Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript investigates the dynamic evolution of the global timber trade network using a Stochastic Actor-Oriented Model and a multidimensional cost framework. The topic is timely, the dataset is rich, and the methodological approach is theoretically relevant to forest economics, trade modeling, and network analysis. understanding how cost factors and country attributes shape trade ties can inform sustainable forest trade policy. The manuscript shows strong potential.  However, there are several issues related to structure, clarity, terminology, statistical reporting, and interpretation of results. Some sections are overly long and important methodological assumptions need clearer justification. Some suggestions for the authors to further improve and advance their work are provided below:

Line 31: The study should more directly and effectively highlight its fundamental gap, namely the absence of a systematic explanation of how cost factors guide the dynamics of the timber trade network. Organizing the literature review to focus on why incorporating both trade costs and resource costs into a dynamic network model is innovative would make the introduction even more effective.

Line 295- 308: Four network structural effects are listed when describing the SAOM methodology: outdegree, reciprocity, Transitive Triplets/3-Cycles, and the GWESP effect. This presentation is somewhat confusing and terminologically inconsistent. In SAOM terminology, typically one would include a transitive triplets effect and possibly a 3-cycles effect to capture triadic closure. GWESP is a term from Exponential Random Graph Models, not commonly used in SAOM, unless the  specifically implemented a GWESP-type effect. If the model indeed includes both a transitive closure effect and a GWESP statistic, you need to clarify why both are included and how they differ in this context. Including multiple highly related measures of triadic closure raises concerns of collinearity and interpretational overlap.

Line 332-335: Several predictor variables (GDPP, LAW, FTA, PRICE, DIS) may be highly correlated. I recommend reporting VIF values or providing a cross-correlation matrix to assess multicollinearity.

Line 375: “This study ingeniously derives price data by calculating the ratio of trade value to trade volüme...” The use of “ingeniously” here is overly strong and somewhat out of place in an academic paper. And these words give me an AI-generated impression. Similarly, I encountered this impression in many of the expressions used in the introduction (e.g.  Many scholars, offers valuable insights ..)

Line 533: Please add an explanation below Table 3 indicating what the asterisks represent. This clarification is included in other tables but is missing here, even though it is particularly important for this one.

Line 420-421: Although data from 2000 to 2024 are available, only three time points (2010, 2015, 2020) were selected for the dynamic network analysis. The rationale for choosing these specific years as waves should be better explained. It needs to be clarified why the analysis begins in 2010 (and not in 2000 or 2005) and ends in 2020 instead of 2024.

Line 600-601: The manuscript frequently states that a hypothesis “is verified” (or “partially verified” in some cases) when describing results. In scientific writing, it is more appropriate to say a hypothesis is “supported” (or not supported) by the data, rather than “verified.” Verification implies absolute proof, which is not achievable with statistical inference. For example, here it is confirmed that greater distances reduce the likelihood of trade ties, consistent with H1a-this should be phrased as “and Hypothesis H1a is supported.” Throughout the results and discussion, I recommend changing wording like “verified” or “passes the significance test, verifying H…” to phrases such as “provides support for Hypothesis H..” or “Hypothesis H… is supported/confirmed by this result.”

Line 633-635: It is stated that timber price (a key resource cost) has a coefficient of 0.05 and is not statistically significant, thus “failing to validate H2a.” This non-significant result is acknowledged, but it would benefit from further reflection. Timber price might be expected to influence trade ties (higher prices could incentivize exporters or deter importers), so why is it insignificant here? The authors should discuss possible reasons. For example, is it because timber prices are relatively homogeneous globally or co-move due to world market forces, thus providing little variation to explain the formation of new ties? Or perhaps because any effect of price is indirect and already captured by other variables (like trade agreements or resource endowment)? A short discussion of why “timber price alone is insufficient to reshape trade ties” (as noted in the abstract) would add insight. This could be placed in the discussion section, tying into literature on commodity price volatility vs. network stability.

