Green Credit Policy, Analyst Attention, and Corporate Green Innovation
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis paper addresses the effects of green credit policy on firms' likelihood of undertaking green innovation. This is an important topic in this era of man-made climate change that threatens future life on the planet. The authors have done a good job in providing a rigorous analysis of the Chinese experience with green credit policy. Theoretical issue are well covered resulting in the formulation of appropriate hypotheses. The methodology is robust and enables the authors to answer a list of relevant questions probing the relationship between green credit policy and firms' responses. The paper is clearly written and communicates well to the reader. I strongly recommend this paper for publication.
Two minor typos - p2,line50 - should be ''emphasize". p3,line98 - should be upper case first letter for "Steininger"
Author Response
Thank you for your recognition of our work. We have taken your suggestions into account and corrected the two writing errors. Thank you again for your feedback, and I wish you a wonderful day.
Reviewer 2 Report
Comments and Suggestions for AuthorsI am very glad to review this manuscript titled “Green Credit Policy, Analyst Attention, and Corporate Green Innovation”. The research question is clear and the structure is relatively complete, but there are several limitations as follows:
- The design of the Difference-in-Differences (DID) model is unreasonable. Firstly, if whether an enterprise receives green credit support is considered as the impact of green credit policies, then the staggered DID model should be constructed. Subsequent research would then need to consider examining parallel trends through event study methods or honest DID, conducting Bacon decomposition, and performing negative weight tests et al. If it is treated as a policy implemented at a specific point in time, where all studied samples receive the policy, then there is no control group, and thus there is no need to construct a DID model. Secondly, the article uses heavily polluting enterprises as the treatment group and non-heavily polluting enterprises as the control group. This design, on the one hand, requires that all heavily polluting enterprises be affected by green credit policies, while non-heavily polluting enterprises are not affected by the policy, which is unlikely. On the other hand, heavily polluting enterprises and non-heavily polluting enterprises are completely different samples, meaning that it is impossible to construct a counterfactual framework. Therefore, it is suggested that the author classify heavily polluting enterprises based on whether they are affected by the policy to construct the DID model.
- Due to the high probability of self-selection bias and reverse causality issues, it is recommended to conduct an endogeneity test.
- If possible, use a mediation effect model to explore the mechanism through which green credit policy influences green innovation.
- Refine the "Recommendations" to better align them with the conclusions of the article and to provide more specificity.
- Models 3-2 and 6-1 seem to be duplicative, and it is necessary to check the subscript of 𝑎𝑓𝑛𝑢𝑚 to see if it represents analyst attention for firm i.
- Some language expressions and content in the article need to be checked, such as line 166.
some details in the language should be checked.
Author Response
Thank you very much for your professional advice. We have made modifications based on your feedback, as follows:
1. The design of the Difference-in-Differences (DID) model is unreasonable. Firstly, if whether an enterprise receives green credit support is considered as the impact of green credit policies, then the staggered DID model should be constructed. Subsequent research would then need to consider examining parallel trends through event study methods or honest DID, conducting Bacon decomposition, and performing negative weight tests et al. If it is treated as a policy implemented at a specific point in time, where all studied samples receive the policy, then there is no control group, and thus there is no need to construct a DID model. Secondly, the article uses heavily polluting enterprises as the treatment group and non-heavily polluting enterprises as the control group. This design, on the one hand, requires that all heavily polluting enterprises be affected by green credit policies, while non-heavily polluting enterprises are not affected by the policy, which is unlikely. On the other hand, heavily polluting enterprises and non-heavily polluting enterprises are completely different samples, meaning that it is impossible to construct a counterfactual framework. Therefore, it is suggested that the author classify heavily polluting enterprises based on whether they are affected by the policy to construct the DID model.
R:Thank you very much for your opinion. Your opinion is very professional. You pointed out that there was an issue with the setting of the processing group in our previous use of the DID model. We realize that our previous version did not provide a clear description of why polluting enterprises were designated as treatment groups, which caused your misunderstanding.
