The Impact of Climate Change on the Risk of Bankruptcy of Agricultural Companies in Poland: Regional Characteristics

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The article is interesting. The author(s) have addressed an interesting and important topic. The research hypotheses are clearly presented and justified in the context of the literature.

The research methods seem to be adequate to the needs. In addition, they have been extensively explained. Therefore, the study should not contain controversy.

However, the specificity of agricultural production raises doubts as to the correctness of the results obtained. First of all, it is important to pay attention to the differences between animal and plant production. In addition, it is worth paying attention to the occurrence of costs and revenues in agriculture. These factors are important in bankruptcy processes, because the occurrence of revenue occurs with a delay in relation to production. Production is mainly dependent on climatic conditions, including air temperaturę, especially in spring and summer. For this reason, I have doubts whether the results shown in the seasons are justified. It is worth for the author(s) to develop this thread and show a greater cause-and-effect relationship between the phenomena studied.

In practice, obtaining low yields in summer under the influence of high temperatures is the cause of low income in autumn and winter from the sale of these crops. This creates a risk of bankruptcy. However, the cause occurred during the summer. A time shift is visible. It seems to me that the author/authors did not take this into account. And if they did, they didn't emphasize it enough in the description. It is worth taking this into account in the text.

I have no more comments. In general, the thesis is written with great care, both from the technical and substantive side. And I recommend posting it after taking into account my comment.

Author Response

1	The article is interesting. The author(s) have addressed an interesting and important topic. The research hypotheses are clearly presented and justified in the context of the literature. The research methods seem to be adequate to the needs. In addition, they have been extensively explained. Therefore, the study should not contain controversy.	Thank you for your positive assessment of the topic, the structure of the article and the way the research was conducted and presented.
2	However, the specificity of agricultural production raises doubts as to the correctness of the results obtained. First of all, it is important to pay attention to the differences between animal and plant production. In addition, it is worth paying attention to the occurrence of costs and revenues in agriculture. These factors are important in bankruptcy processes, because the occurrence of revenue occurs with a delay in relation to production. Production is mainly dependent on climatic conditions, including air temperature, especially in spring and summer. For this reason, I have doubts whether the results shown in the seasons are justified. It is worth for the author(s) to develop this thread and show a greater cause-and-effect relationship between the phenomena studied.	We agree with the observation that companies specializing in animal breeding and those specializing in grain cultivation will react to temperature changes somewhat differently. Unfortunately, most of the entities analyzed are small businesses that use simplified accounting and do not indicate their dominant agricultural activity – most often they report mixed activities. Therefore, we cannot generate separate data sets for entities specializing in animal breeding and those specializing in grain cultivation.
3	In practice, obtaining low yields in summer under the influence of high temperatures is the cause of low income in autumn and winter from the sale of these crops. This creates a risk of bankruptcy. However, the cause occurred during the summer. A time shift is visible. It seems to me that the author/authors did not take this into account. And if they did, they didn't emphasize it enough in the description. It is worth taking this into account in the text.	We agree that there is likely a lag between temperature changes and changes in a company's operating results in subsequent months or quarters. Unfortunately, our company data is annual, aggregating results from all seasons. Therefore, we cannot analyze the impact of temperature in individual months/seasons in detail, or introduce lags for company-specific variables, for example, by one quarter. In our study, we only examine whether, and if so, how temperature fluctuations in individual seasons affect the risk of bankruptcy for a company in a given year. Thank you for your suggestion, and we added appropriate clarification to the text. This and the previous comments are extremely important when detailed data on companies are available. We will use these suggestions in preparing surveys among farms, which will allow for a more precise characterization of the type of business conducted, as in addition to farms specializing in animal breeding or grain cultivation, there is a large group of farms engaged in mixed production. In such cases, it will be possible to apply some classification/grouping, including based on the share of crop production.
4	I have no more comments. In general, the thesis is written with great care, both from the technical and substantive side. And I recommend posting it after taking into account my comment.	Thank you again for your positive assessment and suggestions for future research.

