Climate Change and Corporate Financial Performance
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Authors,
Congratulations on your chosen theme and the effort put into your work.
I hope that the suggestions outlined in the attached document prove to be valuable. I recommend considering a minor revision of the article to address these points, as I believe doing so would improve the overall quality of your paper.
Kind regards,
Comments for author File: 
Comments.pdf
Author Response
Responds to the reviewer’s comments:
Comment 1 (Line 121):
I recommend revising the title of subsection "2.1. Key variables" to "Dependent variable and climate-related key variables." This adjustment will distinguish it from the following section, "2.2 Methodology," which also addresses other control variables. By clarifying the focus on climate-related variables in the first subsection, readers can better discern the purpose of each section.
Response 1:
Yes, we agree. We have renamed the title of subsection 2.1 as "Dependent variable and climate-related key variables."
Comment 2 (Line 121):
I recommend beginning subsection “2.1. Key variables” with the explanation of the dependent variable (referenced in lines 149-155). Moreover, using the phrase “The primary dependent variable” implies the existence of other dependent variables, which is not applicable in this context.
Response 2:
Yes, we agree. We have moved the explanation of the dependent variable at the beginning of subsection 2.1 and removed “primary” from the phrase “the primary dependent variable” as below:
“The dependent variable of interest is Return on Assets (ROA), which is calculated as the ratio of a firm's net income to its total assets' book value. ROA is a widely ac-cepted financial performance metric that reflects a firm’s profitability and efficiency in generating returns from its assets (Peters and Mullen, 2009; Gallego‐Álvarez et al., 2014; Sun et al., 2020; Cevik and Miryugin, 2023). It provides valuable insights into how effectively a company is deploying its resources to generate earnings, irrespective of its size or industry.”
Comment 3 (Line 124):
The study uses 173 firms from 14 provinces, with data collected quarterly from 2018 to 2022, should result in a maximum of 3,460 firm-level observations and 280 province-level observations. However, Table A2 reports a larger number of observations. Also, Table 1 reports 176 firms in panel (1). I suggest the authors verify these discrepancies.
Response 3:
Yes, we agree. We have verified the data and confirmed that our dataset covers 209 firms across 20 provinces. However, due to missing data, 196 firms are finally included in the regression. We have corrected the line 124 as below:
“To investigate the effect of climate change on corporate financial performance in the PRC, this research utilizes a firm-level fixed effects panel model across a dataset comprising 209 Chinese-listed firms from 20 provinces.”
Comment 4 (Line 180-182):
The comments on the descriptive statistics of GDP per capita in lines 180-182 appear premature in the text, as their relevance is elucidated later in lines 238-249.
Response 4:
Yes, we agree. We have moved the comments on the descriptive statistics of GDP per capita to the later part in 238-249 as below:
“Economic development varies across different provinces in the PRC. According to the statistical summary in Table A2, the mean GDP per capita of the provinces reaches 31069.41 but with a substantial standard deviation of 11762.57, highlighting the considerable economic heterogeneity prevalent across provinces within China. Thus, it is important to consider and incorporate specific macroeconomic factors related to each province when analyzing the factors affecting corporate financial performance: provincial GDP per capita and urban population ratio.”
Comment 5 (Line 180-182):
For better readability and contextualization, I recommend that the authors include the monetary unit (e.g., $US) for GDP per capita and specify the minimum and maximum values for the "climate change exposure score" and "climate change management score." For example: "The mean GDP per capita of 31,069.41 (in monetary unit?)... The average exposure variable of 2.13 (spanning from 0-3)... Similarly, the average management score of 1.25 (ranging from 0-5)...
Response 5:
Yes, we agree. We have added the monetary unit for GDP per capita and specified the minimum maximum values for the "climate change exposure score" and "climate change management score" as below:
“The descriptive statistics reveal noteworthy patterns among the selected variables of interest. The average exposure variable of 2.13 (ranging from 0-3), accompanied by a relatively narrow standard deviation of 0.79, suggests an overall high level of climate change vulnerability among firms operating within the Chinese market. Similarly, the average management score of 1.25 (ranging from 0-5), coupled with a standard deviation of 0.94, signifies a spectrum of managerial quality among sampled firms and a low level of climate change management.”
