Do Environmental Administrative Penalties Affect Audit Fees? Results from Multiple Econometric Models

Abstract

The construction of ecological civilization is an important requirement to realize high-quality development of the economy. Over recent decades, natural disasters, global warming, and other ecological problems have occurred frequently, and countries around the world are facing severe environmental challenges. Sustainable development is an effective way to solve these environmental threats, and environmental administrative penalties play a fundamental role in sustainable growth. So, we explore the response of external stakeholders to environmental administrative penalties from the perspective of auditors. The study examines whether environmental administrative penalties affect audit fees, the heterogeneity effects and the mechanisms. The research uses data of heavily polluting listed firms in China that have been revealed by the Institute of Public and Environmental Affairs (IPE). The findings show the following: (1) environmental administrative penalties can increase audit fees; (2) the effective internal control environment can weaken the positive relationship between firms’ environmental administrative penalties and audit fees, and the stronger regional environmental regulation can enhance the positive impact of firms’ environmental administrative penalties on audit fees; (3) further analysis shows that the impact of environmental administrative penalties on audit fees mainly comes from the “risk premium mechanism” rather than the “cost compensation mechanism”, and the response from audit fees can encourage firms to engage in green innovation activities. The above research conclusions provide a certain reference for auditors’ pricing decisions.

1. Introduction

Over recent decades, natural disasters, global warming, and other ecological problems have occurred frequently. With the development of global industrialization, countries around the world are facing severe environmental challenges, which also pose some threats to global economic development, human health, and social stability. How to realize the coordinated development between the economy and the environment has never been more urgent [1,2]. There is no doubt that achieving the aim of sustainable development by adjusting the process of resource allocation and economic construction is important. Sustainable development can meet the need of future generations while achieving contemporary development goals, which also emphasizes economic coordination, energy conservation, and environmental governance. Sustainable growth not only mobilizes the participation of the public, the government, and enterprises in environmental governance, but also realizes the harmonious coexistence of the economy, nature, and society [3]. When it comes to how to realize the construction of sustainable development, although the prior environmental protection, such as green innovation, is significant [4,5], the environmental regulation after the event, such as environmental punishment, is also necessary [6].

According to the latest Environmental Performance Index Report (2020), released by Yale University and Columbia University, China’s environmental performance ranks 120th among 180 countries. Economic growth has resulted in severe environmental degradation since the opening-up policy in 1978, and thus sustainable development is becoming important. On 15 March 2021, at the ninth meeting of the Central Committee on finance and economics, President Xi Jinping chose carbon neutrality and carbon peak as the topic and formulated the goal of achieving carbon peak by 2030 and carbon neutral by 2060. As early as 1999, China’s Environmental Protection Administration set up administrative penalties for environmental protection to attenuate firms’ pollution behaviors. Since then, local environmental protection departments have punished the environmental pollution behaviors of firms according to the law. Furthermore, the report of the 19th CPC National Congress pointed out that establishing an effective environmental penalty system is important. During the 13th Five-Year Plan period, there were 833,000 environmental administrative penalties issued, and the fines for environmental administrative penalties totaled RMB 53.61 billion. Environmental administrative penalties play a fundamental role in sustainable growth.

The Auditing Standards for the Chinese Certified Public Accountants No. 1631-Consideration of Environmental Information in the Audit of Financial Statements, issued by the Chinese Institute of Certified Public Accountants in 2006, requires auditors to pay attention to environmental matters. Presently, only a small amount of literature has researched the reaction of auditors to negative environmental events. These studies mainly focus on environmental information disclosure [7], greenwashing [8], government environmental regulation [9], and environmental liability insurance [10]. The disclosure of environmental administrative penalties, which reveal environmental misconducts of firms, provides an opportunity to measure firm environmental risks. It is important to study how environmental administrative penalties affect auditors’ decisions, especially for audit fees, which is the external embodiment of audit risk and audit cost. However, to the best of our knowledge, there is a paucity of direct empirical evidence on whether audit fees are affected by environmental administrative penalties. Therefore, the study investigates the interface between environmental risks and audit fees, and it examines how environmental administrative penalties shape auditors’ behaviors.

China, the largest developing country and the largest emitter of greenhouse gases in the world, is facing a conflict between economic growth and environmental protection [11]. Firms in China have sufficient incentives to engage in pollution control efforts to improve environmental conditions, and China is the largest emitter of greenhouse gases [12] and the largest investor in energy utility [13]. Besides that, the data of environmental penalties comes from the website of the Institute of Public and Environmental Affairs (IPE). The database regularly collects information on environmental administrative penalties, including fines, reasons, and so on, which lays a foundation for our research in the context of China.

The aim of this study is to examine whether environmental administrative penalties affect audit fees, using data of heavily polluting listed firms in China that have been revealed by the Institute of Public and Environmental Affairs (IPE). In brief, our findings include the following aspects: Firstly, the study finds that environmental administrative penalties are significantly positively associated with audit fees. Secondly, the findings reveal that the effective internal control attenuates the positive association between environmental administrative penalties and audit fees, but the regional environmental regulation enhances the positive impact of firms’ environmental administrative penalties on audit fees. Thirdly, the impact of environmental administrative penalties on audit fees mainly comes from the “risk premium mechanism” rather than the “cost compensation mechanism”, and the response from audit fees can encourage firms to engage in green innovation activities.

One of the main contributions of our research is that we measure environmental administrative penalties from four dimensions including the number of penalties, the fine of penalties, the severity of penalties, and the department of penalties, which is more comprehensive compared with previous studies. The historical studies gauged environmental penalties using dummy variable [14] or the kinds of violation [15]. Secondly, additional evidence that environmental administrative penalties have impacts on firms is found. Although previous studies found that environmental administrative penalties can stimulate the environmental enthusiasm of peer enterprises [16] and increase financing constraints [15], how environmental administrative penalties affect audit fees remains unknown. The study provides empirical evidence on the relationship between environmental administrative penalties and audit behavior. Thirdly, the study sheds new light on what factors influence audit fees. While prior work concerning audit fees has mainly focused on client characteristics [17], auditor characteristics [18,19], and external environmental uncertainty [20,21], research on environmental risks is scarce. So, how environmental risks influence audit fees from the perspective of environmental administrative penalties is depicted.

The content of the paper is as follows: In Section 2, the literature related to the study is presented. Section 3 introduces the theory and research hypotheses. Section 4 introduces the data, variables, and empirical model. Section 5 presents the results. Section 6 examines the mechanism and consequence. Section 7 presents four robustness tests. Finally, Section 8 provides conclusions, recommendations and limitations.

2. Literature Review

2.1. The Consequences of Environmental Administrative Penalties

Consistent with previous studies [22,23], we give the definition of environmental administrative penalties, which mean that when firms are against environmental laws, rules or regulations, governments will punish the firms’ environmental misconducts through fines, shutdown and so on. Not only do the environmental administrative penalties force firms to engage in pollution control efforts, but they also affect firms’ economic value. Additionally, previous literature on the environmental administrative penalties also mainly focuses on these two aspects.

In terms of environment protection, according to environmental economics, there is a presupposition for punished firms to play a role in pollution control, which is that penalties must have substantial impacts on the profit of firms. In other words, if the cost of environmental administrative penalties outweighs the benefit of environmental pollution, firms will intend to comply with environmental protection standards. In China, relative to property penalties, reputation penalties and behavior penalties have more dominant impacts on firms’ profit [24], which stimulates the environmental governance enthusiasm of polluting firms more directly. Chang et al. [25] also found that the Chinese environmental penalty system should be based on profits of firms’ pollution activities rather than the negative impacts on the environment. In addition, generally, studies showed environmental penalties encourage firms to engage in environmental projects when firms are punished [8]. With the similar characteristics between target firms and peer firms, environmental administrative penalties have a deterrent effect on potential violators. When target firms are punished, peer firms will intend to increase environmental protection investment because of the sense of hardship [26,27,28].

In terms of economic value, the existing research includes the following perspectives. The first perspective is corporate reputation. Media is an important public opinion supervision tool, and reputation is one of the intangible resources; Zou et al. [14] pointed out that environmental administrative penalties lead to negative reports, thus reducing market reputation. The second perspective is stock return. Many studies examine the relationship between environmental administrative penalties and stock return. They adopted event study methodology to analyst investors’ reactions to environmental misconducts, and showed that firms’ stock returns decreased when they experienced environmental administrative penalties [29,30,31]. In particular, Xiong et al. [23] extended the research into the area of supply chain, and the negative impact on stock returns also appear in the violators’ supply chain partners, such as customers and suppliers. The third perspective is financing constraints. Zou et al. [15] confirmed the negative effect of environmental administrative penalties on the cost of loan financing, especially after the violation announcement involving wastewater discharge being issued. Ma et al. [32] believed that environmental penalties raise investors’ concerns and further induce a higher cost of corporate bonds.

