Can Climate Risk Disclosure Attract Analyst Coverage? A Study Based on the Dual Perspective of Information Supply and Demand

Abstract

In the context of the intensifying global climate change and its associated risks, the interaction between corporate climate risk disclosure and analyst forecasting behavior has become a pivotal scholarly focus in sustainability research. This study uses a sample of 20,978 firm-year observations from non-financial Chinese A-share listed companies over the period 2007–2021 to examine the impact of corporate climate risk disclosure on analyst coverage, applying ordinary least squares (OLS) regression. The results reveal a positive relationship between corporate climate risk disclosure and analyst coverage. This positive effect is more prominent in firms with lower annual report readability, a higher proportion of independent institutional investors, and in contexts involving team analysts or analysts from large brokerage firms. Mechanism analysis reveals two pathways for increased analyst coverage: increasing institutional investors’ demand for information and reducing analysts’ reliance on on-site research to uncover private information. Further research reveals that severe and chronic risk disclosures attract more analyst coverage than transition risk disclosures. Additionally, climate risk disclosure can significantly reduce analyst forecast dispersion and long-term forecast bias. Overall, this study holds important implications for improving corporate climate risk disclosure practices and enhancing analysts’ role as information intermediaries.

1. Introduction

Climate change and its associated risks have emerged as one of the most pressing global challenges, with their widespread impacts becoming increasingly apparent. The “Climate Change 2023” report from the Intergovernmental Panel on Climate Change (IPCC) highlights that the continuous increase in global greenhouse gas emissions is driving global warming and triggering a series of severe consequences, including extreme weather events, economic losses, and biodiversity degradation [1]. These changes not only pose substantial risks to the natural environment but also have profound implications for the global economy and societal structures, particularly regarding business operations and investment decisions [2]. As climate risks gain growing prominence in global decision making, market participants are increasingly integrating climate change considerations into their decision making processes, emphasizing the crucial importance of climate risk management [3]. Consequently, the potential of climate risk disclosure to attract analysts’ coverage has become a key issue that warrants further investigation.

As an essential information intermediary in the capital market, securities analysts regularly use their professional knowledge and skills to research and analyze relevant information about listed companies and issue forecast reports. Analysts’ judgment offers a vital reference basis for investors’ decision making [4]. Several studies have explored the factors influencing analyst coverage, suggesting that company size, institutional ownership proportion, and media attention significantly affect analyst coverage [5,6,7]. Studies examining the relationship between corporate disclosure content, quality, and analyst coverage show that more detailed disclosure content [8] and higher disclosure quality attract greater analyst coverage [9]. Additionally, some studies focus on the impact of risk-based disclosure in annual reports on analyst coverage. For example, Derouiche et al. [10] find a positive relationship, indicating that firms with higher levels of risk disclosure are more likely to attract analyst coverage. Similarly, He and Lu [11] report that qualitative future supply chain risk disclosures draw more analyst coverage. Clearly, analysts play a crucial role in making investment analysis judgments on the fundamentals of the listed companies they cover and providing research reports to investors.

In an era of heightened attention to climate risk, climate risk significantly impacts listed companies’ fundamentals [12], ultimately influencing their investment value. Moody’s ESG Report 2021 shows that almost all industries face severe climate risk. Climate risk information has become an essential basis for analysts to judge the company’s fundamentals and long-term value. Hence, analysts pay more attention to companies that disclose climate risk more thoroughly. Indeed, research shows that analysts have gradually paid attention to listed companies’ climate risk in their tracking forecasts, utilizing this information to enhance future forecast returns [13]. In addition, the growing investor demand for climate risk information reinforces the attractiveness of corporate climate risk disclosure to capital market information intermediaries. Therefore, understanding how corporate climate risk disclosure behavior affects analysts’ role as information intermediaries is essential. Specifically, whether corporate climate risk disclosure can attract analyst coverage reflects the market’s perception of climate risks and serves as a basis for assessing whether voluntary disclosure of such risks can enhance market efficiency. Extant research does underscore the necessity of integrating climate risks into investment decision making processes. However, substantial challenges persist in practice, which are primarily attributable to portfolio companies’ inadequate disclosure of climate-risk-related information [2]. To meet the growing demand for climate risk information from investors, in June 2017, the Task Force on Climate-Related Financial Disclosure released its report outlining a framework and recommendations for climate-related disclosures to support companies in identifying and assessing the potential climate risks they face. In June 2023, the International Sustainability Standards Board issued Sustainability Disclosure Standard No. 2—Climate-Related Disclosures (IFRS S2). The standard recommends that companies disclose climate-related risks and opportunities in four areas—governance, strategy, risk management, and metrics and objectives—to help investors more comprehensively assess their climate risk exposure and their ability to address it. In March 2024, the U.S. Securities and Exchange Commission (SEC) announced the approval of climate-related disclosure rules for U.S. public companies. These rules, for the first time, will require companies to disclose climate-related information and risks in their annual reports and registration statements. In April of the same year, China’s three major stock exchanges in Shanghai, Shenzhen, and Beijing issued the “Guidelines on Sustainability Reporting for Listed Companies (Trial)”. These guidelines require some companies to disclose their governance, strategy, impacts, risks, and opportunities management related to addressing climate change.

While climate risk disclosure is crucial for investor decision making and appropriately pricing climate risks, many companies have yet to disclose climate risk information voluntarily. Even when they do, the content is often fragmented and incomplete, and it lacks a unified disclosure standard [14]. Users of financial statements, especially non-professional investors, often need help with understanding and assessing the specific impacts of climate risks on businesses covered in financial statements [2]. As essential information intermediaries, analysts play an important role in translating such climate risk information into forecasts and influencing investor judgment [15]. Studies have focused on the relationship between climate risk disclosure and analyst forecast quality. However, few have addressed the relationship between climate risk disclosure and analyst coverage. Given that analyst coverage is the logical starting point for analysts’ behavioral decisions, this study analyzes data on Chinese A-share non-financial listed companies from 2007 to 2021 to examine the impact of corporate climate risk disclosure on analyst coverage and the underlying mechanisms. We also explore the effect of the heterogeneity of roles in the context of different firm and analyst characteristics and whether corporate climate risk disclosure can reduce analysts’ forecast bias and forecast dispersion. Specifically, this study addresses the following research questions. How does the level of climate risk disclosure affect analyst coverage in China’s capital market? What role do investor information demand and analyst information supply play in this context?

This study makes two contributions. First, it explores the information-enhancing role of corporate climate risk disclosure. Analysts have begun incorporating climate-related factors into earnings forecast considerations in Western capital markets. However, this is still in its initial stages, and it shows significant variations across firms and regions [16]. For instance, we do not know whether analysts in China’s capital markets similarly focus on corporate climate risks. Addressing this gap, this study shows that climate risk disclosure enhances analyst coverage through two mechanisms: increasing institutional investors’ demand for information and reducing analysts’ reliance on field research to mine private information. This study verifies the mechanism of climate risk disclosure in analyst coverage of the dual research perspectives of investors’ information demand and analysts’ information supply. It explores the heterogeneity of this effect considering firms’ and analysts’ characteristics. This shed more light on the critical role of climate risk disclosure in improving the information environment of the capital market. Second, this study enhances the literature on the factors influencing analysts’ coverage behavior. While research has explored the effects of annual report risk and supply chain risk disclosure on analyst coverage [10,11], the relationship between corporate climate risk disclosure and analyst coverage remains underexplored. This study contributes to extant findings by focusing on corporate climate risk disclosure and examining the factors that affect analysts’ coverage behavior from both the information supply and demand perspectives. Our results also provide insights for China’s listed companies, helping them better understand the value of climate risk disclosure. Additionally, this research is significant in optimizing analysts’ role as information intermediaries and improving pricing efficiency in capital markets.

