When Confidence Backfires: The Impact of Managerial Overconfidence on Environmental Information Disclosure

Abstract

This paper investigates whether and how managerial overconfidence affects firms’ environmental information disclosure. Using 32,191 firm-year observations from Chinese companies between 2008 and 2022, the study finds that managerial overconfidence negatively impacts environmental information disclosure. These results remain robust across various tests, including alternative overconfidence measures, instrumental variable regressions, and propensity score matching. Mechanism analysis reveals that overconfidence reduces disclosure through overestimation of risk control ability and underestimation of stakeholder importance. Further analysis shows that external governance pressures from government regulation, media scrutiny, and institutional investor monitoring effectively mitigate this negative impact, while also confirming the value relevance of environmental information disclosure and the moderating role of managerial overconfidence. This study clarifies the influence and mechanisms of managerial overconfidence on environmental disclosure in developing countries, highlighting the role of external oversight.

1. Introduction

Environmental degradation poses a significant challenge to China’s sustainable development, threatening both the well-being and health of its population [1]. In response to this pressing issue, the Chinese government has introduced a series of regulations to improve environmental quality and protection efficiency, encouraging companies to disclose information related to their environmental practices [2]. However, apart from key pollutant-emitting enterprises, listed companies still retain considerable flexibility in environmental information disclosure. The substantial discretion exercised by management in these disclosures amplifies the influence of executives on the reported environmental information. This can lead to selective disclosure and other distortions, which may ultimately affect investors’ and stakeholders’ perceptions and decisions regarding the company’s environmental performance.

Some scholars have recognized the potential role of managers in explaining the diversity of environmental practices [3,4]. Previous studies have mainly focused on observable managerial traits such as gender [5], tenure [6], educational background [6], and military experience [7], while psychological traits such as overconfidence have received less attention. Executive overconfidence, typically characterized by an overestimation of capabilities and an underestimation of risks [8,9], has been linked to various firm outcomes, including financial strategies [10,11], innovation [10,12], and socially responsible activities [13,14,15]. While overconfident executives are often seen as more aggressive in driving firm performance, it remains unclear whether they are also proactive in sustainability, especially in environmental disclosure. Against this backdrop, this study aims to investigate whether and how managerial overconfidence affects corporate environmental information disclosure. In particular, we seek to disentangle the mechanisms through which overconfidence may either promote or inhibit disclosure behaviors.

Overconfidence may exert countervailing impacts on environmental information disclosure. On the one hand, overconfident executives tend to overestimate the benefits of their decisions and underestimate the associated risks [10,16]. This implies that they may exaggerate the perceived benefits of disclosing environmental information while downplaying the associated costs [16], thereby potentially increasing environmental information disclosure. Their motivation to outperform their peers may further strengthen this trend. On the other hand, overconfidence can also negatively impact environmental information disclosure. Overconfident executives, who are self-assured in their abilities and inclined to take risks and confront challenges, may exhibit lower reliance on stakeholders [9,11] and perceive less need for the risk-mitigating role of environmental information disclosure [17,18]. Consequently, they might disclose less environmental information.

Using content analysis of 32,191 firm-year observations from Chinese listed companies between 2008 and 2022, this study empirically examines this relationship and explores the underlying mechanisms. We find a significant negative correlation between managerial overconfidence and environmental disclosure. Our results are robust to a series of tests, including alternative measures of overconfidence, instrumental variable regressions, and propensity score matching. Mechanism analysis suggests that overconfident managers tend to reduce disclosure by overestimating their ability to control environmental risks and underestimating stakeholder importance. Further analysis shows that external governance pressures from government regulation, media scrutiny, and institutional investor monitoring can effectively mitigate this negative effect. Our study makes contributions to the existing literature on several fronts.

Firstly, our study is the first to examine the relationship between managerial overconfidence and environmental information disclosure. Our findings indicate that, in contrast to their characteristically aggressive behavior in financial decision-making, overconfident managers tend to adopt a more restrained and passive stance toward environmental information disclosure. Unlike prior studies that primarily focus on specific aspects of environmental disclosure, such as carbon or greenhouse gas emissions [19,20] our research broadens the scope by examining the influence of managerial overconfidence on the overall quality of environmental information disclosure. More importantly, we empirically identify the behavioral mechanisms underpinning this relationship, thereby providing deeper theoretical and practical insights into how overconfidence shapes disclosure behavior.

Secondly, we use content analysis to improve the accuracy of environmental disclosure measurement. Previous studies have predominantly relied on ranking-based measures of corporate environmental information disclosure [21] or binary indicators [20] to reflect whether firms disclose environmental information, particularly in the Chinese context. However, these proxies fail to accurately capture firms’ actual environmental information disclosure practices [1], raising concerns about the robustness and generalizability of prior research findings. To address these limitations, we draw on the recent work of Nguyen et al. (2021) [22] and employ a content analysis technique to assess environmental information disclosure in China. This method is particularly well-suited for analyzing environmental information disclosure in such contexts, as it captures both the depth and breadth of firms’ environmental disclosures, providing a more nuanced and comprehensive understanding.

Thirdly, the findings carry important implications for policymakers and corporate stakeholders. To the best of our knowledge, this study is the first to empirically explore the mechanisms through which managerial overconfidence influences corporate information disclosure, with a particular focus on environmental disclosure. By uncovering the behavioral pathways behind this relationship, our findings provide important insights into how cognitive biases shape disclosure practices and offer a foundation for more targeted policy and governance interventions. Moreover, our analysis takes a comprehensive perspective on external oversight by examining the moderating roles of government regulation, media scrutiny, and institutional investor monitoring. The results suggest that these regulatory forces can effectively curb the adverse impact of managerial overconfidence, highlighting the necessity of a multi-faceted external governance framework to promote greater transparency and accountability in corporate disclosure.

The remainder of the paper is organized as follows: Section 2 presents a literature review; Section 3 develops the hypothesis; Section 4 describes the data and research design; Section 5 reports the empirical results; Section 6 reports a heterogeneity analysis; and finally, Section 7 concludes.

2. Literature Review

2.1. Managerial Overconfidence

According to prospect theory and bounded rationality, individual decision-making under uncertainty is influenced not only by value judgments of potential gains and losses [23], but also by cognitive biases arising from limited rationality [24]. Overconfidence, as a pervasive cognitive bias, comprises two key components: dispositional optimism and miscalibration [25]. Dispositional optimism refers to an unrealistic tendency to expect favorable outcomes under uncertainty. It is closely related to the “better-than-average” effect [26], whereby individuals overestimate their relative abilities, and to the “illusion of control,” where they believe they exert greater influence over uncertain events than they actually do [27]. Miscalibration, on the other hand, reflects individuals’ tendency to underestimate uncertainty when forecasting uncertain outcomes. In essence, overconfidence is characterized by an inflated sense of one’s own abilities (including control over events) and an underestimation of risk [8,9]. As a result, overconfident individuals tend to overestimate the probability of favorable outcomes and underestimate the likelihood of unfavorable consequences arising from their actions [28].

