Analysts’ Green Coverage and Corporate Green Innovation in China: The Moderating Effect of Corporate Environmental Information Disclosure

Abstract

Against the backdrop of China’s growing environmental concerns, investing in green technology innovation is a crucial solution to achieve the goal of “carbon peak and carbon neutrality”. Combining the perspectives of signaling theory and corporate governance theory, we use a sample of Chinese-listed companies from 2008 to 2020 to investigate the influence of analysts’ green coverage (AGC) on corporate green innovation (CGI) and the moderating effect of corporate environmental information disclosure (CEID) based on a textual analysis approach. The results show that AGC can significantly promote the quantity and quality of CGI, and CEID has a positive moderating effect on the process. Moreover, the mechanism analysis reveals that enhancing investors’ value recognition, improving corporate reputation capital, alleviating corporate financing constraints, reducing management agency costs, and curbing managerial myopia are the influence mechanisms of AGC on CGI. Additionally, the positive effect of AGC is more significant for firms located in regions with a favorable institutional environment, firms belonging to heavily polluting industries, and firms that have not adopted continuous innovation strategies.

1. Introduction

One of the biggest economic and social problems that humanity will encounter in the near future is global climate change [1]. The Sustainable Development Goals (SDGs), adopted at the UN summit in 2015, aim to ensure human well-being, protect the planet, and promote global prosperity. Key targets of the SDGs include achieving economic growth and industrial innovation, as well as improving the climate by 2030. These goals represent a significant step forward in the global effort to create a more sustainable and equitable future for all [2]. As one of the world’s largest emitters of carbon, how China can achieve its carbon emission reduction targets through effective green practices is crucial to the achievement of the SDGs. In recent years, China has prioritized the integration of environmental protection goals and technological innovation goals with the goals of sustainable economic development. The 14th Five-Year Plan period presents a crucial opportunity for China to slow down the growth rate of CO₂ emissions through green technology innovation. In this context, firms’ green innovation has become a focus of research both domestically and internationally [1,3,4,5,6], and burgeoning literature has found that highly developed financial markets play a vital role in facilitating firms’ green innovation [7,8,9].

The influence of securities analysts on firms’ general innovation has received considerable attention from scholars [10,11,12]. Based on the research samples of the American capital market, one strand of literature found that analysts’ earnings forecasts and stock recommendations increase managers’ performance pressure, resulting in a significant decrease in the general innovation output of US-listed firms [10,11]. Meanwhile, provided that the Chinese capital market is characterized by high ownership concentration, long-term shareholding of major shareholders, and relatively low stock liquidity, the other strand of literature documented that analysts’ coverage can promote the general innovation performance of Chinese-listed firms through the information effect and the governance effect [12]. As research on green innovation continues to grow, recent studies have shed light on the positive impact of financial analysts on firms’ eco-innovation [13,14].

Previous research on how analyst coverage affects firm (green) innovation has primarily used metrics such as the total number of analysts who publish research reports on a given firm or the total number of research reports published by analysts [10,14]. However, these studies have treated all analysts as a homogeneous group, failing to account for potential differences in expertise, knowledge, and focus among analysts. Meanwhile, the emerging literature highlights the existence of heterogeneity among analysts [15,16], especially the heterogeneous information content of different research reports [17,18]. In fact, the research reports published by analysts on green or innovation topics should be more effective in guiding investors’ value recognition to corporate green innovation and motivating corporate management to continuously invest in green innovation activities. Therefore, we propose a new concept of analysts’ green coverage (hereinafter referred to as AGC), i.e., financial analysts’ coverage of corporate green investment or green technology innovation. Specifically, we use a textual analysis approach to quantify analysts’ research reports to measure AGC. While previous studies by Fiorillo et al. (2022) and Han et al. (2022) [13,14] have made significant contributions to understanding the relationship between analyst coverage and green innovation, they have primarily focused on the quantity of green innovation and overlooked its quality. Therefore, our study aims to provide a more comprehensive analysis of the impact of AGC on both the quantity and quality of corporate green innovation (hereinafter referred to as CGI).

As an informal environmental regulation tool for public participation, the role of environmental information disclosure (hereinafter referred to as EID) in solving environmental problems is becoming more prominent. Drawing on signaling theory and organizational legitimacy theory, previous research has emphasized the role of EID in promoting CGI [19,20,21]. In an environment where investors are highly concerned about corporate environmental risks, securities analysts tend to scrutinize and interpret corporate environmental information to improve the reliability of investment ratings. This, in turn, provides timely feedback to the government to evaluate environmental policies and to investors to make investment decisions [22]. As a source of analysts’ access to environmental information, high-quality corporate environmental information disclosure (hereinafter referred to as CEID) has a critical impact on the relationship between AGC and CGI. Therefore, we incorporate voluntary CEID into the theoretical framework of AGC affecting CGI to explore the potential moderating effect of CEID.

In this study, we utilize a sample of Chinese-listed companies spanning from 2008 to 2020 to explore the influence of AGC on CGI. Specifically, we conduct a textual analysis of analysts’ research reports to measure their coverage of “green” or “innovation”, and we examine the moderating effect of CEID. Some of our findings strengthen the existing literature and contribute to the relevant research in the three aspects as follows. First, based on the textual analysis approach, we dig up the positive effect of analysts’ coverage on CGI performance by measuring analysts’ coverage of firms’ green behavior or innovative activities, which could effectively isolate the most relevant analysts’ coverage that directly affects CGI and thus provide more detailed empirical evidence. Second, benefiting from the data availability of forward citations of green patents and IPC classification codes, we further investigate the effect of AGC on the quality of CGI, which could take a step forward to the existing research. Third, we comprehensively dig up the influence mechanism of AGC on CGI from the perspective of the information effect by strengthening incentives for CGI and the monitoring effect by enhancing external governance. This provides a crucial complement to previous studies.

2. Theoretical Analysis and Hypothesis Development

2.1. Analysts’ Green Coverage and Corporate Green Innovation

In recent decades, governments have become increasingly concerned about environmental quality, leading to the implementation of environmental laws and regulations aimed at controlling pollution. Guided by stakeholder theory, the public and investors are highly concerned about firms’ environmental risks and behavior, which has compelled firms to transition towards a green and sustainable development path, with green technology innovation serving as the primary tool [21]. Simultaneously, there is a growing recognition of the critical role that intermediaries play in fostering eco-innovation [23]. Faced with a large amount of information on environmental regulations issued by local governments, as well as information on environmental protection and green innovation disclosed by listed companies, it is difficult for ordinary investors to reasonably assess the investment value of companies, which results in a loss of investment efficiency in the capital market. As sophisticated and professional experts of financial information, analysts can enhance the visibility of companies [24], by processing forward-looking and immeasurable environmental information and green innovation information and disseminating the information to investors in an easy-to-understand manner [14]. Meanwhile, the finance and accounting literature has emphasized the monitoring effect of analysts on corporate innovation for a long time [25,26,27]. Therefore, we can summarize the mechanisms by which AGC influences CGI as the information effect and the monitoring effect.

2.1.1. The Information Effect of Analysts’ Green Coverage

First, AGC could promote CGI by enhancing investors’ value recognition. Financial analysts are generally considered information lubricants in the capital market, which can provide more incremental information to optimize resource allocation and enhance capital market efficiency [10,11]. In the context of increasing environmental regulations and public concern about environmental issues, the disclosure of environmental and green innovation information by listed companies is not only a way to reduce environmental pollution but also a way to demonstrate companies’ efforts in environmental management and suggest environmental risks [28]. However, retail investors may not accurately understand the value of corporate environmental and green innovation information, while analysts can write in-depth research reports on these companies to interpret and reflect the information in the stock price. This can enhance investors’ value recognition, strengthen their knowledge of information on the advantages of green innovation of companies [29], and improve the efficiency of information transmission in the capital market. Furthermore, the study of Fan and Yao (2022) [22] pointed out that analysts pay attention to environmental and innovation information of firms when making earnings forecasts. Therefore, AGC can not only enhance the effectiveness of the capital market but also encourage management to invest in CGI and achieve sustainable development.

Second, AGC can facilitate CGI by improving corporate reputation capital. Analysts can help companies maintain their green reputation by effectively disseminating the firms’ green information. Reputation theory reveals that a company’s reputation is an intangible asset derived from its daily behavior and from the public’s attraction to it [30]. Therefore, firms are encouraged to increase investment in CGI based on the pursuit of a good reputation. Moreover, the increase in AGC can further strengthen the motivation of companies to build and maintain a good reputation. Corporate reputation can influence the behavior of investors and consumers as well as the market’s perception of corporate image. Specifically, companies with good reputations can achieve higher reputational gains, and external markets will respond more positively to their behavior [31]. Therefore, AGC can help companies build and maintain a good reputation, which in turn can encourage firms to invest in CGI.