Line 665-671: In discussing the interaction between TII and economic development, the phrasing is a bit confusing. The interaction term TII ego-GDP alter has a positive coefficient (0.02). The text states this indicates “TII exerts a significant negative moderating effect on the relationship between economic development level and network evolution.” The word “negative” here is hard to parse, since the coefficient is positive. I believe the intended meaning is that being a net exporter (high TII ego) counteracts the negative baseline effect of partner’s high GDP on tie formation. Recall, the direct GDP effect is -0.52 (line 662), implying that, in general, ties are less likely toward economically advanced countries – possibly because developed countries import less. The positive interaction (0.02) means that for net-exporter countries, this negative effect of partner GDP is reduced. In other words, net exporters prefer economically developed destination countries, effectively reversing the baseline tendency. The authors do convey this later by saying “net timber-exporting countries preferentially target relatively developed economies for their exports”. To avoid confusion, I suggest rephrasing the earlier sentence.

Author Response

Comments 1:Line 31: The study should more directly and effectively highlight its fundamental gap, namely the absence of a systematic explanation of how cost factors guide the dynamics of the timber trade network. Organizing the literature review to focus on why incorporating both trade costs and resource costs into a dynamic network model is innovative would make the introduction even more effective.

Response 1: We sincerely appreciate your valuable suggestion. Following your recommendation, we have revised the introduction and literature review sections to explicitly articulate the innovative significance of integrating both trade costs and resource costs within the dynamic network model, as well as to highlight the systematic explanation this study provides for the evolution mechanisms of the timber trade network. Additionally, we have refined relevant formulations to eliminate ambiguity and ensure clearer logic and more prominent contributions.

 

Comments 2: Line 295- 308: Four network structural effects are listed when describing the SAOM methodology: outdegree, reciprocity, Transitive Triplets/3-Cycles, and the GWESP effect. This presentation is somewhat confusing and terminologically inconsistent. In SAOM terminology, typically one would include a transitive triplets effect and possibly a 3-cycles effect to capture triadic closure. GWESP is a term from Exponential Random Graph Models, not commonly used in SAOM, unless the specifically implemented a GWESP-type effect. If the model indeed includes both a transitive closure effect and a GWESP statistic, you need to clarify why both are included and how they differ in this context. Including multiple highly related measures of triadic closure raises concerns of collinearity and interpretational overlap.

Response 2: We sincerely appreciate your valuable suggestion. Your perspective is highly insightful. The inclusion of the GWESP effect in our model was guided by existing literature; however, we acknowledge that this approach may have overlooked potential issues of collinearity and interpretational overlap arising from multiple highly correlated triadic closure measures. We will carefully address this aspect in our future research.

 

Comments 3: Line 332-335: Several predictor variables (GDPP, LAW, FTA, PRICE, DIS) may be highly correlated. I recommend reporting VIF values or providing a cross-correlation matrix to assess multicollinearity.

Response 3: We sincerely thank you for this valuable suggestion. We fully understand your concern regarding potential multicollinearity among predictor variables, which is crucial for ensuring the robustness of the model results.

It is important to note that the estimation process of the Stochastic Actor-Oriented Model (SAOM) employed in this study—based on the simulation method of moments—fundamentally differs from that of traditional regression models (e.g., OLS). As such, it does not directly compute or report the Variance Inflation Factor (VIF).

Nevertheless, we have accorded significant attention to this issue and have elaborated on the multicollinearity assessment in the revised manuscript. Following established practices in SAOM research (Block et al., 2022; Snijders, 2010, 2017; Ceoldo et al., 2024), we adopted widely recognized diagnostic strategies to evaluate dependencies among variables, including:

(1) Diagnostics during model estimation: We closely monitored two key indicators:

• Overall stability of estimates: All final models demonstrated good convergence (with all absolute t-ratios < 0.1), and parameter estimates remained stable across different model specifications.
• Magnitude of standard errors: We paid particular attention to the standard errors of effect parameters. Severe multicollinearity typically manifests as abnormally large standard errors for one or more effects. In our results, all standard errors fell within reasonable ranges, with no such warning signs observed.

(2) Model specification and comparison: We examined the robustness of key variables by constructing nested models (with variables introduced sequentially). The stability of significance levels and directions of key variables across model comparisons further strengthens our confidence that the results are not substantially affected by severe multicollinearity.

In summary, although VIF is methodologically inapplicable in SAOM, we have systematically evaluated and addressed potential multicollinearity using the above approaches, which represent standard practice in the field.

 

Comments 4: Line 375: “This study ingeniously derives price data by calculating the ratio of trade value to trade volüme...” The use of “ingeniously” here is overly strong and somewhat out of place in an academic paper. And these words give me an AI-generated impression. Similarly, I encountered this impression in many of the expressions used in the introduction (e.g.  Many scholars, offers valuable insights ..)