Previous research has predominantly focused on the economic consequences of managerial overconfidence, thoroughly investigating its effects across multiple dimensions of corporate behavior. Compared to their rational counterparts, overconfident managers exhibit excessive confidence in future cash flows and are more prone to engage in mergers and acquisitions [11]. Since overestimating their ability to generate returns, they tend to pay higher prices for target firms, resulting in greater acquisition premiums [8], which results in inflated goodwill and subsequent impairment charges [29]. Consequently, these behaviors ultimately contribute to poorer merger and acquisition performance [11]. Due to their excessive confidence in their abilities and the prospects of their companies, overconfident managers tend to avoid equity financing when seeking external funding [10,30]. They may also be psychologically less constrained by external regulations, thereby weakening regulators’ efforts to curb earnings management by corporate executives [31]. Additionally, overconfidence alters managers’ estimates of the costs and benefits of tax avoidance, leading them to expect higher net returns and consequently engage in more aggressive tax avoidance [32]. Nevertheless, when firms require external knowledge to drive innovation, overconfident CEOs are more likely to pursue alliances and establish collaborative relationships with other firms [33]. Their inflated belief in the success of innovation efforts and diminished perception of risks also contribute to improved firm innovation performance [12,34]. Collectively, these findings suggest that managerial overconfidence can exert both detrimental and beneficial effects on firm outcomes.

Managerial overconfidence also exerts a profound influence on firms’ information disclosure practices and corporate social responsibility (CSR) activities. Due to the presence of dispositional optimism and miscalibration [25], overconfident managers are inclined to proactively issue forecasts; however, their overly optimistic bias toward their own abilities and the firm’s future performance often leads to upward-biased forecasts, resulting in less accurate predictions [25,35]. Possessing an irrational belief in their own judgment and abilities, these managers are also more susceptible to engaging in financial misreporting [36]. Furthermore, their overestimation of their own abilities and strong sense of self-sufficiency lead them to underestimate the company’s reliance on stakeholder resources and support, resulting in lower levels of engagement in socially responsible activities [14,15]. Compounding these issues, overconfident CEOs are more likely to exhibit socially irresponsible behaviors, which further undermine the firm’s reputation and erode stakeholder trust [13,15]. Collectively, these findings suggest that managerial overconfidence may undermine the quality and reliability of corporate disclosures, while potentially posing challenges to corporate social responsibility and stakeholder relationships.

These research findings suggest that managerial overconfidence influences corporate financial outcomes and information disclosure behavior. However, except for a few efforts (e.g., [19,20]), previous studies have not adequately examined the impact of managerial overconfidence on non-financial information disclosure, particularly environmental disclosure. Moreover, to the best of our knowledge, no existing research has empirically tested the underlying mechanisms through which overconfidence affects such disclosure practices.

2.2. Environmental Information Disclosure

Managers wield considerable influence in shaping environmental practices. The upper echelons theory [37] suggests that executives’ attributes shape their perceptions and decision-making processes. Recent research has increasingly focused on how managers’ unique characteristics, including both observable traits and social and psychological attributes, affect environmental information disclosure and sustainable performance.

Regarding observable traits, these have been shown to notably influence executives’ approaches to environmental practices. Younger CEOs often prioritize wealth maximization, which can detract from sustainable and environmental efforts [38]. In contrast, firms led by newly appointed CEOs and those with MBAs are typically more responsive to carbon disclosure initiatives [6]. However, as CEOs age, their engagement in corporate social responsibility tends to decrease [39]. Additionally, managerial attributes such as financial expertise, research backgrounds, overseas education and experience, and academic experience have been positively linked to environmental disclosure [38,40]. Conversely, CEOs with legal education or military experience tend to have low levels of responsiveness to environmental performance disclosure requests [6,7].

Recent studies have begun exploring how psychological traits influence CSR participation. For instance, narcissistic CEOs tend to have enhanced CSR levels [41], whereas overconfident CEOs tend to have reduced CSR levels due to underestimating company risks [42]. Lee (2021) [20], in a study of South Korean listed companies, found that overconfident CEOs are more likely to prioritize environmental issues and voluntarily disclose greenhouse gas (GHG) emission information. He et al. (2021) [19], using data from listed companies in China’s low-carbon pilot provinces and cities, found that managerial overconfidence negatively impacts the quality of carbon information disclosure. In summary, while a substantial amount of research focuses on the impact of executives’ social and physical attributes, psychological-level investigations remain limited in scope and lack a comprehensive examination of overall corporate environmental information disclosure practices. Our study extends this area of research by examining the impact of executive overconfidence on environmental information disclosure.

3. Hypothesis Development

According to Simon’s (1955) [24] theory of bounded rationality, managers operate under cognitive and informational constraints, often resorting to heuristics rather than fully rational analysis. Overconfidence is one such heuristics-based bias, characterized by an overestimation of one’s own abilities and future outcomes (dispositional optimism), and an underestimation of external risks and uncertainties (miscalibration). Building on this, prospect theory [23] highlights that individuals’ risk preferences are influenced by how situations are framed. Specifically, individuals tend to be risk-averse when facing potential gains and risk-seeking when facing potential losses. Taken together, overconfidence, as a cognitive bias rooted in bounded rationality, may interact with framing effects to shape managerial decisions regarding environmental information disclosure. These interactions can lead to theoretically ambiguous outcomes.

On the one hand, overconfidence may weaken managers’ motivation to disclose environmental information, as they tend to overestimate their own capabilities and downplay stakeholder pressure. Based on prospect theory, individuals tend to make decisions based on potential gains and losses rather than final outcomes. Typically, managers are more inclined to disclose environmental information, as such disclosure serves both as a form of insurance or risk mitigation against environmental risks and as a communication tool with stakeholders, while nondisclosure or minimal disclosure may lead to greater potential losses [4]. However, overconfident managers are less likely to perceive the potential losses associated with such disclosure. Specifically, overconfident managers tend to downplay the impact of adverse environmental events [9] and overestimate their ability to manage uncertainty [43]. As a result, they may view environmental disclosure as unnecessary for risk mitigation, thereby weakening its role in integrated risk management [42]. Furthermore, overconfident executives often assume that internal resources are sufficient for achieving firm goals and thus undervalue stakeholder input [9,11]. This self-sufficiency reduces their responsiveness to stakeholder demands for transparency. For example, Tang et al. (2015) [15] found that overconfident CEOs are less responsive to stakeholder concerns, resulting in lower engagement in corporate social responsibility initiatives.

On the other hand, managerial overconfidence is also likely to increase firms’ environmental information disclosure. This effect can be primarily attributed to overconfident executives’ perception of greater benefits and lower risks associated with such disclosure. Environmental information disclosure helps mitigate potential environmental and legal risks [17,18] and enhances corporate reputation [44], thus creating strategic value. Nevertheless, it can also bring additional costs and risks. When disclosure is viewed through a gain frame, managers may exhibit risk-averse behavior that limits their willingness to disclose, consistent with prospect theory’s notion that individuals are generally risk-averse in the domain of gains. In contrast, overconfident managers tend to overestimate returns and inflate perceived benefits [10], which makes them more willing to disclose environmental information. Their cognitive miscalibration also causes them to underestimate disclosure-related risks [16], such as potential spillovers, thereby reducing perceived costs and further promoting transparency. Based on this, we propose the following null hypothesis:

Hypothesis 1: 

Managerial overconfidence has no effect on environmental information disclosure.