2.2. Influencing Factors of Audit Fees

Simunic [33] put forward the audit pricing model, indicating that firm scale, number of subsidiaries and business complexity impact audit fees. Since then, scholars have conducted extensive research on audit fees and found that client characteristics [17], auditor characteristics [18,19], and external environmental uncertainty [20,21] can also influence audit fees. With the implementation of the ecological civilization establishment, the Auditing Standards for the Chinese Certified Public Accountants No. 1631, Consideration of Environmental Information in the Audit of Financial Statements, requires auditors to pay more attention to environmental matters. Existing studies have discussed the impact of environmental matters on auditors’ behaviors from the perspectives of corporate environmental information disclosure, environmental responsibility, and environmental regulation. Yao et al. [7] took the new audit standards issued by the Chinese Institute of Certified Public Accountants as a natural experiment and believed that the requirements of the new audit standards for environmental information disclosure improve the transparency of enterprise environmental information, thus reducing audit pricing and improving audit quality. Liu et al. [34] viewed the implementation of new environmental protection law (EPL) as a natural experiment and found that new EPL increases environmental and audit risk for heavily polluting firms. To promote environmental governance, the China government implemented corporate environmental liability insurance in 2013. However, firms which purchase environmental liability insurance have more “environmental inertia”, and the corporate environmental risks are higher [14]. Therefore, auditors are more cautious and charge higher audit fees. Huang [8] believed that although shareholders pay more and more attention to corporate environmental risks, some firms’ environmental responsibility responses are ostensible. That is to say, firms cover up negative environmental activities through greenwashing and mislead auditors’ decisions. In addition, firms in different areas of China may face varied social norms, which means that the impact of corporate environmental responsibility on audit fees is asymmetric [35]. Yu et al. [9] examined the relationship between regional environmental regulation and audit fees.

To the best of our knowledge, there is a paucity of studies on the relationship between environmental risks and audit fees. The administrative measure for information disclosure of listed companies, promulgated and implemented in 2007, requires companies to report environmental administrative penalty information, which provides an opportunity to measure firms’ environmental risks and investigate the relationship between environmental risks and audit fees. The article attempts to examines environmental risks by focusing on environmental administrative penalties of firms and explores the impact of environmental administrative penalties on audit fees.

3. Theory and Research Hypotheses

According to Simunic [33], audit fees consist of the costs during audit procedures and the compensation of potential risks. The audit cost includes human resources, time resources, and material resources. The risk premium comes from the possible losses, such as reputation risks and litigation risks. The Auditing Standards for the Chinese Certified Public Accountants No. 1631, Consideration of Environmental Information in the Audit of Financial Statements, emphasizes the importance of assessing both favorable and unfavorable environmental matters throughout the audit process, and unfavorable environmental matters include the economic consequences due to environmental penalties [8]. Thus, environmental administrative penalties may affect audit fees in two ways.

On the one hand, the adverse economic consequences brought by environmental administrative penalties increase firms’ risks and then lead to higher risk premiums. Firstly, environmental administrative penalties bring higher audit fees because of business risks. The punished firms will face rectification requirements such as production restriction, shutdown, or business suspension, and their production capacity may be constrained due to industrial policies [28]. The firms’ business licenses will be revoked, or the firms will be closed directly, which may result in serious uncertainties regarding future operations. In addition, based on the signaling theory, it is believed that customers and investors will be aware of firms’ environmental risks with the disclosure of environmental administrative penalties, and the pessimistic expectations from investors and consumers appear, which influence earning fluctuation [36,37]. So, auditors can evaluate the operational risk via environmental administrative penalties and charge higher audit fees [38]. Secondly, environmental administrative penalties increase reputation risks and auditors charge higher audit fees. Based on reputation theory, reputation is an important intangible asset of a corporation, and environmental performance is linked to corporate reputation [39,40]. Zou et al. [14] also found that environmental violation events can harm corporate reputation. In October 2011, Beijing entered the haze weather period. At the end of the same month, the US embassy released a message via microblogging regarding Beijing PM2.5 exceeding the standard, but the Chinese government never disclosed the PM2.5 data, which aroused media and public discussion. After the “PM2.5” event in 2011, environmental risks attracted extensive attention from the media [41]. The media reports and traces the reasons for firms being punished, because of industry competition and social responsibility [42]. These negative reports can not only increase firms’ negative reputation and auditors’ perceptions about risks, but also may transfer media pressure to auditors who provide services for firms. In such circumstances, the public will pay more attention to auditors, which will lead to the risk of audit failure being discovered, reputation loss [43] and litigation risks [44]. Liu et al. [45] argue that auditors incorporate firms’ reputation risks into their decision of audit fees.

On the other hand, auditors may also spend more resources including human, material, and time to make correct decisions and charge higher cost compensation. Auditors may also spend more audit effort to make correct decisions. When firms are subject to environmental administrative penalties, they may have more potential environmental risks. To deal with these risks, auditors will conduct a more detailed environmental risk assessment, formulate a comprehensive audit plan, select appropriate auditors with professional competence and carry out more detailed tests. In addition, when auditors perceive higher environmental risks, they will allocate more hours to obtain adequate audit evidence, and it is necessary to make good use of experts’ work to clearly recognize firms’ environmental risks throughout the audit process [10]. Accordingly, environmental administrative penalties may force auditors to spend more audit effort and charge higher audit fees.

In sum, to reduce audit risk, auditors may choose to charge higher audit risk premiums to cope with operation risks and reputation risks. In addition, auditors may also choose to spend more effort. In other words, environmental administrative penalties could affect audit fees through the “risk premium mechanism” or “cost compensation mechanism”. Hence, we propose the following predictions:

Hypothesis 1 (H1).

Firms’ environmental administrative penalties lead to higher audit fees.

Hypothesis 1a (H1a).

Firms’ environmental administrative penalties lead to higher audit fees through the “risk premium mechanism”.

Hypothesis 1b (H1b).

Firms’ environmental administrative penalties lead to higher audit fees through the “cost compensation mechanism”.

The internal control is an internal governance factor that affects the relationship between environmental administrative penalties and audit fees. The internal control system includes the internal environment, risk assessment, control activities, communication, and internal supervision. It is an important criterion to judge whether the firm’s internal governance environment is good or not.

Many studies show that effective internal control is the guarantee to improve operation efficiency [46,47]. The adverse economic consequences caused by environmental administrative penalties will be alleviated when the internal control environment is effective. Firstly, the punished firms may suffer rectification or shutdown, which will lead to operation uncertainties, such as the decrease in cash flows, stock return and market value [29,30,32]. Secondly, firms may be subject to industrial constraints and may be required to make industrial adjustments after being punished, especially for heavily polluting companies. However, compared with other firms, it is believed that the ability to deal with market shocks and maintain business safety is stronger for firms with better internal control. Undoubtedly, the effective internal control can alleviate the adverse impact of environmental administrative penalties on firms, which may lead to lower audit fees. Hence, we give the following prediction:

Hypothesis 2 (H2).

The effective internal control weakens the positive impact of firms’ environmental administrative penalties on audit fees.

The regional environmental regulation is an external governance factor affecting the relationship between environmental administrative penalties and audit fees. Environmental regulation can not only impact corporate environmental governance behavior [48], but also audit fees [9].

With the gradual strengthening of environmental regulation, the possibility of firms being punished increases significantly. Stronger environmental regulation also draws more public’s environmental attention, which results in higher earning fluctuation [49]. In addition, under the strict environmental regulation, the media will pay more attention to firms’ pollution behaviors and deliver more negative reports of punished firms, which arouses firms’ reputation risks. Therefore, auditors, the entities to which reputation risks may be transferred, face higher litigation risks and more serious reputation threats when auditing in areas of strict environmental regulation. In addition, auditors need to be more knowledgeable about environmental rules, regulations, and legal provisions to avoid audit failure. At the same time, it is important to understand deeply the environmental risks in the process of production and operation, even making use of experts’ work when necessary [9]. Therefore, in areas with strict environmental regulation, auditors fall into more serious uncertain risks and spend more audit effort, which requires higher audit fees. Therefore, we present the following prediction:

Hypothesis 3 (H3).

Stronger regional environmental regulation enhances the positive impact of firms’ environmental administrative penalties on audit fees.

