, we can see that the mean value of E performance is 12.54, the median is 13 and the standard deviation is 0.49. This indicates that the distribution of E performance of these listed companies is relatively concentrated. The mean value of S performance is 14.62, the median is 6.51, and the standard deviation is 40.97, which is about 3 times the mean. The minimum value is 0, and the maximum value is 250.87. This shows that the S performance of the selected part of the listed companies is different. The mean value of G performance is 5.76, the median is 4.78, and the standard deviation is 2.13. This indicates that the distribution of G performance of these listed companies is relatively concentrated. The mean of the enterprise ROA and Tobin’s Q are 9.84 and 2.33, respectively. The standard deviation of the enterprise ROA and Tobin’s Q are 6.08 and 1.99, respectively, indicating a relatively small variation in the sample. The mean value of LR is 2.87, and the standard deviation is 12.01, which is nearly 6 times the minimum value, indicating that the LR of the selected part of the listed companies are quite varied.

4.2. Correlation Analysis

Table 3. Pearson Correlation Coefficient Matrix.

Variable	ROA	TOBIN’S Q	E	S	G	SIZE	LR	QR	NETBENE	ASSET_TURN
ROA	1
TOBIN’S Q	0.336 (***)	1
E	0.011	0.041 (**)	1
S	−0.098	−0.023 (**)	−0.011	1
G	−0.101 (***)	0.079 (**)	0.075	0.027	1
SIZE	−0.121 (***)	−0.296 (***)	−0.265 (***)	0.408 (***)	0.104 (***)	1
LR	0.413 (***)	0.610 (***)	0.076 (*)	-0.101	0.057	−0.599 (***)	1
QR	0.391 (***)	0.508 (***)	0.107 (**)	−0.096 (**)	0.010	−0.486 (***)	0.802 (***)	1
NETBENE	0.478 (***)	0.265 (***)	−0.024	−0.072	0.171 (**)	−0.062	0.066	0.093	1
ASSET_TURN	0.295 (***)	−0.033	−0.098 (***)	0.009	−0.234 (***)	0.078	−0.165 (***)	−0.021	0.044	1

Note: The values in brackets are the standard errors corresponding to the coefficients, *** represents p < 0.01, ** represents p < 0.05, and * represents p < 0.1.

shows the Pearson correlation coefficient matrix of the independent variable, dependent variable and control variable.

As can be seen from Table 3, E is negatively correlated with S and positively correlated with G, and S is positively correlated with G. This shows that the better the corporate environmental performance is, the worse the corporate social responsibility performance is, and the better corporate governance is. Corporate social performance is positively correlated with governance, which is consistent with many previous research conclusions (e.g., Garcia et al., 2017; Velte, 2017) [4,14]. E is positively correlated with dependent variables (ROA, TOBIN’S Q), S is negatively correlated with dependent variables (ROA, TOBIN’S Q), and G is negatively correlated with ROA and positively correlated with TOBIN’S Q. The control variable SIZE is negatively correlated with the dependent variable (ROA, Tobin’s Q). LR, QR and NETBENE are positively correlated with the dependent variable (ROA, Tobin’s Q). ASSET_TURN is positively correlated with ROA and negatively correlated with Tobin’s Q.

4.3. Regression Analysis of E, S and G

In

Table 4. Estimation of Panel Regression Coefficient.

	(1) Dependent Variable ROA	(2) Dependent Variable Tobin’s Q
Constant term	8.438 (4.136) **	9.744 (0.932) ***
E	−0.280 (0.140) **	0.018 (0.032)
S	−0.184 (0.134)	0.005 (0.030)
G	0.359 (0.166) **	0.044 (0.037)
SIZE	−0.691 (0.374) *	−0.869 (0.084) ***
LR	−1.171 (0.483) **	−0.236 (0.109) **
QR	1.928 (0.533) ***	0.342 (0.120) ***
NETBENE	0.005 (0.000) ***	0.0004 (0.0000) ***
ASSET_TURN	7.187 (0.736) ***	0.327 (0.166)**
Observations	1146	1146
Adjusted R2	68.54%	74.58%
F	9.78	9.14

Note: The values in brackets are the standard errors corresponding to the coefficients, *** represents p < 0.01, ** represents p < 0.05, and * represents p < 0.1.

are shown the estimated results of panel regression analysis.