In this article, the policy shock comes from the "Key Evaluation Indicators for the Implementation of Green Credit" document released by the Chinese government in 2012, which identifies whether listed companies belong to industries with environmental and social risks classified as Class A as restricted industries for green credit. Specifically, A-class enterprises belong to nine industries, including nuclear power generation, hydropower generation, water conservancy and inland port engineering construction, coal mining and washing industry, oil and gas mining industry, ferrous metal mining and beneficiation industry, non-ferrous metal mining and beneficiation industry, non-metallic mining and beneficiation industry, and other mining industries. Therefore, the impact of the policy is only targeted at these 9 industries. So your understanding is correct, when a company belongs to these 9 industries, it is a processing group, and vice versa. When it does not belong to these 9 industries, it is the control group. In addition, the time difference is based on the period before and after 2012.
Based on your feedback, we have realized that there were issues in our previous version. Therefore, we have provided a more detailed description of the Independent Variable in section 3.1 of the revised version, as follows:
Referring to the "Key Evaluation Indicators for the Implementation of Green Credit" document released by the Chinese government in 2012, the industry to which enterprises with environmental and social risks classified as Class A belong is identified as a restricted industry for green credit. Specifically, A-class enterprises belong to nine heavily polluting industries, including nuclear power generation, hydropower generation, water conservancy and inland port engineering construction, coal mining and washing industry, oil and gas mining industry, ferrous metal mining and beneficiation industry, non-ferrous metal mining and beneficiation industry, non-metallic mining and beneficiation industry, and other mining industries. When a company belongs to these 9 industries, it is a processing group, and vice versa. When it does not belong to these 9 industries, it is the control group.
At the same time, we have added references to existing literature that use this approach to increase persuasiveness.
2. Due to the high probability of self-selection bias and reverse causality issues, it is recommended to conduct an endogeneity test.
R:Thank you for your suggestion. We have further improved the discussion on endogeneity issues. Firstly, in section 5.2 of the article, we added that the research question may face endogeneity issues caused by sample selection problems. We mainly adopt PSM-DID to solve this problem. The modifications are as follows:
The impact of green credit policy on corporate green innovation may face endogeneity issues caused by sample selection bias. To address this, we use the PSM-DID method to mitigate endogeneity concerns. The PSM-DID approach combines Propensity Score Matching (PSM) with the Difference-in-Differences (DID) method, matching high-pollution and non-high-pollution enterprises based on their background characteristics before and after the policy implementation. This method enables a more accurate estimation of the causal relationship of the policy.
In addition, we have added section 5.4 on policy exogeneity testing to supplement the explanation of policy exogeneity. Specifically, as follows:
As a quasi natural experiment, it is required that there is no endogeneity issue between policy shocks and experimental groups, that is, policy shocks must meet the conditions of exogeneity. Considering that companies may anticipate the introduction of policies and change their business behavior in advance, so that policy shocks no longer meet exogenous requirements. This study further introduced robustness testing by constructing an interaction term did_one and incorporating it into the regression model. Specifically, the did_one variable is defined as a time dummy variable, representing the year before the implementation of the smart city pilot policy. As shown in Table 5, the empirical results indicate that the coefficient of did_one is not statistically significant, while the coefficient of is still significant and positive. These findings confirm that expected effects do not confuse policy evaluations, further validating the robustness and reliability of the main conclusions.
3. If possible, use a mediation effect model to explore the mechanism through which green credit policy influences green innovation.
R:We fully understand your suggestion. In this article, we use the form of constructing interaction terms to illustrate the moderating effect of analyst attention. In our research, we hope to demonstrate that this is a moderating effect or that it plays a mediating role. Therefore, we can fully understand your suggestion by using the mediation effect model. The reason why we consider using interaction terms for mechanism analysis is mainly based on the DID framework, which makes the operation more convenient. In addition, the mediation model is considered in some studies to introduce new endogeneity issues, although this issue is still controversial. We greatly appreciate your professional advice and fully understand it.