Third, AGC can also promote CGI by alleviating corporate financing constraints. By combining economic and environmental benefits, green innovation can benefit companies more than general corporate innovation. However, unlike general innovation, the innovation process for green technologies is more systematic and complex and is fraught with uncertainty and risk. As a result, CGI may lead to more severe financing constraints for firms, which in turn discourages CGI over the long term [32]. However, existing studies also highlighted the positive role of analyst coverage in improving the efficiency of corporate general innovation financing [10,11]. In the case of CGI, AGC can accelerate the speed and breadth of information transmission, such as the information of CEID and CGI, which can enhance corporate reputation and alleviate the degree of external financing constraints [22]. Additionally, AGC could facilitate the acquisition of green innovation financing through green credit, green bonds, and green insurance [5,33,34], thus promoting the development of CGI activities.

In summary, based on the above three channels, we believe that AGC would promote CGI through the information effect.

2.1.2. The Monitoring Effect of Analysts’ Green Coverage

Building on the information effect, the monitoring effect of analysts in improving the effectiveness of corporate governance has received extensive attention in previous studies [26,35]. From this perspective, we can further analyze the influence of AGC on CGI. By monitoring firms’ behavior and performance related to CGI, analysts can provide valuable feedback to management and investors, and hold firms accountable for their environmental impact. This can enhance the transparency and credibility of CGI activities and encourage firms to adopt more sustainable practices. Additionally, the monitoring effect of analysts can help identify potential risks and challenges associated with CGI and provide guidance on how to mitigate them, which can improve the success rate of green innovation projects. Therefore, the monitoring effect of analysts is an important factor in promoting CGI and achieving the goals of sustainable development.

First, AGC can promote CGI by reducing managerial agency costs. The process of CGI is often accompanied by agency problems [36]. Managers have more discretion in the process of CGI, which could lead to opportunistic behavior and the pursuit of private interests [37]. This behavior can seriously damage the performance of CGI. AGC can help to curb managers’ self-interested behavior and monitor the utilization of CGI investment funds. This can prompt corporate management to improve CGI in response to the environmental legitimacy crisis and obtain corporate legitimacy recognition [22], which in turn promotes the improvement of corporate environmental governance performance. Therefore, AGC can effectively reduce the first type of agency costs in the process of CGI. Additionally, AGC can also reduce the second type of agency costs by disseminating information on corporate environmental efforts and CGI to investors and the public. This can reduce the degree of information asymmetry, leading to more informed investment decisions and improved corporate reputation. By reducing both types of agency costs, AGC can promote greater transparency and accountability in the process of CGI, ultimately leading to more sustainable development.

Second, AGC could also promote CGI by curbing managerial myopia. In fact, there is a large asymmetry between the benefits of successful green innovation projects and the loss of managers’ personal wealth and reputation due to green innovation failures [10]. As a result, managers may be more risk-averse and thus show greater conservatism and myopia in the process of CGI [38]. Meanwhile, these firms with relatively high green innovation performance will face serious systematic undervaluation of their stock value, which will further reduce management’s willingness toward CGI. According to threat rigidity theory, when faced with a threat, corporate managers tend to become emotionally stressed and anxious due to increased stress, and the negative emotion will further lead managers to narrow the scope of information processing, making it impossible for them to effectively process the complex information [39]. Therefore, under the pressure of the capital market, corporate management may prefer to reduce investment in CGI. As information amplifiers in the capital market, analysts can comprehensively demonstrate management’s efforts and investments in environmental protection and green innovation to investors and the public, which can effectively reduce management’s stress and anxiety, improve management’s long-term horizon, and guide investors’ value recognition [11]. As a result, AGC could promote CGI performance by motivating management’s innovation efforts in the long run.

According to the above two channels, we believe that AGC can also promote CGI through the monitoring effect, and propose research Hypothesis 1 as follows.

Hypothesis 1.

Analysts’ green coverage will facilitate corporate green innovation.

2.2. The Moderating Effect of Corporate Environmental Information Disclosure

As mentioned above, reducing the friction in the information dissemination process is the core mechanism of AGC influencing CGI. Considering analysts’ information production process, the sources and quality of corporate environmental information analysts obtain have a crucial influence on the actual effect of their green coverage [19,20,21]. Analysts’ sources of access to corporate environmental information include two main sources. The first source is public information, mainly from CSR reports, sustainability reports, environmental reports, and ESG reports issued by listed firms [22]. Meanwhile, environmental policies and evaluation reports issued by local governments are also the sources of public information for AGC. The second source is private information, which analysts can obtain by communicating privately with managers through corporate site visits and other private channels [40,41,42,43,44]. Given the lower cost of analysts’ access to public environmental information and the predominantly voluntary form of CEID by listed firms, we further analyze the role of CEID on the process of AGC influencing CGI.

Recent studies have investigated the direct incentive effect of EID on CGI. By conducting the DID approach based on the assessment of the PITI for Chinese cities in 2008, one strand of research found that mandatory EID by local governments could significantly promote CGI [21,31,45]. Based on this, another strand of research highlighted the positive effect of voluntary CEID by listed companies on facilitating CGI [20,46,47]. In the process, the core influence mechanisms of CEID on CGI could be summed up as the improvement of CEID and the enhancement of reputation capital. Based on organizational legitimacy theory, the act of CEID can demonstrate the legitimacy of a firm’s existence if they adhere to certain social norms [48]. Therefore, by increasing public access to environmental information and permitting public participation in environmental governance actions, CEID can assist firms in communicating their environmental protection responsibility to the government and the public [49,50], thereby enhancing corporate information transparency. Moreover, CEID can signal CSR, thereby improving the firm’s reputational capital [51]. As a result, with an improved information environment and increased reputation capital, companies can attract more investor attention and analysts’ coverage [22]. In terms of the motivation for CEID, analysts’ demand for environmental information and following behavior will, in turn, stimulate voluntary and active CEID. Therefore, the interaction and matching of corporate environmental information supply and analysts’ environmental information demand are conducive to further promoting CGI. In other words, the positive effect of AGC on CGI will be more pronounced when the CEID is greater.

In accordance with the above theoretical analysis, we propose research Hypothesis 2 as follows.

Hypothesis 2.

Corporate environmental information disclosure could positively moderate the effect on which analysts’ green coverage influences corporate green innovation.

To illustrate the logic of hypothesis development more intuitively, we construct a hypothetical framework, as shown in Figure 1.

3. Research Design

3.1. Data Sources and Sample Selection

In this study, we take Chinese-listed companies from 2008 to 2020 as the research sample. Specifically, the data on corporate green patent, general patent, analyst research report, and media coverage are obtained from the CNRDS database. The data of the total marketization index is from the CMI database. Meanwhile, the data on CEID, corporate finance, and corporate governance are from the CSMAR database and Wind database. In China, listed companies serve as effective representatives of market participants and play a significant role in overall economic development. The sample industries selected for this paper are widely distributed and encompass numerous sectors. Additionally, the data used are all publicly available, which makes the empirical results replicable. Finally, we obtained a research sample consisting of 30,937 observations.

3.2. Empirical Model and Variable Measurement

In order to test the two aforementioned hypotheses, we establish the empirical models as follows.

$$CG{I}_{i,t}={\alpha }_0+{\alpha }_{1}AG{C}_{i,t}+{\displaystyle \sum {\alpha }_j}C{V}_{i,t}+Yea{r}_t+Industr{y}_{i,t}+Cit{y}_{i,t}+{\epsilon }_{i,t}$$

(1)

$$\begin{array}{l}CG{I}_{i,t}={\beta }_0+{\beta }_{1}AG{C}_{i,t}+{\beta }_{2}CEI{D}_{i,t}+{\beta }_{1}AGC*CEI{D}_{i,t}\\ +{\displaystyle \sum {\beta }_j}C{V}_{i,t}+Yea{r}_t+Industr{y}_{i,t}+Cit{y}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(2)

In the above models, CGI represents the dependent variable, AGC represents the independent variable, and CEID represents the moderating variable. The CV is a vector of control variables that could affect CGI, the Year, Industry, and City to capture the corresponding fixed effect. In addition, the subscripts i and t denote the firm and year.