Reviewer 2 Report

Comments and Suggestions for Authors

The study assesses the resilience to bankruptcy risk of enterprises using the Altman and Zmijewski methods, and combines panel regression analysis to investigate the impact of temperature changes on the resilience to bankruptcy risk of over 4,000 agricultural enterprises in Poland from 2016 to 2023 under different temperature and regional macroeconomic conditions.

1. The study uses both the Z-score and X-score models, but fails to conduct an in-depth comparison of their differences in applicability to bankruptcy risk assessment of agricultural enterprises. It is suggested to add an analysis of the limitations of the two models in agricultural applications, explain the reasons for using both models simultaneously, and clarify the interpretation logic when there are differences in model results, so as to improve the scientificity of model selection.
2. In the study, temperature variables are based on long-term average values, but the scientificity of using this long-term average as a reference is not clearly explained, such as whether it matches the historical climate adaptation period of agricultural production in Poland. It is recommended to supplement the comparative analysis of long-term averages in different time windows (e.g., 1990-2010) to verify the sensitivity of the results to the selection of benchmarks, so as to avoid conclusion deviations caused by unreasonable reference periods.
3. There are problems in the paper's sample, such as significant differences in enterprise scales and irregular data due to small enterprises adopting simplified accounting principles. It is suggested to supplement and explain the specific criteria for data screening, such as how to handle the data of small enterprises using simplified accounting principles, whether the rationality of extreme value processing methods (e.g., 1.5 times IQR rule) is verified, and the proportion of enterprises of different scales in the sample and their potential impact on the results, so as to enhance data reliability.
4. Both the Z-score and X-score models are used to assess bankruptcy risk, but the conclusions of the two models on the impact of summer temperatures are conflicting (the Z-score shows a linear improvement, while the X-score shows a weak non-linearity with negligible impact), and the root cause of the conflict is not analyzed in depth.
5. Although the fixed-effects model is adopted and heteroscedasticity and autocorrelation are corrected, cross-period robustness tests are not conducted, such as regression in two sub-periods of 2016-2019 and 2020-2023, nor is it verified whether the results are stable after excluding specific regions, such as the high-variability winter regions with the largest temperature fluctuations. It is recommended to add robustness tests such as sub-sample analysis and replacement of core explanatory variables (e.g., using quarterly temperature standard deviation instead of deviation value) to enhance the reliability of conclusions.

