Introduction

The energy industry provides an important material basis for economic and social development; however, with the rapid development of modern society and the economy, more attention has been paid to the problem of high pollution and high emissions caused by traditional energy utilization (Meng and Huang 2018; Strusnik et al. 2020; Huang et al. 2020b), especially in China (Huang et al. 2020a; Yu et al. 2023). The realization of China’s carbon peak and carbon neutralization goals requires the unremitting efforts of China’s energy industry (Jia and Lin 2021). Energy enterprises assume important social responsibility in ensuring national energy security, solving environmental problems, coping with climate change, and so on.

Global CO2 emissions have increased significantly in recent years. China has led the expansion of the global energy market, accounting for more than three-quarters of the net increase in global energy consumption (Du et al. 2021), making it the largest energy driver to dateFootnote 1. Although the global primary energy consumption decreased by 4.5% in 2020 due to the severe situation around the COVID-19 pandemic, China’s primary energy consumption still maintained the largest growth rate of 2.1%, and became the only country with an increase in oil consumption.Footnote 2 According to the Global Energy Review: CO2 emissions report of 2021 released by the International Energy Agency (IEA), per capita CO2 emission in China has reached 8.4 tonnes, higher than the average of 8.2 tonnes in developed countries. The development of China’s energy industry has attracted much attention from scholars. Previous discussions on the energy industry focus on many areas, such as overcapacity (Du et al. 2020), government subsidies (Luo et al. 2021), corporate governance (Shi 2019), resource allocation (Guo et al. 2021), corporate risk (Wei et al. 2019), and so on. The general notion holds that the production and operation activities of energy enterprises have negative externalities, especially greenhouse gas and other harmful substances produced by fossil energy consumption (Ahmed et al. 2023), significantly affecting the sustainable development of the natural environment and human society. Consequently, energy companies face greater pressure and social responsibility than other industries (Latapí Agudelo et al. 2020).

The idea of corporate social responsibility (CSR) originated from the Industrial Revolution in the 18th century, emphasizing the contribution to the environment (Chen et al. 2019), consumers, and society in the production process. CSR is defined as the company’s interests and social interests beyond the legal requirements (McWilliams and Siegel 2001), which refers to the actions taken by enterprises to promote social welfare, including many employee-friendly, environment-friendly, and investor-friendly behaviors (Becchetti et al. 2018).

Currently, the international reliable CSR database is MSCI ESG (Environmental, Social, and Governance) Ratings, which encompasses the enterprise data of each country, covering environmental, social, stakeholder, and corporate governance aspects. In China, the social responsibility information disclosed by listed companies is gradually being improved. The authoritative evaluation system mainly includes Rankins CSR Ratings (RKS) and the Hexun CSR database. RKS mainly assesses CSR from the Macrocosm (M), content (C), technique (T), and industry (I) aspects and establishes the RKS ESG rating system concerning international standards. Similarly, Hexun’s social responsibility measurement system evaluates listed companies from five indicators: shareholder responsibility; employee responsibility; supplier, customer, and consumer rights and interests’ responsibility; environmental responsibility; and social responsibility. According to the Research Report on Corporate Social Responsibility of China (Li et al. 2019), the social responsibility of Chinese state-owned enterprises is ahead of that of private enterprises and foreign-funded enterprises, and the social responsibility score of the power industry is much higher than that of other industries.

The existing literature on CSR mainly focuses on two aspects: influencing factors and effects. On the one hand, those factors influencing CSR include multiple dimensions. Externally, the political, economic, and cultural background of the region where the enterprise is located are all factors that affect CSR activities (Tilt 2016). For example, the social anti-corruption campaign can significantly improve the performance of CSR (Kong et al. 2021). Internally, companies with more effective risk management are more willing to participate in CSR behavior (Kuo et al. 2021); ethical leadership has a positive effect on CSR (Nguyen et al. 2021); strategic planning and corporate culture are also important factors affecting CSR (Kalyar et al. 2013); green financial help may have gradual negative impacts on environmental and social responsibility (Sinha et al. 2021). Moreover, CSR activities also exert a variety of influences on the development of enterprises. Research shows that CSR significantly improves corporate performance (Javeed and Lefen 2019), reduces the cost of debt (Yeh et al. 2020), and promotes corporate innovation (Javeed et al. 2021). Participation in CSR activities can also help to a good corporate reputation (Han and Lee 2021) and enhance customer trust (Islam et al. 2021), thereby gaining a competitive advantage (Turyakira et al. 2014), which is conducive to the long-term development of enterprises. In addition, social responsibility activities is found to affect stock prices. Huang and Liu (2021) found that companies engaged in more CSR activities faced less risk of stock price crashes under the effect of COVID-19.