3.2.1. Dependent Variable: CGI (Corporate Green Innovation)

Referring to the prior studies [14,45], we construct the CGI variable by extracting the CGI output data from the GPRD in the CNRDS database. Using the green inventory of the IPC presented by the WIPO [52], the GPRD sub-library provides detailed statistics of green patent applications, grants, and citations. Therefore, from the perspective of quantity and quality, we comprehensively measure CGI performance. First, we use a firm’s total number of green patent filings to measure CGI quantity (CGI1) [53]. Second, we use the average number of forward citations of a firm’s green patent filings to measure CGI quality (CGI2). Considering the negative effect of zero green patent filings or zero citations, we further calculate the natural logarithm of the original values (plus one) [54].

3.2.2. Independent Variables: AGC (Analysts’ Green Coverage)

After processing the text data of analysts’ research reports from TDAR in the CNRDS database, we construct the AGC variable using the textual analysis approach introduced in previous studies [17,55]. The detailed processing steps are as follows. First, we extract and label analyst reports that contain keywords such as “energy”, “clean”, “environmental protection”, “green”, “sustainable”, or “ecological”, and define them as analyst reports on the topic of “energy and environmental protection.” Second, we also extract and label analyst reports that contain keywords such as “innovation”, “patents”, “R&D”, or “research and development”, and define them as analyst reports on the topic of “R&D innovation” [56]. Finally, we aggregate the total number of labeled analyst reports and define it as AGC. Consistent with CGI, we also compute the natural logarithm of the original values (plus one) of AGC.

3.2.3. Moderating Variables: CEID (Corporate Environmental Information Disclosure)

Based on the previous research [57], we use the content analysis method to construct a corporate environmental information evaluation index system. Benefiting from the availability of listed companies’ environmental research data in the CSMAR database, we use a set of CEID indicators. Specifically, a second-level indicator equals two if the given company provides quantitative EID, one if the given company provides only qualitative EID, and zero otherwise. Then, we calculate the total scores by summing the above values to obtain the maximum value of the total scores (Max. = 38). Finally, we divide the total scores by 38 and obtain the CEID index. The detailed CEID indicators can be obtained from the CSMAR database (The website URL is http://cn.gtadata.com/. The access date is 1 January 2023).

3.2.4. Control Variables

In accordance with the previous studies, we control for a battery of firm and industry characteristics that may affect CGI [1,58]. First, we control for corporate finance and governance characteristics, including firm size (Size), financial leverage (Lev), firm age (Age), sales growth (Growth), Tobin’s Q (Tobin), cash flow (Cflow), cash holdings (Cash), environmental protection investment (EPInvest), management duality (Duality), largest shareholder holdings (Top1), management remuneration (Remuner), CEO holdings (CEO_Holding), state-owned enterprise (SOE), and media coverage (MC). Second, we control for firms’ general innovation characteristics, including R&D investment (RDS), R&D investment missing values (RDS_MV), and total number of general patent applications (Total_Patent). Third, we also control for the industrial policy, industry characteristics, and regional marketization characteristics. Specifically, we control for government subsidies (Subsidy) and tax benefits (Tax_Benefit) to capture the effect of government industrial policy, heavily polluting industries (Industry_HP) to examine the influence of heavily polluting industries, and control for the marketization total index (MTI) to examine the effect of the regional institutional environment.

4. Empirical Results

4.1. Descriptive Statistics

Table 1. Descriptive statistics.

Variable	Obs.	Mean	Std. Dev.	Min	Median	Max
CGI1	30,937	0.838	1.159	0.000	0.000	4.663
CGI2	30,937	0.354	0.637	0.000	0.000	2.454
AGC	30,937	0.460	0.933	0.000	0.000	3.611
CEID	30,937	0.164	0.168	0.000	0.079	1.000
Size	30,937	22.105	1.298	19.613	21.923	26.059
Lev	30,937	0.426	0.207	0.051	0.420	0.886
Age	30,937	10.652	7.120	−1.000	9.000	31.000
Growth	30,937	0.171	0.417	−0.589	0.107	2.694
Tobin	30,937	2.017	1.278	0.871	1.603	8.366
Cflow	30,937	0.046	0.071	−0.174	0.046	0.249
Cash	30,937	0.200	0.145	0.014	0.158	0.698
EPInvest	30,937	0.290	1.487	0.000	0.000	9.259
RDS	30,937	0.022	0.036	0.000	0.000	0.196
RDS_MV	30,937	0.510	0.500	0.000	1.000	1.000
Total_Patent	30,937	1.346	1.757	0.000	0.000	6.205
Duality	30,937	0.267	0.442	0.000	0.000	1.000
Top1	30,937	0.349	0.149	0.089	0.329	0.750
Remuner	30,937	15.412	0.772	13.087	15.410	17.392
CEO_Holding	30,937	0.055	0.121	0.000	0.000	0.570
SOE	30,937	0.381	0.486	0.000	0.000	1.000
Subsidy	30,937	0.012	0.018	0.000	0.006	0.106
Tax_Benefit	30,937	0.136	0.180	−0.712	0.140	0.879
Industry_HP	30,937	0.279	0.448	0.000	0.000	1.000
MTI	30,937	9.179	1.697	−0.161	9.438	11.494
MC	30,937	3.432	1.439	0.000	3.434	7.302

reports the descriptive statistics. The mean value of CGI1 (corporate green innovation quantity) is 0.838, indicating that, on average, a firm applies for 5.26 green patents (exp(0.838) −1) per year. In terms of corporate green innovation quality (CGI2), each firm receives, on average, 0.88 citations per year for its filed green patents (exp(0.354) −1). Furthermore, the mean value of AGC (analysts’ green coverage) shows that each firm is followed, on average, by 2.22 analyst research reports on “green” or “innovation” (exp(0.460) −1) per year. Meanwhile, the mean value of CEID (corporate environmental information disclosure) is 0.164, and it shows that the overall CEID quality of the sample companies is relatively low. With respect to the main control variables, there is no obvious sample selection bias.

4.2. Baseline Regression Results

Table 2. Baseline regression results.