Author Response

1	The study uses both the Z-score and X-score models, but fails to conduct an in-depth comparison of their differences in applicability to bankruptcy risk assessment of agricultural enterprises. It is suggested to add an analysis of the limitations of the two models in agricultural applications, explain the reasons for using both models simultaneously, and clarify the interpretation logic when there are differences in model results, so as to improve the scientificity of model selection.	We thank you for your comment and suggestion to compare the applicability of the X-score and Z-score models. A relevant analysis comparing the two models has been included in the text on pages 6-7.
2	In the study, temperature variables are based on long-term average values, but the scientificity of using this long-term average as a reference is not clearly explained, such as whether it matches the historical climate adaptation period of agricultural production in Poland. It is recommended to supplement the comparative analysis of long-term averages in different time windows (e.g., 1990-2010) to verify the sensitivity of the results to the selection of benchmarks, so as to avoid conclusion deviations caused by unreasonable reference periods.	We agree that the use of deviation from the long-term mean rather than absolute temperature was not clearly explained. This clarification was outlined in the research questions (p. 2), where we indicated that we were examining the impact of annual and quarterly temperature increases relative to the long-term mean temperature. Researchers use various measures to measure temperature changes, but a relatively common approach is to define these changes as the difference between a given temperature and the long-term average (see below for a few examples we considered). Our weather data are only available from 2004, making it difficult to conduct an analysis using deviations from the proposed long-term mean for the years 1990-2010. Following the suggestion, for robustness checks, calculations were performed in two versions, defining the temperature variable as the difference between a given temperature and the mean temperature in the years: (1) 2004-2023 and (2) 2004-2015. The results for the 2004-2015 average have been inserted into the article's text in the Appendix section. We appreciate your suggestion, as the results of this calculation confirmed the results originally presented.   Article examples:   ·         Dell, M., Jones, B. & Olken, B. (2008) Climate Change and Economic Growth: Evidence from the Last Half Century. Am. Econ. J.: Macroecon. DOI: 10.3386/w14132 ·         Deryugina, T, & Hsiang, S. (2014) Does the environment still matter? Daily temperature and income in the United States. NBER Working Paper 20750. http://www.nber.org/papers/w20 ·         Kolstad, Ch., & Moore, F. (2020) Estimating the Economic Impacts of Climate Change Using Weather Observations. Review of Environmental Economics and Policy, 14(1), 1-24. DOI: 10.1093/reep/rez024 ·         Pretis, F. (2020) Econometric modelling of climate systems: The equivalence of energy balance models and cointegrated vector autoregressions. Journal of Econometrics, 214(1), 256-273. https://doi.org/10.1016/j.jeconom.2019.05.013. ·         Chang, J., Mi, Z., & Wei, Y. (2023) Temperature and GDP A review of climate econometrics analysis. Structural Change and Economic Dynamics, 66, 383–392. doi.org/10.1016/j.strueco.2023.05.009 ·         Song, X, & Fang, T. (2023) Temperature shocks and bank systemic risk: Evidence from China. Finance Research Letters, 51, 103447. https://doi.org/10.1016/j.frl.2022.103447
3	There are problems in the paper's sample, such as significant differences in enterprise scales and irregular data due to small enterprises adopting simplified accounting principles. It is suggested to supplement and explain the specific criteria for data screening, such as how to handle the data of small enterprises using simplified accounting principles, whether the rationality of extreme value processing methods (e.g., 1.5 times IQR rule) is verified, and the proportion of enterprises of different scales in the sample and their potential impact on the results, so as to enhance data reliability.	We agree that the significant differences in company size make econometric analysis difficult. However, we do not select companies based on their size, although it is true that the sector is very diverse. To reduce estimation error, we use the natural logarithm of company assets in the equation to account for differences in company size. Furthermore, to remove many extreme values ​​in the Z-score and X-score results, a 1.5-times IQR rule was applied for extreme values. This procedure is relatively commonly used to organize the dataset. The results of its application are presented in the following histograms for the X-score and Z-score variables. Nevertheless, we greatly appreciate your remarks. Histograms of the variables after removing outliers have been included in the Appendix. We are convinced that this will increase the information content and value of our article. Histogram: raw z-score, before removing extreme values Histogram: z-score, after removing extreme values   Histogram: x-score raw, before removing extreme values   Histogram: x-score, after removing extreme values
4	Both the Z-score and X-score models are used to assess bankruptcy risk, but the conclusions of the two models on the impact of summer temperatures are conflicting (the Z-score shows a linear improvement, while the X-score shows a weak non-linearity with negligible impact), and the root cause of the conflict is not analyzed in depth.	The model estimation results for the Z-score and X-score may not have been presented clearly, so we re-examined the relevant sections of the text. For both the Z-score and X-score measures, most models demonstrate a nonlinear relationship between temperature change and the level of bankruptcy risk. In our opinion, the difference between some models for Summer may be due to a slight difference in the Z-score and X-score distributions. Small, though significantly smaller, differences appear for the coefficients relating to ΔT and ΔT2 for the model with annual data and for Winter, Spring, and Autumn.
5	Although the fixed-effects model is adopted and heteroscedasticity and autocorrelation are corrected, cross-period robustness tests are not conducted, such as regression in two sub-periods of 2016-2019 and 2020-2023, nor is it verified whether the results are stable after excluding specific regions, such as the high-variability winter regions with the largest temperature fluctuations. It is recommended to add robustness tests such as sub-sample analysis and replacement of core explanatory variables (e.g., using quarterly temperature standard deviation instead of deviation value) to enhance the reliability of conclusions.	Thank you for your consideration and suggestion to perform robustness tests. Our research covers a relatively short period, and performing calculations over a four-year period did not yield significant results. Furthermore, the agricultural situation in 2016-2019 and 2020-2023 differed significantly. Table 2 indicates, among other things, that in the first period, the average annual agricultural production was approximately PLN 110 billion, while in 2022-2023 it was close to PLN 200 billion, which resulted, among other things, from changes forced by the COVID-19 pandemic and the war in Ukraine. In response to the second suggestion, we conducted an estimation using the standard deviation of temperature differences as the temperature variable. Significant results were obtained for models with annual temperatures. In these calculations, only the standard deviation of the first power was used, as the second power presents a problem of collinearity and the results are not significant. We present the results of these calculations below. The sign of the coefficient for the temperature variable is identical to that of the basic models. However, this variable (standard deviation) is not consistent with the topic of the current research, because the subject of this research is the effect of the deviation from the mean taking into account the direction of change, and not the standard deviation, which is always positive regardless of the direction of change.     (Z-score) (X-score) assets 0.278* 0.008   (0.160) (0.057) eff_asset 0.038* -1.192***   (0.020) (0.460) roa 0.039* -0.000   (0.022) (0.001) lev 0.125 -3.867***   (0.087) (0.451) short_liab 0.001 -0.000***   (0.001) (0.000) Sd_Tyr -1.829*** 0.457***   (0.241) (0.088) cons 3.503*** -0.508   (1.122) (0.659) Observations 22,397 17,486.000 Groups 4,555 3,565.000 F-stat 24.468 75.524 R-squared 0.080 0.690 Standard deviation measures volatility of temperatures in each year. To confirm the validity of the results, additional calculations were performed using other methods, such as random-effects panel regression and the Arellano–Bond dynamic model. The results of these estimations, confirming the validity of our basic calculations, are included in the article's Appendix. We believe this will enhance the article's informative value.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