Given the important position of the energy industry in the economic system, improving the social responsibility of energy enterprises and maintaining a relative balance between providing energy for economic development and environmental protection is vital. In existing studies, managerial characteristics, as a key factor in enterprise development, has been proven to affect CSR significantly. Studies found that the marital status of the CEO (Cronqvist and Yu 2017; Hegde and Mishra 2019), excessive self-confidence (Zribi and Boufateh 2020), internal debt (Kim et al. 2020), board structure (Bolourian et al. 2021), and gender ratio (Wang et al. 2021) all influence CSR activities. Besides these managerial characteristics, the relationship between executive compensation and CSR/CER (Corporate Environmental Responsibility) has also received extensive study.

Zou et al. (2015) found that top executives’ cash payment positively affects corporate environmental performance, whereas equity ownership has a negative one. Similarly, the study of Shaer et al. (2023) also revealed that CEOs (Chief Executive Officers) are motivated to improve CER when they receive compensation for engagement in environmental activities. Jiang et al. (2021) examined the executives’ compensation restriction imposed by the government on CSR, and the results indicate that CSR performance decreases with increases in the pay restriction. The relationship between executive compensation and CSR may also be an inverted U-shaped relationship, and the threshold level is 18.7 percent in the level of executive compensation (Pareek and Sahu 2024). Contrary to these studies, another line of study aims to explore CSR’s impact on executive compensation. CSR has a positive effect on the financial performance (Jahmane and Gaies 2020). Good CSR performance increases management compensation by improving corporate profitability (Ho et al. 2022), the same effect was also found for environmental performance (Berrone and Gomez-Mejia 2009). However, some studies provide the opposite results. The study of Jian and Lee (2015) pointed out that CEO total compensation is negatively associated with CSR investment. In the same way, the study of Cai et al. (2011) showed that CSR negatively affects both CEOs’ total compensation and cash compensation. According to the literature above, two-way causality between CSR and management compensation may exist. This is partly confirmed by the study of Kao et al. (2018), which shows the presence of a bidirectional relationship between CSR and firm performance.

For China, designing an optimal compensation structure for management in the energy industry to act morally and engage more in social responsibility is of great importance for environmental protection and sustainable development. Therefore, this paper falls in the first category given above, exploring management compensation’s impact on CSR. Considering the potential two-way causality, the main aim of this study is to identify the causal effects of management compensation on CSR using robust methodologies.

This study mainly answers the following questions. 1) What are the causal effects of management compensation incentives on corporate social responsibility and environmental responsibility? 2) What is the difference between salary incentives and equity incentives on corporate social responsibility and environmental responsibility? 3) What is the heterogeneity of different types of sub-industries, different types of enterprise ownership, and different regions? 4) What are the implications of this study for energy enterprises to undertake CSR & CER?

The contributions of this paper are as follows: 1) Unlike most previous studies, which mainly investigate the relationship between CSR and executive pay, this article tries to identify the causal effect of management compensation incentives on CSR using the instrumental variable estimation method on enterprise level. Specifically, we constructed corresponding instrumental variables for salary and equity incentives. This study’s method can effectively solve the endogeneity bias due to omitted variables and reverse causality. 2) Considering that there may be greater externalities in the production process of energy enterprises, the issue of CSR of energy enterprises is particularly important. Therefore, the main object of this study is the CSR of the energy enterprises in China, which is the largest greenhouse gas emitter worldwide.

The rest of the article is structured as follows: the second part presents the theoretical basis and hypothesis underpinning the research, the third part contains an introduction to the data and models, the fourth part demonstrates a report and analysis of empirical regression results, and the fifth part puts the conclusion.