2. Literature Review and Research Hypotheses

2.1. Literature Review

2.1.1. The Economic Consequences of Climate Risk Disclosure

Climate risk carries significant implications for both investors and portfolio companies. It can physically damage a firm’s assets, which may diminish asset values and economic returns [17], thereby impacting the pricing of stocks and bonds [18,19]. Conversely, climate risk also presents investment opportunities for portfolio companies and investors [2]. By integrating climate risks into their investment decision making processes, investors can mitigate their overall portfolio risk [20]. Consequently, corporate climate risk disclosure becomes essential for investors as it enables them to comprehend companies’ specific measures and strategies to tackle climate change [21]. Actively engaging in climate risk disclosure is a vital indicator of a company’s commitment to environmental and social responsibility, and thus can help enhance its corporate image [22]. This practice can improve corporate performance [23], mitigate information asymmetry, and lower the likelihood of future stock price declines [24]. Furthermore, firms that disclose significant climate risks often enjoy reduced equity costs [25]. Climate risk disclosure can also foster corporate green innovation [26,27]. Some scholars have also argued that corporate climate risk disclosure driven by Confucian culture has both green innovation and value creation effects based on an informal system perspective [28]. However, some studies also reveal the possible negative impacts of climate risk disclosure, such as a higher cost of equity capital [3], higher audit fees [1], and increased corporate tax avoidance [29]. In addition, some studies have focused on the impact of climate risk disclosure on analysts’ forecast quality. For instance, some argue that the corporate disclosure of climate change risks and opportunities reduces analysts’ forecast error and forecast dispersion, especially in the categories of governance, risk management, and strategy [13]. Recent empirical evidence further supports this mechanism. Liu and Han [30] demonstrate that climate risk disclosures alleviate information asymmetry between managers and analysts by providing firm-specific climate information, thereby systematically reducing forecast errors and dispersion. This emphasizes the positive role of climate risk disclosure in enhancing market information transparency. However, some studies take a different view, suggesting that corporate climate risk disclosures do not necessarily reduce analyst forecast bias and dispersion. Rather, the reduction only happens when investors perceive corporate climate risk disclosures to be financially material at the industry level, which reduces analyst forecast bias and dispersion by mitigating information asymmetry [15].

2.1.2. Factors Influencing Analyst Coverage

Analysts play a vital role as information intermediaries in capital markets. Their primary task is collecting information about listed companies, conducting in-depth information processing and analysis, and providing research reports to investors to assist them in investment decision making. Bhushan proposed a theoretical model of the supply and demand for analyst services. The author suggested that both investor demand for such services and analysts’ willingness to provide them simultaneously determine the extent of analyst coverage. Moreover, firm characteristics influence both analyst services demand and supply, thereby shaping the level of analyst coverage [5]. Building on this theoretical foundation, a substantial empirical literature has investigated the factors affecting analyst coverage, revealing that it is closely related to firm characteristics. For instance, firms with larger sizes [5], a higher proportion of institutional investor ownership [6], and greater media attention [7] attract significantly more analysts. Furthermore, the corporate information environment, particularly information disclosure practices, significantly influences analyst behavior. Analysts tend to cover companies with timely and detailed information disclosure [8], high disclosure quality [9], and greater transparency [31]. However, some studies reveal that companies with lower annual report readability attract more analyst coverage. This is primarily because the higher costs of processing and interpreting information for such companies induce greater investor demand for analyst services [32]. In addition, empirical studies have examined the impact of specific corporate accounting information disclosure practices on analyst coverage. Specifically, disclosures of core competencies [4], annual report risk factors [10], and supply chain risks [11] significantly enhance analyst coverage. Nevertheless, few studies have directly explored the impact of corporate climate risk disclosures on analyst coverage.

In summary, research on the economic consequences of climate risk disclosure primarily focuses on the corporate level, showing that climate risk disclosure has dual effects on firms. While existing studies have explored the impact of climate risk disclosure on analysts’ forecast accuracy, such as reducing forecast errors and dispersion [13,30], few have examined its influence on analysts’ coverage decisions. Much of the literature centers on the effects of climate risk disclosure on corporate financial performance, with limited investigation into its direct impact on analyst behavior, particularly in terms of analyst coverage decisions. To the best of our knowledge, this study represents the first exploration of this issue within the context of China, thereby addressing a critical gap in the literature. We address this gap by analyzing the effects of climate risk disclosure on analyst coverage in China’s capital market from the dual perspectives of investor information demand and analyst information supply.

2.2. Research Hypothesis

Analysts’ optimal service level is contingent upon investors’ demand for information and analysts’ willingness to supply it [5]. Therefore, this study proposes research hypotheses based on two perspectives: investors’ information demand and analysts’ information supply.

First, we argue that corporate climate risk disclosure can enhance investors’ information demands for analysts’ predictions. Recent asset pricing models have emphasized the significance of climate risk as a long-term risk factor and its importance in stock returns [18]. Studies reveal that enterprises exposed to high climate risks often achieve higher stock returns because such risks heighten market uncertainty as well as the idiosyncratic uncertainty of individual stocks. Under such circumstances, investors require higher returns to compensate for the risks they undertake [2]. The current Chinese stock market exhibits a notable and stable positive climate risk premium, with enterprises with high climate risks typically having higher stock earnings [12]. Xiao et al. [33] also found that Chinese capital market investors do seek compensation for holding stocks of high-emission companies and the potential asset-stranding risks they might face. Therefore, companies with higher carbon risks often have higher stock returns, and high-yield stocks tend to attract more investor attention [34,35]. The European Central Bank, considering climate change, also prefers high-carbon enterprises in its bond purchase portfolio [36]. Additionally, according to the signal transmission theory, disclosing the climate risks enterprises face can provide more information related to enterprise valuation, reduce information asymmetry among stakeholders, improve the stock market’s pricing efficiency, and increase companies’ market value [12,37]. The signal transmission theory suggests that the act of disclosing important information, like climate risk, serves as a signal to the market that helps reduce information asymmetry, leading to more efficient pricing and informed investor decisions. Many investors, especially long-term, large-scale, and environmental, social, and governance (ESG)-focused ones, are willing to take action to address climate risks. They believe that participating in risk management is more effective in resolving climate risk issues than divestment [2]. Thus, enterprises that proactively identify and disclose climate risks are more likely to gain attention and favor from market investors by conforming to the trends of the times and demonstrating strategic considerations for long-term sustainable development [23]. However, climate risks are extensive and heterogeneous, with their sources including physical and transition risks. Physical risks directly affect enterprises’ production and operation environment, while transition risks indirectly influence enterprises’ business models, compliance costs, and market demands. Different industries and enterprises exhibit significant differences when facing climate risks. Consequently, investors need help to effectively capture and deeply analyze the climate risks disclosed by enterprises. Investors cannot quantify the potential financial impacts of climate risks or identify related investment opportunities. As essential information intermediaries in the capital market, analysts have obvious advantages in mining and interpreting such information [38], and they can transform the climate risks disclosed by enterprises into predictions and influence investors’ judgments [15]. According to information asymmetry theory, analysts play a crucial role in reducing information asymmetry between firms and investors by interpreting and transmitting complex information, such as climate risks [30]. In this case, analysts help investors understand the risks better and make more informed decisions. In summary, this study contends that investors will pay attention to enterprises with more climate risk disclosures and require analysts, as information intermediaries, to provide predictions for such enterprises.