	(1)	(2)	(3)	(4)	(5)	(6)
	CGI1	CGI1	CGI1	CGI2	CGI2	CGI2
AGC	0.056 ***	0.057 ***	0.022 *	0.014 **	0.014 **	0.002
	(5.23)	(5.32)	(1.65)	(2.12)	(2.18)	(1.20)
CEID		0.007 ***	0.010 ***		0.005 ***	0.005 ***
		(3.91)	(5.07)		(3.24)	(3.87)
AGC*CEID			0.005 ***			0.002 **
			(3.70)			(2.55)
Size	0.310 ***	0.299 ***	0.300 ***	0.178 ***	0.171 ***	0.172 ***
	(19.40)	(18.74)	(18.83)	(14.66)	(14.32)	(14.37)
Lev	0.196 ***	0.208 ***	0.209 ***	0.118 ***	0.126 ***	0.126 ***
	(3.11)	(3.32)	(3.32)	(2.64)	(2.83)	(2.83)
Age	−0.005 **	−0.005 **	−0.005 **	−0.002	−0.002	−0.002
	(−2.23)	(−2.27)	(−2.34)	(−1.46)	(−1.50)	(−1.53)
Growth	0.026 **	0.029 **	0.028 **	0.007	0.009	0.009
	(2.05)	(2.32)	(2.23)	(0.76)	(0.99)	(0.95)
Tobin	0.019 **	0.018 **	0.019 **	0.016 ***	0.016 ***	0.016 ***
	(2.49)	(2.34)	(2.50)	(3.11)	(2.97)	(3.04)
Cflow	−0.392 ***	−0.420 ***	−0.418 ***	−0.305 ***	−0.322 ***	−0.321 ***
	(−3.93)	(−4.24)	(−4.22)	(−4.33)	(−4.62)	(−4.61)
Cash	0.226 ***	0.235 ***	0.218 ***	0.220 ***	0.226 ***	0.220 ***
	(3.40)	(3.56)	(3.29)	(4.40)	(4.55)	(4.42)
EPInvest	0.025 ***	0.019 ***	0.017 ***	0.014 ***	0.010 **	0.009 **
	(4.80)	(3.50)	(3.19)	(3.16)	(2.18)	(2.04)
RDS	1.118 ***	1.183 ***	1.179 ***	1.376 ***	1.417 ***	1.416 ***
	(3.38)	(3.58)	(3.56)	(5.26)	(5.41)	(5.41)
RDS_MV	0.008	0.014	0.020	0.064 ***	0.068 ***	0.070 ***
	(0.36)	(0.64)	(0.89)	(3.92)	(4.14)	(4.29)
Total_Patent	0.231 ***	0.231 ***	0.232 ***	0.147 ***	0.146 ***	0.147 ***
	(28.62)	(28.52)	(28.58)	(23.22)	(23.16)	(23.14)
Duality	0.004	0.005	0.006	0.025	0.026	0.026
	(0.17)	(0.21)	(0.25)	(1.57)	(1.60)	(1.62)
Top1	−0.221 ***	−0.226 ***	−0.224 ***	−0.155 ***	−0.158 ***	−0.157 ***
	(−2.73)	(−2.79)	(−2.77)	(−2.58)	(−2.64)	(−2.63)
Remuner	0.011	0.007	0.007	0.012	0.010	0.010
	(0.65)	(0.43)	(0.43)	(1.00)	(0.81)	(0.81)
CEO_Holding	−0.028	−0.022	−0.026	0.016	0.020	0.019
	(−0.33)	(−0.25)	(−0.30)	(0.24)	(0.30)	(0.28)
SOE	0.024	0.016	0.017	0.042 *	0.037	0.038
	(0.77)	(0.53)	(0.56)	(1.82)	(1.61)	(1.63)
Subsidy	2.524 ***	2.533 ***	2.525 ***	1.735 ***	1.740 ***	1.737 ***
	(4.75)	(4.78)	(4.78)	(4.06)	(4.08)	(4.08)
Tax_Benefit	0.095 ***	0.093 ***	0.092 ***	0.050 **	0.048 **	0.048 **
	(3.15)	(3.09)	(3.04)	(2.38)	(2.32)	(2.30)
Industry_HP	0.056 *	0.070 **	0.069 **	0.033 *	0.041 **	0.040 **
	(1.96)	(2.40)	(2.37)	(1.85)	(1.99)	(1.98)
MTI	0.020	0.020	0.020	0.008	0.009	0.008
	(1.54)	(1.57)	(1.53)	(0.95)	(0.98)	(0.96)
MC	0.044 ***	0.043 ***	0.043 ***	0.054 ***	0.053 ***	0.053 ***
	(4.33)	(4.23)	(4.19)	(6.91)	(6.85)	(6.83)
Cons	−7.171	−6.899	−6.959	−4.407	−4.236	−4.260
	(−0.00)	(.)	(.)	(.)	(.)	(.)
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes
City FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Adj. R2	0.45	0.45	0.45	0.35	0.36	0.36
N	30,937	30,937	30,937	30,937	30,937	30,937

Notes: t-statistics shown in parentheses. * $p<0.1$, ** $p<0.05$, *** $p<0.01$.

shows the results of the baseline regression. The dependent variables in columns 1 to 3 are CGI1, and columns 4 to 6 are CGI2. As shown in Table 2, the coefficients of AGC ($\beta =0.056$, $t=5.23$, $p<0.01$; $\beta =0.014$, $t=2.12$, $p<0.05$) in columns 1 and 4 are all significant, which verifies Hypothesis 1; that is, AGC can promote CGI. Meanwhile, we can further find that compared with the quality of CGI, the effect of AGC on the quantity of CGI is much stronger.

Columns 2 to 3 and 5 to 6 present the regression result of the moderating effect of CEID. The coefficients of CEID ($\beta =0.007$, $t=3.91$, $p<0.01$; $\beta =0.005$, $t=3.24$, $p<0.01$) in columns 2 and 5 are all significant, indicating that CEID has a direct positive effect on CGI. Therefore, the results of CEID validate the conclusion of the existing literature [20,46,47]. The coefficients of the interaction terms AGC*CEID ($\beta =0.005$, $t=3.70$, $p<0.01$; $\beta =0.002$, $t=2.55$, $p<0.05$) in columns 3 and 6 are all significant, suggesting that CEID can positively moderate the effect of AGC on CGI. Therefore, Hypothesis 2 is also supported.

4.3. Robustness Tests

The baseline regression results may be affected by endogeneity issues. For example, given the availability and richness of information, analysts may tend to focus on listed companies with larger sizes, longer listing years, better performance, more environmental protection investment, and more general innovation stock. At the same time, such listed firms may have a better information environment, fewer financing constraints, and more green innovation output. To control for endogenous concerns, we test the robustness of the previous empirical results.

4.3.1. Substituting the Core Variables

Previous studies have documented that analyst coverage is influenced by firm characteristics such as firm size, profitability, leverage, etc. [35]. In order to exclude the influence of firm characteristics on AGC, we refer to Yu’s method [35] to build the following regression model.

$$AG{C}_{i,t}={\alpha }_0+{\alpha }_{1}Siz{e}_{i,t}+{\alpha }_{2}A{\mathrm{ge}}_{i,t}+{\alpha }_{3}Tobi{n}_{i,t}+{\alpha }_{4}EPInves{t}_{i,t}+{\alpha }_{5}Total\_Paten{t}_{i,t}+{\epsilon }_{i,t}$$

(3)

In Model (3), the detailed definitions of Size, Age, Tobin, EPInvest, and Total_Patent are all exactly the same as the control variables. The residuals obtained from the regression of Model (3) reflect the AGC after excluding the firm characteristics. We denote the residuals as Net_AGC. We then substitute Net_AGC for the explanatory variable. On this basis, we further substitute the dependent variables with the CGI_Grant and CGI_Ratio, respectively. Specifically, we use the total number of green patent grants (CGI_Grant) to measure the quantity of CGI, and the ratio of cited green patents to the total number of green patent applications (CGI_Ratio) to measure the quality of CGI [1,45]. Panel A of

Table 3. Robustness test results.

	Panel A: Substituting Core Variables		Panel B: Poisson Regression		Panel C: Panel Fixed Effect Regression
	(1)	(2)	(3)	(4)	(5)	(6)
	CGI_Grant	CGI_Ratio	CGI1	CGI2	CGI1	CGI2
Net_AGC	0.048 ***	0.010 ***
	(5.26)	(4.74)
AGC			0.061 ***	0.027 **	0.103 ***	0.038 ***
			(2.99)	(2.00)	(11.04)	(5.82)
Size	0.273 ***	0.019 ***	0.507 ***	0.464 ***	0.299 ***	0.152 ***
	(18.29)	(9.41)	(14.77)	(10.44)	(14.92)	(11.39)
Lev	0.184 ***	0.018 *	0.442 ***	0.389 **	0.069 **	0.056
	(3.20)	(1.90)	(2.80)	(2.02)	(2.09)	(1.44)
Age	−0.003 *	−0.000	−0.002	−0.001	−0.350 ***	−0.077 ***
	(−1.72)	(−0.90)	(−0.34)	(−0.18)	(−20.63)	(−7.39)
Growth	0.013	0.006 ***	0.089 ***	0.091 **	0.002 **	0.001 **
	(1.15)	(2.73)	(2.71)	(2.20)	(2.12)	(1.97)
Tobin	0.014 **	0.002 *	0.007	0.017	0.004	0.001
	(2.03)	(1.76)	(0.34)	(0.68)	(0.69)	(0.25)
Cflow	−0.257 ***	−0.061 ***	−0.764 ***	−1.140 ***	−0.040 **	−0.077 ***
	(−2.86)	(−3.64)	(−2.62)	(−3.25)	(−2.54)	(−2.60)
Cash	0.081	0.037 ***	0.062	0.406	0.087	0.040
	(1.38)	(3.21)	(0.30)	(1.58)	(1.63)	(1.18)
EPInvest	0.023 ***	0.002 **	0.001	0.001	0.011 ***	0.003
	(4.91)	(2.10)	(0.17)	(0.13)	(3.09)	(0.95)
RDS	0.572 **	0.471 ***	1.312 **	2.362 ***	0.877 ***	0.941 ***
	(2.02)	(6.64)	(2.12)	(2.90)	(3.54)	(5.13)
RDS_MV	−0.009	0.017 ***	−0.049	0.001	−0.016	0.007
	(−0.46)	(4.23)	(−0.92)	(0.01)	(−0.90)	(0.59)
Total_Patent	0.156 ***	0.030 ***	0.330 ***	0.310 ***	0.108 ***	0.075 ***
	(22.63)	(24.22)	(11.65)	(9.55)	(18.23)	(16.95)
Duality	−0.008	0.006	−0.027	0.041	0.014	0.028 **
	(−0.40)	(1.55)	(−0.55)	(0.62)	(0.76)	(2.13)
Top1	−0.137 *	−0.045 ***	−0.085	−0.121	−0.107	−0.119 *
	(−1.84)	(−4.06)	(−0.48)	(−0.53)	(−0.94)	(−1.65)
Remuner	0.009	0.007 **	0.114 ***	0.148 ***	0.006	0.002
	(0.62)	(2.49)	(3.03)	(3.18)	(0.40)	(0.24)
CEO_Holding	0.006	0.014	0.040	0.199	0.110	0.019
	(0.07)	(0.83)	(0.21)	(0.74)	(1.45)	(0.34)
SOE	0.005	0.013 ***	0.090	0.005	0.068	0.045
	(0.20)	(2.91)	(1.22)	(0.05)	(1.49)	(1.40)
Subsidy	1.868 ***	0.358 ***	4.763 ***	5.768 ***	0.604 **	0.403 **
	(3.87)	(4.07)	(4.05)	(3.35)	(2.51)	(2.03)
Tax_Benefit	0.050 *	0.015 ***	0.019	0.012	0.034	0.021
	(1.81)	(2.72)	(0.24)	(0.13)	(1.45)	(1.36)
Industry_HP	0.058 **	0.004 *	0.251 ***	0.241 **	0.020 **	0.022 *
	(2.23)	(1.94)	(3.25)	(2.30)	(2.39)	(1.77)
MTI	0.023 **	0.003	0.021	0.025	0.002	0.004
	(2.05)	(1.25)	(0.60)	(0.55)	(0.15)	(0.47)
MC	0.036 ***	0.003 **	0.040 **	0.105 ***	0.032 ***	0.030 ***
	(3.85)	(2.36)	(2.02)	(4.05)	(3.84)	(5.31)
Cons	−6.320	−0.520	−31.712	−33.215	−0.862	−1.991 ***
	(−0.00)	(.)	(−0.01)	(.)	(−1.52)	(−5.70)
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes
City FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	No	No	No	No	Yes	Yes
Adj. R2/Pseudo R2 /Within R2	0.42	0.18	0.57	0.49	0.26	0.12
N	30,937	30,937	30,937	30,937	30,937	30,937