For the panel regression model with ROA as the dependent variable, Hausman test statistic is 80.80, and the corresponding p-value is 0.000. Therefore, we adopt the fixed effects model. Table 4 shows that environmental performance (E) is significantly negatively correlated with ROA, and social performance (S) is negatively correlated with ROA but not significantly. Corporate governance performance (G) is positively correlated with ROA. The p-value of the corresponding coefficient significance was 0.030, which was significant at the significance level of 0.05. SIZE and LR are significantly negatively correlated with ROA. QR, NETBENE and ASSET_TURN are significantly positively correlated with ROA.

For the panel regression model with Tobin’s Q as dependent variable, Hausman test statistic was 56.92, and the corresponding p-value was 0.000, so the fixed effects model was used. Table 4 shows that E, S, G and Tobin’s Q are not significantly correlated. SIZE and LR are significantly negatively correlated with Tobin’s Q. QR, NETBENE and ASSET_TURN are significantly positively correlated with Tobin’s Q.

Table 4 shows that: H1a and H1b are rejected, no matter if ROA or Tobin’s Q is used as the dependent variable; H1c is not rejected when ROA is used as dependent variable, while H1c is rejected when Tobin’s Q is the dependent variable.

Generally speaking, the better the environmental performance of enterprises, the less conducive to the improvement of accounting-based financial performance is, which is consistent with the research results of Wang and Li [35]. This means that the increase of enterprise cost caused by environmental protection punishment is far lower than the income obtained by enterprises due to environmental pollution. That is to say, the external negative effect of enterprises’ environmental pollution is not internalized to the level of enterprise cost. It is the public that bears this negative effect. Companies with good environmental performance do not significantly improve market-based financial performance. This shows that although companies with poor environmental performance may have achieved good financial performance, environmental penalties have affected the company’s image in the capital market, and risk aversion investors have chosen to avoid it. Companies with good social responsibility performance have an inhibitory effect on financial performance expressed by ROA, but the effect is not significant. This may mean that within the dimensions of the social responsibility measurement in this research, the corporate social responsibility performance has not received positive feedback from stakeholders or the corporate social responsibility performance lacks sufficient visibility to the extent that stakeholders and capital market have overlooked the corporate social responsibility performance. Therefore, in this case, corporate social responsibility performance only increases the company’s operating costs. A well-governed company helps to improve the financial performance of the company. It shows that good corporate governance is conducive to the standardized operation of the company. Company size (SIZE) is negatively correlated with financial performance, which is interpreted as a positive correlation between company size and the number of environmental penalties that the company has received. This indicates that larger companies are more susceptible to environmental problems and less conducive to the improvement of corporate financial performance. This research conclusion is consistent with Wang and Sarkis [1]. Quick ratio (quick assets/current liabilities) is positively correlated with financial performance. The higher the quick ratio, the lower the company’s leverage ratio is, the lower the financial risk is, and the better the financial performance is. Unlike the quick ratio, the current ratio (current assets/current liabilities) is negatively correlated with financial performance. The higher the current ratio, the worse the corporate financial performance. The 3-year average net profit growth rate (NETBENE) of a company is positively correlated with financial performance, indicating that companies with a higher average net profit growth rate tend to have excellent financial performance.

4.4. Robustness Test

In order to test the reliability of the conclusion, we use tool variables to replace E, S and G for a robustness test. Here, we use E, S and G, which lag a period, as the tool variables of E, S and G to eliminate the interference of endogeneity to the results. After using the tool variables, the empirical research results of model (1) and (2) in Table 4 do not change the sign and significance of the parameter results. Thus confirm the robustness of our main results.

5. Empirical Analysis on ESG and Financial Performance

5.1. How Is ESG Performing?

From Table 2, we can know the descriptive statistics of ESG in environmental, social and corporate governance of the sample enterprises selected in this study. The means (median) of E, S and G are 12.54 (13.00), 14.62 (6.51) and 5.76 (4.78), respectively, and the mean value of E is the greatest, while the mean value of G is the lowest. Next, E, S and G indices are standardized firstly and then weighted according to the weight of 3-2-5, so as to obtain the ESG performance score of the enterprise.

The descriptive statistical results of the standardized E, S and G data and ESG performance data are shown in

Table 5. Descriptive Statistical Results of Standardized E, S, G and ESG Performance.