Hypothesis and theoretical basis

Management compensation is one of the most controversial features of corporate governance (Correa and Lel 2016). In China, the management compensation of listed companies is often a combination of monetary salary and shareholding ratio (Liu et al. 2014), and its incentive effect is affected by the pay gap (Xu et al. 2016) and future compensation structure (Tang 2016). In 2016, the implementation of the “Equity Incentive Management Measures for Listed Companies” standardized the equity incentive behavior of listed companies and endowed listed companies with a certain autonomy. Other research found significant differences in the effects of salary and equity compensation on corporate strategy (Chan and Ma 2017). Salary compensation is a fixed material reward for managers to ensure that their day-to-day needs are met (Zhou et al. 2021), while equity incentive increases the variation range of the present value of compensation and links the management interests with the shareholder interests and corporate performance (Brüggen and Zehnder 2014). Therefore, in the present analysis, we divide the management compensation incentives into two parts: salary and equity incentives. And we evaluates their different effects on CSR performance and proposes two hypotheses.

Previous research shows that the compensation of enterprise managers can promote effective internal control of enterprises (Henry et al. 2011) and affect enterprise performance (Conyon and He 2011; O’Connor and Rafferty 2010), debt cost (Kabir et al. 2013), asset valuation (O’Connor and Rafferty 2010), external risk (Campbell et al. 2007), and socially responsible investment (Jian and Lee 2015).

Because of the motivational effect, executives will work harder with higher pay, demonstrating their effort is worth the pay (Malul et al. 2021; Buck et al. 2008). In the daily operation of the company, executives know more than the shareholders due to asymmetric information. Executives are more likely to over-invest in CSR activities in line with agency theory. There are several reasons for this. First, engaging in more CSR activities can improve executives’ reputations (Hassen and Ghardadou 2020). Second, the impact of environmental risk and negative social events easily signifies the management’s failure (Zou et al. 2015). Third, the determinants of economic performance are multiple, the shareholders cannot place all the blame on managers. Therefore, CSR is a source of agency problems because management may use firm resources to engage in conspicuous CSR investments (Kao et al. 2018). Based on principal-agent theory, we proposed the following overinvestment hypothesis:

Hypothesis 1: Management salary incentives has a positive effect on CSR & CER.

With the continuous improvement and development of China’s securities market, equity incentives have also become an important part of management compensation in China (Liu et al. 2014). The experience of principal-agent theory shows that due to the inconsistency between corporate social responsibility and shareholders’ goal of maximizing stock value (Firth et al. 2006), granting managers appropriate equity to encourage them to pay more attention to corporate finance cannot only solve the principal-agent problem (He 2008) but also help to maintain the interests of shareholders and reach the best decision (Shu and Thomas 2017; Ullah et al. 2021). Specifically, executives’ stock ownership increases the range and likelihood of compensation. For managers, the higher the corporate performance, the higher the stock market value, and the higher their actual income.

The trade-off hypothesis indicates that the costs of CSR reduce profits (Makni et al. 2009), and participating in CSR activities cannot guarantee higher corporate performance (Marsat and Williams 2013; Shahbaz et al. 2020). CSR has few short-term measurable economic benefits (Kao et al. 2018). Holding the company’s shares is equivalent to facing relatively high risks and returns (Zhou et al. 2021), and fulfilling social responsibility is an activity with a large investment and slow return.

However, China’s stock market has a short history, higher volatility, and a “policy market” (Zhang et al. 2023). The stocks on the Chinese stock market are not worth long-term holding. As a result, rational managers invest less in social responsibility. Hence, we put forth the second hypothesis:

Hypothesis 2: Management equity incentives have a negative effect on CSR & CER.

Methodology and data

Model

This study aims to assess the influence of executive compensation incentives on CSR & CER in Chinese energy enterprises. The panel data fixed effects model is employed, considering the data used in the analysis is panel data. The specific setting of the baseline model is shown in Eq. (1).

$$CSR\left( {CER} \right)_{it} = \beta _0 + \beta _1Incentive_{it} + X_{it}^\prime \gamma + \theta _i + \mu _t + \varepsilon _{it}$$
(1)

Where the explained variable CSR(CER)it represents the CSR (or CER) score of the enterprise i in the period t. The core explanatory variable Incentiveit denotes the management salary incentive or equity incentive of enterprise i in period t. Control variable Xit includes the basic characteristics and operating conditions of enterprises. θi and μt represent the individual and time-fixed effects respectively, and εit is an error term.

Although the fixed-effect model can control for the effects of omitted variables that do not change over time, endogeneity problems still exist due to time-variant omitted variables and reverse causality. To deal with such issues, we also conduct the analysis using the instrumental-variable regression method. As the explanatory variable in this study is the executive salary incentive and equity incentive, the instrumental-variable (IV) regression model for these two variables will be discussed separately.