Notes: The t-statistics shown in parentheses. * $p<0.1$, ** $p<0.05$, *** $p<0.01$.

reports the results of substituting the core variables.

4.3.2. Utilizing the Poisson Regression Approach

Considering that the dependent variables (CGI1 and CGI2) are all non-negative integers, running the count model might have a better fitting effect. Therefore, we use the Poisson regression model to re-run the preamble empirical models, and the test results are reported in Panel B of Table 3.

4.3.3. Conducting the Panel Fixed Effect Regression Approach

Given that there are numerous individual fixed effects that are difficult to measure directly and do not vary over time, such as the firm’s innovation culture, we also use the panel fixed effects model to re-estimate Model (1), and the test results are reported in Panel C of Table 3.

In summary, the robustness tests based on the substitution of core variables, the Poisson regression approach, and the panel fixed effects regression approach show that there is an incentive effect of AGC on CGI.

5. Mechanism Analysis

The above results are in line with the hypothesis that AGC can promote CGI. However, the underlying mechanisms remain unclear. As the concept of environmental protection continues to gain traction, investors are increasingly factoring a company’s green practices into their investment decisions [59]. Due to stakeholders’ growing concern for corporate green information, the theoretical analysis argues that AGC can promote CGI through the information effect mechanisms and the monitoring effect mechanisms. In terms of the information effect, AGC could promote CGI by enhancing investors’ value recognition, improving corporate reputation capital, and alleviating corporate financing constraints. Meanwhile, in terms of the monitoring effect, AGC could promote CGI by reducing managerial agency costs and curbing managerial myopia. Therefore, we focus on the above influence mechanisms to investigate the underlying channels through which AGC promotes CGI.

5.1. The Information Effect of Analysts’ Green Coverage

5.1.1. Enhancing Investor Value Recognition

Following the previous literature [60,61], we use the long-term market reaction to measure investors’ recognition of firm value. Specifically, a one-year holding excess return (Ret_1Year) is used to capture the long-term market reaction.

Columns 1 to 3 of

Table 4. Mechanisms analysis results of analysts’ green coverage’s information effect.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Ret_1Year	CGI1	CGI2	Rep_Score	CGI1	CGI2	WWIndex	CGI1	CGI2
AGC	0.033 ***	0.054 ***	0.013	0.133 ***	0.045 ***	0.017 *	−0.002 ***	0.289 *	0.018
	(9.50)	(5.05)	(1.62)	(9.52)	(3.92)	(1.93)	(−6.74)	(1.90)	(1.55)
Ret_1Year		0.030 ***	0.014 **
		(2.74)	(2.55)
AGC*Ret_1Year		0.014 **	0.007 **
		(2.27)	(2.08)
Rep_Score					0.086 ***	0.051 ***
					(6.64)	(5.06)
AGC*Rep_Score					0.021 ***	0.007 **
					(4.60)	(2.19)
WWIndex								−0.897 ***	−0.436 **
								(−3.38)	(−2.30)
AGC* WWIndex								0.338 **	0.091 *
								(2.29)	(1.81)
Size	0.015 ***	0.309 ***	0.178 ***	0.604 ***	0.254 ***	0.155 ***	−0.048 ***	0.273 ***	0.164 ***
	(4.64)	(19.34)	(14.61)	(25.73)	(12.99)	(10.09)	(−124.72)	(12.70)	(10.11)
Lev	0.077 ***	0.195 ***	0.118 ***	0.305 ***	0.133 *	0.088	0.040 ***	0.145 **	0.104 **
	(5.07)	(3.10)	(2.64)	(3.29)	(1.79)	(1.61)	(22.24)	(2.09)	(2.09)
Age	−0.002 ***	−0.005 **	−0.002	0.020 ***	−0.006 ***	−0.003 *	0.001 ***	−0.005 **	−0.002
	(−5.92)	(−2.19)	(−1.44)	(7.21)	(−2.66)	(−1.86)	(11.03)	(−2.13)	(−1.12)
Growth	0.102 ***	0.023 *	0.006	0.208 ***	−0.005	−0.013	−0.047 ***	−0.015	−0.015
	(14.84)	(1.87)	(0.66)	(10.50)	(−0.33)	(−1.25)	(−56.51)	(−0.85)	(−1.11)
Tobin	0.071 ***	0.017 **	0.016 ***	0.074 ***	0.014	0.012 *	0.001 ***	0.030 ***	0.021 ***
	(21.69)	(2.25)	(2.95)	(6.87)	(1.61)	(1.85)	(5.59)	(3.59)	(3.43)
Cflow	0.330 ***	−0.399 ***	−0.307 ***	1.213 ***	−0.567 ***	−0.400 ***	−0.121 ***	−0.498 ***	−0.378 ***
	(8.76)	(−3.99)	(−4.36)	(8.26)	(−4.98)	(−4.78)	(−37.01)	(−4.59)	(−4.82)
Cash	0.056 ***	0.225 ***	0.220 ***	0.091	0.275 ***	0.239 ***	−0.012 ***	0.256 ***	0.244 ***
	(2.95)	(3.39)	(4.39)	(0.94)	(3.57)	(4.00)	(−6.07)	(3.47)	(4.29)
EPInvest	0.004 ***	0.025 ***	0.014 ***	0.009	0.023 ***	0.010 **	−0.000	0.023 ***	0.011 **
	(2.96)	(4.79)	(3.15)	(1.45)	(3.94)	(2.12)	(−1.00)	(4.09)	(2.38)
RDS	0.318 ***	1.114 ***	1.375 ***	1.525 ***	0.861 **	1.387 ***	0.003	0.848 **	1.317 ***
	(3.33)	(3.37)	(5.25)	(3.46)	(2.24)	(4.42)	(0.31)	(2.36)	(4.59)
RDS_MV	0.022 ***	0.007	0.064 ***	−0.019	0.020	0.068 ***	0.002 ***	0.011	0.059 ***
	(3.15)	(0.33)	(3.90)	(−0.59)	(0.83)	(3.69)	(3.65)	(0.47)	(3.44)
Total_Patent	0.007 ***	0.231 ***	0.147 ***	0.078 ***	0.240 ***	0.147 ***	−0.000 ***	0.240 ***	0.148 ***
	(3.93)	(28.62)	(23.21)	(7.76)	(26.45)	(20.55)	(−2.84)	(27.91)	(22.22)
Duality	−0.001	0.004	0.025	−0.020	0.017	0.029	0.001	0.015	0.030 *
	(−0.24)	(0.16)	(1.56)	(−0.67)	(0.69)	(1.56)	(1.55)	(0.65)	(1.67)
Top1	0.076 ***	−0.223 ***	−0.156 ***	0.136	−0.192 **	−0.141 **	−0.012 ***	−0.246 ***	−0.168 ***
	(5.32)	(−2.76)	(−2.60)	(1.30)	(−2.19)	(−2.12)	(−6.04)	(−2.89)	(−2.62)
Remuner	0.010 ***	0.011	0.012	0.162 ***	−0.015	−0.002	−0.004 ***	−0.006	0.005
	(2.59)	(0.62)	(0.98)	(6.52)	(−0.79)	(−0.16)	(−9.57)	(−0.32)	(0.36)
CEO_Holding	0.111 ***	−0.031	0.015	−0.165	−0.034	0.034	−0.008 ***	−0.056	0.014
	(5.01)	(−0.36)	(0.23)	(−1.49)	(−0.35)	(0.45)	(−4.01)	(−0.59)	(0.18)
SOE	−0.013 **	0.024	0.042 *	−0.039	0.034	0.047 *	−0.002 ***	0.029	0.043 *
	(−2.37)	(0.78)	(1.82)	(−1.00)	(1.02)	(1.84)	(−2.64)	(0.90)	(1.74)
Subsidy	−0.105	2.531 ***	1.738 ***	2.064 ***	2.638 ***	1.888 ***	−0.047 ***	2.225 ***	1.697 ***
	(−0.75)	(4.76)	(4.07)	(2.82)	(4.51)	(3.98)	(−3.37)	(3.91)	(3.67)
Tax_Benefit	0.004	0.096 ***	0.050 **	0.167 ***	0.120 ***	0.058 **	−0.016 ***	0.107 ***	0.053 **
	(0.35)	(3.15)	(2.38)	(3.20)	(2.99)	(2.05)	(−13.04)	(3.41)	(2.45)
Industry_HP	0.019 ***	0.057 **	0.033	0.098 **	0.046	0.027	0.002 **	0.047	0.028
	(3.35)	(1.98)	(1.61)	(2.14)	(1.43)	(1.17)	(2.44)	(1.54)	(1.26)
MTI	0.008 **	0.020	0.008	0.017	0.019	0.008	−0.001	0.021	0.008
	(2.01)	(1.53)	(0.94)	(0.66)	(1.28)	(0.78)	(−1.06)	(1.53)	(0.85)
MC	0.009 ***	0.044 ***	0.054 ***	0.023 *	0.057 ***	0.062 ***	0.000 *	0.050 ***	0.059 ***
	(3.84)	(4.31)	(6.90)	(1.71)	(4.85)	(6.83)	(1.71)	(4.39)	(6.67)
Cons	−0.600	−7.166	−4.405	−7.463 ***	−5.595 ***	−3.704 ***	0.102 ***	−6.977 ***	−4.327 ***
	(−0.00)	(.)	(.)	(−33.77)	(−12.94)	(−10.98)	(7.33)	(−17.16)	(−14.22)
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
City FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Adj. R2	0.13	0.45	0.35	0.78	0.48	0.37	0.87	0.47	0.37
N	30,937	30,937	30,937	24,277	24,277	24,277	26,974	26,974	26,974