From analysts’ information supply perspective, the greater the potential benefits of forecasts, the higher the level of analyst coverage of the company. Meanwhile, the lower the information acquisition and analysis cost, the stronger the analysts’ willingness to cover the company [39]. On the one hand, analysts have client-induced incentives to follow corporate climate risk [16]. Analysts may make coverage decisions based on self-interest in the hope that their reports will lead to greater trading volume and thus commission gains or being named a star analyst [40,41]. As investors’ attention to climate change increases, the demand for climate-risk-related information and analysis is also growing. This is especially true for institutional investors, who increasingly incorporate enterprise climate risk into their investment decision making considerations [2]. However, most investors and capital market participants believe that the climate risk information disclosed by enterprises lacks unified standards and is not precise enough, increasing investors’ costs to obtain and interpret the information [14]. Analysts’ interpretation of the information becomes particularly important for companies where the information is difficult to understand or has high processing costs [38]. By providing interpretive reports for companies with high climate risk disclosure, analysts can reduce investors’ information processing costs and help them make more informed investment decisions in the face of climate risk uncertainty. This demand increases the value of analysts’ services, enhancing their professional reputation and client loyalty. Moreover, as capital market intermediaries with professional capabilities and high sensitivity to enterprise information, analysts should be aware of the importance of climate risk information in financial reports for evaluating enterprises’ fundamentals and long-term value. Therefore, this article argues that analysts are likely to cater to investors’ demands and follow companies that disclose climate risks, enhancing analysts’ reputation and commission income.

On the other hand, the disclosure of climate risks by enterprises can significantly increase the supply of public information for enterprises, reduce analysts’ information acquisition costs for analysts, and increase their willingness to supply. Analysts use both public and private information in their predictions, with the former being the primary source of information for analysts’ predictions [42,43]. Considering analysts’ prediction costs, public information offers a more significant cost advantage than private information [31]. Firms with lower coverage costs tend to appeal more to analysts [39]. Studies do indicate that companies providing more information can lower analysts’ data acquisition costs, encouraging them to cover these firms [11]. Climate risk disclosure enables stakeholders, such as investors and analysts, to understand an enterprise’s risk exposure and climate change response measures by increasing public information. According to information asymmetry theory, information imbalances exist between market participants, and firms that disclose more information (such as climate risks) help reduce these imbalances [30]. Therefore, this study argues that the greater the climate risks disclosed by listed companies, the more publicly available information analysts can utilize for their forecasts, thereby reducing analysts’ information acquisition costs and attracting more analyst coverage. Accordingly, this study proposes the following hypotheses:

H1. 

Corporate climate risk disclosures increase analyst coverage.

H2. 

Corporate climate risk disclosures increase analyst coverage by enhancing investors’ demand for forecasts and reducing analysts’ information acquisition costs.

3. Research Design

3.1. Sample Selection and Data Source

This study selects China’s A-share listed companies from 2007 to 2021 as the research sample. The sample period begins in 2007, aligning with the formal implementation of China’s new accounting standards system, which came into effect on 1 January 2007 [30]. The data on climate risk disclosure come from annual reports of listed companies obtained through text analysis and machine learning techniques. We obtained the readability data of annual reports from the text data platform of WinGo Finance (WinGo). In contrast, we acquired other data from the China Stock Market and Accounting Research Database (CSMAR) and the Chinese Research Data Services Platform (CNRDS) database. Notably, analysts’ forecasting behavior needs to be observed after annual report disclosures. Hence, following Lehavy et al. [32], we focus on the analysts’ earnings forecasts after the disclosure of the previous year’s report until the disclosure of the following year’s report. Next, following the literature, the sample is screened as follows: exclude listed companies in the ST and *ST categories and financial and insurance industries; remove samples with any missing values; and winsorize all continuous variables at the top and bottom one percentile of their distributions. Finally, we obtained a pooled cross-sectional dataset that includes 20,978 firm-year observations.

3.2. Definition of Variables

3.2.1. Analyst Coverage (Follow)

Following Li et al. [4], we use the number of analyst teams (Follow) following the company to measure analyst coverage, as defined in Equation (1). Specifically, NA_i,t represents the number of analysts covering firm i between the release of its annual reports for years t and t + 1. The analyst coverage data focus on the period between the release of the company’s annual reports in years t and t + 1. Note that analyst teams typically produce reports collaboratively. Hence, we count the analysts from the same team only once.

Follow_i,t = ln(NA_i,t + 1)

(1)

3.2.2. Climate Risk Disclosure (ClimateR)

Following Du et al. [3], we employ textual analysis and machine learning techniques to measure climate risk disclosure at the firm level. The main steps involve identifying seed words related to “climate risk” and utilizing machine learning to train the annual report corpus. Subsequently, we apply the Word2Vec model, using the Continuous Bag of Words (CBOW) method, to expand the seed word set. Ultimately, the final climate-related word set includes 76 seed words and 98 augmented words (for further details, refer to Appendix A). The climate risk disclosure (ClimateR) metric is calculated as the ratio of the total frequency of climate-risk-related words in the annual reports to the overall word count. A higher ratio indicates a greater level of climate risk disclosure [44], as shown in Equation (2).

ClimateR_i,t = Climate risk words/Words in annual reports × 100

(2)

For further analysis, we subdivide climate risk disclosure into SeriousR, ChronicR, and TransitionR, where SeriousR and ChronicR represent severe and chronic physical risk disclosures, respectively. The indicators for SeriousR, ChronicR, and TransitionR are computed similarly to ClimateR [3]. In addition, we conducted an annual trend analysis of ClimateR, revealing a steady upward trajectory in corporate climate risk disclosure (see Figure A1 in Appendix B). This trend aligns closely with the findings presented by Liu and Han [30].

3.2.3. Control Variables

Following the literature [4,10], this study controls for firm size (Size), financial leverage (Lev), growth (Growth), profitability (Roa), earning volatility (Evo), book-to-market ratio (BM), nature of equity (Soe), firm age (Age), shareholding concentration (Top1), and audit quality (Big4) as the firm characteristic variables that may affect analyst coverage. According to the data selection methodology employed in this study, the annual reports of listed companies are published in the subsequent year. Consequently, the analyst-related data are also from the following year, meaning that the dependent variable in this study is derived from the subsequent year, while the independent and control variables are from the current year. This temporal separation helps mitigate, to some extent, the potential endogeneity issue [4].

Table 1. Variable definitions.