“According to the statistical summary in Table A2, the mean GDP per capita of the provinces reaches CNY 31069.41 but with a substantial standard deviation of 11762.57, highlighting the considerable economic heterogeneity prevalent across provinces within China.”
Comment 6 (2.2. Methodology):
The authors specify the model used but do not detail the estimation method, leaving readers unaware of how the analysis was conducted.
Response 6:
Yes, we agree. We have added the explanation about the estimation methods and detailed estimation steps in the subsection “2.2 Methodology” as below:
“In our analysis, we employed a fixed-effects regression model to control for unobserved heterogeneity across the firms. This is particularly important in the context of evaluating the impact of climate change, as it isolates the effect of climate factors from firm-specific and time-invariant influences. δi represents a firm fixed effect, while ui,t is the error term.
This analysis initially regresses on the entire sample of 209 firms across 20 Chinese provinces to assess the overall impact of climate change factors on corporate financial performance. This comprehensive analysis provides a broad overview of how climate variables influence firms across diverse regions. To further refine our understanding, we also perform separate analyses for firms located in coastal areas and those in high-income regions. Coastal firms are likely to experience distinct climate-related challenges, such as rising sea levels and increased storm frequency, which may affect their financial performance differently from firms in inland areas. Similarly, firms in high-income regions may have more resources and infrastructure to adapt to climate impacts, leading to variations in their financial responses compared to those in lower-income areas. By examining these subgroups, we aim to identify specific regional and economic differences in how climate change factors affect corporate performance, providing a more detailed and context-sensitive analysis.”
Besides, we also used the Hausman test to verify our estimation method is appropriate as below:
To further confirm the appropriate estimation approach for our analysis, we conducted a Hausman test, a widely recognized method for assessing the choice between fixed effects (FE) and random effects (RE) models. The Hausman test examines whether the differences in coefficients between FE and RE models are systematic. Our findings indicate strong statistical evidence rejecting the null hypothesis that these differences are not systematic. Therefore, based on the Hausman test results, we conclude that the fixed effects model is more appropriate for our data. This decision is supported by the assumption that individual-specific effects are correlated with the explanatory variables under consideration, validating the use of fixed effects to account for such correlations effectively.
Table 4. Hausman test results: Fixed-effect model vs. Random-effect model
|
Test of H0: Difference in coefficients not systematic |
|
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 36.65 |
|
Prob > chi2 = 0.0000 |
Source: Author’s calculation based on Stata
Comment 7 (Lines 200-202):
The sentence “Managementi,t-1, which represents the firms’ management score of climate change risks, will substitute the exposure score to assess the effect of climate change management on ROA.” seems to indicate that the authors will replace the climate change exposure score with the climate change management score in their analysis; however, in Tables 1 and 2, both variables are included in the model. I suggest the authors verify this discrepancy.
Response 7:
Yes, we agree. We have removed the statement “Managementi,t-1 will substitute the exposure score to assess the effect of climate change management on ROA.”
Comment 8 (Tables 1 and 2):
(i) Remove footnotes (a) and (b) as they do not seem necessary.
(ii) Replace "score" and "Score*GDP per capita" with "management" to be consistent with the model in Equation (1).
(iii) Provide an explanation for the use of interaction terms Exposure*GDP per capita and Score*GDP per capita.
(iv) Address the finding that "covid" has increased ROA, as this seems unusual and warrants investigation to see if it aligns with existing literature.
(v) In each group of variables, include the lag for clarity. For example, in Table 1, specify Climate Change Factors (L=1), Macroeconomic Factors (L=1), and Firm-Specific Factors (L=1). In Table 2, Climate Change Factors (L=8), Macroeconomic Factors (L=1), and Firm-Specific Factors (L=1).
(vi) As a personal preference, I recommend removing the row that repeats ROA and instead including "Dependent: ROA" in the caption. This adjustment would streamline the table and enhance its clarity. However, I understand that this is a subjective preference, and I defer to your judgment on the matter.
(vii) Also, as a personal preference, I propose removing the symbol % from the independent variables and providing clarification in Table A1, where the variables are described.
Response 8:
Yes, we agree.
(i) We have removed footnotes (a) and (b).