In this study, the authors examined the impact of climate change o the risk of bankruptcy of agricultural companies in Poland by using panel regression analysis. Although this study could contribute to the existing literature, the authors should consider the following points to improve the manuscript quality.

1. I would recommend the authors to provide the concrete and clear research questions in the first section, possibly in the end of the third paragraph to enhance the readability of the manuscript.
2. In the second paragraph, fourth line, the authors cited 8 publications (7 to 14) to justify one argument. Is this fully necessary to cite that many papers here?
3. The contributions of the research need further justification. In the last paragraph of the first section, the authors stated that “the study for the first time determines the level of bankruptcy risk for agricultural enterprises in Poland using the Altman Z-score and Zmijewski X-score models simultaneously.” To fully justify this contribution, the authors should consider explaining the advantages of using these models. To do so, a brief comparison between this novel method with other commonly used risk measuring methods can help.
4. The titles of 2.1 to 2.3 can probably undermine the readability of the manuscript. All of them have “temperature increases and business performance”, and it reads like section 2.2 and section 2.3 are part of section 2.1 by just from their section titles. The authors should consider modify the titles to enhance the readability of the sections.
5. In the third line of Page 6, what is the meaning of “XX century”? Is this a typo or an abbreviation? The authors should consider providing further explanations.
6. The table 1 is too large. Please change its width to fit the paper. This applies to Table 2, 3, and 5.
7. In Figure 1, only Opole data ranges from 2005 to 2015. The authors need to explain the reason of this difference. This applies to Figure 2 as well.
8. The authors need to unify the template used in the paper. It can be seen that a different template was used from Page 18.
9. The equation numbers are not fully correct. Please revise them.
10. In the conclusion, the academic and practical contributions should be both mentioned, based on the comparison with previous literature. Also, the research limitation discussions should be enhanced and future research directions should be provided.

Thanks!