First, for executive salary incentive, we construct the instrumental variable by interacting the total salary payment of all enterprises in the industry and the rank of the financial performance in the first five years for each enterprise. Return on Assets (ROA) was used to measure the financial performance. The construction of this instrumental variable borrowed ideas from a similar instrumental variable developed in (Nakamura and Steinsson 2014; Ma and Meng 2022), which is a kind of “Bartik” approach transformation. The reason for this instrumental variable design is that the total salary payment of an industry is relatively exogenous for an individual enterprise. When the total executive payment of an industry increases, the salary payment of an enterprise with higher initial financial performance will increase more. This is a common assumption when using “Bartik” instrumental variable strategy. The discussion of the exogeneity of the instrumental variables is further discussed in the robustness test.

Although the initial financial performance of an enterprise may affect the current CSR, this has been fully absorbed by enterprise fixed effects as enterprise-level time-invariant factors. Meanwhile, even though the initial financial performance has a time-variant influence on CSR, the interaction term between initial firm financial performance and time-fixed effect can control this confounding. The estimation equation of first-stage regression is as follows:

$$Salary_{it} = \beta _0 + \beta _1IV\_S_{it} + X_{it}^\prime \gamma + \theta _i + \mu _t + \rho \left( {Z_i \,*\, \mu _t} \right) + \varepsilon _{it}$$
(2)

Where Salaryit is the endogenous variable, executive salary incentive, IV_Sit is the instrumental variable, the interaction term between the total salary payment of the industry and the rank of the financial performance in the first five years for each enterprise, Zi * μt is the interaction term between initial firm financial performance and time-fixed effect, other variables are similar as the baseline model.

The estimation equation of second-stage regression is as follows:

$$CSR\left( {CER} \right)_{it} = \beta _0 + \beta _1\widehat{{Salary}}_{it} + X_{it}^\prime \gamma + \theta _i + \mu _t + \rho \left( {Z_i \,*\, \mu _t} \right) + \varepsilon _{it}$$
(3)

\(\widehat{{Salary}}_{it}\) is the fitted value of the endogenous variable, executive salary incentive, other variables are similar as before.

Second, we used a similar method for executive equity incentive to construct the instrumental variable as a salary incentive. The instrumental variable in this case is the interaction term between the shareholding ratio of management in the industry and the rank of the ratio of controlling shareholders’ share ownership in the first five years of each enterprise. The shareholding ratio of management in an industry generally affects the equity incentive of an enterprise. Similarly, the confounding influences of initial share ownership of an enterprise on CSR have been absorbed by fixed effects and interactive fixed effects. The first stage regression model for executive equity incentive is as follows:

$$Equity_{it} = \beta _0 + \beta _1IV\_E_{it} + X_{it}^\prime \gamma + \theta _i + \mu _t + \rho \left( {Z_i \,*\, \mu _t} \right) + \varepsilon _{it}$$
(4)

Where Equityit is the endogenous variable, executive equity incentive, IV_Eit is the instrumental variable of executive equity incentive, Zi * μt is the interaction term between initial share ownership of an enterprise and time fixed effect, other variables are similar as the baseline model.

The estimation equation of second-stage regression is as follows:

$$CSR\left( {CER} \right)_{it} = \beta _0 + \beta _1\widehat{{Equity}}_{it} + X_{it}^\prime \gamma + \theta _i + \mu _t + \rho \left( {Z_i \,*\, \mu _t} \right) + \varepsilon _{it}$$
(5)

Where \(\widehat{{Equity}}_{it}\) is the fitted value of the endogenous variable, executive equity incentive, other variables are similar as Eq. (4).

Data sources

The research focuses on China’s energy enterprises: enterprise-level data pertaining to eight sub-sectors, including electric power, power generation equipment, electrical grid, gas, oil and natural gas, coal, energy equipment, and environmental protection, are obtained in the Wind databaseFootnote 3 and used as research samples. The CSR & CER score comes from the social responsibility reports of listed companies released by Hexun (http://www.hexun.com). The evaluation system investigates five aspects and sets 13 secondary indicators and 37 tertiary indicators to evaluate CSR & CER. All other variables are derived from the China Stock Market Accounting Research Database (CSMAR, https://www.gtarsc.com). The CSMAR database covers the main fields of China’s economy and finance, can provide relatively comprehensive and accurate information about listed companies, and has been widely recognized in academic research.