Response 4: We are grateful for your suggestion. In response, we have thoroughly revised the manuscript to eliminate overly assertive tones and any AI-generated characteristics, with all modifications marked as tracked changes.

 

Comments 5: Line 533: Please add an explanation below Table 3 indicating what the asterisks represent. This clarification is included in other tables but is missing here, even though it is particularly important for this one.

Response 5: We thank the reviewer for this valuable suggestion. Accordingly, we have incorporated the recommended annotations into Table 3.

 

Comments 6: Line 420-421: Although data from 2000 to 2024 are available, only three time points (2010, 2015, 2020) were selected for the dynamic network analysis. The rationale for choosing these specific years as waves should be better explained. It needs to be clarified why the analysis begins in 2010 (and not in 2000 or 2005) and ends in 2020 instead of 2024.

Response 6: We sincerely appreciate this insightful suggestion. The selection of three time points (2010, 2015, 2020) for the dynamic network analysis was based on several methodological and practical considerations. First, the SAOM framework requires a minimum of three waves of data to reliably estimate network dynamics, while maintaining consistent country coverage across all periods. Expanding the number of periods would have substantially reduced the number of countries present in all waves, thereby limiting the scope and comparability of the analysis. Second, our approach aligns with established empirical precedents in which three-wave designs have been widely adopted in longitudinal network studies. Finally, the 2020 cutoff reflects data availability constraints—the Global Forest Resources Assessment data were only updated through 2020 at the time of our empirical analysis. We have added clarification on these points in the revised manuscript. We fully acknowledge the value of your suggestion and intend to explore longer time series with more observation points in future research.

 

Comments 7: Line 600-601: The manuscript frequently states that a hypothesis “is verified” (or “partially verified” in some cases) when describing results. In scientific writing, it is more appropriate to say a hypothesis is “supported” (or not supported) by the data, rather than “verified.” Verification implies absolute proof, which is not achievable with statistical inference. For example, here it is confirmed that greater distances reduce the likelihood of trade ties, consistent with H1a-this should be phrased as “and Hypothesis H1a is supported.” Throughout the results and discussion, I recommend changing wording like “verified” or “passes the significance test, verifying H…” to phrases such as “provides support for Hypothesis H..” or “Hypothesis H… is supported/confirmed by this result.”

Response 7: We are grateful for your suggestion. In response, we have revised the relevant expressions in the manuscript to enhance their compliance with academic conventions.

 

Comments 8:Line 633-635: It is stated that timber price (a key resource cost) has a coefficient of 0.05 and is not statistically significant, thus “failing to validate H2a.” This non-significant result is acknowledged, but it would benefit from further reflection. Timber price might be expected to influence trade ties (higher prices could incentivize exporters or deter importers), so why is it insignificant here? The authors should discuss possible reasons. For example, is it because timber prices are relatively homogeneous globally or co-move due to world market forces, thus providing little variation to explain the formation of new ties? Or perhaps because any effect of price is indirect and already captured by other variables (like trade agreements or resource endowment)? A short discussion of why “timber price alone is insufficient to reshape trade ties” (as noted in the abstract) would add insight. This could be placed in the discussion section, tying into literature on commodity price volatility vs. network stability.

Response 8: We thank the reviewer for the valuable suggestion, which has been instrumental in enhancing the quality of our manuscript. Following this recommendation, we have expanded the Discussion section by adding a mechanistic analysis of why price factors alone are insufficient to reshape trade relationships.

 

Comments 9:Line 665-671: In discussing the interaction between TII and economic development, the phrasing is a bit confusing. The interaction term TII ego-GDP alter has a positive coefficient (0.02). The text states this indicates “TII exerts a significant negative moderating effect on the relationship between economic development level and network evolution.” The word “negative” here is hard to parse, since the coefficient is positive. I believe the intended meaning is that being a net exporter (high TII ego) counteracts the negative baseline effect of partner’s high GDP on tie formation. Recall, the direct GDP effect is -0.52 (line 662), implying that, in general, ties are less likely toward economically advanced countries – possibly because developed countries import less. The positive interaction (0.02) means that for net-exporter countries, this negative effect of partner GDP is reduced. In other words, net exporters prefer economically developed destination countries, effectively reversing the baseline tendency. The authors do convey this later by saying “net timber-exporting countries preferentially target relatively developed economies for their exports”. To avoid confusion, I suggest rephrasing the earlier sentence.