Author Response

1	I would recommend the authors to provide the concrete and clear research questions in the first section, possibly in the end of the third paragraph to enhance the readability of the manuscript.	Thank you for your recommendation. The research questions are included on page 2.
2	In the second paragraph, fourth line, the authors cited 8 publications (7 to 14) to justify one argument. Is this fully necessary to cite that many papers here?	Thank you for your remark. One article has been removed. This may seem like an overestimation, however, given its relevance to our research topic, we'd like to retain the remaining articles. These articles were cited to demonstrate the impact not only of temperature but also of rainfall variability on agricultural outcomes.
3	The contributions of the research need further justification. In the last paragraph of the first section, the authors stated that “the study for the first time determines the level of bankruptcy risk for agricultural enterprises in Poland using the Altman Z-score and Zmijewski X-score models simultaneously.” To fully justify this contribution, the authors should consider explaining the advantages of using these models. To do so, a brief comparison between this novel method with other commonly used risk measuring methods can help.	We thank you for your comment and suggestion to compare the applicability of the X-score and Z-score models. A relevant analysis comparing the two models has been included in the text on pages 6-7.
4	The titles of 2.1 to 2.3 can probably undermine the readability of the manuscript. All of them have “temperature increases and business performance”, and it reads like section 2.2 and section 2.3 are part of section 2.1 by just from their section titles. The authors should consider modify the titles to enhance the readability of the sections.	Thank you for your feedback. The subtitles have been modified.
5	In the third line of Page 6, what is the meaning of “XX century”? Is this a typo or an abbreviation? The authors should consider providing further explanations.	Thank you for your remark. In Poland, Roman numerals are used to indicate centuries. This notation has been corrected.
6	The table 1 is too large. Please change its width to fit the paper. This applies to Table 2, 3, and 5.	We agree that the tables look too wide. Unfortunately, this layout was introduced by the editorial office.
7	In Figure 1, only Opole data ranges from 2005 to 2015. The authors need to explain the reason of this difference. This applies to Figure 2 as well.	Thank you for your remark. An appropriate explanation has been provided in the notes below both figures.
8	The authors need to unify the template used in the paper. It can be seen that a different template was used from Page 18.	We agree that there are different templates. As with tables, the template is provided by the editorial office.
9	The equation numbers are not fully correct. Please revise them.	Thank you for your remark. The number of the third equation has been changed.
10	In the conclusion, the academic and practical contributions should be both mentioned, based on the comparison with previous literature. Also, the research limitation discussions should be enhanced and future research directions should be provided.	Thank you for your suggestion. The conclusion section was improved, and the next stage of research has been presented. We believe this will enhance the article's informative value.

Reviewer 4 Report

Comments and Suggestions for Authors

This paper presents an investigation into the impact of climate change, specifically temperature increases, on the bankruptcy risk of agricultural enterprises in Poland. The study is well-structured, methodologically sound, and makes a valuable contribution to the literature by incorporating regional climatic and macroeconomic variables.. However, several aspects could be strengthened.

1. The long-term average temperature is defined as the 2004–2023 average. Given that the study period is 2016–2023, please justify the choice of this baseline and discuss whether using a rolling or sector-specific baseline might affect the results.
2. The use of the 1.5-IQR rule for outlier removal is noted. However, the paper does not report how many observations were removed or discuss the potential impact of this removal on the representativeness of the sample.
3. While the Hausman test supports fixed effects, the paper does not discuss whether alternative specifications (e.g., random effects or dynamic panel models) were considered, especially given the presence of autocorrelation and heteroskedasticity.
4. The paper states that the sample consists of over 4,000 agricultural enterprises, but Table 1 shows varying observation counts per year (e.g., 556 in 2016 vs. 2444 in 2017). Please explain the reason for this fluctuation and how it may affect the panel data analysis. The authors should add more recently published related papers about agricultural development as complementary references, such as: 10.3390/land14040682.
5. The ambiguous effect of summer temperature on bankruptcy risk (e.g., positive linear effect in Z-score vs. weak parabolic in X-score) warrants deeper discussion. Are there agronomic or regional factors that might explain this inconsistency?
6. The paper identifies specific temperature thresholds beyond which risk increases (e.g., +6°C in winter for Z-score). Please provide confidence intervals or sensitivity analyses to support these threshold estimates.
7. While both models show a parabolic relationship, the direction and magnitude of coefficients sometimes differ (e.g., Table 4 vs. Table 5). A discussion on why these models might yield different insights would be useful.