Considering the data availability and release time, we limit this study’s time range to 2010–2021. In the process of data-matching and processing, we obtained 3272 sample data after excluding ST, * ST, and PT companies, as well as excluding the pre-listing data.

Variable description

Based on the needs of the research, the variables are set as shown in Fig. 1; specific explanatory and descriptive statistical analyses of all variables are displayed in Table 1.

Fig. 1
figure 1

Variable setting diagram.

Full size image
Table 1 Descriptive statistics of variables.
Full size table

The explained variables are CSR & CER, and their highest scores are 100. The data come from Hexun.com, including five sub-indicators: shareholder responsibility, employee responsibility, supplier, customer and consumer rights and responsibility, environmental responsibility, and social responsibility. The highest score of sub-indicators is also 100, accounting for 30%, 15%, 15%, 20%, and 20% of the total weighted score, respectively. The core explanatory variable is management incentives, including salary and equity incentives. Managers include directors, supervisors, and senior managers. Salary incentive is estimated by the logarithm of the annual compensation of management personnel. Equity incentive is measured by the shareholding ratio of management personnel.

To control the influence of omitted variables, we include a series of enterprise-level control variables with reference to previous literature, which are total liability ratio (Espahbodi et al. 2016), rate of return on equity (Ratti et al. 2023), the year after the company goes public (Espahbodi et al. 2016), number of employees (Wang et al. 2021), CEO duality and shares held by the top ten shareholders of company (Tsang et al. 2021), value of TobinQ, proportion of the largest shareholder and the number of senior managers (Zhou et al. 2021).

In order to eliminate doubts about multicollinearity, the study conducted a correlation test for variables, and Table 2 shows the correlation coefficient matrix between variables. Most of the coefficients shown in Table 2 are less than 0.4, which indicates that there is no obvious multicollinearity problem.

Table 2 Correlation coefficient matrix.
Full size table

Regression result and analysis

Baseline regression

Table 3 lists the regression results of management compensation incentives (core explanatory variable) and CSR score (explained variable). Models (1) and (2) are the baseline model, which is the common panel data fixed effects model. Models (3) and (4) are the IV estimation method

Table 3 Regression results of management compensation incentives to CSR.
Full size table

The results in model (1) show that the effects of management salary incentives on CSR score is positive and significant, while the effects of management equity incentives on CSR score is negative and significant in model (2).

Models (3) and (4) provide the estimation results of the instrumental variable method of the impact of salary incentive and equity incentive on CSR, respectively. From the results of the first stage estimation, the estimation coefficients of the instrumental variables of salary and equity in the two models are both very significant. At the same time, the test results of weak instrumental variables show that the instrumental variables in the two models are not weak, indicating that they are suitable instrumental variables. From the estimation results of the second stage, the salary incentive is significantly positive and the equity incentive is significantly negative, which is consistent with the results of the benchmark model. However, the difference is that the absolute value of the estimated coefficient of management incentive in the two instrumental variable models is higher than that in the benchmark model. This also suggests that the role of management incentives may be underestimated before the endogenous problem is controlled.

Model (3) reports that the regression coefficient of salary incentive is 15.213, suggesting that every 1% increase in total management salary can increase the CSR score by about 15. Based on the model (4), the regression coefficient of equity incentives is −0.774, revealing that each 1% increase in management shareholding will reduce the CSR score by 0.774. The negative effect of the change in equity incentives is less than the promotion effect of salary incentives. As for the result of control variables, the dual role of CEO and chairman has a stable negative effect on the CSR score. In other words, the concurrent employment of these two positions will reduce the performance of CSR. Furthermore, the regression coefficient of the total liability ratio is significant at the 1% level, indicating that the more debt a company has, the less it invests in CSR investment.

The economic activities of energy enterprises are more likely to harm the environment. Therefore, environmental responsibility is one of the most important aspects of energy enterprises’ social responsibility. Therefore, after verifying the effects of management compensation incentives on total social responsibility, the score “environmental responsibility” sub-index score (CER) is taken as the explained variable, and regression analysis is performed again. The results are summarised in Table 4.