Response 9: We sincerely appreciate your attentive feedback. A clerical error was identified in our interpretation of the coefficient for the interaction term between TII and economic development level, where a positive effect was incorrectly described as negative. This has been corrected in the manuscript. Additionally, we have refined the relevant explanations to improve clarity and better highlight the key findings.

Reviewer 2 Report

Comments and Suggestions for Authors

Determinants of the Global Timber Trade Network Evolution A Stochastic Actor-Oriented Model Analysis

This manuscript presents a substantial and timely contribution to the study of global timber trade networks (GTTN). By employing the Stochastic Actor-Oriented Model (SAOM) and integrating structural network mechanisms with trade costs and resource characteristics, the authors offer valuable insights into the evolution of international timber trade relationships from 2000 to 2024. The focus on sustainable governance—through forest management capacity, certification incentives, and trade policies—aligns well with current research priorities in the forestry and sustainability fields.

The manuscript is methodologically solid, well-motivated, and empirically relevant. Its findings concerning asymmetric trade behaviors among net exporters and the influence of network structures (e.g., triadic closure) are particularly noteworthy. The study has clear implications for global forest governance and sustainable trade practices.

However, there are several areas where the manuscript could be strengthened to enhance clarity, analytical depth, and policy relevance.

 

Major Comments

1. Clarity and Structural Flow

• Some explanations of complex network mechanisms could be clearer and more intuitive. Reducing technical jargon where possible would increase accessibility.
• Consider reorganizing sections so that the main empirical findings are presented before detailed methodological explanations, which may help guide the reader more effectively.

1. Elaboration on Policy Implications

• The manuscript offers policy-relevant insights but does not fully articulate practical recommendations. The authors are encouraged to clarify how structural mechanisms—such as triadic closure—might inform the design of trade agreements, certification schemes, or forest management policies.
• The discussion of certification systems and blockchain traceability would benefit from more explicit analysis of how these tools could influence trade relationships in practice.

1. Inclusion of Ecological and Environmental Variables

• The exclusion of ecological factors, such as climate change impacts on forest resources, is a notable gap. Integrating such variables in future work would strengthen the ecological grounding of the analysis.
• Even a brief discussion of how ecological uncertainties may affect network evolution or trade costs would help contextualize the results.

1. Granularity and Heterogeneity

• The country-level focus is appropriate for macro-level analysis, but may mask important differences at the regional or firm level.
• Future research could examine how sub-national actors or industry players contribute to network formation, potentially revealing additional mechanisms or strategic behaviors.

1. Technical Details and Reproducibility

• More detailed explanation of data sources, preprocessing, and model specifications would improve transparency and facilitate reproducibility.
• Providing supplementary material (if possible) with model parameters or code snippets would further support robustness.

1. Limitations and Future Directions

• The manuscript acknowledges certain limitations, such as data granularity and missing ecological variables, but these points could be expanded.
• The authors may consider outlining specific future research avenues, such as the incorporation of environmental models, firm-level datasets, or applying the approach to other natural resource sectors.

 

 

Author Response

Comments 1: Clarity and Structural Flow

Some explanations of complex network mechanisms could be clearer and more intuitive. Reducing technical jargon where possible would increase accessibility.

Consider reorganizing sections so that the main empirical findings are presented before detailed methodological explanations, which may help guide the reader more effectively.

Response 1:We sincerely thank you for your suggestions. In accordance with your recommendations, we have revised the explanations of complex network mechanisms in method section to enhance readability and reorganized the structure by adjusting the sequence of sections.

 

Comments 2: Elaboration on Policy Implications

The manuscript offers policy-relevant insights but does not fully articulate practical recommendations. The authors are encouraged to clarify how structural mechanisms—such as triadic closure—might inform the design of trade agreements, certification schemes, or forest management policies.

The discussion of certification systems and blockchain traceability would benefit from more explicit analysis of how these tools could influence trade relationships in practice.

Response 2: We sincerely thank you for your highly valuable suggestions, which have significantly contributed to enhancing the quality of our manuscript. Following your recommendations, we have revised the policy recommendations in the Conclusion section to strengthen their actionable nature.

 

Comments 3: Inclusion of Ecological and Environmental Variables

The exclusion of ecological factors, such as climate change impacts on forest resources, is a notable gap. Integrating such variables in future work would strengthen the ecological grounding of the analysis.