4. Refine the "Recommendations" to better align them with the conclusions of the article and to provide more specificity.
R:Thank you for your suggestion. We have rewritten the suggestion section of the article as follows:
The results of this study indicate that green credit policies have a significant positive effect on high-pollution industries and large enterprises, particularly state-owned enterprises. Therefore, policymakers should further optimize green credit policies to enhance their targeting. For high-pollution industries, the guidance of green credit should be strengthened by establishing clear environmental standards and corporate performance evaluation systems to ensure that loan funds are directed toward enterprises genuinely committed to green transformation. Additionally, the government can collaborate with financial institutions to establish dedicated green credit funds, providing long-term, low-interest green loans to eligible enterprises, thereby reducing their financing costs and enhancing their motivation for green innovation.
For small and medium-sized enterprises (SMEs) and non-state-owned enterprises, which often face greater financial pressure and financing constraints, the government should provide more tailored support. For example, a green financing guarantee fund could be established to offer loan guarantees for SMEs lacking sufficient collateral, thereby lowering their financing barriers. At the same time, tax incentive policies could be introduced, such as allowing additional deductions for R&D expenses or granting direct tax reductions for companies meeting green innovation standards, encouraging them to increase investment in green technologies. Furthermore, the government could guide large enterprises to collaborate with SMEs in green innovation, facilitating technology sharing and joint innovation to help smaller firms overcome technological and financial constraints.
This study also finds that analyst attention plays a crucial role in promoting corporate green innovation. Therefore, policymakers can take measures to increase capital market attention to green innovation. For instance, a green enterprise rating system could be established, incorporating green innovation performance into public company disclosure requirements and publishing corporate green innovation reports regularly. These initiatives would raise awareness among investors and analysts about green innovation. Additionally, the government could encourage financial institutions and securities analysts to conduct specialized research on green innovation firms, channeling market resources toward green enterprises and further enhancing the effectiveness of green credit policies.
Moreover, it is recommended to expand the coverage of green credit policies beyond traditional high-pollution industries to other sectors with significant environmental impacts. For example, the manufacturing, transportation, and construction industries are also major contributors to carbon emissions and resource consumption. Extending green credit policies to these sectors would further promote the green transition of the entire economic system. The government can tailor industry-specific green credit standards and incentive mechanisms to ensure that the policy effectively drives green innovation across various sectors.
5. Models 3-2 and 6-1 seem to be duplicative, and it is necessary to check the subscript of ????? to see if it represents analyst attention for firm i.
R:We have checked this error and made the necessary modifications. Thank you for your suggestion.
6. Some language expressions and content in the article need to be checked, such as line 166.
R:Thank you for your suggestion, we have made the necessary modifications.
Thank you again for your professional advice. Your suggestions have helped us achieve great improvement. Wishing you a happy life!
Reviewer 3 Report
Comments and Suggestions for AuthorsThe paper provides an insightful examination of the impact of green credit policies on corporate green innovation in China, with a particular focus on high-pollution enterprises. Using data from A-share listed companies from 2007 to 2023, the study effectively highlights the moderating role of analyst attention in the effectiveness of these policies. While the research presents a strong foundation, several areas require attention and enhancement to improve clarity, methodological rigor, and completeness. Below are my detailed comments and suggestions:
Data and Supplementary Information:
Kindly produce a supplementary document showing the data used, along with the corresponding companies. This will enhance transparency and allow for better validation of results.
Methodology:
Line 68: Provide more details on the Difference-in-Differences (DID) approach. An overview of the method and its application would benefit from referencing other studies that have successfully employed this approach.
Line 161: Consider presenting the expression described here as a formal mathematical formula for clarity and precision.
Line 166: Clarify what the "XXXX Approach" refers to; if it is a specific established method, provide proper citation and context.
Line 175: The term “DID = treat × post” should be expressed as a standalone formula, rather than being embedded within the text.
Line 180: Please express the described expression formally as a formula for consistency and ease of interpretation.
Line 158: Justify the use of the Wind and CSMAR databases. Discuss why these databases were chosen and how using alternative databases might yield different results.