(ii) We have replaced "Score" and "Score*GDP per capita" with "Management" and "Management*GDP per capita".
(iii) We have provided an explanation for the use of interaction terms Exposure*GDP per capita and Score*GDP per capita: “To delve deeper into the differential sensitivity of corporate financial performance to climate factors based on the economic context of the region., we augment our baseline specification with interaction terms of climate factors (climate exposure and management) with GDP per capita, respectively.”
(iv) We have added the explanation for the finding that "covid" has increased ROA as follow: “Surprisingly, this study found that COVID-19 had a positive impact on the ROA of publicly listed companies in China. This may be attributed to China's implementation of strict and early containment measures, which allowed factories to continue operat-ing. Notably, from late 2020 to 2021, many orders were redirected to China as other countries were significantly affected by COVID-19.”
(v) We have added the lag for each group of variables in Tables 1 and 2.
(vi) We have removed the row that repeats ROA and instead included "Dependent: ROA" in the caption.
(vii) We removed the symbol % from the independent variables and provided clarification in Table A1.
Comment 9 (Line 383):
I noticed that most data sources in Table A1 are absent from the references list.
Response 9:
Yes, we agree. We have added the data sources as references of Table A1:
“2. The data is available at: https://capitaliq.com
- The ESG data from FTSE Russel is available at: https://data.ftse.com
- The data is available at: https://www.ceicdata.com”
Comment 10 (Line 434-436):
I noticed that 'Hale et al.' is not listed in alphabetical order in the references list.
Response 10:
Yes, we agree. We have moved the reference “Hale et al.” in alphabetical order in the reference list.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsComments for author File: 
Comments.docx
Author Response
Responds to the reviewer’s comments:
Comment 1:
The authors are encouraged to enrich the empirical literature that is summarized in the introduction section.
Response 1:
Yes, we agree. We have added more empirical literature regarding the effect of climate change exposure and management on corporate financial performance in the Introduction section, such as Huang et al. (2018), Almaghrabi (2023), Ntim and Soobaroyen (2013), etc.
Comment 2:
The authors should present a specific approach (REM, FEM, GLS, GMM, . . .) to employ for equation (1) and explain why it is selected.
Response 2:
Yes, we agree. We have added the explanation about the estimation methods, selection rationale and detailed estimation steps in the subsection “2.2 Methodology” as below:
“In our analysis, we employed a fixed-effects regression model to control for unobserved heterogeneity across the firms. This is particularly important in the context of evaluating the impact of climate change, as it isolates the effect of climate factors from firm-specific and time-invariant influences. δi represents a firm fixed effect, while ui,t is the error term.
This analysis initially regresses on the entire sample of 209 firms across 20 Chinese provinces to assess the overall impact of climate change factors on corporate financial performance. This comprehensive analysis provides a broad overview of how climate variables influence firms across diverse regions. To further refine our understanding, we also perform separate analyses for firms located in coastal areas and those in high-income regions. Coastal firms are likely to experience distinct climate-related challenges, such as rising sea levels and increased storm frequency, which may affect their financial performance differently from firms in inland areas. Similarly, firms in high-income regions may have more resources and infrastructure to adapt to climate impacts, leading to variations in their financial responses compared to those in lower-income areas. By examining these subgroups, we aim to identify specific regional and economic differences in how climate change factors affect corporate performance, providing a more detailed and context-sensitive analysis.”
Comment 3:
Some diagnostic tests for the model need to be conducted and presented in the paper.
Response 3:
Yes, we agree. We have presented the stationary test and Hausman test in section 3 “Empirical Analysis and Discussion” as below:
Prior to conducting the regression analysis, we employ the Levin–Lin–Chu (LLC) test and the Fisher augmented Dickey–Fuller (Fisher ADF) test to assess the stationarity of the variables. The outcomes reject the null hypothesis of unit roots for each variable, indicating that the variables are stationary at their levels. Therefore, all variables are included in the econometric model at their levels.