Even a brief discussion of how ecological uncertainties may affect network evolution or trade costs would help contextualize the results.

Response 3: We sincerely appreciate your valuable suggestion. Your perspective is highly insightful, and we fully recognize the importance of incorporating ecological factors into our research model. In subsequent studies, we will implement your recommendations to enhance the scholarly contribution of our research.

 

Comments 4：Granularity and Heterogeneity

The country-level focus is appropriate for macro-level analysis, but may mask important differences at the regional or firm level.

Future research could examine how sub-national actors or industry players contribute to network formation, potentially revealing additional mechanisms or strategic behaviors.

Response 4: We sincerely appreciate your insightful suggestion. Your perspective is highly innovative and aligns well with our ongoing research endeavors to investigate the driving mechanisms behind the dynamic evolution of China's timber trade network at the sub-national provincial level. In our subsequent studies, we will carefully examine how to systematically incorporate industry participants into our research framework to explore the micro-level formation mechanisms of timber trade networks.

 

Comments 5：Technical Details and Reproducibility

More detailed explanation of data sources, preprocessing, and model specifications would improve transparency and facilitate reproducibility.

Providing supplementary material (if possible) with model parameters or code snippets would further support robustness.

Response 5: We sincerely appreciate your valuable suggestions. In response, we have supplemented the manuscript with more detailed descriptions of the data sources, preprocessing procedures, and model specifications. Furthermore, we have included a statement in the manuscript to confirm that the datasets and model code used in this study will be made available upon request to interested readers. These revisions aim to enhance the methodological transparency and reproducibility of our research.

 

Comments 6：Limitations and Future Directions

The manuscript acknowledges certain limitations, such as data granularity and missing ecological variables, but these points could be expanded.

The authors may consider outlining specific future research avenues, such as the incorporation of environmental models, firm-level datasets, or applying the approach to other natural resource sectors.

Response 6: We sincerely appreciate your valuable suggestions, which have been instrumental in shaping our future research directions. In accordance with your recommendations, we have expanded the discussion in the manuscript regarding limitations such as data granularity and missing ecological variables, while also elaborating on potential avenues for future research. These revisions further enhance the rigor and completeness of our study.

Reviewer 3 Report

Comments and Suggestions for Authors

The paper is interesting and is also in the interest of the readers. To highlight its contributions, the paper should be revised for the following aspects.

 

1) In abstract, the multidimensions can be detailed. 

2) What is export–import paradox, its theoretical background should be given, do authors mean Lerner paradox? Since what authors mean is the focus of nations to preserve their resources, it is not clear if it is a paradox. A better term or revision is needed with improved explanation. It is stated that "Resource-rich countries exhibit an export–import paradox, actively expanding exports while restricting imports to safeguard resource sovereignty" Isn't it the opposite, resource abundant nations tend to preserve resources by encouraging imports instead of limiting them? 

3) In the introduction section, this section is very well formulated with strong references to global events including Covid-19. Could author discuss timber trade more in context of various economic policies and business cycle relations, and also, can authors extend the sustainability issues more? 

4) In literature section, the flow can be improved, references could be extended and a more sharp literature can be formed. Repetetive concepts be checked. In hypothesis formation, Gravity model can be also used, since the model already allows extended forms with different factors that influence trade other than economic size and distance. In fact, factors included are existence of multinational corporations and intratrade, cultural affinity, borders, and many other forms of restrictions, and FTA's helping on shaping trade and also by eliminating distances other than kilometers. Hence, the paper assumes that traditional gravity omitted many factors, which is not the case. 

5) In methodology, equations are very primitive. Original sources and an equation editor can be used for improvement of presentability. Also equation numbers must be added. Referencing should be added with justification of added variables / factors in equations, not only in the introduction paragraph. 

6) In Data section, units of data, transformations, distributional characteristics (skewness kurtosis JB test), why logarithms for some variables why not for others, and by leading to lin-log form for some variable relations, while not for others should be justified. Interpretations should be checked for these aspects. For instance, it is stated that "To ensure the data scale aligns with the model's requirements, logarithmic transfor- 426
mations were applied to the data on per capita forest stock volume, the number of chain- 427
of-custody certifications, inter-country distances, and per capita GDP."
which does not really explain why it was needed for them, and why not for others. In fact, a better approach would be to apply log to all variables to avoid unit effects and to control nonlinearity in addition to achieve better distributional properties in accordance with the method used. 