Symbol	Variable	Definition
Dependent Variable
Follow	Analyst coverage	Natural logarithm of one plus the number of analyst teams covering the company from its annual report for year t to its annual report for year t + 1
Independent Variables
ClimateR	Climate risk disclosure	The frequency of the term “climate risk” as a percentage of the total word frequency in the annual report
SeriousR	Serious risk disclosure	The frequency of the term “severe risk” as a percentage of the total word frequency in the annual report
ChronicR	Chronic risk disclosure	The frequency of the term “chronic risk” as a percentage of the total word frequency in the annual report
TransitionR	Transition risk disclosure	The frequency of the term “transaction risk” as a percentage of the total word frequency in the annual report
Control Variables
Size	Firm size	Natural logarithm of a company’s total assets
Lev	Financial leverage	A company’s total liabilities divided by its total assets
Growth	Growth	A company’s revenue growth rate
Roa	Profitability	A company’s net profit after tax divided by its average total assets
Evo	Earning volatility	The standard deviation of a company’s return on equity over the past three years
BM	Book-to-market ratio	Shareholders’ equity divided by the company’s market value
Soe	Nature of equity	Equals 1 if the equity holder is a state-owned company and 0 otherwise
Age	Years of company listing	Natural logarithm of a company’s listing age plus 1
Top1	Top shareholder’s shareholding concentration	Largest shareholder’s shareholding ratio
Big4	Audit quality	Equals 1 if the company is audited by the “Big Four” and 0 otherwise

provides the definitions of the above variables.

3.3. Model Setting

We follow the existing literature [4,10] and build the following baseline OLS linear regression model to test the hypothesis (H1):

Follow_i,t = β₀ + β₁ ClimateR_i,t + Control_i,t + Industry + Year + ε

(3)

Here, Follow_i,t is an explanatory variable representing an indicator of analyst coverage from the release of the firm’s annual report in year t to its yearly report in year t + 1. ClimateR_i,t is an explanatory variable representing an indicator of the extent of a firm’s climate risk disclosure in year t. Industry and Year represent industry- and year-fixed effects, respectively.

4. Empirical Results

4.1. Descriptive Statistical Analysis and Difference Tests

Panel A in

Table 2. Descriptive statistics and difference testing.

Panel A: Descriptive Statistics
Variable	N	Mean	Std. Dev.	Min	P25	Med	P75	Max
Follow	20,978	1.976	0.908	0.693	1.099	1.946	2.708	3.850
ClimateR	20,978	0.174	0.150	0.014	0.074	0.130	0.220	0.816
SeriousR	20,978	0.002	0.005	0.000	0.000	0.000	0.000	0.030
ChronicR	20,978	0.002	0.005	0.000	0.000	0.000	0.003	0.031
TransitionR	20,978	0.169	0.147	0.013	0.072	0.126	0.213	0.803
Size	20,978	22.467	1.304	20.081	21.528	22.291	23.233	26.366
Lev	20,978	0.447	0.197	0.064	0.294	0.447	0.597	0.865
Growth	20,978	0.228	0.462	−0.488	0.015	0.142	0.313	3.146
Roa	20,978	0.053	0.055	−0.129	0.022	0.046	0.079	0.234
Evo	20,978	0.053	0.074	0.002	0.014	0.029	0.058	0.477
BM	20,978	0.316	0.148	0.060	0.207	0.292	0.402	0.751
Soe	20,978	0.431	0.495	0.000	0.000	0.000	1.000	1.000
Age	20,978	2.309	0.616	1.078	1.804	2.390	2.834	3.321
Top1	20,978	35.885	15.028	9.810	23.880	34.110	46.290	74.980
Big4	20,978	0.076	0.265	0.000	0.000	0.000	0.000	1.000
Panel B: Difference Testing
Variable	High climate risk disclosure group			Low climate risk disclosure group			Mean difference
	N		Mean	N		Mean
Follow	10,489		1.9987	10,489		1.9531	0.0456 ***

Note: The t-test is used to test the differences in means; *** represents significance at the 1% level.

shows the descriptive statistics. The minimum, maximum, mean, and median values of Follow are 0.693, 3.850, 1.976, and 1.946, respectively; this is similar to the values reported by Li et al. [4]. The standard deviation is 0.908, indicating significant differences in analyst coverage between firms. Next, the mean value of ClimateR is 0.174, while the median is 0.130. This implies that most firms have lower than average climate risk disclosure, in line with Du et al.’s [3] findings. The statistical results of other variables are consistent with existing studies.

Panel B presents the results of the difference-in-difference test for analyst coverage between the high and low climate risk disclosure groups, which are divided by the median corporate climate risk disclosure. Clearly, the number of analyst teams in the high climate risk disclosure group is significantly higher than that in the low climate risk disclosure group. This indicates that analysts follow firms with high climate risk disclosure more than firms in the low climate risk disclosure group, thereby providing preliminary support to our hypothesis (H1).

4.2. Benchmark Regression Results

Table 3. Climate risk disclosure and analyst coverage.

Variable	(1)	(2)
	Follow	Follow
ClimateR	0.2810 ***	0.1142 ***
	(5.98)	(2.88)
Size		0.4385 ***
		(71.45)
Lev		−1.1387 ***
		(−24.73)
Growth		0.0132
		(1.09)
Roa		4.7871 ***
		(40.34)
Evo		−0.3180 ***
		(−4.18)
BM		−1.9370 ***
		(−41.51)
Soe		−0.1141 ***
		(−8.76)
Age		−0.2388 ***
		(−22.79)
Top1		−0.0036 ***
		(−10.02)
Big4		0.0522 **
		(2.55)
Constant	1.9373 ***	−6.1824 ***
	(56.14)	(−53.72)
Industry	Yes	Yes
Year	Yes	Yes
Adj.R2	0.0245	0.3763
N	20,978	20,978

Note: The values in parentheses are t-values adjusted for robust standard errors. ** and *** denote significance at the 5% and 1% levels, respectively.

presents the regression results of model (3). The regression results in columns (1) (includes industry- and year-fixed effects) and (2) (includes control variables and industry- and year-fixed effects) show a significant positive relationship between corporate climate risk disclosure and analyst coverage; that is, the greater the corporate climate risk disclosure, the greater analyst coverage. On average, for every one standard deviation increase in corporate climate risk disclosure (0.150), the number of analysts following increases by about 0.0189 standard deviations, suggesting an economically significant effect. Thus, corporate climate risk disclosure significantly contributes to analyst coverage, thereby supporting our research hypothesis (H1).

4.3. Robustness Tests

4.3.1. Endogeneity Test

(1) Instrumental variable approach. Following Zeng et al. [45], this study selects the mean value of climate risk disclosure of firms in the same industry in the same year other than the firms themselves (ClimateR_A) as an instrumental variable and conducts a two-stage regression test. In the first regression stage, we regress ClimateR using the control variables of the model (3) together with the instrumental variable ClimateR_A. In column (1) of

Table 4. Robustness tests considering endogeneity.

Variable	Instrumental Variable Approach		PSM	Heckman’s Two-Stage Model
	First Stage	Second Stage		First Stage	Second Stage
	(1)	(2)	(3)	(4)	(5)
	ClimateR	Follow	Follow	ClimateR_D	Follow
ClimateR_A	0.5743 ***
	(11.49)
ClimateR(IV)		1.3539 ***
		(3.27)
OtherClimateR				0.5576 ***
				(3.55)
ClimateR			0.1205 **		0.1248 ***
			(2.19)		(3.12)
IMR					−0.1855 ***
					(−2.89)
Controls	Yes	Yes	Yes	Yes	Yes
Year/Industry	Yes	Yes	Yes	Yes	Yes
Adj.R2/Pseudo-R2	0.3024	0.3469	0.3680	0.1706	0.3766
N	20,970	20,970	10,462	20,952	20,952

Note: The values in parentheses are t-values adjusted for robust standard errors. ** and *** denote significance at the 5% and 1% levels, respectively.