4. Methodology

4.1. Data and Sample

Our sample contains data of non-financial heavily polluting listed firms in China’s Shanghai and Shenzhen A-share market for the period ranging from 2011 to 2019. The heavily polluting firms are not only the source of pollution, but are also more sensitive to environmental regulation. Compared with non-heavily polluting listed firms, more heavily polluting listed firms are subject to environmental penalties. The definition of heavily polluting listed companies is based on the guidelines for environmental information disclosure of listed companies (Exposure Draft), published by the Ministry of environmental protection in 2010, and the classified management directory of environmental protection verification industry of listed companies (HBH [2008] No. 373). Considering the formal implementation of the revised measures for environmental administrative penalties in March 2010, 2011 is the starting year in the study. Additionally, because the penalty data may be incomplete after the year of 2019, which may lead to bias of our results, the samples stop in 2019. The article obtains financial data from China Stock Market and Accounting Research (CSMAR). The data of environmental administrative penalties comes from the website of the Institute of Public and Environmental Affairs (IPE).

The study implements the following process in the sample selection: Firstly, we delete the samples of special treatment (*ST, ST) and particular transfer (PT) listed companies. Secondly, we delete the missing sample. Thirdly, observations with leverage ratios greater than one are also eliminated, and to control the influence of extreme values, all continuous variables are winsorized at the levels of 1% and 99%. After these processes, the final firm-year observations are 5513 for 843 firms. Additionally, the STATA16.0 software is used in the study.

4.2. Variables

4.2.1. Dependent Variable—Fees

The audit fees present an obvious right-skewed distribution. In order to ensure the data meets the characteristics of normal distribution, consistent with the study of Wu and Ye (2020) [15], the natural logarithm is taken for the audit fees (RMB 10,000).

4.2.2. Independent Variable—Penalty

The Institute of Public and Environmental Affairs (IPE) publishes the information of firms’ environmental administrative penalties. The study uses the “content analysis method” and follows previous research [28] to measure firms’ environmental administrative penalties from the number of penalties (Pen-Num), the fine of penalties (Pen-Fine) (in RMB 10,000), the severity of penalties (Pen-Sev) and the department of penalties (Pen-Dep). The severity and department of penalties are analyzed through the scoring standard of the measures for enterprise environmental credit evaluation (for Trial Implementation) (The scoring standards come from these websites: https://sthjt.ah.gov.cn/public/21691/119957661.html accessed on 8 January 2021; http://www.yixing.gov.cn/doc/2019/02/01/684149.shtml accessed on 1 February 2019; http://www.gongzhuling.gov.cn/hbt/zwgk/tzgg/201812/t20181219_7752287.html accessed on 19 December 2018) which was issued by China in 2013. The score of penalty severity is calculated through the number of penalties of each type, multiplied by the score of each penalty type, then summing them. The score of penalty department is calculated through the number of penalties, multiplied by the score of each penalty department, then summing them. The specific scoring standard of each penalty type is shown in

Table A1. The score standard of different penalty types.

The Type of Penalties	Score
warning	1
Rectification	1
Fines < 10,000 yuan	1
10,000 yuan <= Fines < 50,000 yuan	2
50,000 yuan <= Fines < 100,000 yuan	3
100,000 yuan <= Fines < 200,000 yuan	4
Fines >= 200,000 yuan	6
Stop construction	6
Production restriction	6
Close	6
Confiscation of illegal income	6
Suspension of license	6
Stop production for rectification	12
Revocation of license	12
Administrative detention	12
Environmental crime	12

Note: In China, the types of environmental administrative penalties include warning, rectification, fine, stop construction, and so on; we score them depending on the severity according to the scoring standard of the measures for enterprise environmental credit evaluation (for Trial Implementation).

. The specific scoring standard of different penalty departments is shown in

Table A2. The score standard of different penalty departments.

The Departments of Penalties	Score
County Environmental Protection Bureau	1
County Government	2
Prefecture Level Municipal Environmental Protection Bureau	3
Prefecture Level municipal government	4
Provincial Environmental Protection Bureau	5
Provincial Government	6
Ministry of environmental protection	7
the State Council	8

Note: In China, the environment penalty departments include County Environmental Protection Bureau, County Government, Prefecture Level Municipal Environmental Protection Bureau, Prefecture Level municipal government, Provincial Environmental Protection Bureau, Provincial Government, Ministry of environmental protection, and the State Council, and we give the score from 1 to 8.

. In order to eliminate the influence of dimensions and more conveniently compare the difference of influence degree between various environmental punishment dimensions, according to the regression coefficient, we standardize them and use the standardized variables in regression analysis.

4.2.3. Control Variables

According to the existing literature [50,51], we select some control variables from two aspects. From the perspective of firms, we choose the enterprise value (TobinQ), asset-liability ratio (Lev), return on total assets (Roa), growth ratio of operating revenue (Growth), the shareholding of top ten shareholders (TOP10), loss or not (Loss), CEO duality (Duality), nature of property rights (Soe), and number of board of directors (Board). In addition, from the perspective of audit firms, we choose the audit opinion of the previous period (Lopin) and audit tenure (Tenure). In addition, we control the industry-fixed effect (Industry FE), time-fixed effect (Year FE), and province-fixed effect (Province FE). The specific variable definitions are shown in

Table A3. Definitions of variables.

Variables	Definition
Fee	The log of annual audit fees plus one.
Pen-Num-ori	The number of environmental administrative penalties in the current year
Pen-Fine-ori	The fine of environmental administrative penalties in the current year
Pen-Sev-ori	The severity of environmental administrative penalties in the current year
Pen-Dep-ori	The department of environmental administrative penalties in the current year
Pen-Num	The Standardization of Pen-Num-ori
Pen-Fine	The Standardization of Pen-Fine-ori
Pen-Sev	The Standardization of Pen-Sev-ori
Pen-Dep	The Standardization of Pen-Dep-ori
TobinQ	(total issued market value + total liabilities) / total assets
Lev	Total liabilities/total assets
Roa	Net profits /total assets
Growth	Growth rate of operating revenue
TOP10	The number of shares held by top ten shareholders /total number of shares
Loss	A dummy variable that equals 1 when the firm has operating losses, and 0 otherwise
Duality	A dummy variable that equals 1 when the general manager and chairman are the same person, and 0 otherwise
Soe	A dummy variable that equals 1 when the firm is Soe, and 0 otherwise
Board	The log of the number of directors in the board of directors plus one
LOpin	A dummy variable that equals 1 when the previous report issues a non-standard audit opinion, and 0 otherwise
Tenure	Tenure is calculated with the natural logarithm of years of audit services provided for listed companies plus 1
Year	Dummy variable coded 1 for the specific year, and 0 otherwise
Industry	Dummy variable coded 1 for the specific industry, and 0 otherwise
Province	Dummy variable coded 1 for the specific province, and 0 otherwise

.

4.2.4. Moderating Variables

Internal control (IC): We select the internal control index in the Dibo database and divide the internal control index by 100. The larger value shows the enterprise’s internal control is more effective.

Regional environmental regulation intensity (ERI): Consistent with the prior study of Ren et al. [52], ERI is calculated by the discharge of industrial wastewater, industrial SO₂ and industrial smoke, and the greater the value of ERI shows the environmental regulation intensity is weaker.

4.3. Empirical Model

In order to test H1, the empirical model is as follows, according to Simunic [33].

$$F{\mathrm{ee}}_{i,t}={\beta }_0+{\beta }_{1}Penalt{y}_{i,t}+{\displaystyle \sum {\beta }_i}Control{s}_{i,t}+{\xi }_{i,t}$$

(1)

where $F{\mathrm{ee}}_{i,t}$ represents audit fees of firm i in year t, $Penalt{y}_{i,t}$ represents the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev) and the department of penalties (Pen-Dep) of firm i in year t.

In order to test H2 and H3, the empirical models are as follows.

$$F{\mathrm{ee}}_{i,t}={\beta }_0+{\beta }_{1}Penalt{y}_{i,t}+{\beta }_{2}I{C}_{i,t}+{\beta }_{3}Penalt{y}_{i,t}\times I{C}_{i,t}+{\displaystyle \sum {\beta }_i}Control{s}_{i,t}+{\xi }_{i,t}$$

(2)

$$F{\mathrm{ee}}_{i,t}={\beta }_0+{\beta }_{1}Penalt{y}_{i,t}+{\beta }_{2}ER{I}_{i,t}+{\beta }_{3}Penalt{y}_{i,t}\times ER{I}_{i,t}+{\displaystyle \sum {\beta }_i}Control{s}_{i,t}+{\xi }_{i,t}$$

(3)

Among them, $I{C}_{i,t}$ refers to the internal control of firm i in year t and $ER{I}_{i,t}$ refers to the regional environmental regulation intensity of firm i in year t.

