Regional Finance and Environmental Outcomes: Empirical Evidence from Kazakhstan’s Regions

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this paper, the authors examine the impact of regional finance on the transition to a low-carbon economy in Kazakhstan’s regions. They employ an applied panel-data econometric approach to analyze the relationship between regional financial variables and environmental outcomes across four regions over the period 2010–2024. The data are drawn from official national statistical sources. The authors conclude that deeper regional credit markets are associated with lower emissions, whereas budgetary expansion linked to carbon-intensive activities and fiscal centralization constrains regional progress toward low-carbon development.

I would like to offer the following recommendations to improve the paper.

Research aim and conceptual clarity
The stated aim is to analyze the impact of finance on the transition to a low-carbon economy. This framing implies directional change, causal relationships, and a clear link to decarbonization rather than to pollution levels alone. However, the authors estimate a static random-effects model in which emissions are related to bank lending, regional budgets, and unemployment. The analysis does not incorporate dynamics, transition paths, or counterfactual scenarios. Moreover, the concept of “low-carbon development” is not clearly defined. In the literature, this concept is typically operationalized using emissions intensity, decoupling indicators, energy structure, or carbon intensity rather than emissions levels alone.

Literature review
The literature review adopts an unconventional approach by relying heavily on bibliometric maps based on keyword clustering. While this classification of topics is informative, it does not substitute for positioning the paper within a specific scholarly debate. The authors should more clearly explain what previous studies have done, what their main findings are, and how the present paper contributes to or extends those findings. As currently written, the review catalogs research themes rather than situating the paper within an identifiable research conversation.

Methodology and interpretation of results
The empirical results should be interpreted as robust associations rather than as causal effects.

Endogeneity is not addressed. Bank lending and emissions are likely jointly determined: cleaner regions may attract more credit, while more polluted regions may face tighter credit constraints. Similarly, regional budgets are endogenous to economic and industrial structure.

Model choice (FE vs. RE) is not justified. The authors report results from a random-effects model without presenting fixed-effects estimates or reporting a Hausman test. A clear rationale for choosing random effects over fixed effects is needed.

Lack of dynamics. The model does not include lagged variables or adopt a dynamic panel framework (e.g., ARDL, PMG, or related approaches). As a result, the specification remains static and cannot capture adjustment processes over time.

Small cross-section. The analysis is based on only four regions observed over 2010–2024. Consequently, inference relies almost entirely on time variation, and standard errors may be optimistic. The authors should explicitly acknowledge the limited external validity and the inability to generalize the findings.

Three-dimensional cube and quadrant analysis. These tools are visually appealing and pedagogically useful, but they do not add inferential power beyond the regression analysis and should not be presented as substitutes for econometric identification.

Conclusions and policy implications
The conclusions are directionally strong but empirically modest. While the estimations support empirical associations between finance and emissions, they do not identify causal mechanisms. As a result, several policy recommendations extend beyond what can be inferred from the econometric evidence.

Comments on the Quality of English Language

The manuscript would benefit from a thorough revision of the writing. There is excessive and inconsistent hyphenation throughout the text (e.g., lines 31–35). In addition, some citations appear to be incorrect or incomplete; for instance, the citation reported in Figure 4 is inaccurate and should be corrected.

Author Response

Response to Reviewer 1

Comment 1. Research aim and conceptual clarity
The stated aim is to analyze the impact of finance on the transition to a low-carbon economy. This framing implies directional change, causal relationships, and a clear link to decarbonization rather than to pollution levels alone. However, the authors estimate a static random-effects model in which emissions are related to bank lending, regional budgets, and unemployment. The analysis does not incorporate dynamics, transition paths, or counterfactual scenarios. Moreover, the concept of “low-carbon development” is not clearly defined. In the literature, this concept is typically operationalized using emissions intensity, decoupling indicators, energy structure, or carbon intensity rather than emissions levels alone.

Response: We thank the reviewer for this important comment. We agree that the use of the term “low-carbon transition” implies an evolution of adaptation strategies and the underlying processes that go beyond what a fixed data set can capture. Consequently, we have updated the study’s objective and reframed the work’s value to emphasize robust links between financial elements and regional environmental outcomes, abandoning the study of changes over time. More specifically, the study now focuses on differences between regions in terms of emissions metrics relevant to sustainable growth within existing economic systems. The manuscript clarifies that our study does not model development shifts or hypothetical situations; instead, it identifies robust links between finance and environmental burdens. Furthermore, the revised paper provides a clear definition of low-carbon development and explicitly addresses the limitations associated with the use of overall emissions data rather than output-based measures.

We agree that emissions intensity is a more appropriate indicator of low-carbon development. Accordingly, we supplemented our findings with the calculation of the emissions-to-GDP ratio, which reflects the degree of environmental efficiency and the relative disparity between economic activity and emissions (Figure 10).