Table 4 Regression results of management compensation incentives to CER.
Full size table

Similar to the analysis of CSR, in Table 4, models (1) and (2) are common fixed-effect model estimates, and (3) and (4) are instrumental variable method estimates. As can be seen from the results in Table 4, salary incentive is significantly positive, while equity incentive is significantly negative. The estimation results of the instrumental variable method model are consistent with the common fixed effect. The results of the first-stage instrumental variable estimation and weak instrumental variable test also show that the instrumental variable is suitable in this case. However, from the absolute value of the estimation coefficient, the estimation result of salary incentive and equity incentive in the instrumental variable method model are both greater than that of the common fixed effect model. These results are very similar to the results of CSR in Table 3.

Based on the regression results in Table 4, management compensation incentives exert a significant effect on CER, but the absolute values of regression coefficients of core explanatory variables are smaller than those in Table 3. The results of model (3) imply that the regression coefficient of salary incentive is 5.645, and the regression coefficient of management shareholding is −0.245 in the model (4). The implication is that each 1% increase in the total management salary can increase the CER score by 5.645, and for each 1% increase in management shareholding, the CER score will be decreased by 0.245.

Heterogeneity

The industry heterogeneity analysis is undertaken. We divided all industries into two types. One is the electricity and environment industry, including electric power, electrical grid, and environmental protection industry. The other is the conventional energy and equipment manufacturing industry, including gas, oil and natural gas, coal, power generation equipment, energy equipment industry.

Table 5 gives estimates of industry heterogeneity. According to the estimated results of models (1) and (2), the interaction terms of salary incentive and whether it is the electricity and environment industry (EL_EN) are not significant in the two models, which indicates that the impact of salary incentive on CSR and CER is the same for the two types of industries. However, in models (3) and (4), the interaction terms of equity incentive and electricity and environment industry dummy variables (EL_EN) are both significantly negative. This means that the impact of equity incentives on CSR and CER differs between the electricity and environmental industry and the traditional energy and equipment manufacturing industry. The negative impact is even more pronounced in the electricity and environmental industry.

Table 5 Heterogeneity analysis based on industry.
Full size table

China’s state-owned enterprises (SOE) are socially oriented and bear more social responsibilities (Monkkonen et al. 2019) and employment burdens in the process of economic and social development (Li 2008). To determine the heterogeneous effect of management incentives on CSR in enterprises with different ownership types, the sample enterprises are divided into two parts according to the state holding status in state-owned enterprises and non-state-owned enterprises. The regression results are listed in Table 6.

Table 6 Heterogeneity analysis based on ownership.
Full size table

It can be seen from the results in Table 1 that the interaction terms between state-owned enterprises and salary incentives are significantly positive. This shows that compensation incentives have a greater positive effect on CSR for state-owned enterprises. The main reason may be that the government is unable to distinguish the losses of state-owned enterprises from the burden of social responsibility or managers’ mismanagement due to information asymmetry. In addition, as shareholders of state-owned enterprises, the government and the state do not seek to maximize corporate performance (Chang and Wong 2009). Therefore, salary incentive does not cause the management to pay attention to the operating profits of the enterprise, managers of state-owned enterprises are more willing to increase their CSR investments to enhance their reputation.

The coefficients of interaction terms in models (3) and (4) are not significant, which shows that the equity incentives for the two types of enterprises are similar.

China, as the largest developing country, has a vast territory, complex terrain, uneven population distribution, and significant regional differences in economic level. Taking all these factors into account, the sample enterprises are also divided into two parts according to the region where the enterprises are located. As China’s eastern coastal areas are relatively developed, the level of economic and social development is relatively high. Generally speaking, China can be divided into the eastern coastal developed areas and inland areas. The results of regional heterogeneity analysis are shown in Table 7. We find that the interaction terms of management incentives and regional dummy variables are not significant in all models. This shows that there is no regional difference in the impact of salary incentive and equity incentive on CSR or CER in China. Regardless of the relatively developed coastal areas or other areas, manager incentives significantly impact CSR or CER.

Table 7 Heterogeneity analysis based on region.
Full size table

Robustness test

We conducted a series of tests to further investigate the robustness of the results. The details are as follows:

In addition to the control variables in the baseline regression model, more control variables are added to the model in order to control more characteristics. These variables include the proportion of independent directors (PID), the cash ratio (CR), and the growth rate of owners’ equity (GRO). According to the regression results in Table 8, after adding control variables, the regression coefficient and significance of management compensation incentives are consistent with the previous regression results, suggesting that our regression results are robust.