 

Author Response

1	The long-term average temperature is defined as the 2004–2023 average. Given that the study period is 2016–2023, please justify the choice of this baseline and discuss whether using a rolling or sector-specific baseline might affect the results.	Thank you for your remarks. We used to estimate the long-term average all the data periods available to us. We believe that the long-term average should be based on the longest period possible, so that deviations from the average show better the differences from “normal” state. As a robustness check, we estimate also the models with deviations from 2004-2015 average, i.e., the period for which we have climate data and precede our panel model estimation period.
2	The use of the 1.5-IQR rule for outlier removal is noted. However, the paper does not report how many observations were removed or discuss the potential impact of this removal on the representativeness of the sample.	Thank you for your remark. In case of Z-score, 4808 observations were deleted as extreme values (1115 below and 3693 above the threshold). In case of X-score, 2340 observations were deleted (1137 below and 1203 above the threshold). We believe that this procedure does not influence the representativeness of the sample, which includes more than 22 thousand observations in case of Z-score and above 17 thousand observations in case of X-score.
3	While the Hausman test supports fixed effects, the paper does not discuss whether alternative specifications (e.g., random effects or dynamic panel models) were considered, especially given the presence of autocorrelation and heteroskedasticity.	Thank you for your suggestion. We estimated the basic models (Tables 4 and 5 in the text) using random-effects estimation and the Arellano-Bond estimator. The results are included in the Appendix section of the article. They are consistent with most of the models analyzed. We believe this will enhance the informative value of the article.
4	The paper states that the sample consists of over 4,000 agricultural enterprises, but Table 1 shows varying observation counts per year (e.g., 556 in 2016 vs. 2444 in 2017). Please explain the reason for this fluctuation and how it may affect the panel data analysis. The authors should add more recently published related papers about agricultural development as complementary references, such as 10.3390/land14040682.	Company observations vary from year to year due to the availability of data in the EMIS database, which is beyond the authors' control. Data for some of the companies studied covered periods shorter than 8 years, and some of them began and ended their operations in different years. We report this at the end of the article in the section on some of the weaknesses we noticed in the study results. Thank you for pointing out this article, which is relevant to our research. We used it in the introduction section.
5	The ambiguous effect of summer temperature on bankruptcy risk (e.g., positive linear effect in Z-score vs. weak parabolic in X-score) warrants deeper discussion. Are there agronomic or regional factors that might explain this inconsistency?	Thank you for your comment. The ambiguous effect of temperature changes in summer differs somewhat from results obtained in other seasons. In our opinion, the difference between the summer outcomes for the Z-score and X-score results from a difference in the distribution of these two variables. This observation is noted in the text, and histograms of the distributions of both variables are also included in the Appendix.
6	The paper identifies specific temperature thresholds beyond which risk increases (e.g., +6°C in winter for Z-score). Please provide confidence intervals or sensitivity analyses to support these threshold estimates.	Thank you for your remark. The review of this comment revealed an inconsistency, which has been corrected. The threshold values ​​have been entered in the table along with their level of statistical significance.
7	While both models show a parabolic relationship, the direction and magnitude of coefficients sometimes differ (e.g., Table 4 vs. Table 5). A discussion on why these models might yield different insights would be useful.	Thank you for your remark. A note has been added that some differences in the magnitudes of some coefficients in the X-score and Z-score models result, among other things, from a slight difference in the distributions of the two variables. We believe this will enhance the article's informative value.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have completed most of the revisions; however, I still have the following suggestions:

1. The paper indicates that the impact of summer temperature changes on bankruptcy risk is ambiguous, but only attributes this ambiguity to differences in the distributions of Z-score and X-score without conducting an in-depth analysis in combination with the seasonal characteristics of agricultural production in Poland. It is recommended to supplement the following content in the section on seasonal differences in 5. Discussion: ① The typical characteristics of agricultural production in Poland during summer. For instance, summer is the grain filling period, where moderate temperature increases may promote crop growth, but whether extreme high temperatures are mitigated by irrigation facilities? ② The potential adaptive behaviors of enterprises in response to summer temperature changes, such as whether they reduce risks by adjusting crop varieties or increasing irrigation investment, thereby offsetting the negative impacts of rising temperatures. This will provide a more practical basis for the conclusion that the impact of summer temperature changes is ambiguous, rather than confining the analysis merely to the statistical level.
2. The paper calculates the seasonal temperature thresholds through models but fails to explain the practical climatic significance of these thresholds. It is suggested to add the following content in the threshold analysis section of 4. Results: ① The proportion of times that annual/quarterly temperatures in each region exceeded the thresholds during the sample period from 2016 to 2023; ② Taking a typical region (e.g., the region with the greatest temperature variability) as an example, illustrate the corresponding relationship between the years when temperatures exceeded the thresholds and the enterprise bankruptcy risk indicators (decrease in Z-score, increase in X-score) in that region. This transformation will convert the thresholds from abstract statistical values into perceptible risk early warning lines.
3. The paper mentions that the sample includes a small number of large enterprises and a large number of small enterprises but does not analyze whether enterprise size affects the intensity of temperature impacts on bankruptcy risk (e.g., whether small enterprises are more sensitive to temperature changes due to limited resources). It is recommended to add a sub-sample regression by enterprise size in 4. Results: ① Divide the sample into three groups (micro-sized, small-sized, and medium-sized and above) based on enterprise asset size; ② Conduct regression analysis of temperature deviation - bankruptcy risk for each group respectively, supplement the core coefficients of each group in the table, and provide a brief analysis in 5. Discussion to enhance the depth of the results.

Author Response

1. The paper indicates that the impact of summer temperature changes on bankruptcy risk is ambiguous, but only attributes this ambiguity to differences in the distributions of Z-score and X-score without conducting an in-depth analysis in combination with the seasonal characteristics of agricultural production in Poland. It is recommended to supplement the following content in the section on seasonal differences in 5. Discussion: ① The typical characteristics of agricultural production in Poland during summer. For instance, summer is the grain filling period, where moderate temperature increases may promote crop growth, but whether extreme high temperatures are mitigated by irrigation facilities? ② The potential adaptive behaviors of enterprises in response to summer temperature changes, such as whether they reduce risks by adjusting crop varieties or increasing irrigation investment, thereby offsetting the negative impacts of rising temperatures. This will provide a more practical basis for the conclusion that the impact of summer temperature changes is ambiguous, rather than confining the analysis merely to the statistical level.

Response:

Thank you for your valuable suggestions on how to improve the quality of the article.

In response to the suggestions, we included information about the specifics of summer work in Polish agriculture, and we also justified this with conclusions from the article indicating adaptations to rising summer temperatures, including the use of modern irrigation systems, increasing the genetic diversity of plants, and using modern agrotechnics.

 

1. The paper calculates the seasonal temperature thresholds through models but fails to explain the practical climatic significance of these thresholds. It is suggested to add the following content in the threshold analysis section of 4. Results: ① The proportion of times that annual/quarterly temperatures in each region exceeded the thresholds during the sample period from 2016 to 2023; ② Taking a typical region (e.g., the region with the greatest temperature variability) as an example, illustrate the corresponding relationship between the years when temperatures exceeded the thresholds and the enterprise bankruptcy risk indicators (decrease in Z-score, increase in X-score) in that region. This transformation will convert the thresholds from abstract statistical values into perceptible risk early warning lines.

Response:

In response to your suggestion, we calculated the ratio of temperature exceedances in each region to all cases analyzed. This information was commented on, and the data are presented in an expanded Table 6.

1. The paper mentions that the sample includes a small number of large enterprises and a large number of small enterprises but does not analyze whether enterprise size affects the intensity of temperature impacts on bankruptcy risk (e.g., whether small enterprises are more sensitive to temperature changes due to limited resources). It is recommended to add a sub-sample regression by enterprise size in 4. Results: ① Divide the sample into three groups (micro-sized, small-sized, and medium-sized and above) based on enterprise asset size; ② Conduct regression analysis of temperature deviation - bankruptcy risk for each group respectively, supplement the core coefficients of each group in the table, and provide a brief analysis in 5. Discussion to enhance the depth of the results.