4. Methodology

4.1. Sample Selection

This study uses the sample of Chinese listed firms on the Shenzhen and Shanghai stock exchanges from 2008 to 2022. We begin our analysis in 2008, as this is when the Chinese stock exchanges mandated listed firms to issue CSR reports. We obtained environmental information disclosure, corporate governance, and financial variables from CSMAR. Environmental information disclosure data is also obtained from the Chinese Corporate Social Responsibility (CCSR) database. We matched the sample from these two datasets and used 32,191 firm-year observations after excluding observations with missing data. We also trim all continuous variables at the 1% level to mitigate the impact of outliers.

4.2. Measure of Managerial Overconfidence

Various methods have been employed to measure executive overconfidence. Prior studies have utilized indicators such as relative executive salaries [8,45], ownership of company stock options [10,11], discrepancies between manager-forecasted and actual earnings [46,47], overinvestment behaviors [36,48], and media coverage [8,11,35,49], among others.

Following Hayward and Hambrick (1997) [8], we use top executives’ relative salary as the primary proxy. Using relative salary as a proxy for overconfidence is theoretically grounded in both the self-importance and power perspectives. First, from the self-importance viewpoint, some CEOs exhibit systematic overestimation of their own abilities, which may stem either from verifiable past achievements or, more predominantly, from deeply ingrained and relatively stable personality traits of self-importance. The substantial gap between a CEO’s compensation and that of other executives serves as the most salient external manifestation of this self-importance [50]. Hayward and Hambrick (1997) [8] argue that the higher the CEO’s relative pay, the stronger their sense of self-importance, which in turn increases the likelihood of overconfident behavior. Furthermore, psychological research has demonstrated that individuals commonly exhibit an “illusion of control” [27]; accordingly, the more prominent the CEO’s relative position within the firm, the stronger this illusion tends to be, thereby elevating their propensity toward overconfidence. Second, from the power perspective, a CEO’s relative compensation reflects their dominance and influence within the firm. Brown and Sarma (2007) [49] find that higher relative pay is typically associated with greater power status. Fast et al. (2012) [51] confirm that experiencing power leads to overconfident decision-making. In other words, power not only grants CEOs the ability to shape corporate decisions but also enhances their confidence in their own judgments, thereby fostering the emergence of overconfidence.

Due to data constraints in Chinese companies, detailed salary information for top management is often unavailable for much of our sample period. Instead, we follow the approach of Huang et al. (2011) [47], using the ratio of the sum of the salaries of the top three managers to the sum of all managers’ salaries as an indicator of overconfidence among top executives. A higher ratio suggests greater overconfidence among top management. Considering the Chinese business environment, this alternative metric provides a reasonable proxy, as in typical Chinese companies, the top two or three managers (e.g., general manager and party chief) are responsible for making critical day-to-day operational decisions, functioning similarly to a Western-style CEO. Other studies, such as He et al. (2019) [52], He et al. (2021) [19], and Zhang and Song (2025) [45], also adopt the same measurement approach. To ensure the robustness of the results, we additionally conduct regressions using alternative proxies derived from earnings forecasts, overinvestment, and shareholdings.

4.3. Measure of Corporate Environmental Information Disclosure

To measure environmental information disclosure more precisely, we adopt content analysis, building on previous studies [22,53]. All necessary data for the coding process is drawn from the China Stock Market and Accounting Research (CSMAR) database and further verified using the Chinese Corporate Social Responsibility (CCSR) database, along with firms’ CSR and annual reports. Subsequently, we follow Nguyen et al. (2021) [22] to construct eight dimensions for measuring environmental information disclosure: (i) report clarity, (ii) management, (iii) liabilities, (iv) costs, (v) investment, (vi) performance, (vii) reliability, and (viii) compliance with the regulation. These 8 dimensions include 16 indicators, each assigned a score based on the extent of qualitative and quantitative environmental disclosure. Scores are assigned to the 16 indicators according to the type of information disclosed. For example, a point of “1” is assigned when a firm discloses non-monetary information, “2” for monetary information, and “0” when no environmental information is disclosed. Variations in quantitative information result in different maximum scores for individual items, with the optimal disclosure score determined to be 24 (see

Table A1. Dimensions of environmental information disclosure.

Dimensions	Evaluation Contents	Optimal Score
Environmental report clarity	EID section is independently and professionally reported in corporate annual report	1
	EID section is independently and professionally reported in corporate social responsibility report	1
Environmental Management	Disclosure of important environmental events	1
Environmental Liabilities	The current level of pollutant emissions	2
Environmental Costs	The current level of natural resources consumption	2
Environmental Investments	The current level of different types of environmental investment	2
	The current level of research and development	2
Environmental Performance	The current level of reduction of emission	2
	The current level of energy saving	2
	The current level of meeting requirements of green objects	2
Reliability	Have passed necessary environmental certification	1
	Received honorary environmental performance award	1
	Received environmental subsidy	1
Compliance with System	Following environmental policies	1
	Following three simultaneous systems (Design, Construction, Operation)	1
	Paying for sewage fees	2
Total		24

for details). The total score divided by the optimal disclosure score serves as a metric for evaluating the extent of environmental information disclosure.

4.4. Empirical Model

We employ the following ordinary least square regression model to explore the relationship between managerial overconfidence and environmental information disclosure.

$${\mathrm{E}\mathrm{I}\mathrm{D}\text{\_}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y}}_{\mathrm{i},\mathrm{t}}={\mathsf{\alpha }}_0{+{\mathsf{\alpha }}_1\mathrm{O}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{f}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\alpha }}_{\mathrm{j}}\sum {\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{j},\mathrm{t}}+\mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{ }+\mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{y}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(1)

Whereas OverConf is managerial overconfidence and EID_Intensity stands for corporate environmental information disclosure. Following prior studies, we controlled for various variables that may influence the relationship between managerial overconfidence and environmental information disclosure. First, in line with prior studies [12,15], we systematically controlled for CEO-level variables, including CEO duality (IsDuality), CEO gender (Gender), CEO age (CEOage), and CEO Degree (Degree). Second, drawing from the research of Nguyen et al. (2021) [22], we incorporated controls for board-level variables, specifically board size (Bsize) and board independence (Indep). Third, at the firm level, we considered variables such as firm size (Size), ownership structure (SOE), firm leverage (LEV), return on assets (ROA), ownership percentage of the largest shareholder (First), intangible assets (Intang), audit firm (Big4), and firm competition (HHI). To ensure the robustness of our analysis, we also included controls for industry and year effects in the regression model [7,53]. Standard errors are clustered by industry. Table A1 and

Table A2. Variable definitions.