Variable	Mean	Median	Standard Deviation	Maximum	Minimum
Standardization E	−6.55 × 10−8	0.310	1.000	0.408	−20.009
Standardization S	1.588 × 10−7	−0.109	1.000	8.159	−0.298
Standardization G	5.963 × 10−5	−0.098	1.000	10.335	−2.001
ESG	0.011	0.108	0.524	4.103	−5.536

.

According to the standardized descriptive statistics results of E, S and G, the mean value of governance performance of enterprises is the highest. The main reason is that listed companies in the Yangtze River Delta have the most sufficient information disclosure in corporate governance in China. From the perspective of the variation range of E, S and G data after standardization, the variation range of S is the smallest, mainly because the company does not disclose much information on social responsibility. The ESG performance of all samples ranges from −5.536 to 4.103, with a mean of 0.011, a median of 0.108 and a standard deviation of 0.524, indicating the sample has a large variability of ESG performance.

We draw a bar graph to summarize the data of ESG and its components overtime in Figure 1. From Figure 1, the mean value of ESG performance rises slowly from 2015 to 2020. The mean value did not change much in five years and was relatively stable, which to some extent showed the stability of the evaluation system of ESG as a whole. The mean values of standardization of E, S and G show a general growth trend during 2015–2020, which indicates that the E, S and G performance of sample enterprises is getting better and better.

5.2. Regression Analysis of ESG

Table 6. Estimation of Panel Regression Coefficient.

	(3) Dependent Variable ROA	(4) Dependent Variable Tobin’s Q
Constant term	6.020 (4.040)	9.556 (0.907) ***
ESG	−0.707 (0.428) *	0.101 (0.096)
SIZE	−0.528 (0.366)	−0.843 (0.082) ***
LR	−1.095 (0.484) **	−0.232 (0.109) **
QR	1.854 (0.534) ***	0.338 (0.120) ***
NETBENE	0.005 (0.000) ***	0.0004 (0.0000) ***
ASSET_TURN	7.233 (0.737) ***	0.318 (0.166)*
Observations	1146	1146
Adjusted R2	68.34%	74.61%
F	9.82	9.41

Note: The values in brackets are the standard errors corresponding to the coefficients, *** represents p < 0.01, ** represents p < 0.05, and * represents p < 0.1.

shows the empirical results of ESG performance and corporate financial performance.

For the panel regression model with ROA as the dependent variable, Hausman test statistic was 51.21, and its corresponding p-value was 0.000, so the fixed effect model was used. Table 6 shows that ESG is positively correlated with ROA, and the corresponding coefficient significance p-value is 0.0988, which is close to the significance level of 0.1. SIZE is negatively correlated with ROA, but not significantly. LR is significantly negatively correlated with ROA. QR, NETBENE and ASSET_TURN are significantly positively correlated with ROA.

For the panel regression model with Tobin’s Q as the dependent variable, Hausman test statistic was 51.80, and its corresponding p-value was 0.000, so the fixed effect model was used again. Table 6 shows that ESG and Tobin’s Q have no significant positive correlation. SIZE and LR are significantly negatively correlated with Tobin’s Q. QR, NETBENE and ASSET_TURN are significantly positively correlated with Tobin’s Q.

Table 6 shows that ESG has less significant negative impact on the financial performance when ROA is used as dependent variable, while ESG has no significant impact on the financial performance when Tobin’s Q is used as dependent variable.

In addition, we also considered other weights, for example, the ESG weights of 2-1-7, 1-1-8 and 3-3-4. We found the impact of financial performance is not significant at 0.1 under these weights, no matter if financial performance is accounting-based or market-based.

We compare our conclusion with Velter’s findings [14], which indicate that ESG has positive impact on ROA but no impact on Tobin’s Q, and, after analyzing the three different components of ESG, governance performance has the strongest impact on financial performance. The difference lies in the impact of ESG on ROA, which is not significant in our setting. This difference may be related to the fact that China’s ESG is at the initial stage of knowledge and exploration, lacking the legal and regulatory system and relevant institutional constraints requiring enterprises to disclose information, which leads to a lack of motivation for ESG information disclosure in enterprises. In recent years, Chinese enterprises have gradually improved their awareness of ESG, and more enterprises have begun to disclose ESG information. Chen and Xie [22] mentioned that ESG disclosure attracts ESG investors, and ESG investors play a positive moderating role in the connection between ESG ratings and financial performance. In this way, ESG performance will have an obvious effect on promoting financial performance. This conclusion may be related to the characteristics of China’s ESG that are still in the initial stage of understanding and exploration. Enterprises generally have insufficient awareness of ESG. Enterprises’ participation in ESG practice involves symbolism and opportunism, and enterprises do not participate in ESG practice substantively. That is specifically manifested in several aspects, such as the low willingness of enterprises to disclose, the authenticity and credibility of disclosed data need to be improved, the lack of verification by third-party data institutions and the lack of a complete legal system and information disclosure system support.