Line 212: Modify “didafnum_c” by removing the underscore and making the “C” a subscript (i.e., didafnumC) for consistency in notation.
Figures, Tables, and Equations:
Figure 1: Label the two graphs as “a” and “b” and provide appropriate captions for each to enhance clarity.
Equation Numbers: Ensure equations are numbered sequentially throughout the manuscript (e.g., Eqn 1, Eqn 2, etc.) without breaking them into sections.
Table 5: Change “did” under variables to “DID” for consistency in terminology.
Terminology and Consistency:
Line 156: Explain the meaning of ST, *ST, and PT. Additionally, provide an example of missing relevant indicators that were used to eliminate some of the data.
Section 7.3: Use clearer and more precise terms for the phrases “Excess Cash or Not” and “Firm hold excess cash” to maintain academic rigor and avoid ambiguity.
Literature Review:
The manuscript lacks a comprehensive literature review on green credit policy. Including a detailed discussion of previous studies and their findings will strengthen the theoretical foundation and contextualize your contribution.
Limitations: Acknowledging the limitations of the study and the methods employed will enhance the integrity of the research. This section should provide valuable context for interpreting the results and suggest potential areas for future research.
Author Response
Thank you very much for your professional advice. We have made modifications based on your suggestions, as follows:
Data and Supplementary Information:
Kindly produce a supplementary document showing the data used, along with the corresponding companies. This will enhance transparency and allow for better validation of results.
R: Thank you very much for your suggestion. The data used in this article is mainly from the WIND database（ https://www.wind.com.cn ）And CSMAR database（ https://data.csmar.com/ ）Due to these two being commercial databases, we may not have the authority to share them. In addition, regarding the company name, due to the large number of 3000 sample companies in this article, it is difficult to present them. Of course, we fully understand your opinion and have added explanations for the data in the revised version, as follows:
This study selects A-share listed companies from 2007 to 2023 as the research sample. The sample is processed according to the following criteria: excluding abnormally traded listed companies classified as ST (Special Treatment), *ST (Special Treatment with risk warning), and PT (Particularly Traded). These classifications are used to identify compa-nies that are facing significant financial difficulties or have been placed under special su-pervision by regulatory authorities, which may affect their reliability for research purposes. Additionally, samples with missing relevant indicators, such as missing financial per-formance data (e.g., total assets, revenue, or net income) or incomplete corporate govern-ance information (e.g., board composition or ownership structure), were removed from the sample. After data processing, a total of 51,170 annual observations were obtained. All data were sourced from the Wind and CSMAR databases, which are widely recognized and extensively used in academic and industry research.The Wind and CSMAR data-bases were selected for several reasons. First, both databases provide comprehensive and accurate financial, market, and corporate governance data for A-share listed companies, making them reliable sources for analyzing Chinese listed firms. Second, these databases are frequently used in academic studies in China, ensuring that the results are comparable to existing literature. Wind offers a wide range of financial and economic indicators, while CSMAR provides detailed data on corporate governance and firm-level information, which are essential for this study.
Methodology:
Line 68: Provide more details on the Difference-in-Differences (DID) approach. An overview of the method and its application would benefit from referencing other studies that have successfully employed this approach.
R: Based on your suggestion, we have added a description of the DID method in section 3.3. Model Construction, as follows
The Difference-in-Differences (DID) method is a widely used empirical research tech-nique in econometrics, particularly suited for evaluating the impact of policy changes on specific groups. This method estimates the policy effect by comparing the changes before and after the policy implementation between the treatment group (affected by the policy) and the control group (unaffected by the policy). The basic idea is to compare the changes in the treatment and control groups before and after the policy intervention, thereby con-trolling for unobservable fixed factors that may influence the variable being studied.
In this study, the DID method is used to assess the impact of the green credit policy on corporate green innovation. By dividing firms into high-pollution enterprises (treat-ment group) and non-high-pollution enterprises (control group), and observing the changes before and after the policy implementation, the specific impact of the green credit policy can be effectively identified.