Variable	Definition
EID_Intensity	See Table A1 for details.
Overconf_salary	The sum of the top three executives’ compensation divided by the sum of all executives’ compensation.
Overconf_med	Equals 1 if a firm’s Overconf_salary exceeds the industry-year median, and 0 otherwise.
Overconf_forecast	Equals 1 if the number of earnings over forecast instances exceeds those under forecast instances during the sample period.
Overconf_invest	Equals 1 if a firm’s investment intensity is in the top 20% within the same industry-year.
Overconf_stock	Equals 1 if the net stock purchase ratio ranks in the top 20% of the sample in a given year.
Size	The natural logarithm of total assets.
SOE	Indicator variable equals to 1 if firm’s ownership belongs to state, and 0 otherwise.
LEV	Ratio of total liabilities to total assets.
ROA	Ratio of net profits to total assets.
First	Percentage of the ownership of largest shareholder.
Bsize	Total number of directors on board.
Indep	Ratio of the number of independent directors to the total number of board members.
Intang	Ratio of intangible assets to total assets.
Big4	Indicator variable equals to 1 if a firm is audited by Big 4 auditor, and 0 otherwise.
HHI	The squared sum of the ratio of operating revenue to industry operating revenue for each company in the industry.
IsDuality	Indicator variable equals to 1 if a firm indicates a combined CEO and chairman position, and 0 otherwise.
Gender	Indicator variable equals to 1 if the CEO of the company is a man,0 otherwise.
CEOage	The age of CEO in years.
Degree	1 = Secondary school and below, 2 = Junior college, 3 = Bachelor’s degree, 4 = Master’s degree, 5 = Ph.D. degree, 6 = Other (e.g., honorary doctorate, correspondence), 7 = MBA/EMBA.

provide detailed definitions for each variable.

5. Empirical Results

5.1. Descriptive Statistics

Table 1. This table delineates the descriptive statistics. Panel A provides summary statistics for the full sample over the study period, while Panel B presents yearly descriptive statistics of EID_Intensity, enabling temporal comparisons across years. Our sample includes 32,191 valid firm-year observations originating from Chinese A-share listed companies from the Shenzhen and Shanghai stock exchanges from 2008 to 2022. To address outliers, we winsorized each continuous variable at the 1% level in both tails. Environmental information disclosure data is obtained from the CCSR and CSMAR databases, while other data are sourced from CSMAR.

Panel A: Descriptive Statistics for the Full Sample
Variable	N	Mean	SD	Min	Median	Max
EID Intensity	32,191	0.164	0.186	0	0.083	0.708
Overconf_salary	32,191	0.604	0.175	0.279	0.583	1
Overconf_med	32,191	0.494	0.500	0	0	1
Overconf_forecast	16,733	0.313	0.464	0	0	1
Overconf_invest	32,191	0.193	0.395	0	0	1
Overconf_stock	30,803	0.022	0.148	0	0	1
Size	32,191	21.85	1.310	19.30	21.69	26.05
SOE	32,191	0.308	0.462	0	0	1
LEV	32,191	0.406	0.210	0.048	0.393	0.940
ROA	32,191	0.039	0.066	−0.280	0.041	0.207
First	32,191	34.24	14.66	9.090	32.02	74.82
Bsize	32,191	2.236	0.177	1.792	2.303	2.773
Indep	32,191	37.49	5.250	33.33	33.33	57.14
Intang	32,191	0.046	0.050	0	0.033	0.322
Big4	32,191	0.0560	0.229	0	0	1
HHI	32,191	0.124	0.130	0.020	0.079	0.743
IsDuality	32,191	0.309	0.462	0	0	1
Gender	32,191	0.929	0.257	0	1	1
CEOage	32,191	49.86	6.691	33	50	66
Degree	32,191	3.645	1.264	1	4	7
Panel B: Descriptive Statistics by Year
Year	N	Mean	SD	Min	Median	Max
2008	1052	0.085	0.136	0.000	0.000	0.667
2009	1092	0.082	0.135	0.000	0.000	0.708
2010	1380	0.090	0.151	0.000	0.000	0.708
2011	1599	0.093	0.156	0.000	0.000	0.708
2012	1697	0.121	0.167	0.000	0.042	0.708
2013	1682	0.123	0.165	0.000	0.042	0.708
2014	1727	0.120	0.164	0.000	0.042	0.708
2015	1917	0.118	0.161	0.000	0.042	0.708
2016	2101	0.128	0.173	0.000	0.042	0.708
2017	2448	0.146	0.174	0.000	0.083	0.708
2018	2522	0.178	0.185	0.000	0.125	0.708
2019	2688	0.189	0.190	0.000	0.125	0.708
2020	2990	0.177	0.179	0.000	0.125	0.708
2021	3464	0.239	0.198	0.000	0.208	0.708
2022	3832	0.269	0.201	0.000	0.250	0.708
Total	32191	0.164	0.186	0.000	0.083	0.708

presents the descriptive statistics. In Panel A, we provide summary statistics for the full sample over the study period. The average value of environmental information disclosure is 0.164, with a range from 0 to 0.708. This range is consistent with the findings of previous studies [22]. On average, 30.8% of the firms’ ownership is state-owned, and the largest shareholder holds 34.24% of the shares. The mean values for firm size, ROA, leverage, and HHI are 21.85, 0.039, 0.406, and 0.124, respectively. The average board size is 8.36 members, with a minimum of 5 and a maximum of 15 directors, consistent with the Chinese corporate governance code, which mandates a board size of 5 to 15 directors. The average value of board independence is 37.49%, which meets the requirements of Chinese corporate governance, which requires at least one-third of the board to be independent. On average, 5.6 percent of firms get their financial audit from one of the big four accounting firms. In our sample, 92.9 percent of CEOs are male on average, and 30.9 percent of CEOs also serve as board chairman. The average age of a CEO is 49.86 years old and ranges from a minimum of 33 years to a maximum of 66 years old. Panel B provides yearly descriptive statistics for the variable EID_Intensity, facilitating temporal comparisons over the sample period. We observed that this variable increases over time, with a median value of zero in the early years of the sample. This indicates a growing response from the business community to environmental issues, likely driven by market forces or policy initiatives.

5.2. Main Results

Table 2. Baseline regression results. This table presents the impact of managerial overconfidence on environmental information disclosure. EID_Intensity measures the extent of corporate environmental disclosure, with higher values indicating more comprehensive and detailed disclosure. Overconf_salary serves as a proxy for managerial overconfidence based on executive compensation data, where higher values reflect a greater degree of overconfidence. The variable EID_Intensity is obtained from the CCSR database and manually coded, whereas Overconf_salary is constructed based on data from the CSMAR database through calculation. Definitions of variables are provided in Table A1 and Table A2. T-statistics are reported in parentheses, with significance levels indicated by *** for 1%, ** for 5%, and * for 10% (two-tailed).