Table 8 Robustness test when adding control variables.
Full size table

Considering the differences in the size of enterprise management personnel, the core explanatory variables are changed into per unit salary incentive and per unit equity incentive, and the regression is repeated. The results are shown in Table 9. The regression coefficients of the core explanatory variables are significant and match the baseline regression.

Table 9 Robustness test when replacing the core explanatory variable.
Full size table

The rating of CER in Hexun changed in 2018. Therefore, we removed the post-2018 sample and reestimated the model. The results are shown in Table 10. It can be seen that the estimated results of all models are similar to the baseline model, which indicates that changes in CER scores are unlikely to affect the estimated results.

Table 10 Robustness test of subsample in the period 2010–2018.
Full size table

Some companies have missing values in some years. The sample in the analysis is an unbalanced sample. If the sample loss is not random, there may be estimation bias caused by sample self-selection. Therefore, we removed some samples containing missing values and used the balanced panel data for estimation. The results are shown in Table 11, which clearly shows that the estimates agree with the baseline model. This also indicates that sample self-selection may have little effect on the estimated results.

Table 11 Robustness test of regression using balanced sample.
Full size table

In order to further verify the exogeneity of the instrumental variables, we simultaneously add both endogenous and instrumental variables to the reduced form model. The results are provided Table 12. It can be seen that in all models, the estimated coefficients of endogenous variables, management salary incentive and equity incentive, are significant, while the instrumental variables are not. To some extent, this shows that the instrumental variables can only affect the explained variables through endogenous variables. In other words, the instrumental variables are relatively exogenous in this model.

Table 12 Exogenous test of instrumental variables.
Full size table

Conclusions

The study of energy CSR & CER under the background of global carbon neutrality is of great significance to economic and social development. Based on the data of China’s listed energy companies from 2010 to 2021, the effects of management compensation incentives, including salary incentives and equity incentives, on CSR & CER are explored. To solve the endogeneity problem, the paper employed the identification strategy of the instrumental variable estimation method.

The econometric results show that management compensation incentives can significantly affect CSR and CER. The degree of influence and the significance are heterogeneous among enterprises in different sub-industries, different ownership properties, and different regions. Firstly, the effects of management salary incentives and equity incentives on CSR and CER are not the same. Salary incentives can significantly promote the implementation of CSR and CER: however, the higher the equity incentive, the worse the performance of CSR and CER. Secondly, there is significant heterogeneity in the effects of management compensation incentives on CSR& CER. Based on the heterogeneity analysis of enterprise ownership, it is found that the salary incentives have a greater impact on CSR&CER of state-owned enterprises, although the impact on CER is not significant. There is no difference in the impact of equity incentives on state-owned and non-state-owned enterprises. Based on the heterogeneity analysis of sub-industries, it is found that there is no difference in the impact of salary incentives on CSR & CER of sub-industries. However, we found that the impact of equity incentives on CSR and CER differs between the electricity and environmental industry and the traditional energy and equipment manufacturing industry. And the negative impact is even more pronounced in the electricity and environmental industry. The analysis based on regional heterogeneity did not find regional differences in the impact of compensation incentives on CSR & CER. Finally, the robustness of the regression results is verified by a series of robustness tests.

The CSR & CER of energy enterprises is of great significance to social development. Enterprises, especially new energy enterprises, should focus on the salary incentive of management personnel and appropriately reduce their equity holdings to improve energy enterprises’ social & environmental responsibility and increase their positive externalities. Besides, weakening equity incentives can help enterprises in the electricity and environmental industries to undertake more corporate social & environmental responsibility. Moreover, the negative effect of management shareholding in state-owned enterprises cannot be ignored. The government should pay attention to the development of state-owned energy enterprises, limit the incentive system such as management shareholding ratio through policies, and replace it with mainly salary incentives.

Considering future research, this study’s research paradigm and experience can be applied to the discussion of CSR & CSR in other emerging manufacturing industries, such as biopharmaceutical, information technology, and other industries. Besides, from practical experience, the relationship between government and business has a certain impact on energy enterprises’ social and environmental responsibilities. Therefore, this perspective can also be considered in future corporate social responsibility analysis to obtain fresh empirical evidence, especially in discussions in the energy industry.