, the regression coefficient of ClimateR_A is significantly positive. In the second stage of the regression, model (3) is re-tested with the predicted value of ClimateR, ClimateR (IV), as an explanatory variable. Column (2) of Table 4 shows that ClimateR (IV) is significantly and positively related to analyst coverage at the 1% level. Thus, the main conclusions hold after controlling for endogeneity issues.

(2) Propensity score matching method. To avoid the endogeneity problem caused by omitted variables, we refer to the process of Wang and Li [46] and adopt the propensity score matching method. First, we divide the sample into high and low climate risk disclosure groups based on the median of the explanatory variable, climate risk disclosure. Next, we utilize the control variables of the model (3) and direct economic losses (Losses) associated with climate disasters in the provinces where the firms operate as covariates. We match the data using a 1:1 nearest-neighbor matching method with a caliper range of 0.05 and obtain 10,462 valid samples. Finally, model (3) is used to re-regress using the matched samples. The results in column (3) in Table 4 show that the regression coefficient of climate risk disclosure is significantly positive at the 5% level, consistent with the main findings.

(3) Heckman’s two-stage model. Firms may exhibit selectivity when disclosing climate risks, which could lead to sample self-selection issues in the central hypothesis. We employ the Heckman two-stage method to perform the correction test. In the first stage of Heckman’s probit regression model, we define the explanatory variable as a dummy variable, ClimateR_D. This variable is determined based on whether the firm’s climate risk disclosure, ClimateR, exceeds the sample median. If it surpasses the median, ClimateR_D equals 1, indicating a high level of climate risk disclosure; otherwise, this dummy variable equals 0. Additionally, we include the proportion of firms within the same industry that exhibit higher climate risk disclosure (OtherClimateR) as an exogenous instrumental variable in the first-stage model to compute the inverse Mills ratio (IMR). We then add the IMR calculated in the first stage to the second-stage model for the correction test. Column (5) of Table 4 presents the results of Heckman’s second-stage regression. The regression coefficient of ClimateR remains significantly positive, suggesting that the finding that climate risk disclosure attracts analyst following still holds after controlling for the sample self-selection problem.

4.3.2. Other Robustness Tests

(1) Replacement of analyst coverage indicators. First, following the literature [4], we use “the number of analysts’ research report releases that follow the firm” (Report) and “the number of brokerage institutions that follow the firm” (Sinstitution) to re-measure the analyst coverage indicator. Specifically, we define Report as the natural logarithm of one plus the number of research reports published after the release of the firm’s annual report in period t until the release of its annual report in period t + 1 and Sinstitution as the natural logarithm of one plus the number of brokerage firms following the firm after the release of its annual report in period t until the release of its annual report in period t + 1. Second, to minimize the disruption of analyst coverage by other significant events between the balance sheet date and the annual report release, we refer to Wang et al. [47] and restrict the analyst coverage data to the period from the release of the annual report in year t to the balance sheet date in year t. Accordingly, we recalculate the number of analyst teams (Follow_New). The corresponding test results presented in columns (1) to (3) of

Table 5. Other robustness tests.

Variable	Replacement of Analyst Coverage Indicators			Replacement of Climate Risk Disclosure Indicators	Fixed Effects Models
	(1)	(2)	(3)	(4)	(5)
	Report	Sinstitution	Follow_New	Follow	Follow
ClimateR	0.1372 ***	0.1017 ***	0.1120 ***		0.3251 ***
	(2.74)	(2.83)	(2.86)		(3.89)
Adj_ClimateR				0.0919 **
				(2.29)
Controls	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	NO
Year	Yes	Yes	Yes	Yes	Yes
Firm	NO	NO	NO	NO	Yes
Adj.R2	0.3617	0.3809	0.3539	0.3762	0.6239
N	20,978	20,978	20,454	20,978	20,978

Note: The values in parentheses are t-values adjusted for robust standard errors. ** and *** denote significance at the 5% and 1% levels, respectively.

show that the baseline conclusion holds: corporate climate risk disclosures attract analysts to follow them.

(2) Replacement of climate risk disclosure indicators. Specific inherent industry differences in climate risk disclosure of enterprises in different industries may affect the results. Therefore, to eliminate the impact of industry-specific differences on the regression results, this study adjusts corporate climate risk disclosure by using the annual industry average. The adjusted indicator, Adj_ClimateR, is a substitute for the explanatory variable in the robustness tests. The results are shown in column (4) of Table 5. The regression coefficient of Adj_ClimateR is significantly positive at the 5% level, indicating that our findings are still valid.

(3) Fixed effects models. Our current model setup may omit firm-level influences that do not change over time. To address this issue, we add firm-individual fixed effects to model (3) and adjust for clustering at the firm level. As shown in column (5) of Table 5, the coefficient of ClimateR is significantly positive at the 1% level, consistent with the findings above.

(4) Placebo test. Other unobserved factors, such as institutional changes, like the signing of the Paris Agreement in 2015, may have influenced the relationship between corporate climate risk disclosure and analyst coverage [28]. Therefore, following Wang et al. [47], we utilize a placebo test to rule out this possibility. We randomly generate a dummy independent variable that matches the value range of the climate risk disclosure variable, ClimateR, across all samples. We incorporate this dummy variable into the original sample and perform regression analysis based on model (3). We repeat this process 1000 times. Figure 1 shows the results, where the absolute values of the t-statistics for the estimated coefficients of climate risk disclosure and analyst coverage in the vast majority of these 1000 random samples fall within 0.5, suggesting that climate risk disclosure does not significantly affect analyst coverage. The results of the placebo test validate the robustness of the study’s conclusions.

4.4. Heterogeneity Analysis

This section discusses the impact of firm- and analyst-level heterogeneity on the above relationships based on the two perspectives of investor information demand and analyst information supply.

4.4.1. Impact of Firm-Level Heterogeneity: Information Demand Perspective

Firm characteristics can significantly affect analyst coverage behavior [5]. Accordingly, we classify the overall sample according to the type of institutional investors and the readability of annual reports. The purpose is to examine the impact of corporate climate risk disclosure on analyst coverage from the perspective of investors’ information demand and uncover the underlying driving factors behind this relationship. To test the coefficient variability between different groups, this study adopts the self-sampling method (Bootstrap) to conduct a 2000 sampling analysis.

(1) Types of institutional investors. Research indicates that compared to short-term investors, long-term investors place greater emphasis on the potential effects of climate change on a firm’s future financial performance, reputation, and operational stability [2]. Following the literature [48], we consider securities investment funds, social security funds, and Qualified Foreign Institutional Investors as independent institutional investors, which uphold the investment concepts of value, long-term, and responsible investments. The remaining institutional investors are considered non-independent institutional investors. We use the median of the proportion of independent institutional investors to total institutional investors to divide the whole sample into the high and low independent institutional groups. The results in columns (1) and (2) of Panel A of

Table 6. Heterogeneity tests.