Variables	EID_Intensity	EID_Intensity
Overconf_salary	−0.179 ***	−0.063 ***
	(−18.01)	(−4.78)
Size		0.053 ***
		(11.77)
SOE		0.019 ***
		(3.71)
LEV		−0.019 *
		(−1.78)
ROA		0.195 ***
		(11.90)
First		0.000 *
		(2.00)
Bsize		0.027 ***
		(3.55)
Indep		−0.000
		(−1.37)
Intang		0.044
		(0.79)
Big4		0.064 ***
		(5.47)
HHI		−0.005
		(−0.23)
IsDuality		−0.006 **
		(−2.12)
Gender		0.007
		(1.16)
CEOage		0.000
		(1.53)
Degree		−0.002 **
		(−2.20)
Constant	0.272 ***	−1.043 ***
	(45.27)	(−9.83)
Observations	32,191	32,191
Year FE	YES	YES
Industry FE	YES	YES
Adjusted R-squared	0.212	0.363

presents the key findings of our analysis. In column (1), we estimate a baseline regression including only the managerial overconfidence variable, while controlling for industry and year fixed effects. The results reveal a statistically significant negative relationship between Overconf_salary and environmental information disclosure (coefficient = −0.179, p < 0.01). In column (2), after adding all control variables, the coefficient of Overconf_salary remains negative and statistically significant at the 1% level (−0.063). Economically, a one standard deviation increase in Overconf_salary is associated with an approximate 0.011 point decrease in the environmental disclosure index (calculated as −0.063 × 0.175). Given that the mean value of the environmental disclosure index is 0.164, this represents about a 6.7% reduction relative to the average level of environmental disclosure, suggesting a substantial and meaningful negative impact. It is also important to note that the minimum unit of our EID_Intensity variable is 1/24. Accordingly, even a seemingly small change of 0.042—given the estimated coefficient of −0.063—corresponds to a substantive shift in corporate disclosure practices on one or two specific dimensions. For instance, such a change may reflect a movement from “no disclosure” to “partial disclosure,” or from “qualitative disclosure” to “quantitative disclosure” for a particular item. Overall, these results indicate that managerial overconfidence significantly impairs the quality of corporate environmental information disclosure.

5.3. Robustness Checks

5.3.1. Alternative Measurements and Firm Fixed Effect

To assess the robustness of our findings, we adopt several alternative proxies for managerial overconfidence. First, we refine the original measure Overconf_salary by constructing a new variable, Overconf_med, which captures relative overconfidence based on industry-year medians. Specifically, Overconf_med is coded as 1 if a firm’s Overconf_salary exceeds the median level within the same industry and year, and 0 otherwise.

Second, we employ Overconf_forecast, a proxy based on earnings forecast bias, to capture overconfidence in managerial expectations. Prior literature suggests that overconfident managers tend to overestimate future performance [35], rendering the gap between forecasted and actual earnings a valid indicator. Following Huang et al. (2011) [47], we classify an executive as overconfident if the number of over-forecasting instances (i.e., forecasted earnings exceeding actual earnings) surpasses the number of under-forecasting instances during the sample period.

Third, building on prior studies that link investment behavior to overconfidence [36,48], we construct an investment-based measure, Overconf_invest. Following Campbell et al. (2011) [54], we define Overconf_invest as a dummy variable equal to 1 if a firm’s investment intensity—measured by capital expenditures divided by beginning-of-year net property, plant, and equipment (PPE)—falls within the top quintile of its industry-year distribution, and 0 otherwise. This indicator reflects the tendency of overconfident managers to overinvest relative to industry peers.

Finally, we include Overconf_stock, a proxy based on executives’ trading behavior, as proposed by Ahmed and Duellman (2013) [48]. We compute the net change in shareholdings as the difference between year-end and beginning-of-year holdings, then normalize this change by the beginning-of-year level. Firms with a positive net purchase are ranked by this normalized change, and Overconf_stock is coded as 1 for firms in the top quintile (i.e., the highest 20%) and 0 otherwise. This measure captures executives’ excessive confidence in their firm’s future performance as implied by aggressive stock accumulation.

In addition, we conduct a robustness check by replacing the industry fixed effects in the baseline regression with firm fixed effects, while retaining the year fixed effects. As shown in Panel A of

Table 3. Robust tests. Panel A presents robustness checks using alternative independent variables and firm fixed effects. Columns (1) to (4) report regression results based on different proxies for managerial overconfidence, including Overconf_med, Overconf_forecast, Overconf_invest, and Overconf_stock. The variable EID_Intensity is obtained from the CCSR database and manually coded, while all measures of managerial overconfidence are derived from the CSMAR database through computational processing. Definitions of these variables are provided in Table A2. Column (5) replaces industry fixed effects with firm fixed effects to control for time-invariant firm-specific characteristics. Panel B shows results from the instrumental variable regression’s first and second stages. The instrument (IV) is constructed as the average managerial overconfidence level within each region–industry year group. Panel C illustrates the results of propensity score matching, where “High_OC” denotes the binary variable indicating whether the executive overconfidence level falls within the top 25% based on industry year segmentation. T-statistics are reported in parentheses, with significance levels indicated by *** for 1%, and * for 10% (two-tailed).

Panel A: Alternative Measurements and Firm Fixed Effect
	EID_Intensity
Variables	(1)	(2)	(3)	(4)	(5)
Overconf_med	−0.018 ***
	(−5.09)
Overconf_forecast		−0.014 ***
		(−2.92)
Overconf_invest			−0.011 ***
			(−3.87)
Overconf_stock				−0.007 *
				(−1.83)
Overconf_salary					−0.028 ***
					(−3.03)
Controls	YES	YES	YES	YES	YES
Constant	−1.093 ***	−1.147 ***	−1.138 ***	−1.143 ***	−0.263 ***
	(−11.12)	(−10.44)	(−12.63)	(−11.76)	(−3.97)
Observations	32,191	16,733	32,191	30,803	32,191
Year FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
Firm FE					YES
Adjusted R-squared	0.362	0.376	0.360	0.357	0.680
Panel B: Instrumental Variable Approach
Variables		1st Stage—Overconf_salary		2nd Stage—EID_Intensity on Overconf_salary
Overconf_salary				−0.056 *
				(−1.74)
IV		0.875 ***
		(47.45)
Controls		YES		YES
Constant		0.888 ***
		(20.89)
F-statistic		2251.87
Observations		32,191		32,191
Year FE		YES		YES
Industry FE		YES		YES
Adjusted R-squared		0.262		0.219
Panel C: Using the Propensity Score Matching Method
Variables		Probit	Full		Matched
		(1)	(2)		(3)
Overconf_salary			−0.056 *
			(−1.74)
Controls		YES	YES
Constant		4.471 ***	−1.043 ***		−0.867 ***
		(8.17)	(−9.83)		(−8.85)
Observations		32,189	32,191		15,874
Year FE		YES	YES
Industry FE		YES	YES
Pseudo R2		0.0480
Adjusted R-squared		0.262	0.219

, the results remain consistent: regardless of whether alternative proxies for managerial overconfidence or alternative fixed effects are employed, overconfident executives are still significantly associated with lower levels of environmental information disclosure. These findings reinforce the robustness of our main results across different model specifications and variable constructions.