5. Results

5.1. Descriptive Statistics

Summary statistics are presented in

Table 1. Descriptive statistics.

Variable	N	Mean	Sd	Min	Max
Fee	5513	13.681	0.634	12.585	15.825
Pen-Num-ori	5513	0.179	1.320	0.000	29.000
Pen-Fine-ori	5513	2.027	23.415	0.000	576.000
Pen-Sev-ori	5513	0.564	4.821	0.000	108.000
Pen-Dep-ori	5513	0.478	3.851	0.000	87.000
Pen-Num	5513	0.000	1.000	−0.136	21.838
Pen-Fine	5513	0.000	1.000	−0.087	24.513
Pen-Sev	5513	−0.000	1.000	−0.117	22.286
Pen-Dep	5513	−0.000	1.000	−0.124	22.469
TobinQ	5513	1.925	1.176	0.852	7.251
Lev	5513	0.443	0.200	0.057	0.876
Roa	5513	0.038	0.055	−0.164	0.205
Growth	5513	0.196	0.656	−0.669	4.825
TOP10	5513	0.564	0.148	0.217	0.910
Loss	5513	0.104	0.306	0.000	1.000
Duality	5513	0.202	0.401	0.000	1.000
Soe	5513	0.467	0.499	0.000	1.000
Board	5513	2.274	0.176	1.792	2.773
LOpin	5513	0.976	0.152	0.000	1.000
Tenure	5513	1.959	0.716	0.693	3.178

Note: This table presents summary statistics of the variables used. The table of all variable definitions is in Appendix A.

. Firstly, in terms of audit fees, the minimum value is 12.585, the maximum value is 15.825, and the standard deviation is 0.634. Secondly, for the original values, the standard deviations of the number of penalties (Pen-Num-ori), the fine of penalties (Pen-Fine-ori), the severity of penalties (Pen-Sev-ori), and the department of penalties (Pen-Dep-ori) are 1.320, 23.415, 4.821, and 3.851, respectively, indicating that there are great differences in environmental administrative penalties among heavily polluting listed firms in China. In particular, there are significant differences between fine of penalties (Pen-Fine-ori) values, which is also the most common method of environmental punishment in our country. Additionally, the mean of the department of penalties (Pen-Dep-ori) is 0.478, which shows that the probability of firms being punished by higher-level environmental regulatory departments is low. Additionally, the variables after standardizing are the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev) and the department of penalties (Pen-Dep), with a mean value of 0 and standard deviation of 1. The descriptive statistics for the control variables are also in line with expectations.

5.2. Correlation Analysis

Table 2. Correlation matrix.

	Fee	Pen-Num	Pen-Fine	Pen-Sev	Pen-Dep
Fee	1
Pen-Num	0.075 ***	1
Pen-Fine	0.074 ***	0.847 ***	1
Pen-Sev	0.068 ***	0.966 ***	0.893 ***	1
Pen-Dep	0.086 ***	0.924 ***	0.783 ***	0.892 ***	1

Note: *** denotes statistical significance at the 1% level. Limited by space, the test results of correlation coefficient of control variables are not reported, and the data are available from authors.

is the Pearson correlation coefficient. The results show that there is a significant positive correlation between environmental administrative penalties and audit fees, which provides preliminary evidence for H1. At the same time, the correlation coefficient between the control variables is appropriate, indicating that there is no serious multicollinearity problem between variables in the study.

5.3. Tests Results

5.3.1. The Effect of Environmental Administrative Penalties on Audit Fees

Firstly, the direct effects of environmental administrative penalties on audit fees are tested. The regression results for the direct effects are shown in

Table 3. The effect of environmental administrative penalties on audit fees.

	(1)	(2)	(3)	(4)
	Fee	Fee	Fee	Fee
Pen-Num	0.027 **
	(2.540)
Pen-Fine		0.026 **
		(2.136)
Pen-Sev			0.025 **
			(2.327)
Pen-Dep				0.031 ***
				(3.097)
TobinQ	−0.118 ***	−0.118 ***	−0.118 ***	−0.118 ***
	(−7.760)	(−7.764)	(−7.762)	(−7.742)
Lev	0.968 ***	0.969 ***	0.969 ***	0.968 ***
	(10.407)	(10.407)	(10.414)	(10.408)
Roa	1.389 ***	1.384 ***	1.386 ***	1.385 ***
	(4.279)	(4.266)	(4.270)	(4.270)
Growth	−0.057 ***	−0.057 ***	−0.057 ***	−0.056 ***
	(−3.030)	(−3.018)	(−3.035)	(−2.993)
TOP10	0.832 ***	0.831 ***	0.833 ***	0.829 ***
	(6.865)	(6.858)	(6.863)	(6.844)
Loss	0.013	0.012	0.012	0.012
	(0.354)	(0.328)	(0.337)	(0.334)
Duality	−0.037	−0.037	−0.037	−0.036
	(−1.185)	(−1.185)	(−1.188)	(−1.165)
Soe	−0.018	−0.017	−0.018	−0.019
	(−0.470)	(−0.444)	(−0.450)	(−0.487)
Board	0.442 ***	0.443 ***	0.442 ***	0.443 ***
	(4.584)	(4.591)	(4.583)	(4.602)
LOpin	0.051	0.051	0.051	0.051
	(0.599)	(0.596)	(0.598)	(0.595)
Tenure	0.033	0.033	0.033	0.033
	(1.610)	(1.595)	(1.613)	(1.612)
Constant	12.126 ***	12.125 ***	12.125 ***	12.124 ***
	(40.386)	(40.383)	(40.379)	(40.436)
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513
Adj. R2	0.360	0.360	0.360	0.361

Note: robust standard errors clustered at the firm level are in parentheses. *** and ** denote statistical significance at the 1% and 5% levels (two-tailed), respectively.

. The results show that the coefficients of the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev) are positive and marginally significant at the level of 5%. The coefficient between the department of penalties (Pen-Dep) and audit fees (Fee) is positive and marginally significant at the level of 1%. The results show that environmental administrative penalties will increase the audit fees, which supports H1. We can also find the following: the audit fees will increase 4.26% (=0.027 × 1/0.634) if the number of penalties increases by one standard deviation; the audit fees will increase 4.10% (=0.026 × 1/0.634) if the fine of penalties increases by one standard deviation; the audit fees will increase 3.94% (= 0.025 × 1/0.634) if the severity of penalties increases by one standard deviation; the audit fees will increase 4.89% (=0.031 × 1/0.634) if the department of penalties increases by one standard deviation. On the one hand, environmental administrative penalties bring adverse effects on the sustainable operation of firms, leading to greater reputational pressure and auditors charging a higher risk premium. On the other hand, auditors may also spend more effort and charge higher cost compensation when firms are punished, but the specific mechanism still needs to be discussed further.

5.3.2. The Effect of Internal Control

To measure the impact of internal control, we establish Equation (2). The regression results are shown in

Table 4. The effect of internal control.

	(1)	(2)	(3)	(4)
	Fee	Fee	Fee	Fee
Pen-Num	0.057 ***
	(4.959)
Pen-Num×IC	−0.006 **
	(−2.150)
Pen-Fine		0.058 ***
		(5.111)
Pen-Fine×IC		−0.006 *
		(−1.923)
Pen-Sev			0.052 ***
			(5.719)
Pen-Sev×IC			−0.005 **
			(−2.068)
Pen-Dep				0.054 ***
				(4.684)
Pen-Dep×IC				−0.004 *
				(−1.777)
IC	0.011	0.011	0.011	0.011
	(1.313)	(1.258)	(1.286)	(1.325)
TobinQ	−0.117 ***	−0.117 ***	−0.117 ***	−0.117 ***
	(−7.681)	(−7.686)	(−7.684)	(−7.665)
Lev	0.970 ***	0.970 ***	0.971 ***	0.969 ***
	(10.396)	(10.396)	(10.401)	(10.392)
Roa	1.332 ***	1.332 ***	1.331 ***	1.328 ***
	(4.042)	(4.041)	(4.037)	(4.032)
Growth	−0.057 ***	−0.057 ***	−0.057 ***	−0.056 ***
	(−3.035)	(−3.030)	(−3.040)	(−3.002)
TOP10	0.827 ***	0.826***	0.828 ***	0.825 ***
	(6.837)	(6.831)	(6.838)	(6.820)
Loss	0.021	0.020	0.021	0.021
	(0.601)	(0.568)	(0.589)	(0.593)
Duality	−0.036	−0.036	−0.036	−0.035
	(−1.164)	(−1.166)	(−1.171)	(−1.147)
Soe	−0.019	−0.018	−0.018	−0.020
	(−0.486)	(−0.464)	(−0.470)	(−0.502)
Board	0.441 ***	0.441 ***	0.441 ***	0.442 ***
	(4.577)	(4.581)	(4.575)	(4.595)
LOpin	0.033	0.033	0.033	0.032
	(0.370)	(0.370)	(0.373)	(0.363)
Tenure	0.033	0.033	0.033	0.033
	(1.588)	(1.574)	(1.592)	(1.583)
Constant	12.079 ***	12.08 2 ***	12.079 ***	12.078 ***
	(40.157)	(40.159)	(40.142)	(40.202)
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513
Adj. R2	0.361	0.361	0.360	0.361

Note: robust standard errors clustered at the firm level are in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels (two-tailed), respectively.