7) In figures, sources should be stated, (authors own calculations with...).
Similar to Table references, Figure references should be in capital letters in the text. 

8) For results, is it possible to report goodness of fit measures as models progress from Model 1 to 6? 

9) Too close to zero parameter estimates in all models and even insignificance for interaction terms, their interpratations can be improved in the text. 

10) Limitations, policy implications and future directions subsections would bring the importance of the paper. Also, conclusion should highlight the contribution.  

11) Extending references with more recent studies from 2024, 2025, 2026.

Author Response

Comments 1：In abstract, the multidimensions can be detailed. 

Response 1:We sincerely thank you for this valuable suggestion. Accordingly, we have revised the abstract to explicitly specify the dimensions covered by the term "multidimensions", which include trade costs, resource costs, network structure, and trade structure.

 

Comments 2：What is export–import paradox, its theoretical background should be given, do authors mean Lerner paradox? Since what authors mean is the focus of nations to preserve their resources, it is not clear if it is a paradox. A better term or revision is needed with improved explanation. It is stated that "Resource-rich countries exhibit an export–import paradox, actively expanding exports while restricting imports to safeguard resource sovereignty" Isn't it the opposite, resource abundant nations tend to preserve resources by encouraging imports instead of limiting them? 

Response 2:We sincerely thank you for this valuable suggestion. Upon careful reconsideration, we agree that the term "export–import paradox" may not precisely capture the phenomenon described in our context. Accordingly, we have revised the manuscript to refer to this pattern as the "phenomenon of export expansion and import contraction," which more accurately reflects the empirical observations and theoretical framing of our study. Your feedback has been instrumental in enhancing the conceptual clarity of our work.

Comments 3： In the introduction section, this section is very well formulated with strong references to global events including Covid-19. Could author discuss timber trade more in context of various economic policies and business cycle relations, and also, can authors extend the sustainability issues more? 

Response 3: We sincerely thank you for your valuable suggestions, which have provided clear direction for our subsequent in-depth research. Following your recommendations, we have refined the literature review in Section 2 to include a detailed discussion on how economic policies—represented by trade agreements—influence the evolution of the timber trade network. We acknowledge that the current study has limitations in addressing the compounded effects of various economic policies and business cycle fluctuations, as well as in exploring sustainability issues more comprehensively. In future research, we will systematically develop a dual-impact framework that integrates economic policies (including industrial policies, tariff structures, and trade agreements) with business cycles—particularly fluctuations during the post-pandemic global economic recovery—to examine their combined effects on timber trade. Additionally, we will extend the discussion on sustainability to encompass dimensions such as forest carbon sinks, biodiversity conservation, and circular economy principles, thereby constructing a holistic eco-economic synergistic analytical perspective. This will allow us to clarify the transmission pathways of cost factors under the joint influence of policies, cyclical fluctuations, and sustainability requirements.

 

Comments 4：In literature section, the flow can be improved, references could be extended and a more sharp literature can be formed. Repetetive concepts be checked. In hypothesis formation, Gravity model can be also used, since the model already allows extended forms with different factors that influence trade other than economic size and distance. In fact, factors included are existence of multinational corporations and intratrade, cultural affinity, borders, and many other forms of restrictions, and FTA's helping on shaping trade and also by eliminating distances other than kilometers. Hence, the paper assumes that traditional gravity omitted many factors, which is not the case. 

Response 4: We sincerely thank you for your valuable suggestions, which have been instrumental in enhancing the quality of our manuscript. In accordance with your recommendations, we have thoroughly revised the literature review section by incorporating additional references, consolidating redundant concepts, and refining the formulation of hypotheses based on extended forms of the gravity model.

 

Comments 5：In methodology, equations are very primitive. Original sources and an equation editor can be used for improvement of presentability. Also equation numbers must be added. Referencing should be added with justification of added variables / factors in equations, not only in the introduction paragraph. 

Response 5:We sincerely appreciate your thorough and valuable suggestions. In response to your feedback, we have revised the manuscript by:

1. Citing original references for all equations,
2. Utilizing equation editor for proper mathematical formatting,
3. Numbering all equations systematically,
4. Providing justification for variable inclusion at each equation.

These revisions have significantly improved the academic rigor and reproducibility of our methodology section.