5.3.2. Instrumental Variable Approach

To further mitigate endogeneity concerns, we construct an instrumental variable (IV) based on the average level of managerial overconfidence within region–industry year groups. Firms are classified according to their industry and provincial location, and we compute the average overconfidence score for each group year. This IV captures exogenous variation in overconfidence arising from the broader regional–industry context, rather than firm-specific characteristics. The identification assumption is that local managerial behavior is shaped by shared institutional, cultural, or market forces, while the instrument remains exogenous to firm-level environmental disclosure after controlling for firm fundamentals and fixed effects.

We first estimate a reduced-form regression of managerial overconfidence on the constructed IV, followed by a two-stage least squares (2SLS) regression to assess its effect on environmental disclosure. The results, presented in Panel B of Table 3, show that the instrumented overconfidence variable remains significantly negatively associated with environmental information disclosure, consistent with our baseline results. These findings strengthen the causal interpretation of our main results and mitigate concerns regarding potential reverse causality or omitted variable bias.

5.3.3. Propensity Score Matching

To address potential selection bias between firms with high and low managerial overconfidence, we employ a propensity score matching (PSM) method. Firms in the top quartile of overconfidence within the same industry year are classified as the treatment group, while those in the lower three quartiles constitute the control group. We estimate a probit model incorporating firm-level control variables to predict the probability of treatment assignment. Based on the estimated propensity scores, treated firms are matched to control firms using one-to-one nearest-neighbor matching without replacement and a caliper of 0.01. Subsequently, the main regression model is re-estimated on the matched sample to verify the robustness of our findings.

The results, presented in Panel C of Table 3, show that the coefficient of managerial overconfidence remains significantly negative in the matched sample, consistent with our full-sample estimates. This suggests that the observed relationship is robust to potential selection bias and is unlikely to be driven by systematic differences in observable firm characteristics.

5.4. Mechanism Analysis

5.4.1. Overestimation of Risk Control Ability

Overconfident managers often overestimate their capabilities in dealing with uncertainties [43], leading them to undervalue the role of environmental information disclosure in corporate risk management. If the relationship between managerial overconfidence and environmental information disclosure is driven by an overestimation of risk management capabilities, this effect is likely to be more pronounced among executives with an environmental background. Due to their relevant expertise or experience, such executives may exhibit greater confidence in their ability to identify and manage environmental risks internally, thereby reducing their reliance on external disclosure as a risk mitigation tool.

To test this “ability overestimation” mechanism, we construct an indicator for executives with an environmental background. Specifically, we identify executives whose résumés include keywords such as “environment”, “environmental protection”, “new energy”, “clean energy”, “ecology”, “low carbon”, “sustainability”, “energy saving”, or “green”. We count the number of such executives in each firm and classify firms into high and low groups based on the sample median. The results, reported in Panel A of

Table 4. Mechanism Analysis. Panel A presents the overestimation of ability mechanism, which is measured by the number of executives with an environmental background. The environmental background variable is manually constructed based on executives’ resumes. Specifically, an executive is considered to have an environmental background if their resume contains keywords such as “environment,” “environmental protection,” “new energy,” “clean energy,” “ecology,” “low carbon,” “sustainability,” “energy saving,” or “green.” Based on the median number of executives with an environmental background within the executive team, firms are divided into a High group (above the median) and a Low group (below the median). Panel B shows the underestimation of stakeholder importance mechanism, which is captured by the degree of market competition. The Herfindahl–Hirschman Index (HHI) is obtained from the CSMAR database. Specifically, samples with a Herfindahl index below the median are classified as the High group, while those above the median are classified as the Low group. The row labeled “Differences” reports the coefficients and statistical significance of the differences between the High and Low groups for each pair of columns. The significance levels are tested by performing 1000 bootstrap resamples, reporting empirical p-values. T-statistics are presented in parentheses. Statistical significance is denoted by *** for 1%, ** for 5%, and * for 10% levels (two-tailed). These settings apply to all panels in this table.

Panel A: Overestimation of Risk Control Ability
Variables	Environmental Background
	High	Low
Overconf_salary	−0.064 ***	−0.058 ***
	(−3.57)	(−4.38)
Controls	YES	YES
Constant	−1.109 ***	−0.998 ***
	(−7.88)	(−11.01)
Differences	0.006 **
Observations	9784	22,407
Year FE	YES	YES
Industry FE	YES	YES
Adjusted R-squared	0.377	0.356
Panel B: Underestimation of Stakeholder Importance
Variables	Market Competition Intensity
	High	Low
Overconf_salary	−0.068 **	−0.061 ***
	(−2.87)	(−5.22)
Controls	YES	YES
Constant	−1.116 ***	−0.929 ***
	(−13.42)	(−4.81)
Differences	0.007 *
Observations	15,995	16,196
Year FE	YES	YES
Industry FE	YES	YES
Adjusted R-squared	0.370	0.366

, show that the negative effect of managerial overconfidence on environmental disclosure is significantly stronger in firms with more environmentally experienced executives. This finding supports the view that overconfident managers with relevant domain knowledge are more prone to overestimating their own ability, thereby reinforcing the adverse impact of overconfidence on disclosure behavior.

5.4.2. Underestimation of Stakeholder Importance

Overconfident managers tend to overestimate the sufficiency of internal firm resources while underestimating the importance of external inputs and feedback, leading to reduced attention on the needs of external stakeholders [9,11]. In contrast, firms operating in more competitive markets face more complex challenges and have relatively limited access to resources [55], which compels overconfident managers to rely more heavily on stakeholder support. However, if overconfident managers systematically underestimate the importance of stakeholders, then firms with overconfident executives are likely to disclose less environmental information in response to intensified competition—compared to similarly competitive firms without overconfident executives. This amplified effect highlights a behavioral bias stemming from their misperception of stakeholder relevance.

Building on this reasoning, we use industry-level competition as a proxy for stakeholder pressure and measure industry concentration using the Herfindahl–Hirschman Index (HHI). Firms are classified into high- and low-competition groups based on the sample median of HHI. Panel B of Table 4 presents the subgroup analysis results, showing that the negative effect of managerial overconfidence on environmental disclosure is more pronounced in highly competitive industries, thereby providing evidence consistent with the mechanism of managerial underestimation of stakeholder importance.

6. Further Analysis

Building on the evidence that managerial overconfidence suppresses environmental information disclosure, this study further investigates the role of external governance mechanisms—including regulatory oversight from the government, media scrutiny, and monitoring by institutional investors—in moderating this adverse effect. Subsample analyses are conducted to examine whether and how these external pressures influence the relationship between overconfidence and disclosure behavior.