. The results show that the coefficients of Pen-Num×IC, and Pen-Sev×IC are negative and marginally significant at the level of 5%, while the coefficients of Pen-Fine×IC and Pen-Dep×IC are significantly negative at the level of 10%. It indicates that the increase in audit fees of punished firms with weak internal control is more obvious, which proves H2. An effective internal control environment is a favorable “barrier” for firms to deal with uncertainties. Additionally, it is effective to weaken the adverse economic consequences from environmental administrative penalties, maintain the normality of operation and reduce audit fees. So, it is necessary to assess whether the internal control is effective or not, and auditors should treat them differently when making decisions regarding audit fees and especially pay more attention to firms in which the internal control is poor.

5.3.3. The Effect of the Regional Environmental Regulation

To verify the effect of the regional environmental regulation, we establish Equation (3). The regression results are shown in

Table 5. The effect of regional environmental regulation.

	(1)	(2)	(3)	(4)
	Fee	Fee	Fee	Fee
Pen-Num	0.052 ***
	(4.313)
Pen-Num×ERI	−0.031 ***
	(−3.519)
Pen-Fine		0.054 ***
		(4.022)
Pen-Fine×ERI		−0.032 ***
		(−2.964)
Pen-Sev			0.052 ***
			(4.241)
Pen-Sev×ERI			−0.031 ***
			(−3.496)
Pen-Dep				0.047 ***
				(3.661)
Pen-Dep×ERI				−0.024 **
				(−2.157)
ERI	0.010	0.008	0.009	0.008
	(0.356)	(0.286)	(0.312)	(0.265)
TobinQ	−0.118 ***	−0.118 ***	−0.118 ***	−0.118 ***
	(−7.753)	(−7.756)	(−7.753)	(−7.739)
Lev	0.968 ***	0.968 ***	0.970 ***	0.968 ***
	(10.412)	(10.408)	(10.429)	(10.410)
Roa	1.387 ***	1.382 ***	1.385 ***	1.384 ***
	(4.275)	(4.263)	(4.270)	(4.270)
Growth	−0.057 ***	−0.057 ***	−0.057 ***	−0.056 ***
	(−3.025)	(−3.009)	(−3.035)	(−2.988)
TOP10	0.832 ***	0.831 ***	0.833 ***	0.829 ***
	(6.870)	(6.856)	(6.870)	(6.846)
Loss	0.013	0.011	0.012	0.013
	(0.364)	(0.312)	(0.343)	(0.349)
Duality	−0.036	−0.036	−0.036	−0.036
	(−1.172)	(−1.177)	(−1.175)	(−1.158)
Soe	−0.019	−0.019	−0.019	−0.019
	(−0.478)	(−0.479)	(−0.479)	(−0.483)
Board	0.446 ***	0.446 ***	0.445 ***	0.445 ***
	(4.628)	(4.637)	(4.622)	(4.621)
LOpin	0.051	0.051	0.052	0.051
	(0.596)	(0.592)	(0.600)	(0.595)
Tenure	0.034	0.033	0.034	0.034
	(1.630)	(1.619)	(1.638)	(1.634)
Constant	12.119 ***	12.119 ***	12.118 ***	12.118 ***
	(40.390)	(40.380)	(40.382)	(40.413)
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513
Adj. R2	0.361	0.361	0.361	0.361

Note: robust standard errors clustered at the firm level are in parentheses. *** and ** denote statistical significance at the 1% and 5% levels (two-tailed), respectively.

, which show that the coefficients of Pen-Num×ERI, Pen-Fine×ERI, and Pen-Sev×ERI are significantly negative at the level of 1%. The coefficient of Pen-Dep×ERI is significantly negative at the level of 5%, which may be due to the low probability of firms being punished by higher-level environmental regulatory departments. The greater the value of ERI, the weaker the regional environmental regulation intensity is showed to be. So, these results indicate that if firms being punished are located in areas with more intensive environmental regulation, the audit fees will be higher, which proves H3. A possible reason is that regional environmental regulation increases the uncertainty of the external environment of firms, which will increase the adverse economic consequences of environmental administrative penalties, such as stronger market fluctuation and more negative media reports. In addition, when working in areas with weak environmental regulation, auditors need to be vigilant, learn more about environmental regulations to increase their professional knowledge, and spend more effort.

6. Further Tests

6.1. Mechanism Analysis

Our results show that environmental administrative penalties can increase audit fees. Based on the foregoing theoretical analysis, we point out there may be two mechanisms: the risk premium mechanism and the cost compensation mechanism. We draw on Baron and Kenny [53] to establish the following test equations to verify the mechanism between environmental administrative penalties and audit fees. Additionally, the $Mediato{r}_{i,t}$ represents the intermediary variables.

$$Mediato{r}_{i,t}={\beta }_0+{\beta }_{1}Penalt{y}_{i,t}+{\displaystyle \sum {\beta }_i}Control{s}_{i,t}+{\xi }_{i,t}$$

(4)

$$F{\mathrm{ee}}_{i,t}={\beta }_0+{\beta }_{1}Penalt{y}_{i,t}+{\beta }_{2}Mediato{r}_{i,t}+{\displaystyle \sum {\beta }_i}Control{s}_{i,t}+{\xi }_{i,t}$$

(5)

6.1.1. The Risk Premium Mechanism

Environmental administrative penalties may affect audit fees by increasing operational risks and reputation risks, and we test the risk premium mechanism from these two aspects.

• (1) The intermediary effect of operational risks

Not only do environmental administrative penalties reduce the market competitiveness and the inflow of operating cash, but also firms need to adjust industrial structure after being punished, especially heavily polluting enterprises, while the process is costly and risky. Therefore, environmental administrative penalties may increase firms’ operational risks. To verify the intermediary effect of operational risks, the standard deviation of individual stock returns considering cash dividend for three consecutive years (t − 1 year to t + 1 year) is used to denote the operational risk, which is represented by Risk. The greater the value, the greater operational risk. The mechanism test results are shown in