To further confirm the appropriate estimation approach for our analysis, we conducted a Hausman test, a widely recognized method for assessing the choice between fixed effects (FE) and random effects (RE) models. The Hausman test examines whether the differences in coefficients between FE and RE models are systematic. Our findings indicate strong statistical evidence rejecting the null hypothesis that these differences are not systematic. Therefore, based on the Hausman test results, we conclude that the fixed effects model is more appropriate for our data. This decision is supported by the assumption that individual-specific effects are correlated with the explanatory variables under consideration, validating the use of fixed effects to account for such correlations effectively.
Table 4. Hausman test results: Fixed-effect model vs. Random-effect model
|
Test of H0: Difference in coefficients not systematic |
|
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 36.65 |
|
Prob > chi2 = 0.0000 |
Source: Author’s calculation based on Stata
Comment 4:
Because the paper is not so long, I suggest that the appendixes should be placed in the main text of the manuscript.
Response 4:
Yes, we agree. We have removed the appendixes part and moved the tables in the main text of the manuscript as Table 1, 2 and 3.
Comment 5:
Limitations of the study and suggestions for further research should be added in the conclusion of the manuscript.
Response 5:
Yes, we agree. We have added the limitations the study and suggestions for further research in the “Conclusion” part as below:
“A key limitation is that the data only includes publicly listed companies, which are typically large firms. China also has numerous small, non-listed companies that are not covered in this study. Future research may expand the sample size to include these smaller firms for a more comprehensive analysis. A sufficiently large sample enables us to conduct more practical and detailed studies across different industries.”
Comment 6:
Citations and references should be used in compliance of the Journal’s rules.
Response 6:
Yes, we agree. We have doubled-checked and revised accordingly to ensure citations and references are in compliance with the journal’s requirements.
Comment 7:
Some citations (Moyo & Wingard, 2015) could not be found in the references
Response 7:
Yes, we agree. We have double-checked and added the missing reference in the “Reference” list.
Responds to the reviewer 4’s comments:
Comment 1 and 2:
The introduction is not convincing, no statistical support the evidences of volatility, research gap and contribution.
Literature is not convincing at all. Just combine many previous studies altogether. The literature should lead to the research gap of the study.
Response 1 and 2:
Yes, we agree. We have enriched our literature on the channels through which climate change affects corporate financial performance and empirical evidence regarding the effect of climate change exposure and management on corporate financial performance.
Based on the literature, we have resummarized the research gap and refined our contribution as below:
“The literature examining the potential economic impacts of climate change is large and growing rapidly. However, significant gaps remain. Empirical research investi-gating firm-level impacts of climate change in China remains relatively scarce. In ad-dition, existing empirical literature based on firm-level data often uses a country-level climate vulnerability measure (Acevedo et al., 2020). Such a measure may not be suf-ficiently accurate for firms’ climate exposure assessment, especially in large countries like the PRC. While difficulties in gathering data on firm-level climate exposures re-main, it needs to be borne in mind that firm-based analyses using country-level climate exposure measures should be interpreted with caution (Bernstein et al. 2019; Hong et al. 2019; Choi et al. 2020). Additionally, existing firm-level research overlooks the sub-stantial regional economic disparities and the varying levels of sensitivity to climate change within the countries. This is crucial because economic development varies widely across regions within many countries, profoundly impacting the economic performance of companies operating at the regional level.
To fill these research gaps, this paper empirically investigates the impact of firms’ climate change exposure on corporate financial performance using a firm-level panel dataset of 173 firms over the period Q1 2018 to Q2 2022 in PRC. In particular, this paper exploits a climate change exposure and management score measured at the firm level, which is a more comprehensive measure than that used by Wu et al. (2022). More spe-cifically, this study utilizes two measures of firms’ climate change risks: (i) the Financial Times Stock Exchange (FTSE) climate change exposure score and (ii) the FTSE climate change management score. Our results show a significant negative impact of firms’ climate change exposure on their rate of return, especially in the long term. The effect is more substantial for firms in coastal areas and high-income provinces. The empirical results imply that activities to reduce climate change exposure (such as integrating climate risk considerations into business models and implementing emission reduction initiatives) could improve firm performance.
The study makes two contributions to the existing literature. First, this paper adds to the limited body of empirical literature in China concerning the impact of climate change exposure and management on firm corporate financial performance. Specifically, the study encompasses 209 companies from 20 different provinces across China. These provinces vary in economic development levels and exposure to environmental changes. We estimate subpanels for economically developed versus less developed regions, as well as coastal versus non-coastal areas. This approach allows for a more comprehensive and detailed analysis, providing practical insights that are more nuanced and region-specific.