ESG places companies in a broader external environment, reflecting the concerns of a wider range of stakeholders, especially natural and social aspects, which is conducive to the transformation of enterprises into a micro-basis for global sustainable development. At the same time, enterprises need to carry out more trade-offs and sustainability management of the diversified needs from all stakeholders. If the legal system is backward, supervision is absent, the cost of punishment for environmental damage and neglect of social responsibility is very low, or environmental friendliness and social responsibility behavior are not sufficiently encouraged, companies will not devote their limited resources to ESG.

The conclusion between ESG and financial performance, suggests that China still has a long way to go in terms of the positive connections among stakeholder rights, sustainability management and economic benefits.

5.3. Robustness Tests

We use the variable replacement method to test the robustness. The first is about the replacement of enterprise performance variables. We use return on equity (ROE) as a substitute variable for ROA to test the relationship between ESG performance and financial performance. The empirical research results of model (3) in Table 6 do not change the sign and significance of the parameter results. Thus confirm the robustness of our main results. The second is the replacement of ESG variables. We replace ESG variables with the ESG rating data of SynTao Green Finance (STGF). ESG variables are assigned to 1 to 10 according to the ESG rating of STGF from D to A+. ESG rating data of STGF is from Wind database. The empirical research results of model (4) in Table 6 do not change the sign and significance of the parameter results. Thus confirm the robustness of our main results.

6. Conclusions and Suggestions

This paper studies the impacts of E, S, G and ESG performance on financial performance of Listed Companies in the Yangtze River Delta. We use a sample consisting of 191 Listed Companies in the Yangtze River Delta from 2015 to 2020 with 1146 annual observations values.

From the impacts of E, S, G on accounting-based financial performance (ROA), this study shows that: E has a significant negative impact on accounting-based financial performance; G has a significant positive impact on accounting-based financial performance; S has no significant impact on financial performance; G and E have more impact on financial performance according to the regression coefficients; ESG has less significant impact on financial performance.

From the impacts of E, S and G on market-based financial performance (TOBIN’S Q), this study shows that all of these three has no significant impact on TOBIN’S Q; ESG has no significant impact on financial performance.

These results are related to the characteristics of China’s ESG at the initial stage of understanding and exploration, lacking legal and regulatory systems and related institutional constraints that require companies to disclose information. These results are also related to uncoordinated development of E, S, G in China. The supervision of E and G is more and earlier, and environmental and governance information disclosure is relatively comprehensive, but social responsibility has a lack of sufficient attention.

Based on our study, we provide some suggestions for the development of ESG in China. First, it is necessary to raise the awareness of the enterprises to participate in ESG practice activities and investors’ awareness of green investment. The government should encourage enterprises to substantively participate in ESG practice activities, so that enterprises regard ESG not as a “green washing” label but as an objective and reliable information tool to promote the long-term green development of Chinese enterprises. Encourage investors to choose more socially conscious investments and increase long-term investment value to promote the long-term green development of Chinese enterprises. Second, the full implementation of ESG disclosure rules in line with international advanced standards should be launched first. According to the most advanced ESG disclosure standards in the world, the standard ESG disclosure rules should be specified, and enterprises are encouraged to actively disclose ESG information and set up a demonstration for enterprises in the developed regions like Yangtze River Delta. Encourage and support third-party green assessment agencies to evaluate and rank corporate ESG as well as guide the whole society’s green investment and green consumption concepts. At the same time, encourage enterprises to make continuous progress in environmental, social responsibility and corporate governance. Third, the government can further improve the related incentive and restraint mechanisms to promote the ESG system. For example, certain incentives can be given to listed companies with high ESG scores in terms of bidding, procurement, tax reduction and exemption, as well as convenient conditions for IPO, refinancing, green credit, etc. Sanctions should be imposed on listed companies with low ESG scores and companies that do not disclose information as required or disclose false information.