Panel A: Impact of Firm-Level Heterogeneity: Information Demand Perspective
Variable	(1)	(2)	(3)	(4)
	Low independent institutional group	High independent institutional group	Low readability of annual reports group	High readability of annual reports group
	Follow	Follow	Follow	Follow
ClimateR	0.1036 **	0.2025 ***	0.1485 ***	0.0284
	(2.00)	(3.62)	(2.88)	(0.45)
Controls	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Adj.R2	0.3274	0.3713	0.3805	0.3652
N	10,454	10,454	10,475	10,475
p-value	0.009		0.001
Panel B: Impact of Analyst-Level Heterogeneity: Information Supply Perspective
Variable	(1)	(2)	(3)	(4)
	Individual analysts	Analyst teams	Large platform analysts	Small platform analysts
	Follow	Follow	Follow	Follow
ClimateR	0.0344	0.1649 ***	0.0952 ***	0.0627 **
	(0.90)	(4.14)	(2.62)	(1.97)
Controls	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Adj.R2	0.3585	0.3554	0.3562	0.2987
N	20,978	20,978	20,661	20,661

Note: The values in parentheses are t-values adjusted for robust standard errors. ** and *** denote significance at the 5% and 1% levels, respectively.

show that the regression coefficient of climate risk disclosure is significantly positive at the 1% and 5% levels in the high and low independent institutional groups, respectively. Thus, corporate climate risk disclosure significantly impacts analyst coverage of both groups of firms. Crucially, the regression coefficients of climate risk disclosure for the two groups are significantly different at the 1% level. Thus, the effect of climate risk disclosure on analyst coverage is more significant in the high rather than the low independent institutional group. Hence, climate risk disclosure attracts more analysts to follow firms with higher long-term investor ownership.

(2) Annual report readability. Studies indicate that when the readability of corporate financial reports is low, investors exhibit greater reliance on analyst reports [31]. Because investors find it difficult to identify and assess climate-related risks and opportunities from management information disclosures when the readability of corporate annual reports is low, they rely more on information provided by third parties, such as analysts’ forecasts, to make decisions [49]. We divide the sample into lower and higher readability groups based on the annual industry median for annual readability. Annual readability is assessed using the WinGo readability index, which is constructed based on deep learning algorithms. The process is as follows. First, each word is represented as a dense, fixed-length, real-valued vector through Word Embedding, ensuring that semantically similar words are represented similarly in the vector space. Second, drawing on the optimization techniques of Hierarchical Softmax and Negative Sampling, the generation probability of each sentence is computed. Finally, the logarithmic mean of the generation probabilities of all sentences is used as the document’s readability measure. A higher readability score indicates that word pair sequences occur more frequently in the corpus, making the text easier to comprehend; conversely, a lower readability score suggests poorer readability. As shown in columns (3) and (4) of Panel A of Table 6, the regression coefficient of climate risk disclosure is significantly positive at the 1% level in the group with low readability. In contrast, the regression coefficient for climate risk disclosure is not statistically significant in the group with higher readability. Furthermore, the regression coefficients of climate risk disclosure for the two groups are significantly different at the 1% level. Thus, the effect of climate risk disclosure on analyst coverage is more significant in the group with low readability than the group with high readability of annual reports.

4.4.2. Impact of Analyst-Level Heterogeneity: Information Supply Perspective

Analyst coverage is affected by the information collection cost [39]. Therefore, from an information supply perspective, the attractiveness of corporate climate risk disclosures for analysts varies according to team characteristics. Based on this perspective, this section further explores the differential impact of climate risk disclosures in attracting different types of analyst coverage.

(1) Analyst teams. Integrating climate risk is complex and costly for analysts [16]. Analyst teams can distribute workloads through division of labor and resource sharing, leveraging their collective expertise to improve the accuracy of forecasts. They may have an advantage when addressing tasks requiring interdisciplinary analysis, such as climate risk disclosure [50]. In contrast, individual analysts, lacking the support of a team, face higher costs and more significant challenges when processing complex information. These limitations result in individual analysts paying lower attention to corporate climate risk disclosure. Therefore, this difference in information processing costs makes it more likely that complex information disclosed by firms will attract coverage from teams of analysts. In contrast, individual analysts may be under-responsive to similar information. The results in columns (1) and (2) of Panel B of Table 6 show that corporate climate risk disclosures are significantly attractive to analyst teams, with a regression coefficient of 0.1649, which is significant at the 1% level. In contrast, the regression coefficient for climate risk disclosure is not significant for individual analysts. This highlights the advantages of analyst teams in dealing with complex information, suggesting that teamwork helps to reduce information processing costs and enhances analysts’ ability to interpret complex information.

(2) Analyst platforms. Large brokerage firms maintain a leading position in ESG investing because they possess excellent resources and well-established research teams. They can more effectively follow and interpret corporate climate risks. In contrast, smaller brokerage firms, constrained by limited resources and lower client demand, pay relatively less attention to such information. This study classifies brokerage firms into large and small categories based on the annual median size of brokerages tracking a given firm. We measure brokerage size using the natural logarithm of the sum of “the number of active analysts plus 1”, which captures the differences in resource allocation and market coverage across brokerage firms. The results in columns (3) and (4) of Panel B of Table 6 show that firms’ climate risk disclosure significantly positively impacts the coverage behavior of analysts from both large and small brokerage firms. However, their response significantly differs. Analysts from large brokerage firms are more responsive to corporate climate risk disclosure, with a regression coefficient of 0.0952 significant at the 1% level, suggesting that corporate investment in climate risk disclosure can significantly improve analyst coverage. This response reflects the advantages of large brokerage firms regarding resource integration, depth of analysis, and response to complex information. Meanwhile, analysts from small brokerage firms show a relatively weak response, with a coefficient of 0.0627, which is significant at the 5% level. Thus, analysts from large brokerage firms can integrate and analyze corporate climate risks more effectively, thus showing a greater willingness to follow up.

4.5. Mechanism Analysis

According to the previous section, climate risk disclosure affects analysts’ coverage behavior by influencing investors’ information demand and analysts’ willingness to supply information. Therefore, we must test the possible information demand and supply mechanisms. Following Niu et al. [51], a four-stage mediation mechanism model is used for regression testing to validate our hypothesis (H2):

M_i,t = β₀ + β₁ ClimateR_i,t + Control_i,t + Industry + Year + ε

(4)

Follow_i,t = β₀ + β₁ M_i,t + Control_i,t + Industry + Year + ε

(5)

Follow_i,t = β₀ + β₁ M_i,t + β₂ ClimateR_i,t + Control_i,t + Industry + Year + ε

(6)

M_i,t denotes the mediating variables, which represent investors’ information demand and analysts’ information supply mechanism. The definitions of other variables are consistent with model (3).

4.5.1. Investors’ Information Demand

Climate risk disclosure is the best choice for enterprises seeking long-term strategic development. Through such disclosures, enterprises can optimize the risk management process, enhance corporate reputation and image, increase investor confidence, and attract more investors’ attention and favor. However, investors may find it difficult to effectively assess and deeply analyze the climate risks disclosed by enterprises. Therefore, investors expect analysts to be able to evaluate and quantify the impact of these risks on investors and portfolio companies based on the climate risk information companies disclose. Meanwhile, institutional investors are a company’s major shareholders, with sizeable influence and deal size compared to individual investors. Studies have found that when institutional investors hold significant positions in a particular stock, an increasing number of analysts tend to follow the company and issue research reports, motivated by considerations of trading volume and service fees, to meet the demands of institutional investors. Therefore, corporate climate risk disclosure may attract more analysts to follow a company by increasing investors’ demand for analysts’ interpretation of information. Accordingly, we further examine the mechanisms by which institutional investors’ information needs play a role in the impact of corporate climate risk disclosures on analyst coverage. Institutional investors’ information demand is represented by their shareholding ratio (Hold), with a higher shareholding ratio indicating greater information demand [52].