Secondly, prior research has primarily focused on the effect of GHG emission-related risks or policies on corporate financial performance. By extending beyond these areas, this study broadens the scope of investigation to include climate change exposure and management, thereby providing a more comprehensive understanding of the factors influencing financial outcomes for firms. This holistic approach offers deeper insights into the multifaceted relationship between climate change and corporate financial performance.”
Comment 3:
Methodolgy is not correct and less information of the panel regression, cointrgration. No control variable which might lead omited variable bias.
Response 3:
Yes, we agree. We have added the explanation about the estimation methods and detailed estimation steps in the subsection “2.2 Methodology” as below:
“In our analysis, we employed a fixed-effects regression model to control for unobserved heterogeneity across the firms. This is particularly important in the context of evaluating the impact of climate change, as it isolates the effect of climate factors from firm-specific and time-invariant influences. δi represents a firm fixed effect, while ui,t is the error term.
This analysis initially regresses on the entire sample of 209 firms across 20 Chinese provinces to assess the overall impact of climate change factors on corporate financial performance. This comprehensive analysis provides a broad overview of how climate variables influence firms across diverse regions. To further refine our understanding, we also perform separate analyses for firms located in coastal areas and those in high-income regions. Coastal firms are likely to experience distinct climate-related challenges, such as rising sea levels and increased storm frequency, which may affect their financial performance differently from firms in inland areas. Similarly, firms in high-income regions may have more resources and infrastructure to adapt to climate impacts, leading to variations in their financial responses compared to those in lower-income areas. By examining these subgroups, we aim to identify specific regional and economic differences in how climate change factors affect corporate performance, providing a more detailed and context-sensitive analysis.”
Besides, we also used the Hausman test to verify our estimation method is appropriate as below:
To further confirm the appropriate estimation approach for our analysis, we conducted a Hausman test, a widely recognized method for assessing the choice between fixed effects (FE) and random effects (RE) models. The Hausman test examines whether the differences in coefficients between FE and RE models are systematic. Our findings indicate strong statistical evidence rejecting the null hypothesis that these differences are not systematic. Therefore, based on the Hausman test results, we conclude that the fixed effects model is more appropriate for our data. This decision is supported by the assumption that individual-specific effects are correlated with the explanatory variables under consideration, validating the use of fixed effects to account for such correlations effectively.
Table 4. Hausman test results: Fixed-effect model vs. Random-effect model
|
Test of H0: Difference in coefficients not systematic |
|
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 36.65 |
|
Prob > chi2 = 0.0000 |
Source: Author’s calculation based on Stata
In our regression, we have included a range of control variables:
- Macroeconomic factors: provincial GDP per capita, Covid contingency index, Urban population ratio
- Firm-specific factors: Total asset, Revenue growth, Financial leverage, and Firm age
Comment 4:
The results are weak
Response 4:
We have elaborated on the interpretation of the estimated coefficients of interest more clearly in the revised paper. In particular, the finding of a significant and robust effect of exposure to climate change on firms’ ROA over time implies that such risks need to be more comprehensively incorporated into the investment strategies of firms, including through enhanced climate risk hedging and insurance mechanisms. This can also help to reinforce firms’ own commitments to actions aimed at reducing climate change risks and exposures. Our findings in the case of coastal firms provide evidence of the need for more urgent action given the larger relative negative impacts on ROA due to climate exposure.
Comment 5:
I have not seen any market volatility in this analysis.
Response 5:
This is a very much appreciated comment. Wider market volatility can indeed affect corporate performance measures such as firms’ ROA. In our methodological framework, such volatility can be regarded as a trend determining factor, and as such is captured in our time fixed effects. .
Responds to the reviewer 1’s comments:
Comment 1 and 3:
The authors should refine the contribution part. More specifically, the authors can list relevant literature for comparison, which is a good way to highlight the contribution of the article.
Try to provide more insightful analysis of previous studies and some of their shortcomings.