Notes: t-statistics shown in parentheses. * $p<0.1$, ** $p<0.05$, *** $p<0.01$.

report the channel test results of AGC’s enhancing investors’ value recognition. The coefficient of AGC ($\beta =0.033$, $t=9.50$, $p<0.01$) in column 1 is significant, which confirms that AGC can induce investors’ value recognition of the following firm. Therefore, the hypothesis of analysts’ enhancing investors’ value recognition is supported [62,63]. The coefficients of the interaction terms AGC*Ret_1Year ($\beta =0.014$, $t=2.27$, $p<0.05$; $\beta =0.007$, $t=2.08$, $p<0.05$) in columns 2 and 3 are all significant, which indicates that AGC could promote CGI by enhancing investors’ value recognition. In other words, the test results support the investor value recognition channel.

5.1.2. Improving Corporate Reputation Capital

Based on the ranking system of domestic and international corporate reputation lists, such as Fortune Magazine’s Most Admired Respected U.S. Companies, and the corporate reputation evaluation system developed by Guan and Zhang (2019) [64], we calculate the reputation capital of listed companies. Specifically, we select 12 indicators to construct corporate reputation capital from the consumer and social perspective, the creditor perspective, the shareholder perspective, and the firm perspective. From the consumer and social perspective, we select the ranking of assets, sales, net profit, and value in the industry; from the creditor perspective, we select the leverage ratio, current ratio, and long-term debt ratio; from the shareholder perspective, we select the EPS, dividend per share, and whether the company is audited by a “Big Four” accounting firm; from the firm perspective, we select the sustainable growth rate and the ratio of independent directors. Then, we perform the factor analysis based on the above indicators and obtain the company’s reputation capital score (Rep_Score).

Columns 4 to 6 of Table 4 report the channel test results of AGC’s improving firms’ reputation capital. The coefficient of AGC ($\beta =0.133$, $t=9.52$, $p<0.01$) in Column 4 is significant, indicating that AGC can directly improve the firms’ reputation capital, thus supporting the analysts’ improving firms’ reputation capital hypothesis [65,66]. The coefficients of the interaction terms AGC*Rep_Score ($\beta =0.021$, $t=4.60$, $p<0.01$; $\beta =0.007$, $t=2.19$, $p<0.05$) in columns 5 and 6 are all significant, indicating that AGC could promote CGI by improving corporate reputation capital. Therefore, the corporate reputation capital channel is also supported.

5.1.3. Alleviating Corporate Financing Constraints

Using the WW index (WWIndex) based on dynamic structural estimation methods [67], we measured the corporate financing constraints and examined the channel of AGC in alleviating corporate financing constraints. The construction method of the WW index is more consistent with the financing constraint concept itself, and the positive correlation between equity financing constraint and debt financing constraint is also more consistent with the general intuition [68].

Columns 7 to 9 of Table 4 report the channel test results of AGC’s alleviation of corporate financing constraints. The coefficient of AGC ($\beta =-0.002$, $t=-6.74$, $p<0.01$) in column 7 is significant, indicating that AGC can directly reduce corporate financing constraints, thus supporting the hypothesis of analysts’ alleviating corporate financing constraints [69,70]. The coefficients of the interaction terms AGC*WWIndex ($\beta =0.338$, $t=2.29$, $p<0.05$; $\beta =0.091$, $t=1.81$, $p<0.10$) in columns 8 and 9 are all significant, which can verify that AGC can promote CGI by alleviating the firms’ financing constraints. Therefore, the test results support the alleviation of corporate financing constraints channel.

5.2. The Monitoring Effect of Analysts’ Green Coverage

5.2.1. Reducing Managerial Agency Costs

As a necessary product of the firm’s normal operation and contractual incompleteness, managerial perquisites have a certain degree of reasonableness. However, once the normal amount is exceeded, managerial perquisites are often transformed into managerial private interests [71,72]. Recent literature has highlighted that excess perquisites can be an effective proxy for managerial agency costs [73,74]. Therefore, we use the excess perquisites (Experk) to measure managerial agency costs. Specifically, the calculation model is as follows.

$$\begin{array}{l}\frac{Per{k}_{i,t}}{Asse{t}_{i,t-1}}={\alpha }_0+{\alpha }_1\frac{1}{Asse{t}_{i,t-1}}+{\alpha }_2\frac{\Delta Sal{e}_{i,t}}{Asse{t}_{i,t-1}}+{\alpha }_3\frac{PP{E}_{i,t}}{Asse{t}_{i,t-1}}+{\alpha }_4\frac{Inventor{y}_{i,t}}{Asse{t}_{i,t-1}}\\ +{\alpha }_{5}lnEmploye{e}_{i,t}+{\epsilon }_{i,t}\end{array}$$

(4)

In Model (4), Perk represents management perquisites (equal to G&A expenses less management compensation, amortization of long-term amortized expenses, and amortization of intangible assets), Asset represents firm assets; $\Delta \mathrm{Sale}$ represents changes in sales; PPE represents property, plant, and equipment; Inventory represents firm inventory; and lnEmployee represents the number of employees. Then, by regressing model (6) by industry and year, we obtain the regression residuals, and we name the residuals management excess perquisites (Experk).