Table 7. Mechanism analysis: investors’ information demand.

Variable	(1)	(2)	(3)
	Hold	Follow	Follow
ClimateR	3.8421 ***		0.0916 **
	(4.00)		(2.33)
Hold		0.0059 ***	0.0059 ***
		(21.05)	(20.97)
Control	Yes	Yes	Yes
Industry	Yes	Yes	Yes
Year	Yes	Yes	Yes
Sobel Z		−3.336 ***
Bootstrap (1000 times) confidence interval		[−0.0336, −0.0094]
Adj.R2	0.4864	0.3894	0.3895
N	20,978	20,978	20,978

Note: The values in parentheses are t-values adjusted for robust standard errors. ** and *** denote significance at the 5% and 1% levels, respectively.

presents the results for the “information demand” mechanism through which corporate climate risk disclosure affects analyst coverage. Column (1) shows that the regression coefficient of ClimateR is positive at the 1% level, indicating that corporate climate risk disclosure increases institutional investors’ information demand. Columns (2) and (3) show that the coefficients of Hold are significantly positive, suggesting that institutional investors’ information demand can significantly promote analyst coverage. Furthermore, the Sobel Z-test statistic is −3.336 and significant at the 1% level. Meanwhile, the confidence intervals of the Bootstrap mediation effect do not include zero. Thus, corporate climate risk disclosure attracts more analyst coverage by increasing institutional investors’ information demand.

4.5.2. Analysts’ Information Supply

Field research is the most direct and reliable way for analysts to obtain private information about companies. By providing more publicly available information, corporate climate risk disclosure reduces analysts’ need to dig for private information. Based on the literature [53], we measure analysts’ field research at the brokerage level by tracking the count of analyst visits. Additionally, we define a binary variable (Visit) to indicate whether analysts have conducted field research. Specifically, Visit equals 1 if field research has been conducted and 0 otherwise. The number of analyst visits (Vc) is the natural logarithm of one plus the total number of times analysts visit a company in a given year. We then examine the impact of corporate climate risk disclosure on analysts’ field research.

Table 8. Mechanism analysis: analysts’ information supply.

Panel A: Whether There Are Analyst Field Visits
Variable	(1)	(2)	(3)
	Visit	Follow	Follow
ClimateR	−0.4392 ***		0.1711 ***
	(−5.89)		(4.38)
Visit		0.1289 ***	0.1296 ***
		(36.34)	(36.50)
Control	Yes	Yes	Yes
Industry	Yes	Yes	Yes
Year	Yes	Yes	Yes
Sobel Z		−4.673 ***
Bootstrap (1000 times) confidence interval		[−0.0558, −0.0220]
Adj.R2	0.1364	0.4113	0.4118
N	20,978	20,978	20,978
Panel B: Number of Analyst Field Visits
Variable	(1)	(2)	(3)
	Vc	Follow	Follow
ClimateR	−0.1934 ***		0.1564 ***
	(−7.63)		(3.96)
Vc		0.2164 ***	0.2186 ***
		(20.37)	(20.55)
Control	Yes	Yes	Yes
Industry	Yes	Yes	Yes
Year	Yes	Yes	Yes
Sobel Z		−6.358 ***
Bootstrap (1000 times) confidence interval		[−0.0439, −0.0236]
Adj.R2	0.1295	0.3883	0.3887
N	20,978	20,978	20,978

Note: The values in parentheses are t-values adjusted for robust standard errors. *** denotes significance at the 1% level.

presents the results. Column (1) of Panels A and B reveals that climate risk disclosure is negatively related to analysts’ field research at the 1% level. Thus, corporate climate risk disclosures reduce the frequency of analyst visits. By lowering the costs associated with information acquisition for analysts, these disclosures ultimately lead to increased analyst coverage. Furthermore, the Sobel Z-values are −4.673 and −6.358, respectively, and significant at the 1% level. Additionally, the confidence intervals for the Bootstrap mediation effect do not include zero, indicating that the mediation mechanism test is valid. Thus, climate risk disclosure attracts more analyst coverage by reducing analysts’ information acquisition cost. The above results support our hypothesis (H2).

4.6. Further Analysis

Next, this subsection delves deeper into the relationship between different types of climate risk disclosures and analyst coverage from both corporate and analyst perspectives. Additionally, it explores whether climate risk disclosure impacts the quality of analyst forecasts.

4.6.1. Different Types of Climate Risk Disclosure and Analyst Coverage

Physical climate risks refer to the direct costs caused by extreme weather events, such as typhoons, floods, and earthquakes. Meanwhile, transition climate risks arise from increased environmental awareness among society, stakeholders, and policymakers and the shift to a low-carbon economy to address climate change. This transition often involves management, technological, and market changes, which may pose significant financial challenges for companies [3]. This study replaces the ClimateR variable in model (3) with disclosures for various climate risks and re-runs the regressions to explore how different climate risk disclosures influence analysts’ behavior.

Table 9. Different types of climate risk disclosures and analyst coverage.

Variable	(1)	(2)	(3)	(4)
	Follow	Follow	Follow	Follow
SeriousR	3.9395 ***			3.2782 ***
	(3.77)			(3.05)
ChronicR		3.3992 ***		2.3772 **
		(3.60)		(2.43)
TransitionR			0.1072 ***	0.0881 **
			(2.68)	(2.18)
Controls	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Adj.R2	0.3765	0.3764	0.3763	0.3768
N	20,978	20,978	20,978	20,978

Note: The values in parentheses are t-values adjusted for robust standard errors. ** and *** denote significance at the 5% and 1% levels, respectively.

presents the results. Columns (1) to (3) display the effects of severe (SeriousR), chronic (ChronicR), and transition risk disclosures (TransitionR) on analyst coverage, respectively. Meanwhile, column (4) examines the combined effect of all three types of risk disclosures on analyst coverage. In all models, the coefficients for SeriousR, ChronicR, and TransitionR are significantly positive, indicating that disclosures of severe, chronic, and transition risks significantly influence analyst coverage. Column (4) further reveals that after controlling for related variables, the regression coefficient for SeriousR is 3.2782, for ChronicR it is 2.3772, and for TransitionR it is 0.0881. Thus, severe and chronic risks have stronger impacts on analyst coverage than transition risk disclosures. This finding is consistent with the research of Ben-Amar et al. [15], who concluded that transition risk disclosures have a relatively minor influence on investors and analysts.