I also expect a detailed motivation of the chosen methods of analysis. The authors simply state what they have chosen to use without explaining whether they are the most appropriate of the many possibilities, and without discussing their strengths or weaknesses compared to competing potential alternatives.
Response 1 and 3:
Yes, we agree. We have enriched our literature on the channels through which climate change affects corporate financial performance and empirical evidence regarding the effect of climate change exposure and management on corporate financial performance.
Based on the literature, we have resummarized the research gap and refined our contribution as below:
“The literature examining the potential economic impacts of climate change is large and growing rapidly. However, significant gaps remain. Empirical research investi-gating firm-level impacts of climate change in China remains relatively scarce. In ad-dition, existing empirical literature based on firm-level data often uses a country-level climate vulnerability measure (Acevedo et al., 2020). Such a measure may not be suf-ficiently accurate for firms’ climate exposure assessment, especially in large countries like the PRC. While difficulties in gathering data on firm-level climate exposures re-main, it needs to be borne in mind that firm-based analyses using country-level climate exposure measures should be interpreted with caution (Bernstein et al. 2019; Hong et al. 2019; Choi et al. 2020). Additionally, existing firm-level research overlooks the sub-stantial regional economic disparities and the varying levels of sensitivity to climate change within the countries. This is crucial because economic development varies widely across regions within many countries, profoundly impacting the economic performance of companies operating at the regional level.
The study makes two contributions to the existing literature. First, this paper adds to the limited body of empirical literature in China concerning the impact of climate change exposure and management on firm corporate financial performance. Specifically, the study encompasses 209 companies from 20 different provinces across China. These provinces vary in economic development levels and exposure to environmental changes. We estimate subpanels for economically developed versus less developed regions, as well as coastal versus non-coastal areas. This approach allows for a more comprehensive and detailed analysis, providing practical insights that are more nuanced and region-specific.
Secondly, prior research has primarily focused on the effect of GHG emission-related risks or policies on corporate financial performance. By extending beyond these areas, this study broadens the scope of investigation to include climate change exposure and management, thereby providing a more comprehensive understanding of the factors influencing financial outcomes for firms. This holistic approach offers deeper insights into the multifaceted relationship between climate change and corporate financial performance.”
Comment 2:
In the introduction it is not clear ‘what was done’ to address the research objective? What is the theory? What are the methodology and methods implicated in this paper? These need to be clearly mentioned in the introduction
Response 2:
Yes, we agree. We have specified “what was done”, methodology, data and results in the “Introduction” part as below:
“To fill these research gaps, this paper empirically investigates the impact of firms’ climate change exposure and management on corporate financial performance using a fixed effects panel model. This study covers a firm-level panel dataset of 209 firms over the period Q1 2018 to Q2 2022 in PRC. More specifically, this study utilizes two measures of firms’ climate change factors: (i) the Financial Times Stock Exchange (FTSE) climate change exposure score and (ii) the FTSE climate change management score, which is a more comprehensive measure than that used by Wu et al. (2022). Our results show a significant negative impact of firms’ climate change exposure on their rate of return, especially in the long term. The effect is more substantial for firms in coastal areas and high-income provinces. The empirical results imply that activities to reduce climate change exposure (such as integrating climate risk considerations into business models and implementing emission reduction initiatives) could improve firm performance.”
As regards the theoretical considerations, conceptually our paper is related to traditional asset pricing theory and the role of climate change as a risk factor for the financial performance of firms. We have not elaborated on this aspect in the paper as the focus of the paper and the related research question are empirical in nature.
Comment 4:
I also expect a detailed motivation of the chosen methods of analysis. The authors simply state what they have chosen to use without explaining whether they are the most appropriate of the many possibilities, and without discussing their strengths or weaknesses compared to competing potential alternatives.
Response 4:
Our motivation for the study therefore is linked to addressing the research gaps outlined. As regards the selected methodology, we feel that a panel regression with fixed effects is the most appropriate given our research question. In particular, our approach permits an econometrically sound assessment of average elasticities, with flexibility to compare across sample subsets and also control sufficiently for heterogeneity across firms and global factors affecting corporate performance. Other approaches could include panel VARs, which would be more appropriate for examining dynamic impacts over time rather than average effects, or GARCH models for modelling volatility effects.