6.1. The Impact of Environmental Regulation

Governmental environmental regulation, as a formal institutional pressure, signals the government’s commitment to environmental issues and conveys clear policy expectations to firms, making it a pivotal driver of corporate environmental information disclosure [56]. Empirical evidence shows that stringent regulatory enforcement substantially elevates disclosure levels [57]. Overconfident managers—who typically downplay external risks and regulatory requirements—are inclined to restrict disclosure. However, in regions with high regulatory intensity, the strong focus and enforcement by local authorities create a robust external constraint, compelling even overconfident executives to moderate their behavior and enhance disclosure practices. Conversely, in areas where regulatory oversight is weak, the lack of effective external supervision may allow overconfident managers to arbitrarily curtail the scope of their disclosures.

To assess the moderating role of environmental regulation, we partition the sample into high- and low-regulatory-pressure groups. Regulatory pressure is proxied by the share of environment-related keywords (e.g., “environmental protection”) in local government work reports, measured as the ratio of these keywords’ word count to the report’s total word count [58]. Government work reports, as legally binding policy documents summarizing past achievements and outlining future priorities, directly reflect policymakers’ attention allocation and resource commitment at a given point in time. We then re-estimate Model (1) separately for each subgroup. The results in Panel A of

Table 5. Heterogeneity analysis. This table investigates the influence of external regulatory forces, encompassing environmental regulatory pressure from the government, reputational pressure exerted by media attention, and monitoring pressure from institutional investors. In Panel A, environmental regulatory pressure is proxied by the frequency of environment-related keywords (e.g., “environmental protection”) appearing in local government work reports, measured as a proportion of the total word count. These data were sourced from official local government work reports and manually coded to construct the variable. A higher frequency reflects greater governmental emphasis on environmental issues, thereby indicating stronger environmental regulatory pressure. In Panel B, media scrutiny is measured using the Janis–Fadner Coefficient of Imbalance, which captures the directional tone of media coverage. The media coverage data are obtained from the Chinese Research Data Services (CNRDS) database. Higher values reflect more favorable reporting and thus indicate weaker media scrutiny and pressure. In Panel C, institutional investor monitoring is proxied by institutional ownership, with data obtained from the CSMAR database. The sample is divided into High and Low groups based on the median values of each indicator. For environmental regulatory pressure and institutional ownership, observations with values above the median are classified as High; for media scrutiny, classification is reversed due to its inverse relationship with oversight pressure. Across all panels, the High group represents greater external oversight pressure. The “Differences” row reports the differences in coefficients and their statistical significance between columns (1) and (2), and columns (3) and (4). T-statistics, reported in parentheses, are based on standard errors clustered at the industry level. ***, ** indicate significance at the 1%, 5% levels, respectively.

Panel A: The Effect of Environmental Regulation
Variables	Environmental Regulatory Pressure
	High	Low
Overconf_salary	−0.058 ***	−0.067 ***
	(−3.04)	(−7.44)
Controls	YES	YES
Constant	−1.011 ***	−1.077 ***
	(−8.32)	(−11.00)
Differences	−0.009 ***
Observations	16,433	15,758
Year FE	YES	YES
Industry FE	YES	YES
Adjusted R-squared	0.359	0.370
Panel B: The Effect of Media Scrutiny
Variables	Media Scrutiny and Pressure
	High	Low
Overconf_salary	−0.050 **	−0.073 ***
	(−2.11)	(−10.51)
Controls	YES	YES
	−0.973 ***	−1.102 ***
Constant	(−6.65)	(−12.70)
Differences	−0.023 ***
Observations	15,546	15,499
Year FE	YES	YES
Industry FE	YES	YES
Adjusted R-squared	0.315	0.385
Panel C: The Effect of Institutional Investors
Variables	Institutional Ownership
	High	Low
Overconf_salary	−0.055 ***	−0.078 ***
	(−3.99)	(−4.36)
Controls	YES	YES
Constant	−1.192 ***	−0.680 ***
	(−15.38)	(−6.05)
Differences	−0.023 ***
Observations	16,070	16,071
Year FE	YES	YES
Industry FE	YES	YES
Adjusted R-squared	0.392	0.300

show that the negative effect of managerial overconfidence on disclosure is materially weaker in the high-pressure group. The difference in coefficients between the high- and low-pressure subsamples is statistically significant at the 1% level, confirming that stronger environmental regulatory pressure mitigates the adverse impact of overconfident managers on their disclosure behavior.

6.2. The Impact of Media Scrutiny

Prior studies have shown that media scrutiny can constrain managerial misconduct and promote better governance practices, ultimately enhancing the quality of information disclosure [59]. Specifically, given concerns about career prospects and compensation, managers are highly sensitive to their personal reputations. Negative media coverage can exert reputational pressure on executives, which may in turn prompt overconfident managers to restrain their behavior. We therefore expect that increased, especially negative, media attention imposes greater oversight, which helps curb the opportunistic impact of executive overconfidence.

We measure media supervision using the Janis–Fadner coefficient (the J-F coefficient is calculated as follows: when e > c, J-F = (e² − e × c)/t²; when e < c, J-F = (e × c − e²)/t²; and when e = c, J-F = 0. Here, e denotes the number of positive media reports, c the number of negative media reports, and t the total number of reports (i.e., e + c)), ranging from −1 to 1, where higher values indicate more favorable coverage and weaker pressure. As shown in Panel B of Table 5, the negative effect of overconfidence on environmental disclosure is weaker in firms facing stronger media pressure (i.e., lower index values), highlighting the moderating role of media scrutiny. Moreover, bootstrap analysis with 1000 replications confirms that the differences between these groups are statistically significant at the 1% level.

6.3. The Impact of Institutional Investor Monitoring

Institutional investors serve as a key external governance force that may effectively constrain managerial overconfidence. Prior studies suggest that higher institutional ownership is often associated with improved monitoring of management. For example, Hartzell and Starks (2003) [60] found that firms with greater institutional ownership exhibit stronger pay-for-performance sensitivity, which helps curb managerial opportunism. Similarly, Chung et al. (2002) [61] show that institutional investors help reduce earnings management by limiting discretionary accounting choices, thereby enhancing the quality of financial reporting.

Given their access to extensive resources and professional expertise, institutional investors are well-positioned to assess corporate strategies and risk profiles, making them more capable of restraining overconfident managerial behavior. Conversely, in firms with lower institutional ownership, weaker external oversight may allow overconfident managers to disclose environmental information in a more passive or selective manner.

Specifically, institutional investor governance is measured by the shareholding ratio of institutional investors. Firms with institutional ownership below the sample median are classified as the low governance group, while those above the median are assigned to the high governance group. The empirical results presented in Panel C of Table 5 show that the negative relationship between managerial overconfidence and environmental disclosure is less pronounced in firms with weaker external governance, consistent with our expectations. Furthermore, we conducted a bootstrap test with 1000 replications, and the results indicate that the differences between the two groups are statistically significant at the 1% level. These findings provide strong support for the hypothesis that institutional investor monitoring can mitigate the adverse effects of managerial overconfidence.