Line 161: Consider presenting the expression described here as a formal mathematical formula for clarity and precision.
R: Based on your feedback, we have made modifications and used formulas to express it.
Line 166: Clarify what the "XXXX Approach" refers to; if it is a specific established method, provide proper citation and context.
R: Thank you for your feedback. This is a typographical error, and we have made the necessary corrections by adding a citation.
Line 175: The term “DID = treat × post” should be expressed as a standalone formula, rather than being embedded within the text.
R: We made modifications and used formulas to express it.
Line 180: Please express the described expression formally as a formula for consistency and ease of interpretation.
R: Thank you for your feedback. We have continued with the revisions.
Line 158: Justify the use of the Wind and CSMAR databases. Discuss why these databases were chosen and how using alternative databases might yield different results.
R: Thank you for your feedback. We have provided additional instructions on using these two databases in section 3.1 Research Sample and Data Sources, as follows:
All data were sourced from the Wind and CSMAR databases, which are widely recog-nized and extensively used in academic and industry research.The Wind and CSMAR databases were selected for several reasons. First, both databases provide comprehensive and accurate financial, market, and corporate governance data for A-share listed compa-nies, making them reliable sources for analyzing Chinese listed firms. Second, these da-tabases are frequently used in academic studies in China, ensuring that the results are comparable to existing literature. Wind offers a wide range of financial and economic in-dicators, while CSMAR provides detailed data on corporate governance and firm-level in-formation, which are essential for this study.
Line 212: Modify “didafnum_c” by removing the underscore and making the “C” a subscript (i.e., didafnumC) for consistency in notation.
R: Thank you for the reminder. We have made revisions to maintain consistency before and after.
Figures, Tables, and Equations:Figure 1: Label the two graphs as “a” and “b” and provide appropriate captions for each to enhance clarity.
R: We have labeled the graph with a and b.
Equation Numbers: Ensure equations are numbered sequentially throughout the manuscript (e.g., Eqn 1, Eqn 2, etc.) without breaking them into sections.
R: Thank you for your suggestion. We have modified the numbering of the equation.
Table 5: Change “did” under variables to “DID” for consistency in terminology.
R: Thank you for your suggestion. We have continued with the modifications
Terminology and Consistency:Line 156: Explain the meaning of ST, *ST, and PT. Additionally, provide an example of missing relevant indicators that were used to eliminate some of the data.
R: We have supplemented the explanation here, as follows: This study selects A-share listed companies from 2007 to 2023 as the research sample. The sample is processed according to the following criteria: excluding abnormally traded listed companies classified as ST (Special Treatment), *ST (Special Treatment with risk warning), and PT (Particularly Traded). These classifications are used to identify companies that are facing significant financial difficulties or have been placed under special supervision by regulatory authorities, which may affect their reliability for research purposes. Additionally, samples with missing relevant indicators, such as missing financial performance data (e.g., total assets, revenue, or net income) or incomplete corporate governance information (e.g., board composition or ownership structure), were removed from the sample. After data processing, a total of 51,170 annual observations were obtained.
Section 7.3: Use clearer and more precise terms for the phrases “Excess Cash or Not” and “Firm hold excess cash” to maintain academic rigor and avoid ambiguity.
R: We have made supplementary modifications to the content of 7.3, as follows:
In this study, surplus cash is defined as the difference between the cash held by a firm and its normal cash requirements. Firms that hold cash greater than the industry average are classified as having surplus cash. We group the sample firms based on whether they possess surplus cash and examine the changes in their green innovation before and after the implementation of the green credit policy.
Literature Review:
The manuscript lacks a comprehensive literature review on green credit policy. Including a detailed discussion of previous studies and their findings will strengthen the theoretical foundation and contextualize your contribution.
R: Based on your feedback, we have added three paragraphs to the introduction for a comprehensive literature review, as follows:
Limitations: Acknowledging the limitations of the study and the methods employed will enhance the integrity of the research. This section should provide valuable context for interpreting the results and suggest potential areas for future research.