Comment 5:
The author(s) need to work more on the section methodology and make it understandable. Which are the advantages of employed methodology compared to other techniques?
Response 5:
Yes, we agree. We have added the explanation about the estimation methods and detailed estimation steps in the subsection “2.2 Methodology” as below:
“In our analysis, we employed a fixed-effects regression model to control for unobserved heterogeneity across the firms. This is particularly important in the context of evaluating the impact of climate change, as it isolates the effect of climate factors from firm-specific and time-invariant influences. δi represents a firm fixed effect, while ui,t is the error term.
This analysis initially regresses on the entire sample of 209 firms across 20 Chinese provinces to assess the overall impact of climate change factors on corporate financial performance. This comprehensive analysis provides a broad overview of how climate variables influence firms across diverse regions. To further refine our understanding, we also perform separate analyses for firms located in coastal areas and those in high-income regions. Coastal firms are likely to experience distinct climate-related challenges, such as rising sea levels and increased storm frequency, which may affect their financial performance differently from firms in inland areas. Similarly, firms in high-income regions may have more resources and infrastructure to adapt to climate impacts, leading to variations in their financial responses compared to those in lower-income areas. By examining these subgroups, we aim to identify specific regional and economic differences in how climate change factors affect corporate performance, providing a more detailed and context-sensitive analysis.”
Besides, we also used the Hausman test to verify our estimation method is appropriate as below:
To further confirm the appropriate estimation approach for our analysis, we conducted a Hausman test, a widely recognized method for assessing the choice between fixed effects (FE) and random effects (RE) models. The Hausman test examines whether the differences in coefficients between FE and RE models are systematic. Our findings indicate strong statistical evidence rejecting the null hypothesis that these differences are not systematic. Therefore, based on the Hausman test results, we conclude that the fixed effects model is more appropriate for our data. This decision is supported by the assumption that individual-specific effects are correlated with the explanatory variables under consideration, validating the use of fixed effects to account for such correlations effectively.
Table 4. Hausman test results: Fixed-effect model vs. Random-effect model
|
Test of H0: Difference in coefficients not systematic |
|
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 36.65 |
|
Prob > chi2 = 0.0000 |
Source: Author’s calculation based on Stata
Comment 6:
The empirical analysis although interesting, lack the interpretation and economic meaning. I would strongly suggest that the authors add more intuitive explanation.
Response 6:
We have elaborated on the interpretation of the estimated coefficients of interest more clearly in the revised paper. In particular, the finding of a significant and robust effect of exposure to climate change on firms’ ROA over time implies that such risks need to be more comprehensively incorporated into the investment strategies of firms, including through enhanced climate risk hedging and insurance mechanisms. This can also help to reinforce firms’ own commitments to actions aimed at reducing climate change risks and exposures. Our findings in the case of coastal firms provide evidence of the need for more urgent action given the larger relative negative impacts on ROA due to climate exposure.
Comment 7:
There are no robustness tests have been used to check the validity of the results.
Response 7:
As regards the robustness, we introduced a lag of one period for the climate change variables and other variables to mitigate the endogeneity concerns. Furthermore, to account for the lagged impact of climate change on firm performance, we also implemented a specification wherein the climate change variables are lagged by eight periods (i.e., eight quarters, or two years). This dual regression framework allows us to comprehensively examine the potential time-delayed effects of climate change factors on corporate financial performance and also serves as a robustness check.
Comment 8:
The authors need to present the major limitations and implications on your analysis and how they open the doors for future research.
Response 8:
Yes, we agree. We have added the limitations of the study and suggestions for further research in the “Conclusion” part as below:
“A key limitation is that the data only includes publicly listed companies, which are typically large firms. China also has numerous small, non-listed companies that are not covered in this study. Future research may expand the sample size to include these smaller firms for a more comprehensive analysis. A sufficiently large sample enables us to conduct more practical and detailed studies across different industries.”
Comment 9:
Regarding the quality of writing and communication, English has presented some minor problems in Language quality and style. The writing of the paper should be made clearer and the manuscript needs a language revision.
Response 9:
Yes, we agree. We have thoroughly reviewed our manuscript and corrected any errors in English expression.
Author Response File: Author Response.pdf