Response:

In response to your suggestion, the sample was divided into three subgroups: large, medium-sized, and small enterprises. Then, for each of them, a basic group of models, i.e., from 1 to 5, was estimated (the estimation results are included in the Appendix, as Tables A7-A12). The results were commented on, indicating, among other things, that the relationship between bankruptcy risk and temperature rise is most noticeable in the case of large and small enterprises and least noticeable in the case of medium-sized enterprises.

We would like to thank you again for your valuable suggestions, which we used to improve this article and determine directions for further research.

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank the authors' effort and time in revising the manuscript. They have addressed majority of my comments, and here are several minor points for them to consider:

1. I would recommend the authors to provide more relevant literature when building their hypothesis. For example, in the 5th of Section 2.1, there is one study cited for 11 lines of description. This also applies to the Section 2.2 and Section 2.3. To develop a solid hypothesis, more literature should be provided.

2. I would recommend the authors to provide the literature in the Section 6. For example, they should compare the previous literature with this study to highlight the novelty in the end of this paper. Also, when discussing the research gap, previous literature can be compared to justify the arguments.

3. The notations in the texts should have consistent format. For example, the notations after equation (3) have different formats. For Macro_ij, the authors use equation format. However, for other notations, plain texts are used. Please unify the whole manuscript's notation format.

Thanks!

Author Response

1. I would recommend the authors to provide more relevant literature when building their hypothesis. For example, in the 5th of Section 2.1, there is one study cited for 11 lines of description. This also applies to the Section 2.2 and Section 2.3. To develop a solid hypothesis, more literature should be provided.

Response:

Thank you for your valuable suggestions on how to improve the quality of the article.

In response to your suggestions, we have expanded the arguments for the proposed hypotheses in sections 2.1, 2.2, and 2.3 and also included new references.

2. I would recommend the authors to provide the literature in the Section 6. For example, they should compare the previous literature with this study to highlight the novelty in the end of this paper. Also, when discussing the research gap, previous literature can be compared to justify the arguments.

Response:

In response to the suggestion in section 6, we have included references to relevant literature to compare our results with the conclusions obtained so far.

3. The notations in the texts should have consistent format. For example, the notations after equation (3) have different formats. For Macro_ij, the authors use equation format. However, for other notations, plain texts are used. Please unify the whole manuscript's notation format.

Response:

In response to your suggestion, we have standardized the format of recording variables in the text.

Thanks!

We would like to thank you again for your valuable suggestions, which we used to improve this article and determine directions for further research.

Reviewer 4 Report

Comments and Suggestions for Authors

The authors notes that: "Thank you for pointing out this article, which is relevant to our research. We used it in the introduction section.". However, I did not see any relevant literature citations.

Author Response

The authors notes that: "Thank you for pointing out this article, which is relevant to our research. We used it in the introduction section.". However, I did not see any relevant literature citations.

Response:

First and foremost, we would like to apologize for the fact that when entering the article into the system, a version was mistakenly inserted that had been prepared before the comments from the proposed article were incorporated.

In the current version, we have taken into account the important findings, primarily the fact that the overall rise in temperatures has improved rice growing conditions and led to increased rice yields in northern regions of China, making the northern regions increasingly attractive for agriculture.

We would like to thank you again for your valuable suggestions, which we used to improve this article and determine directions for further research.

Round 3

Reviewer 4 Report

Comments and Suggestions for Authors

The authors notes that: "Thank you for pointing out this article, which is relevant to our research. We used it in the introduction section.". However, I did not see any relevant literature citations, such as 10.3390/land14040682.

Author Response

We are grateful for the comments we received in previous rounds. In accordance with the Reviewer’s comment, in the revised version, we have supplemented the article with additional literature items, including a work relevant to our research indicated by the Reviewer.

We would like to thank the Reviewer again for the valuable suggestions, which we used to improve this article and determine directions for our further research. 

Author Response File: Author Response.docx