The main results remain qualitatively unchanged when adjusting emissions for GDP or population, indicating that the identified relationships are not solely due to regional effects. The relationship between emissions and gross regional product serves as an intensity indicator (emissions intensity (emissions-to-GDP ratio)), which reflects both the environmental performance and the carbon footprint of regional economies. Unlike overall emissions metrics, this calculation demonstrates how effectively economic growth avoids negative environmental impacts and is therefore widely used in studies of low-carbon development and sustainable growth.

 

Comment 2. Literature review
The literature review adopts an unconventional approach by relying heavily on bibliometric maps based on keyword clustering. While this classification of topics is informative, it does not replace the need to position the paper within a specific scholarly debate. The authors should more clearly explain what previous studies have done, what their main findings are, and how the present paper contributes to or extends those findings. As currently written, the review catalogs research themes rather than situating the paper within an identifiable research conversation.

Response: Thank you for your comments and recommendations. We have included a key explanation of the role of this study in the modeling section (Table 4), which details how the scientific discussions influenced the study's findings (subject rationale, choice of methods, and factors underlying the modeling). Furthermore, in accordance with your recommendations, we have expanded the conclusions in the "Literature Review" section.

 

Comment 3. Methodology and interpretation of results
Comment 3.1. The empirical results should be interpreted as robust associations rather than as causal effects.

Comment: We agree with the reviewer's assessment. In the updated version of the manuscript, we now clearly understand that the observed results reflect relationships, not causation. Consequently, the wording suggesting causation has been amended in the abstract, results section, and concluding remarks.

The article title has been changed to: Regional Finance and Environmental Outcomes: Evidence from Kazakhstan’s Regions

Comment 3.2. Endogeneity is not addressed. Bank lending and emissions are likely jointly determined: cleaner regions may attract more credit, while more polluted regions may face tighter credit constraints. Similarly, regional budgets are endogenous to economic and industrial structure.

Response: Recognizing that bank loans, local finances, and pollution levels might influence each other, we refrain from establishing cause-and-effect relationships because of restricted data and a limited number of regions studied. Rather, this work aims to pinpoint consistent connections among these factors. This constraint openly addressed in the updated version of our document.

While we recognize that bank loans, local finance, and pollution levels might influence each other, we refrain from establishing causal relationships due to data limitations and the small number of regions studied. Instead, this paper aims to identify robust relationships between these factors. This limitation is explicitly addressed in the updated version of our paper.

We agree that emissions intensity is a more appropriate indicator of low-carbon development. Accordingly, we supplement our findings by calculating the emissions-to-GDP ratio, which reflects the degree of environmental efficiency and the relative disparity between economic activity and emissions (Figure 10). The main results remain qualitatively unchanged when adjusting emissions for GDP or population, indicating that the identified relationships are not solely driven by regional effects. The relationship between emissions and gross regional product serves as an intensity measure (emissions intensity (emissions-to-GDP ratio)), which reflects both environmental performance and the carbon footprint of regional economies. Unlike overall emissions figures, this calculation demonstrates how effectively economic growth avoids negative environmental impacts and is therefore widely used in studies of low-carbon development and sustainable growth.

While methods such as instrumental variables or generalized method of moments can solve endogeneity issues, their effective application requires a significant amount of observational data (instrumental variables and generalized method of moments are not applicable at N = 4). Given that our dataset includes only four regions, these methods will not yield robust statistical inferences. Therefore, we avoid attempting to establish causality and instead view the results as reflecting the underlying relationships.

As part of our work on these comments, we additionally estimated the model using lagged variables. This is a standard procedure for partially eliminating endogeneity when it is necessary to exclude the influence of current emissions on current financial indicators.

The results confirmed our basic findings: the coefficients on the unemployment rate and budget expenditures remained statistically significant and retained their signs. The relationship between emissions and investment proved less stable over time, emphasizing their short-term impact. Despite the natural decrease in explanatory power (R^2), this modification of the model only confirms the validity of our main analytical conclusions."

 

Variable	Coefficient	Standard error	z-statistics	p-value
Constant	154 020	134 690	1,144	0,253
Unemployed population, thousands of people (x36₍t−1₎)	2 271,0	1 046,0	2,171	0,030
Bank loans to the economy (x106₍t−1₎)	−0,085	0,039	−2,180	0,029
Regional budget (x66₍t−1₎)	0,066	0,073	0,912	0,362

 

Variable	Base RE	RE with lags	Conclusion
x36 (unemployment)	+, *	+ , * (t−1)	The effect is stable and has increased over time.
x106 (bank loans)	− , ***	− , * (t−1)	Sustained reduction effect.
x66 (budget)	+ , ***	insignificant	Short-term effect, not persistent.
R²	0.76	0.32	Expectedly lower (stricter model).

Comment 3.3. Model choice (FE vs. RE) is not justified. The authors report results from a random-effects model without presenting fixed-effects estimates or reporting a Hausman test. A clear rationale for choosing random effects over fixed effects is needed.

Response: In response to the reviewer's comments, this version includes both fixed- and random-effects analyses, as well as a Hausman test. The results of this test justify the use of a random-effects model. A detailed explanation of our modeling decisions has been added to the methods section.