R: Based on your feedback, we have added section 8.3 to discuss limitations, as follows:
While this study provides valuable insights into the impact of green credit policies on corporate green innovation, several limitations should be acknowledged. First, the study focuses on A-share listed companies in China, which may limit the generalizability of the findings to other countries or industries. The business environment, regulatory frame-works, and access to green financing may differ significantly across regions, affecting the applicability of the results to non-Chinese contexts.
Second, this study primarily relies on firm-level data from publicly listed companies, which may not fully capture the behavior of smaller firms or those not listed on the stock market. Small and medium-sized enterprises (SMEs), which may face different challenges and incentives in accessing green credit, are underrepresented in the sample. Future stud-ies could include a broader range of firms, including SMEs and privately held companies, to examine how green credit policies affect different types of firms.
Lastly, while the study examines the moderating role of analyst attention, other ex-ternal factors such as market competition, consumer behavior, and industry-specific reg-ulations may also influence the effectiveness of green credit policies. Future research could explore these factors in greater depth and analyze their interaction with green credit poli-cies to provide a more comprehensive understanding of the drivers of green innovation.
Thank you again for your suggestion. Wishing you a happy life!
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsAfter the authors' revision, the article became clearer, nevertheless, since not every company in the nine industries as a control group may have green credit behaviors, and of course, it is not excluded that the companies that do not have access to green credit are trying to get it, it is recommended to evaluate the nine industries by dividing them into treatment and control groups according to the presence of access to green credit or lack of it. If this is due to missing data or other reasons, this can be stated in the shortfall.
Author Response
Thank you very much for your valuable feedback and for helping improve the clarity of the paper. We appreciate your suggestion to divide the nine heavily polluting industries into treatment and control groups based on whether they have access to green credit, rather than simply categorizing them based on industry classification.
We fully understand your concern that not every company in the nine industries may have access to green credit, and some companies in the control group may still be trying to obtain green credit. In our study, due to limitations in the availability of data, we were unable to differentiate between companies that have actual access to green credit and those that do not. This is indeed a potential shortcoming of our study, which we acknowledge.
In response to your suggestion, we have clarified this limitation in the revised version of the paper, stating that the data limitations prevented us from further distinguishing between enterprises that actually received green credit support and those that did not within the heavily polluting industries. We have also added a note in the revised manuscript indicating that this issue can be a potential avenue for future research, where researchers can attempt to better capture green credit access using more granular data, if available.
We hope this revised explanation addresses your concerns, and we appreciate your suggestion to improve the rigor of the model. We believe the current approach, while not perfect, still provides meaningful insights into the effect of green credit policies on heavily polluting industries.
The revised content is as follows:
In addition, in analyzing the impact of the green credit policy, this study classified enterprises in the nine heavily polluting industries into treatment and control groups based on industry classification, referring to the policy documents. However, due to limitations in the data, this study did not further distinguish whether specific enterprises within these industries actually received green credit support. As a result, there may be instances where some control group enterprises also engage in green credit behaviors, or have not received green credit but are attempting to obtain it. Due to the lack of detailed information regarding whether specific enterprises received green credit, this study did not account for this variable difference. Future research is recommended to improve this analysis by using more detailed data.
Thank you again for your feedback. Wishing you a happy life！
Reviewer 3 Report
Comments and Suggestions for AuthorsThe revisions made to the manuscript have significantly enhanced its overall quality, addressing key concerns and refining the clarity, structure, and rigour of the content. I am confident that the paper now meets the standards required for publication in this journal.
Author Response
We would like to sincerely thank the reviewers for their insightful comments and constructive suggestions. Your feedback has greatly improved the clarity and rigor of our paper. We appreciate the time and effort you dedicated to reviewing our work. Your thoughtful recommendations have provided us with valuable directions for enhancing the study. We believe the revisions have strengthened the overall quality of the manuscript, and we are grateful for your assistance in this process. Once again, thank you for your invaluable contributions.