Table 6. The risk premium mechanism: operational risks.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Risk	Fee	Risk	Fee	Risk	Fee	Risk	Fee
Pen-Num	0.009 ***	0.025 **
	(3.038)	(2.404)
Pen-Fine			0.006 ***	0.025 **
			(2.658)	(2.031)
Pen-Sev					0.007 ***	0.023 **
					(2.761)	(2.209)
Pen-Dep							0.010 ***	0.029 ***
							(3.461)	(2.931)
Risk		0.210 ***		0.211 ***		0.211 ***		0.209 ***
		(4.765)		(4.787)		(4.780)		(4.742)
TobinQ	0.042 ***	−0.127 ***	0.042 ***	−0.127 ***	0.042 ***	−0.127 ***	0.042 ***	−0.126 ***
	(6.602)	(−8.209)	(6.597)	(−8.216)	(6.600)	(−8.212)	(6.609)	(−8.189)
Lev	0.156 ***	0.935 ***	0.156 ***	0.936 ***	0.156 ***	0.936 ***	0.156 ***	0.935 ***
	(4.871)	(10.142)	(4.877)	(10.140)	(4.879)	(10.148)	(4.868)	(10.143)
Roa	0.232 *	1.340 ***	0.230 *	1.335 ***	0.231 *	1.337 ***	0.231 *	1.337 ***
	(1.852)	(4.119)	(1.833)	(4.105)	(1.842)	(4.109)	(1.841)	(4.110)
Growth	0.011	−0.059 ***	0.011	−0.059 ***	0.011	−0.060 ***	0.011	−0.059 ***
	(1.525)	(−3.188)	(1.534)	(−3.178)	(1.520)	(−3.193)	(1.558)	(−3.152)
TOP10	0.139 ***	0.803 ***	0.139 ***	0.802 ***	0.139 ***	0.803 ***	0.138 ***	0.801 ***
	(4.184)	(6.655)	(4.178)	(6.648)	(4.184)	(6.653)	(4.163)	(6.636)
Loss	0.008	0.011	0.008	0.010	0.008	0.011	0.008	0.010
	(0.595)	(0.309)	(0.570)	(0.285)	(0.579)	(0.293)	(0.578)	(0.291)
Duality	0.006	−0.038	0.006	−0.038	0.006	−0.038	0.006	−0.037
	(0.515)	(−1.237)	(0.509)	(−1.237)	(0.511)	(−1.241)	(0.531)	(−1.218)
Soe	−0.011	−0.016	−0.011	−0.015	−0.011	−0.015	−0.011	−0.017
	(−0.966)	(−0.413)	(−0.924)	(−0.389)	(−0.942)	(−0.395)	(−0.982)	(−0.430)
Board	0.028	0.436 ***	0.028	0.437 ***	0.028	0.436 ***	0.029	0.437 ***
	(1.098)	(4.566)	(1.104)	(4.573)	(1.098)	(4.565)	(1.113)	(4.583)
LOpin	−0.000	0.051	0.000	0.051	−0.000	0.051	−0.000	0.051
	(−0.003)	(0.608)	(0.001)	(0.605)	(−0.002)	(0.607)	(−0.007)	(0.604)
Tenure	0.005	0.032	0.005	0.032	0.005	0.032	0.005	0.032
	(0.835)	(1.561)	(0.826)	(1.546)	(0.839)	(1.564)	(0.837)	(1.563)
Constant	0.085	12.108 ***	0.083	12.108 ***	0.084	12.107 ***	0.084	12.106 ***
	(1.019)	(40.496)	(0.999)	(40.495)	(1.008)	(40.492)	(1.009)	(40.543)
Industry	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513	5513	5513	5513	5513
Adj. R2	0.152	0.366	0.152	0.366	0.152	0.365	0.153	0.366

Note: robust standard errors clustered at the firm level are in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels (two-tailed), respectively.

. The results in columns (1), (3), (5), and (7) show that the coefficients of the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev), and the department of penalties (Pen-Dep) are significantly positive. The results of columns (2), (4), (6), and (8) show that the coefficients of Risk are also significantly positive, which indicate that the operational risk has a certain intermediary role between environmental administrative penalties and audit fees through the “risk premium mechanism”, and thus hypothesis H1a is verified.

• (2) The intermediary effect of reputation risks

Environmental administrative penalties may also affect audit fees by increasing reputation risks. With the of construction of ecological civilization, enterprises which are subject to environmental administrative penalties also face more negative media coverage. At the same time, these negative reports draw more attention to auditors from the public, which increases the risk of reputation loss and litigation. Therefore, auditors take reputation risks into consideration when making decisions. To measure the intermediary role of reputation risks, we refer to Liu et al. [45], wherein the natural logarithm of the number of negative news reports plus 1 is denoted as the reputation risk, which is represented by Media. The mechanism test results are shown in

Table 7. The risk premium mechanism: reputation risks.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Media	Fee	Media	Fee	Media	Fee	Media	Fee
Pen-Num	0.065 ***	0.016 *
	(4.365)	(1.693)
Pen-Fine			0.060 ***	0.016
			(4.757)	(1.447)
Pen-Sev					0.063 ***	0.014
					(4.525)	(1.484)
Pen-Dep							0.069 ***	0.019 **
							(5.225)	(2.082)
Media		0.174 ***		0.174 ***		0.174 ***		0.174 ***
		(12.377)		(12.397)		(12.386)		(12.349)
TobinQ	0.022	−0.122 ***	0.022	−0.122 ***	0.022	−0.122 ***	0.023	−0.122 ***
	(0.931)	(−8.928)	(0.922)	(−8.932)	(0.928)	(−8.931)	(0.951)	(−8.911)
Lev	0.998 ***	0.795 ***	0.999 ***	0.795 ***	1.000 ***	0.795 ***	0.997 ***	0.795 ***
	(6.535)	(9.446)	(6.547)	(9.443)	(6.551)	(9.449)	(6.531)	(9.447)
Roa	4.340 ***	0.635 **	4.326 ***	0.632 **	4.333 ***	0.632 **	4.327 ***	0.634 **
	(7.450)	(2.158)	(7.417)	(2.150)	(7.433)	(2.150)	(7.427)	(2.157)
Growth	−0.059 **	−0.047 **	−0.058 **	−0.047 **	−0.059 **	−0.047 **	−0.057 *	−0.046 **
	(−2.008)	(−2.525)	(−1.987)	(−2.517)	(−2.033)	(−2.527)	(−1.927)	(−2.503)
TOP10	0.831 ***	0.688 ***	0.829 ***	0.687 ***	0.831 ***	0.688 ***	0.825 ***	0.686 ***
	(4.345)	(6.047)	(4.322)	(6.043)	(4.337)	(6.047)	(4.323)	(6.030)
Loss	0.310 ***	−0.041	0.308 ***	−0.042	0.309 ***	−0.041	0.308 ***	−0.041
	(4.569)	(−1.179)	(4.533)	(−1.199)	(4.546)	(−1.193)	(4.541)	(−1.191)
Duality	0.023	−0.041	0.023	−0.041	0.023	−0.041	0.024	−0.040
	(0.395)	(−1.410)	(0.395)	(−1.410)	(0.394)	(−1.414)	(0.416)	(−1.396)
Soe	0.169 **	−0.048	0.172 **	−0.047	0.171 **	−0.047	0.168 **	−0.048
	(2.482)	(−1.316)	(2.518)	(−1.302)	(2.502)	(−1.303)	(2.469)	(−1.329)
Board	0.592 ***	0.339 ***	0.593 ***	0.340 ***	0.592 ***	0.339 ***	0.595 ***	0.340 ***
	(3.922)	(3.829)	(3.926)	(3.833)	(3.922)	(3.827)	(3.943)	(3.842)
LOpin	−0.042	0.059	−0.042	0.059	−0.043	0.059	−0.043	0.058
	(−0.435)	(0.721)	(−0.438)	(0.718)	(−0.438)	(0.721)	(−0.439)	(0.717)
Tenure	0.033	0.027	0.032	0.027	0.033	0.028	0.033	0.028
	(0.896)	(1.427)	(0.875)	(1.418)	(0.901)	(1.430)	(0.901)	(1.429)
Constant	0.058	12.116 ***	0.055	12.116 ***	0.057	12.115 ***	0.050	12.115 ***
	(0.133)	(42.881)	(0.125)	(42.868)	(0.130)	(42.869)	(0.114)	(42.920)
Industry	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513	5513	5513	5513	5513
Adj. R2	0.273	0.430	0.273	0.430	0.273	0.429	0.274	0.430

Note: robust standard errors clustered at the firm level are in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels (two-tailed), respectively.

. The results in columns (1), (3), (5), and (7) show that the coefficients of the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev), and the department of penalties (Pen-Dep) are significantly positive. Columns (2), (4), (6), and (8) show that the regression coefficients of Media are significantly positive. These results indicate that the reputation risk plays an intermediary role between environmental administrative penalties and audit fees through the “risk premium mechanism”, and thus hypothesis H1a is supported.