We are grateful to the reviewer for their attention to the issue of choosing a panel data specification. Following this recommendation, we conducted a comparative analysis of fixed- and random-effects models. A Hausman test revealed a χ² statistic of 0.97 with a p-value of 0.808. Since the null hypothesis regarding the appropriateness of using random effects was not rejected, we can conclude that there is no correlation between the individual characteristics of the regions and the set of explanatory variables. Therefore, the random-effects (RE) model is consistent within this study and was retained as our main specification. Here are the results:

Table Y. Hausman Test

Indicator	Result
χ²	0,971
Degrees of freedom	3
p-value	0,808

 

Comment 3.4. Lack of dynamics. The model does not include lagged variables or adopt a dynamic panel framework (e.g., ARDL, PMG, or related approaches). As a result, the specification remains static and cannot capture adjustment processes over time.

Response: While we acknowledge that the transition to low carbon emissions involves ongoing change, data limitations prevent us from using analytical methods such as GMM, PMG, or ARDL. Therefore, our updated paper clearly states that the model illustrates fixed relationships rather than reflecting evolving processes.

Comment 3.5. Small cross-section. The analysis is based on only four regions observed over 2010–2024. Consequently, inference relies almost entirely on time variation, and standard errors may be optimistic. The authors should explicitly acknowledge the limited external validity and the inability to generalize the findings.

Response: Given the limited scope of the study areas, we understand that the results of this study apply only to Kazakhstan's regions. This updated version clearly outlines the limitations of the generalizability of the results and emphasizes how preliminary the findings remain.

Comment 3.6. Three-dimensional cube and quadrant analysis. These tools are visually appealing and pedagogically useful, but they do not add inferential power beyond the regression analysis and should not be presented as substitutes for econometric identification.

Response. It's acknowledged that using three-dimensional cube and quadrant analyses fails to yield econometric identification. Consequently, the updated manuscript clarifies their role, defining them as supportive, descriptive elements supplementing the regression analysis instead of serving as instruments for inference. Comments on the findings have been corrected in Appendix B (Three-dimensional cube) and Figure 12 (quadrant analysis).

Comment 4. Conclusions and policy implications
The conclusions are directionally strong but empirically modest. While the estimations support empirical associations between finance and emissions, they do not identify causal mechanisms. As a result, several policy recommendations extend beyond what can be inferred from the econometric evidence.

Response. Considering the additions made based on the reviewer's comments and recommendations, the revised version of the article reflects the findings provided in the conclusion. Specifically, the Literature Review section includes conclusions reflecting the contribution of this research to substantiating the relevance of the study and the potential for further research. The Methodology section explains the limitations that shaped the research approach and the findings.

Comment 5. The manuscript would benefit from a thorough revision of the writing. There is excessive and inconsistent hyphenation throughout the text (e.g., lines 31–35). In addition, some citations appear to be incorrect or incomplete; for instance, the citation reported in Figure 4 is inaccurate and should be corrected.

Response. The manuscript was revised. The citation and reference links were revised, checked, and updated (in case of need)

Reviewer 2 Report

Comments and Suggestions for Authors

The article is very interesting and the authors used the statistical tool appropriately. The strategic implications part could be improved if the authors consider the proposal adequate. The research is valuable the more adequate the implementation solutions are. Currently, a significant part of these revenues goes to the national budget and the National Fund, instead of remaining at the regional level to support the ecological transformation. The Karaganda and East Kazakhstan regions should benefit from dedicated strategic recommendations. Using the “regional sustainability quadrants” model, the article could propose sets of specific recommendations for each identified group. Please indicate the software used in the statistical modeling (R?). How was the data accuracy testing performed?

Author Response

Comment 1. Using the “regional sustainability quadrants” model, the article could propose sets of specific recommendations for each identified group.

Response: Thank you for your recommendations. We have taken them into account, compiled recommendations for each group, and created Table 5 of Targeted Recommendations for Kazakhstan's Regions according to the Sustainability Quadrant of Regions model.

 

Comment 2. Please indicate the software used in the statistical modeling (R?).

Response: There was used R-Studio 3.5.0

 

Comment 3. How was the data accuracy testing performed?

Response: Data accuracy and validation. A systematic verification process evaluated data accuracy. Variables originated from reputable national and local statistical agencies, guaranteeing substantial organizational trustworthiness. Initially, the data underwent review to detect absent entries, typing mistakes, and measurement discrepancies. Subsequently, descriptive measures and temporal graphs aided in pinpointing unusual observations, sudden shifts, and unlikely figures.
To maintain logical soundness, interconnected metrics were compared for sensible agreement (such as emissions concerning gross regional product and populace dimensions). As needed, variables were adjusted or converted to lessen magnitude impacts and enhance uniformity among areas and periods. Lastly, different modeling approaches and variable descriptions were calculated as sensitivity tests to confirm that core findings weren’t influenced by data irregularities. Taken together, these actions secure the exactness and dependability of the research dataset.

A brief description of the data accuracy check is included in the Methods section.