Columns 1 to 3 of

Table 5. Mechanism analysis results of analysts’ green coverage monitoring effect.

	(1)	(2)	(3)	(4)	(5)	(6)
	ExPerk	CGI1	CGI2	Myopia	CGI1	CGI2
AGC	−0.001 ***	0.071 ***	0.024 ***	−0.002 ***	0.039 ***	0.008 *
	(−4.72)	(6.57)	(2.87)	(−3.46)	(2.89)	(1.72)
ExPerk		−0.786 ***	−0.56 9 **
		(−2.68)	(−2.46)
AGC*ExPerk		0.135 **	0.042 *
		(2.19)	(1.75)
Myopia					−0.276 ***	−0.189 ***
					(−2.82)	(−2.73)
AGC*Myopia					0.198 **	0.080 **
					(2.43)	(2.11)
Size	−0.006 ***	0.323 ***	0.186 ***	−0.001	0.307 ***	0.175 ***
	(−14.37)	(19.03)	(14.46)	(−0.86)	(19.23)	(14.47)
Lev	0.010 ***	0.203 ***	0.130 ***	0.007	0.215 ***	0.133 ***
	(5.67)	(3.05)	(2.71)	(1.44)	(3.43)	(3.02)
Age	0.000 ***	−0.002	−0.001	0.001 ***	−0.004 *	−0.002
	(6.74)	(−0.86)	(−0.52)	(8.04)	(−1.86)	(−1.03)
Growth	−0.000	0.033 **	0.016	−0.010 ***	0.022 *	0.007
	(−0.37)	(2.45)	(1.60)	(−9.28)	(1.72)	(0.69)
Tobin	0.000	0.035 ***	0.025 ***	−0.001	0.019 **	0.016 ***
	(0.86)	(4.23)	(4.30)	(−0.81)	(2.46)	(3.04)
Cflow	0.023 ***	−0.363 ***	−0.290 ***	0.012	−0.392 ***	−0.314 ***
	(6.77)	(−3.35)	(−3.81)	(1.39)	(−3.92)	(−4.46)
Cash	0.004 *	0.075	0.150 **	−0.019 ***	0.225 ***	0.228 ***
	(1.69)	(0.98)	(2.55)	(−3.75)	(3.41)	(4.62)
EPInvest	−0.000 *	0.021 ***	0.012 **	−0.001 ***	0.025 ***	0.014 ***
	(−1.78)	(3.54)	(2.37)	(−2.58)	(4.69)	(3.12)
RDS	0.142 ***	1.469 ***	1.507 ***	−0.018	1.149 ***	1.387 ***
	(11.26)	(4.25)	(5.50)	(−0.89)	(3.46)	(5.28)
RDS_MV	0.005 ***	0.065 ***	0.017	0.005 ***	0.006	0.057 ***
	(6.74)	(2.68)	(0.97)	(3.11)	(0.27)	(3.55)
Total_Patent	0.002 ***	0.254 ***	0.159 ***	−0.001 *	0.227 ***	0.143 ***
	(10.50)	(28.56)	(23.13)	(−1.72)	(28.23)	(22.76)
Duality	−0.001	0.003	0.028	−0.002	0.003	0.026
	(−0.88)	(0.13)	(1.63)	(−1.09)	(0.12)	(1.59)
Top1	0.004	−0.269 ***	−0.189 ***	−0.001	−0.244 ***	−0.181 ***
	(1.62)	(−3.17)	(−2.97)	(−0.17)	(−3.02)	(−3.06)
Remuner	0.006 ***	0.006	0.005	−0.001	0.011	0.013
	(11.26)	(0.32)	(0.34)	(−1.29)	(0.66)	(1.05)
CEO_Holding	−0.002	−0.059	0.009	−0.022 ***	−0.036	0.014
	(−0.66)	(−0.62)	(0.13)	(−4.05)	(−0.42)	(0.22)
SOE	0.002 **	0.033	0.051 **	0.010 ***	0.022	0.041 *
	(2.01)	(1.04)	(2.11)	(4.42)	(0.72)	(1.77)
Subsidy	−0.005	2.048 ***	1.593 ***	−0.023	2.515 ***	1.727 ***
	(−0.36)	(3.64)	(3.50)	(−0.67)	(4.70)	(4.02)
Tax_Benefit	−0.001	0.020	0.003	−0.001	0.093 ***	0.049 **
	(−0.96)	(0.61)	(0.11)	(−0.30)	(3.05)	(2.33)
Industry_HP	−0.002 ***	0.053 *	0.030	0.011 ***	0.053 *	0.030
	(−2.93)	(1.71)	(1.34)	(5.37)	(1.83)	(1.48)
MTI	0.001	0.027 *	0.012	−0.002 *	0.017	0.005
	(1.33)	(1.93)	(1.23)	(−1.71)	(1.33)	(0.64)
MC	0.000	0.030 ***	0.047 ***	−0.003 ***	0.043 ***	0.053 ***
	(0.35)	(2.70)	(5.55)	(−4.47)	(4.18)	(6.83)
Cons	0.020 *	−5.545 ***	−4.185 ***	0.129 ***	−7.963 ***	−4.455 ***
	(1.84)	(−12.95)	(−13.04)	(5.86)	(−24.26)	(−17.27)
Industry FE	Yes	Yes	Yes	Yes	Yes	Yes
City FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Adj. R2	0.18	0.46	0.36	0.15	0.45	0.35
N	26,846	26,846	26,846	30,253	30,253	30,253

Notes: t-statistics shown in parentheses. * $p<0.1$, ** $p<0.05$, *** $p<0.01$.

report the channel test results of AGC’s reduction of management agency costs. The coefficient of AGC ($\beta =-0.001$, $t=-4.72$, $p<0.01$) in column 1 is significant, indicating that AGC can directly reduce the management agency costs, thus supporting the hypothesis of analysts’ reduction of management agency costs [75,76]. The coefficients of the interaction terms AGC*Experk ($\beta =0.135$, $t=2.19$, $p<0.05$; $\beta =0.042$, $t=1.75$, $p<0.10$) in columns 2 and 3 are all significant, which can verify that AGC can promote CGI by reducing management agency costs. In a word, the test results support the management agency’s cost reduction channel.

5.2.2. Curbing Managerial Myopia

Using textual analysis and machine learning methods, Hu et al. (2021) [77] proposed a novel textual metric that directly quantifies managerial myopia. Compared with other proxy variables that suffer from high noise, the textual measure proposed by Hu et al. (2021) [77] is able to capture managerial myopia more effectively and accurately. Therefore, with reference to previous research [77,78], we use the textual analysis approach based on the MD&A content in the companies’ annual reports to measure managerial myopia (Myopia). The raw data of managerial myopia are obtained from the WinGO Finance text data platform.

Columns 4 to 6 of Table 5 report the channel test results of AGC’s curbing managerial myopia. The coefficient of AGC ($\beta =-0.002$, $t=-3.46$, $p<0.01$) in column 4 is significant, indicating that AGC can directly curb the firm’s managerial myopia [79]. The coefficients of the interaction terms AGC*Myopia ($\beta =0.198$, $t=2.43$, $p<0.05$; $\beta =0.080$, $t=2.11$, $p<0.05$) in columns 5 and 6 are all significant, which can confirm that AGC can promote CGI by curbing managerial myopia. Therefore, the test results also support the curbing managerial myopia channel.

To sum up, the three information effect channels and two monitoring effect channels proposed in the theoretical analysis have all been empirically verified.

6. Heterogeneity Analysis

To extend our research findings, we further conduct the heterogeneity analysis from the perspective of regional institutional environment, industry characteristics, and corporate innovation strategy. First, a good regional institutional environment (i.e., highly developed market intermediaries, sound legal environment) is beneficial to the positive role of securities analysts in the capital market and facilitates CGI performance [80,81]. Thus, we examine the influence of regional institutional environment heterogeneity. Second, due to strict regulation by local governments and focused attention from capital market analysts and investors, listed firms belonging to heavily polluting industries have a stronger incentive to seek organizational legitimacy by investing more in CGI activities [46,82]. Therefore, we also examine the influence of industry characteristics. Third, corporate innovation strategy is important for the sustainability of green innovation projects [83,84]. Firms that adopt continuous innovation strategies may achieve larger and higher quality CGI outputs [85]. Accordingly, we further examine the influence of a firm’s continuous innovation strategy.