6.1.2. The Cost Compensation Mechanism

Environmental administrative penalties may also force auditors to spend more audit efforts, which affect audit fees. Based on the research of Zhu et al. [10], the audit cost is measured by the days between the current annual audit report date and financial statement date, which is denoted by Time. The mechanism test results are shown in

Table 8. The cost compensation mechanism: audit cost.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Time	Fee	Time	Fee	Time	Fee	Time	Fee
Pen-Num	0.127	0.027 **
	(0.599)	(2.506)
Pen-Fine			0.073	0.026 **
			(0.435)	(2.111)
Pen-Sev					0.102	0.025 **
					(0.507)	(2.295)
Pen-Dep							0.263	0.031 ***
							(1.209)	(3.012)
Time		0.002 ***		0.002 ***		0.002 ***		0.002 ***
		(3.171)		(3.181)		(3.176)		(3.148)
TobinQ	−0.903 **	−0.116 ***	−0.904 **	−0.116 ***	−0.903 **	−0.116 ***	−0.899 **	−0.116 ***
	(−2.332)	(−7.635)	(−2.337)	(−7.637)	(−2.333)	(−7.636)	(−2.321)	(−7.618)
Lev	4.754 **	0.960 ***	4.762 **	0.960 ***	4.760 **	0.961 ***	4.735 **	0.959 ***
	(2.000)	(10.327)	(2.003)	(10.327)	(2.002)	(10.334)	(1.992)	(10.328)
Roa	−38.035 ***	1.457 ***	−38.079 ***	1.452 ***	−38.058 ***	1.454 ***	−38.012 ***	1.452 ***
	(−4.370)	(4.548)	(−4.376)	(4.535)	(−4.373)	(4.538)	(−4.369)	(4.537)
Growth	0.314	−0.058 ***	0.316	−0.057 ***	0.314	−0.058 ***	0.320	−0.057 ***
	(0.650)	(−3.067)	(0.654)	(−3.054)	(0.650)	(−3.072)	(0.662)	(−3.028)
TOP10	0.242	0.832 ***	0.249	0.831 ***	0.246	0.832 ***	0.199	0.829 ***
	(0.091)	(6.866)	(0.093)	(6.860)	(0.092)	(6.864)	(0.075)	(6.846)
Loss	2.825 ***	0.008	2.820 ***	0.007	2.822 ***	0.007	2.824 ***	0.007
	(2.644)	(0.215)	(2.640)	(0.189)	(2.642)	(0.198)	(2.645)	(0.196)
Duality	1.089	−0.039	1.087	−0.038	1.088	−0.039	1.099	−0.038
	(1.250)	(−1.248)	(1.248)	(−1.248)	(1.249)	(−1.252)	(1.261)	(−1.229)
Soe	−5.284 ***	−0.009	−5.275 ***	−0.008	−5.279 ***	−0.008	−5.303 ***	−0.010
	(−5.081)	(−0.229)	(−5.076)	(−0.202)	(−5.078)	(−0.209)	(−5.098)	(−0.247)
Board	2.518	0.438 ***	2.521	0.438 ***	2.519	0.438 ***	2.524	0.439 ***
	(1.046)	(4.530)	(1.047)	(4.537)	(1.046)	(4.528)	(1.048)	(4.548)
LOpin	−5.075 **	0.061	−5.071 **	0.060	−5.074 **	0.061	−5.086 **	0.060
	(−2.482)	(0.712)	(−2.481)	(0.709)	(−2.481)	(0.711)	(−2.486)	(0.707)
Tenure	−0.130	0.033	−0.130	0.033	−0.129	0.033	−0.130	0.033
	(−0.254)	(1.626)	(−0.255)	(1.611)	(−0.253)	(1.630)	(−0.256)	(1.629)
Constant	93.050 ***	11.959 ***	93.014 ***	11.958 ***	93.033 ***	11.958 ***	93.095 ***	11.958 ***
	(13.661)	(39.996)	(13.647)	(39.986)	(13.654)	(39.986)	(13.664)	(40.050)
Industry	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513	5513	5513	5513	5513
Adj. R2	0.125	0.363	0.125	0.363	0.125	0.363	0.125	0.363

Note: robust standard errors clustered at the firm level are in parentheses. *** and ** denote statistical significance at the 1% and 5% levels (two-tailed), respectively.

. The results in columns (1), (3), (5), and (7) show that the coefficients of the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev), and the department of penalties (Pen-Dep) are positive but not significant. The results of columns (2), (4), (6), and (8) show that the coefficients of the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev), the department of penalties (Pen-Dep), and the coefficients of audit cost (Time) are significantly positive. In addition, the results of the Sobel test show that the Z statistical values are not significant, which shows the audit cost (Time) does not play an intermediary role between environmental administrative penalties and audit fees. Therefore, hypothesis H1b is not verified.

6.2. Is Audit Fees Functional for Environmental Governance?

Prior research found that green innovation is popular in dealing with environmental deterioration and achieving sustainable development [54,55], which also presents firms’ effort and determination in environmental protection. Therefore, the study observes whether the increase in audit fees because of environmental administrative penalties can encourage firms to participate in green innovation activities. In previous studies, R&D and patents are both used to measure innovation. However, when focusing on green innovation, the data of green R&D is not available in China, and we use green patents to measure green innovation, consistent with Zhang et al. [12]. Meanwhile, green innovation activities are risky and costly, and we lag two periods for green innovation variables, which is denoted by Green_pat_t+2. According to existing studies [12,56], we select some control variables including the asset (Size), asset–liability ratio (Lev), return on total assets (Roa), the proportion of independent directors (Ind), and nature of property rights (Soe), and we also control the industry-fixed effect (Industry FE), time-fixed effect (Year FE), and province-fixed effect (Province FE). The regression results are in

Table 9. The function of audit fees on green innovation.

	(1)	(2)	(3)	(4)
	Green_patt+2	Green_patt+2	Green_patt+2	Green_patt+2
Fee	−0.015	−0.013	−0.014	−0.016
	(−0.237)	(−0.202)	(−0.212)	(−0.251)
Pen-Num	−0.689 **
	(−2.028)
Pen-Num×Fee	0.055 **
	(2.205)
Pen-Fine		−0.793 **
		(−2.328)
Pen-Fine×Fee		0.061 **
		(2.536)
Pen-Sev			−0.759 **
			(−2.049)
Pen-Sev×Fee			0.060 **
			(2.243)
Pen-Dep				−0.611
				(−1.451)
Pen-Dep×Fee				0.050 *
				(1.673)
Size	0.215 ***	0.216 ***	0.215 ***	0.214 ***
	(5.646)	(5.605)	(5.622)	(5.677)
Lev	−0.086	−0.081	−0.084	−0.084
	(−0.608)	(−0.569)	(−0.591)	(−0.595)
Roa	0.554	0.540	0.550	0.555
	(1.523)	(1.479)	(1.509)	(1.526)
Ind	0.195	0.197	0.196	0.190
	(0.486)	(0.488)	(0.486)	(0.473)
Soe	0.124 **	0.127 **	0.126 **	0.123 **
	(2.202)	(2.239)	(2.230)	(2.185)
_cons	−4.443 ***	−4.500 ***	−4.469 ***	−4.406 ***
	(−5.676)	(−5.704)	(−5.693)	(−5.652)
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes
N	3722	3722	3722	3722
r2_a	0.165	0.163	0.165	0.166

Note: robust standard errors clustered at the firm level are in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels (two-tailed), respectively.

, which shows that the coefficients of Pen-Num×Fee, Pen-Fine×Fee, Pen-Sev×Fee, and Pen-Class×Fee are significantly positive. The results indicate that the audit fee has a positive regulatory effect on environmental administrative punishment and green innovation; in other words, the social audit can punish poor environmental performance and encourage firms to improve environmental governance behaviors by increasing audit fees.

7. Robustness Tests

We conduct five methods to further test the robustness of our results, and these robustness tests are reported in Appendix B.

7.1. Heckman Two-Stage Model

The relationship between environmental administrative penalties and audit fees may be the result of sample self-selection. In order to alleviate the endogenous problem, a Heckman two-stage model is used in the study. The model includes two stages. The first stage estimates the probability of environmental administrative penalties to obtain the inverse Mills ratio (IMR). The first stage mainly uses the probit model for estimation, whereby a dummy variable is established according to the mean of the environmental administrative penalty and divided into two groups. The group greater than the mean value is recorded as 1, otherwise it is 0. Additionally, we choose the enterprise value (TobinQ), asset-liability ratio (Lev), return on total assets (Roa), and growth ratio of operating revenue (Growth) as the characteristics which influence whether the firm is subject to environmental administrative punishment or not. In the second stage, the inverse Mills ratio (IMR) estimated in the first stage is added to the main regression model. The results are shown in

Table A4. Heckman two-stage model.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Dum-Num	Fee	Dum-Fines	Fee	Dum-Sev	Fee	Dum-Dep	Fee
Pen-Num		0.026 **
		(2.491)
Pen-Fine				0.025 **
				(2.043)
Pen-Sev						0.024 **
						(2.285)
Pen-Dep								0.031 ***
								(3.058)
Constant	−2.684 ***	11.982 ***	−3.765 ***	9.646 **	−2.745 ***	12.019 ***	−2.753 ***	12.167 ***
	(−8.253)	(5.446)	(−6.627)	(2.544)	(−8.271)	(5.494)	(−8.301)	(5.497)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	5422	5422	5258	5258	5422	5422	5422	5422
Adj. R2	0.126	0.360	0.134	0.360	0.131	0.360	0.129	0.360

Note: robust standard errors clustered at the firm level are in parentheses. *** and ** denote statistical significance at the 1% and 5% levels (two-tailed), respectively.