6.4. The Analysis of Economic Consequences

The empirical evidence regarding the value relevance of environmental information disclosure remains inconclusive. On the one hand, studies such as Clarkson et al. (2013) [62] and Plumlee et al. (2015) [63] report a positive relationship between the quality of voluntary environmental disclosure and firm value. Plumlee et al. (2015) [63] further demonstrate that this positive association operates through two key channels: improved cash flows and a lower cost of equity capital. On the other hand, several studies have failed to find significant evidence supporting a value-enhancing effect of environmental disclosure [64,65,66]. To further investigate this issue in the context of China, we construct Model (2) to test whether environmental information disclosure is indeed value-relevant—that is, whether it has a significant positive effect on firm value—and to examine the role managerial overconfidence plays in this relationship.

$${\mathrm{T}\mathrm{a}\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{Q}}_{\mathrm{i},\mathrm{t}+1}={\mathsf{\beta }}_0{+{\mathsf{\beta }}_1\mathrm{E}\mathrm{I}\mathrm{D}\text{\_}\mathrm{I}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_{\mathrm{j}}\sum {\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}\mathrm{s}}_{\mathrm{j},\mathrm{t}}+\mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{y}+\mathrm{Y}\mathrm{e}\mathrm{a}\mathrm{r}+{\mathsf{\mu }}_{\mathrm{i},\mathrm{t}}\mathrm{ }$$

(2)

Column (1) of

Table 6. Economic consequences. This table presents the value relevance of environmental information disclosure. TabinQ is used as a measure of firm value, calculated as market value divided by total assets, with data obtained from the CSMAR database. The definitions of other variables are provided in Table A2. T-statistics are reported in parentheses, with significance levels denoted by *** for 1%, ** for 5%, and * for 10% (two-tailed).

Variables	TabinQ	TabinQ
EID_Intensity	0.155 *	1.304 ***
	(1.83)	(4.20)
Overconf_salary		0.643 ***
		(6.30)
EID_Intensity × Overconf_salary		−1.973 ***
		(−3.89)
Size	−0.354 ***	−0.348 ***
	(−24.87)	(−25.50)
SOE	0.092	0.110 **
	(1.66)	(2.10)
LEV	−0.057	−0.058
	(−0.60)	(−0.59)
ROA	2.282 ***	2.326 ***
	(3.30)	(3.35)
First	−0.004 ***	−0.004 ***
	(−4.75)	(−5.24)
Bsize	−0.104	−0.059
	(−0.96)	(−0.51)
Indep	0.007 *	0.007 *
	(1.81)	(1.76)
Intang	0.580	0.575
	(1.31)	(1.29)
Big4	0.282 ***	0.257 ***
	(9.40)	(8.27)
HHI	−0.133	−0.133
	(−0.47)	(−0.45)
IsDuality	−0.021	−0.022
	(−1.33)	(−1.22)
Gender	−0.021	−0.013
	(−0.59)	(−0.35)
CEOage	0.001	0.002
	(0.52)	(0.73)
Degree	0.037 ***	0.038 ***
	(3.42)	(3.40)
Constant	9.607 ***	8.962 ***
	(24.98)	(22.31)
Observations	26,078	26,078
Year FE	YES	YES
Industry FE	YES	YES
Adjusted R-squared	0.248	0.249

reports the regression results based on Model (2). The coefficient of EID_Intensity is significantly positive at the 10% level, indicating that environmental information disclosure substantially enhances firm value. This finding supports the value relevance of environmental disclosure. Specifically, a one standard deviation increase in environmental information disclosure corresponds to an increase in Tobin’s Q of approximately 0.029, representing a 1.38% improvement relative to the sample mean of 2.093. Such a sizable positive effect suggests that firms with stronger environmental disclosure practices tend to achieve significantly higher market valuations, reflecting investors’ favorable assessment of environmental transparency. Given that the previous analysis shows managerial overconfidence reduces environmental disclosure, we further include the interaction term EID_Intensity × Overconf_salary in Model (2), with the results presented in Column (2). The coefficient of the interaction term is significantly negative at the 1% level, suggesting that although environmental disclosure contributes positively to firm value, this beneficial effect is weakened when managerial overconfidence is high. This result extends the previous findings by uncovering a moderating role of overconfidence in shaping the value implications of disclosure.

7. Conclusions

Environmental information disclosure is a fundamental component of corporate environmental governance and is essential for achieving sustainable development. Using panel data on Chinese A-share listed firms, this study investigates the relationship between executive overconfidence and environmental information disclosure. The empirical results show that executive overconfidence significantly harms environmental information disclosure. Further analysis reveals two primary mechanisms underlying this effect: the overestimation of risk control capabilities and the underestimation of stakeholder importance. In addition, the findings suggest that external governance pressures from government regulation, media scrutiny, and institutional investor monitoring can effectively moderate the negative impact of executive overconfidence. Moreover, we further investigate the value relevance of environmental information disclosure and explore how managerial overconfidence moderates this relationship.

This study contributes to the literature by employing a content analysis-based measure to capture the overall quality of environmental information disclosure and by uncovering the behavioral mechanisms through which managerial overconfidence affects such disclosure. Building on these theoretical insights, our findings offer several practical implications for strengthening environmental governance. Policymakers can move beyond reinforcing mandatory disclosure requirements by introducing mechanisms such as third-party ESG report verification, mandatory participation of environmental consultants in key decisions, and the establishment of cognitively diverse advisory committees. Boards of directors can play a complementary role by forming ESG-focused committees, appointing independent directors with environmental expertise, and incorporating psychological assessments into executive recruitment and evaluation. Media outlets may enhance transparency by developing platforms that aggregate and benchmark ESG disclosures, thereby exerting reputational pressure. Institutional and activist investors can monitor linguistic cues—such as excessive certainty in executive communications—to detect overconfidence and respond through voting, engagement, and shareholder proposals. Together, these coordinated strategies can foster a multi-actor governance framework that promotes responsible disclosure and sustainable corporate conduct.

This study has several limitations that also offer opportunities for future research. First, the sample is limited to Chinese A-share listed firms, which may constrain the generalizability of the findings to other cultural or institutional contexts. For example, in more individualistic cultures like the United States, overconfident executives might be less likely to disclose environmental information due to stronger self-enhancement tendencies, whereas collectivist cultures may mitigate this effect through greater emphasis on social responsibility. Institutional factors, such as the strength of environmental regulation enforcement and capital market efficiency, may also shape disclosure behaviors in ways not captured here.

Second, the environmental information disclosure score employed in this study is based on a finite-scale content coding approach rather than a continuous measure. While this method offers transparency and replicability, it may restrict the sensitivity of the analysis. Future research could enhance this measurement by incorporating more nuanced dimensions of disclosure or employing machine learning techniques to extract deeper textual insights from sustainability reports.

Third, in line with the extant literature, we measure overconfidence using observable indicators. Although these proxies are widely accepted and have been commonly used in prior studies, they may not fully capture the psychological traits or nuanced expressions of executive overconfidence. Future research may explore more sophisticated identification strategies. For example, neuroeconomic experimental approaches could provide more direct and precise measures of managerial overconfidence, helping to address existing limitations in measurement and enhance our understanding of its implications.