4.6.2. Climate Risk Disclosure and Analyst Forecast Quality

Given that climate risk disclosure attracts greater analyst coverage, does it also influence the quality of their forecasts? This section examines the effect of climate risk disclosure on analysts’ earnings forecast bias and forecast dispersion using the following regression model:

Error_i,t+1 / Fdisp_i,t+1 = β₀ + β₁ ClimateR_i,t + Control_i,t + Industry + Year + ε

(7)

Error_i,t+1 = |Mean(Feps_i,j,t+1) − Aeps_i,t+1|/(|Aeps_i,t+1| + 0.5)

(8)

Fdisp_i,t+1 = Std(Feps_i,j,t+1)|/|Mean(Feps_i,j,t+1)|

(9)

Here, i represents the firm and t represents the year. Error_i,t+1 denotes the earnings forecast bias of analysts, while Fdisp_i,t+1 represents the dispersion in analysts’ earnings forecasts. Compared to model (3), model (7) incorporates additional control variables, including the number of analysts covering the firm (Follow) and the proportion of institutional investor ownership (Hold). Error_i,t+1, and Fdisp_i,t+1 are defined as shown in Equations (8) and (9), respectively. Feps_i,j,t+1 represents the forecasted earnings per share (EPS) of firm i for year t + 1 by analyst j. Following Wang et al. [54], if analyst j made multiple forecasts before the company announced the actual EPS for year t + 1, we use the first forecast the analyst issued after the release of the annual report for year t. Std(Feps_i,j,t+1) and Mean(Feps_i,j,t+1) refer to the standard deviation and mean of all analysts’ forecasts of the firm’s EPS for year t + 1. Aeps_i,t+1 represents the actual EPS of firm i in year t + 1. Error_i,t+1 reflects the deviation between analysts’ forecasted and the actual EPS for firm i in year t + 1. A more considerable value indicates lower forecast accuracy. Fdisp_i,t+1 measures the dispersion among analysts’ earnings forecasts for firm i in year t + 1, with a higher value indicating more significant disagreement among analysts.

Research shows that high-quality information disclosure facilitates analysts in accessing more valuable public information, thereby reducing forecast bias. Earnings forecast dispersion primarily stems from analysts’ differing interpretations of the same information and their varying efforts in mining information from private channels. Analysts’ extraction of information from private channels significantly amplifies the disparities in information quality among them [31]. Therefore, as listed companies publicly disclose climate risk information, analysts’ reliance on private channels will likely diminish, reducing earnings forecast dispersion.

Table 10. Climate risk disclosure and analyst forecast quality.

Panel A: Climate Risk Disclosure and Forecast Bias
Variable	(1)	(2)	(3)
	Errort+1	Errort+2	Errort+3
ClimateR	−0.0108	−0.0444 *	−0.0462 *
	(−0.76)	(−1.95)	(−1.68)
Control	Yes	Yes	Yes
Industry	Yes	Yes	Yes
Year	Yes	Yes	Yes
Adj.R2	0.0770	0.0713	0.0708
N	17,477	15,902	13,970
Panel B: Climate Risk Disclosure and Forecast Dispersion
Variable	(1)	(2)	(3)
	Fdispt+1	Fdispt+2	Fdispt+3
ClimateR	−0.0646 ***	−0.0467 **	−0.0527 **
	(−2.86)	(−2.04)	(−2.09)
Control	Yes	Yes	Yes
Industry	Yes	Yes	Yes
Year	Yes	Yes	Yes
Adj.R2	0.1087	0.1091	0.1030
N	17,477	15,902	13,970

Note: The values in parentheses are t-values adjusted for robust standard errors. *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively.

presents the regression results of model (7). As shown in column (1) in Panel A, the regression coefficient of climate risk disclosure is statistically insignificant, suggesting that corporate climate risk disclosure does not reduce short-term forecast bias among analysts. Column (1) in Panel B reports a regression coefficient of −0.0646 for climate risk disclosure, which is significant at the 1% level. This suggests that enhanced corporate climate risk disclosure helps reduce the dispersion in analysts’ earnings forecasts; this happens due to the increase in publicly available climate risk information.

Analysts’ long-term earnings forecasts are more likely to reflect their comprehensive evaluation of factors influencing a company’s long-term development [55,56]. Most analysts’ forecasts of company fundamentals focus on the next two to three years. Climate risks, such as regulatory policy changes, resource cost fluctuations, and market demand shifts, typically significantly impact a company’s medium- to long-term operations and profitability. Studies reveal that assessing climate change’s financial impact on a company is both complex and costly. Analysts are likelier to incorporate climate change factors into their analyses only when issuing longer-term and more detailed earnings forecasts [16]. Consequently, corporate climate risk disclosure has a limited impact on analysts’ short-term earnings forecasts but may significantly enhance their long-term forecasts’ accuracy. This study investigates the potential impact of climate risk disclosure on the quality of analysts’ long-term forecasts. Columns (2) and (3) in Panels A and B of Table 10 report the relationship between corporate climate risk disclosure and analysts’ forecast quality for t + 2 and t + 3, respectively. The results demonstrate that climate risk disclosure significantly reduces long-term forecast errors and dispersion. Thus, climate risk disclosure improves analysts’ ability to accurately assess a company’s future long-term performance.

5. Conclusions

Using data from China’s A-share listed companies from 2007 to 2021, this study examines the impact of corporate climate risk disclosure on analyst coverage and explores the underlying mechanisms. The results demonstrate a significant positive relationship between corporate climate risk disclosure and analyst coverage, which holds under a battery of robustness checks. Mechanism analysis reveals that climate risk disclosure increases institutional investors’ information demands, reduces analysts’ information acquisition costs, and attracts more analysts to follow the company. Furthermore, distinguishing between the climate risk categories disclosed by firms, chronic and severe risk disclosures are more attractive for analysts to follow than transformational risk disclosures. The positive effect of climate risk disclosure on analyst coverage is more pronounced among firms with low readability of annual reports and a higher share of independent institutional investors and in the context of team and large brokerage analysts. In addition, climate risk disclosure reduces analysts’ earnings forecast dispersion and long-run earning forecast bias and effectively supports analysts’ more accurate assessment of firms’ performance prospects over a more extended period.

Based on these findings, we propose the following policy recommendations. First, regulators must strengthen Chinese enterprises’ standardization of climate risk disclosure. Based on China’s “Guidelines on Sustainability Reporting for Listed Companies (for Trial Implementation)”, more unified and detailed disclosure standards should be formulated to clarify the disclosure requirements of enterprises in terms of chronic, severe, and transformational risks, and a mandatory disclosure policy should be gradually implemented. Second, institutional investors and securities analysts should be encouraged to pay more attention to corporate climate risk information. Institutional investors should prioritize enterprises with high-quality climate risk disclosures in their investment decisions through policy incentives and regulatory guidance and thus create a benign market feedback mechanism for transparent information. Meanwhile, regulators and institutions should strengthen training and support for securities analysts to enhance their ability to analyze and interpret climate risk information, especially regarding chronic and severe risks. This can enable analysts to more accurately assess the long-term development potential of enterprises.

Future Research Prospects

This study provides valuable insights into the influence of climate risk disclosure on analyst coverage, offering findings that can be informative for other emerging economies. However, several limitations should be noted. First, climate risk disclosure is shaped by institutional environments and regulatory intensity, both of which exhibit considerable variation across countries. Therefore, future research should aim to conduct cross-country comparative studies to investigate the heterogeneous effects of climate risk disclosure on analyst coverage behavior. Such research would enhance the external validity of the findings and contribute to a deeper understanding of this important issue [30]. Second, the study excludes financial companies from the sample, yet the financial sector may have unique mechanisms regarding climate risk disclosures and their impact on analyst coverage [57]. Therefore, future research should extend its focus to include financial companies and further explore their role and influence in climate risk disclosures. Finally, the study primarily relies on text analysis of annual reports to measure climate risk disclosures, which may not fully capture all relevant climate-related information. To address this limitation, future research could integrate data from annual reports, CSR reports, and ESG reports, utilizing advanced natural language processing techniques to more accurately capture the multidimensional nature of climate risk disclosures, thus improving the depth and accuracy of the analysis.