, which show that after adding the inverse Mills ratio (IMR) estimated in the first stage, the coefficients of the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev), and the department of penalties (Pen-Dep) are significantly positive. The above results show that the conclusion is still robust after considering the possible sample self-selection.

7.2. Reselect Time Window

The impact of environmental administrative penalties on audit fees may come from the cumulative result of penalties from several years. Therefore, the test takes the cumulative environmental administrative penalties for two years as the observation interval to analyze the impact of environmental administrative penalties on audit fees. The results (in

Table A5. Reselect time window.

	(1)	(2)	(3)	(4)
	Fee	Fee	Fee	Fee
Pen-Num-year2	0.035 ***
	(3.289)
Pen-Fine-year2		0.035 ***
		(2.879)
Pen-Sev-year2			0.033 ***
			(3.177)
Pen-Dep-year2				0.041 ***
				(4.154)
Constant	12.202 ***	12.199 ***	12.198 ***	12.196 ***
	(39.545)	(39.543)	(39.539)	(39.597)
Controls	Yes	Yes	Yes	Yes
N	4670	4670	4670	4670
Adj. R2	0.362	0.362	0.361	0.363

Note: robust standard errors clustered at the firm level are in parentheses. *** denotes statistical significance at the 1% level (two-tailed), respectively.

) indicate that there is still a significant positive correlation between environmental administrative penalties and audit fees, which are consistent with main findings.

7.3. Variable Replacement: Replace the Paper Media with Network Media

With the development of network technology, network media has gradually replaced paper media, because network media has wider dissemination and social influence. Reputation risks are measured by using the number of negative reports of network media instead of the number of negative reports of paper media, according to Zou et al. [15]. The specific results are shown in

Table A6. Variable replacement: replace the paper media with network media.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	Int-Media	Fee	Int-Media	Fee	Int-Media	Fee	Int-Media	Fee
Pen-Num	0.063 ***	0.014
	(5.539)	(1.451)
Pen-Fine			0.051 ***	0.015
			(4.976)	(1.370)
Pen-Sev					0.055 ***	0.013
					(4.819)	(1.411)
Pen-Dep							0.070 ***	0.016 *
							(6.476)	(1.740)
Int-Media		0.211 ***		0.211 ***		0.211 ***		0.211 ***
		(13.033)		(13.069)		(13.051)		(12.997)
Constant	2.605 ***	11.576 ***	2.598 ***	11.577 ***	2.600 ***	11.575 ***	2.599 ***	11.576 ***
	(7.376)	(40.696)	(7.350)	(40.698)	(7.363)	(40.691)	(7.365)	(40.730)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513	5513	5513	5513	5513
Adj. R2	0.305	0.436	0.304	0.436	0.304	0.436	0.306	0.436

Note: robust standard errors clustered at the firm level are in parentheses. *** and * denote statistical significance at the 1% and 10% levels (two-tailed), respectively.

. The results are also robust. Therefore, reputation risks play an intermediary role between environmental administrative penalties and audit fees, which shows our results are robust.

7.4. Alternative Interpretation

Auditor reputation can also convey corporate reputation to some extent. When a firm is subjected punishment, the firm may employ auditors with higher reputation to deliver positive information to investors. Therefore, the high audit fees caused by environmental administrative penalties may come from auditor reputation premium.

Table A7. Alternative interpretation.

	(1)	(2)	(3)	(4)
	Big10	Big10	Big10	Big10
Pen-Num	0.079
	(1.551)
Pen-Fine		0.061
		(1.291)
Pen-Sev			0.064
			(1.334)
Pen-Dep				0.108
				(1.483)
Constant	−1.125	−1.137	−1.133	−1.120
	(−1.146)	(−1.157)	(−1.154)	(−1.141)
Controls	Yes	Yes	Yes	Yes
N	5507	5507	5507	5507
Pseudo R2	0.118	0.118	0.118	0.119

shows that there is no significant positive correlation between environmental administrative penalties and auditor reputation; in other words, environmental administrative penalties do not significantly improve the employment demand of firms for high reputation auditors, which rules out this alternative hypothesis. The main findings are still robust.

7.5. Fixed Effect Test

In the baseline regression, we use the pooled OLS regression method, but this test cannot exclude the influence of unobservable factors. Therefore, we further use the fixed effect test. The results are in

Table A8. Fixed Effect Test.

	(1)	(2)	(3)	(4)
	Fee	Fee	Fee	Fee
Pen-Num	0.010 ***
	(3.001)
Pen-Fine		0.009 **
		(2.391)
Pen-Sev			0.009 **
			(2.552)
Pen-Dep				0.007 *
				(1.918)
Constant	13.735 ***	13.739 ***	13.736 ***	13.736 ***
	(61.531)	(61.548)	(61.525)	(61.523)
Controls	Yes	Yes	Yes	Yes
N	5513	5513	5513	5513
Adj-R2	0.068	0.067	0.067	0.067

Note: robust standard errors clustered at the firm level are in parentheses. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels (two-tailed), respectively.

, which show that there is still a significant positive correlation between environmental administrative penalties and audit fees.

8. Conclusions, Recommendations and Limitations

8.1. Main Conclusions

The construction of ecological civilization is a responsibility all over the world. As the largest developing country, China also makes its contributions. In 2021, China set carbon peak and carbon neutral as the aim to achieve sustainable development. Besides that, environmental administrative penalties also play a fundamental role in ecological protection.

Although audit fees have gained considerable attention from many scholars, research on how environmental administrative penalties affect audit fees is scarce. The study attempts to fill the gap. Using data of heavily polluting listed firms in China, which have been revealed by the Institute of Public and Environmental Affairs (IPE) from 2011–2019, the study examines the relationship between environmental administrative penalties and audit fees from four dimensions including the number of penalties, the fine of penalties, the severity of penalties, and the department of penalties. The findings show the following:

(1) Environmental administrative penalties are significantly positively associated with audit fees. In other words, the impacts of the number of penalties, the fine of penalties, the severity of penalties, and the department of penalties on audit fees are all significantly positive.

(2) Effective internal control can attenuate the positive association between environmental administrative penalties and audit fees, but regional environmental regulation can enhance the positive impact of firms’ environmental administrative penalties on audit fees.

(3) The impact of environmental administrative penalties on audit fees mainly comes from the “risk premium mechanism” rather than the “cost compensation mechanism”, and the response from audit fees can encourage firms to engage in green innovation activities.

8.2. Recommendations

Presently, China is in an important period of economic transition, and achieving high-quality economic development has become the requirement of the times. Environmental protection is vital to economic development, human health, and social stability, and environmental administrative penalties are an important means of environmental protection. According to our findings, we show the following recommendations.

(1) For auditors, with the development of the sustainable and green economy, auditors, who are the important external stakeholders, should punish the poor environmental performance of enterprises by increasing audit fees and thus encourage firms to decrease their environmental pollution. In addition, our research also exhibits that the internal control and regional environmental regulation can regulate the relationship between environmental administrative punishment and audit fees. In particular, auditors need to take firms’ internal control and regional environmental regulation into account when making decisions and treat them differently.

(2) For firms, when firms receive environmental administrative penalties, there will be many adverse economic consequences from environmental penalties, such as business risk, reputation risk and higher audit fees. Therefore, it is necessary for firms to invest more in environmental governance and consider the long-term benefit.

(3) For governments, our findings point out that auditors can give proper response to firms being punished by requiring higher audit fees and encouraging firms to engage in green innovation, which means that environment governance can be effective by mobilizing stakeholders. So, when carrying out environmental regulation, the government should combine the strength of stakeholders in the capital market and show a “1 + 1 > 2” effect on environmental protection.

8.3. Limitations and Future Work

Although our study considers various conditions, there are also several limitations.

(1) The research reflects the relationship between environmental administrative penalties and audit fees in China. However, China, the largest developing country, is different from other countries in terms of economic system and market characteristics, especially compared to developed ones. Thus, evidence from other countries would help to obtain more general results. We expect future research to develop the study by examining non-Chinese listed firms.

(2) The article selects the number of penalties (Pen-Num), the fine of penalties (Pen-Fine), the severity of penalties (Pen-Sev), and the department of penalties (Pen-Dep) as the characteristics of environmental administrative penalties. There may be many meaningful features that we have not considered, so other characteristics of penalties should be explored in the future.

(3) The data of environmental administrative penalties come from manual collection, due to the difficulty and workload of data acquisition, and thus only the heavily polluting enterprises are researched in the study. When the data acquisition is more convenient, the sample can be extended to all listed companies.