6.1. Regional Institutional Environment

Referring to Fan et al. (2011) [86], we use the LawIndex from the CMI database to measure the quality of the regional institutional environment. More specifically, the LawIndex is a component of the marketization total index (MTI). On the basis of the annual median of the LawIndex, we divide the sample into two subsamples.

The results of the heterogeneity test based on the LawIndex are shown in Panel A of

Table 6. Heterogeneity analysis results.

Panel A: The influence of regional institutional environment
	LawIndex_High	LawIndex_Low	LawIndex_High	LawIndex_Low
	(1)	(2)	(3)	(4)
	CGI1	CGI1	CGI2	CGI2
AGC	0.063 ***	0.036 **	0.015 *	0.002
	(4.96)	(2.20)	(1.81)	(1.16)
Control variables	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
City FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Adj. R2	0.49	0.43	0.39	0.32
N	13,712	17,225	13,712	17,225
SUR Coefficient Diff.	p = 0.059		p = 0.128
Panel B: The influence of heavily polluting industries
	Industry_HP	Industry_Other	Industry_HP	Industry_Other
	(5)	(6)	(7)	(8)
	CGI1	CGI1	CGI2	CGI2
AGC	0.109 ***	0.030 **	0.041 ***	0.002
	(6.73)	(2.25)	(3.28)	(0.83)
Control variables	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
City FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Adj. R2	0.51	0.47	0.38	0.38
N	8628	22,309	8628	22,309
SUR Coefficient Diff.	p = 0.000		p = 0.000
Panel C: The influence of corporate continuous innovation strategy
	InnoStrategy_Con	InnoStrategy_Other	InnoStrategy_Con	InnoStrategy_Other
	(9)	(10)	(11)	(12)
	CGI1	CGI1	CGI2	CGI2
AGC	0.013 *	0.062 ***	0.001	0.009 **
	(1.77)	(4.93)	(1.08)	(1.98)
Control variables	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
City FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Adj. R2	0.49	0.43	0.39	0.33
N	11,448	19,489	11,448	19,489
SUR Coefficient Diff.	p = 0.001		p = 0.076

Notes: t-statistics shown in parentheses. * $p<0.1$, ** $p<0.05$, *** $p<0.01$.

. The coefficient of AGC ($\beta =0.063$, $t=4.96$, $p<0.01$) in column 1 is significant and is larger than the coefficient of AGC ($\beta =0.036$, $t=2.20$, $p<0.05$) in column 2, indicating that the positive effect of AGC on CGI is more significant for firms located in the high legal environment quality region. The results in columns 3 and 4 can also draw the same conclusion.

6.2. Heavily Polluting Industries

Following the previous studies [46,82], we set an indicator variable that can measure whether a firm belongs to a heavily polluting industry (Industry_HP). In detail, the Industry_HP variable equals one when the industry code is B06, B07, B08, B09, B10, C15, C17, C18, C19, C22, C25, C26, C28, C29, C30, C31, C32, C33 or D44, and zero otherwise.

The results of the heterogeneity test based on Industry_HP are reported in Panel B of Table 6. The coefficient of AGC ($\beta =0.109$, $t=6.73$, $p<0.01$) in column 5 is significant, and it is larger than the coefficient of AGC ($\beta =0.030$, $t=2.25$, $p<0.05$) in column 6, indicating that the positive effect of AGC on CGI is more significant for firms belonging to the highly polluting industries. Columns 7 and 8 can also confirm the same conclusion.

6.3. Continuous Innovation Strategy

Based on the previous research [87,88], we construct an indicator variable that can capture whether a firm has adopted continuous innovation strategies (InnoStrategy_Con). In detail, we consider a firm to have adopted a continuous innovation strategy if it has maintained a steady growth in R&D investment over the past three years and assign the variable InnoStrategy_Con a value of one and zero otherwise.

The results of the heterogeneity test based on InnoStrategy_Con are shown in Panel C of Table 6. The coefficient of AGC ($\beta =0.062$, $t=4.93$, $p<0.01$) in column 10 is significant, and it is larger than the coefficient of AGC ($\beta =0.013$, $t=1.77$, $p<0.10$) in column 9, indicating the incentive effect of AGC on CGI is more significant for firms that have not adopted continuous innovation strategies. Columns 11 and 12 can also validate this conclusion. In other words, firms’ continuous innovation strategy has a substitution effect on the process of AGC affecting CGI.

7. Conclusions

Recently, the influence of analysts’ coverage on firms’ innovation has received increasing focus from scholars and practitioners. However, previous studies have primarily focused on financial analysts’ coverage and its impact on financial performance, with limited attention given to the influence of analysts’ coverage on environmental performance. To address this gap, we investigate the influence of AGC on CGI in China, using the textual analysis approach. Our results find that AGC has a significantly positive correlation with CGI. Furthermore, we divide CGI into two dimensions: CGI quality and CGI quantity. The results indicate that, compared with the quality of CGI, AGC has a stronger promotion effect on the quantity of CGI. These findings suggest that in the area of environmental performance, analysts’ coverage can play a crucial guiding role in CGI. Moreover, we also find that CEID has a positive moderating effect on the relationship between AGC and CGI. This suggests that firms with good quality CEID will benefit from AGC and promote more CGI. We propose two influence mechanisms of AGC on CGI: the information effect and the monitoring effect. The information effect enhances investors’ value recognition and improves corporate reputation capital to alleviate corporate financing constraints. The monitoring effect reduces management agency costs and inhibits management myopia. These influence mechanisms help explain why AGC has a positive impact on CGI. Finally, we perform a cross-sectional regression analysis on the main effects, considering regional institutional environment, industry characteristics, and corporate innovation strategy. Our results show that the incentive effect of AGC influencing CGI is more significant for firms operating in regions with a favorable institutional environment, firms belonging to heavily polluting industries, and firms that have not adopted continuous innovation strategies. This highlights the importance of considering firm, industry, and regional characteristics when examining the effect of AGC on CGI.

Our research has three practical implications. First, regulators such as the China Securities Regulatory Commission should continuously improve the institutional construction of the securities analyst industry, actively guide securities analysts to better play their information disclosure and information transmission roles, and focus on strengthening the function of securities analysts in disseminating corporate non-financial information to investors, and further enhance the efficiency of capital market information. Second, analysts can proactively communicate with the management of listed companies, actively expand the channels for obtaining information on CGI, improve the reliability of investment ratings, and thus effectively promote the implementation of CGI strategies. Third, listed companies should continue to strengthen the voluntary disclosure of environmental information, and proactively demonstrate to external information users their efforts to reduce environmental risks and fulfill their environmental social responsibility, thereby gaining value recognition from investors and promoting the performance of CGI.

Nevertheless, there are some drawbacks to our research. First, the study was conducted within a single emerging market context (i.e., mainland China). Therefore, in order to generalize the findings to other countries, it is necessary to understand the specific characteristics of these countries. Further research could replicate this study in other emerging markets or developed countries. Based on this, future research could also consider conducting comparative research on countries with different institutional contexts [89]. Second, the analysts’ green coverage (AGC) used in the study mainly takes into account the total number of analysts’ research reports on “green” or “innovation” for a given company, which only captures AGC in a relatively crude way. In fact, the specific information content of analysts’ research reports that reflect firms’ green technologies, green investments, and sustainable innovations may more accurately capture AGC. Future research could explore more comprehensive measures of AGC over a long time. Finally, the emergence of the modern concept of CSR has led to the development of comprehensive CSR indicators to measure the advantages and disadvantages of corporate business activities, replacing the simple economic indicators used in the early stages [90]. Li et al. (2018) [91] documented a significantly positive correlation between media supervision and CSR performance. However, it is currently unknown whether analyst coverage, particularly on specific content, has an influence on CSR performance. As a result, future research can explore several ways to promote the fulfillment of CSR to achieve sustainable development goals in a similar research context to this paper. For instance, it would be beneficial to investigate the potential interactions between analyst reports and CSR reports, with a particular focus on how CSR implementation strategies are modified in response to the specific information brought to light by financial analysts.

Our research provides several marginal contributions to the existing research on the relationship between AGC and CGI in emerging markets, using the textual analysis of analyst research reports’ coverage of “green” or “innovation”. Our findings confirm the strategic value of AGC in facilitating CGI performance. Furthermore, we reveal that the positive impact of AGC is strengthened by the quality of CEID, which can highlight the important role of analysts serving as the information amplifiers of the capital market. These empirical results have some practical implications for governments to optimize policies, for analysts to improve their performance, and for firms to adjust their green innovation and environmental disclosure strategies.