 

Comments 6：In Data section, units of data, transformations, distributional characteristics (skewness kurtosis JB test), why logarithms for some variables why not for others, and by leading to lin-log form for some variable relations, while not for others should be justified. Interpretations should be checked for these aspects. For instance, it is stated that "To ensure the data scale aligns with the model's requirements, logarithmic transfor- 426
mations were applied to the data on per capita forest stock volume, the number of chain- 427
of-custody certifications, inter-country distances, and per capita GDP."
which does not really explain why it was needed for them, and why not for others. In fact, a better approach would be to apply log to all variables to avoid unit effects and to control nonlinearity in addition to achieve better distributional properties in accordance with the method used. 

Response 6:We sincerely thank you for your valuable comments. Following your suggestions, we have expanded the Data section to provide detailed explanations regarding data units, transformation methods, and distributional characteristics. It is important to note that our dataset comprises not only country-level attributes (e.g., GDP per capita, forest stock per capita) but also relational variables between countries (e.g., FTA networks, common language networks, distance networks). This structural particularity necessitates a different data processing approach compared to conventional regression models. Your question regarding the selective application of logarithmic transformations is well-founded. During our empirical analysis, we also carefully considered this issue. We observed that certain variables—such as the dependent variable NET, common language network (binary values of 0 or 1), TII, LAW, and PRICE—exhibit values primarily in the two-digit range. In contrast, variables like FSP, GDPP, COC, and DIS assume considerably larger magnitudes. To maintain approximate scale consistency and comply with model requirements, we followed established research(Jesse,2016; Zhou et al., 2022; ) by applying logarithmic transformations specifically to per capita forest stock volume, chain-of-custody certifications, inter-country distances, and per capita GDP.

Your insightful suggestion is highly valuable for our future research. We will adopt more systematic data treatment methods in subsequent studies to mitigate unit effects and better control for potential nonlinear relationships.

 

Comments 7: In figures, sources should be stated, (authors own calculations with...).
Similar to Table references, Figure references should be in capital letters in the text. 

Response 7:We sincerely appreciate your constructive suggestion. In response, we have revised the manuscript to explicitly clarify the data sources and standardized the formatting of references to figures and tables to uppercase. These improvements have enhanced the clarity and consistency of our presentation.

 

Comments 8: For results, is it possible to report goodness of fit measures as models progress from Model 1 to 6? 

Response 8: We sincerely appreciate your valuable comment regarding the importance of goodness-of-fit in evaluating model results. We fully acknowledge that assessing model fit is crucial for validating the SAOM framework.

In response to your suggestion, we would like to clarify that the SAOM methodology—specifically as implemented through the RSiena package—does provide goodness-of-fit diagnostics to evaluate how well the model reproduces the observed network dynamics. The following key diagnostics were examined in our study:

Overall Model Convergence: The estimation procedure provides an overall maximum convergence ratio. As recommended in the methodological literature, values below 0.25 indicate satisfactory convergence. All models reported in our manuscript meet this criterion.

Parameter-Level Convergence (t-ratios): For each estimated effect, RSiena reports a t-ratio (estimate divided by standard error). A well-fitting model is generally expected to have absolute t-ratios below 0.1 for most parameters, and certainly below 0.3, with none substantially exceeding 1.0. All key parameters in our models satisfy these conditions, indicating no significant convergence issues.

Based on these established diagnostics, we confirm that the models presented in our manuscript demonstrate adequate goodness of fit.

Thank you again for raising this important point, which has allowed us to more fully articulate the robustness of our empirical approach.

 

Comments 9: Too close to zero parameter estimates in all models and even insignificance for interaction terms, their interpratations can be improved in the text. 

Response 9: We sincerely appreciate your valuable suggestions. In accordance with your recommendations, we have expanded the Discussion section to include explanations for the relatively small regression coefficients of certain variables and the non-significance of others. Regarding your specific comment on the non-significant interaction term coefficient, we have carefully reconsidered this issue. The non-significance of the TII ego and CER alter interaction term in the current model may likely stem from a convergence of factors such as data quality and methodological limitations. This finding nonetheless holds intrinsic value, as it suggests that the moderating effect of the trade imbalance index on the relationship between forest certification level and timber trade network dynamics may be limited at the macro level, or that its influence operates in more complex and indirect ways. Further investigation using more granular data and more powerful analytical methods in future studies will be essential to elucidate this relationship. We have incorporated relevant additions in both the Result and Discussion sections to address these points.

 

Comments 10: Limitations, policy implications and future directions subsections would bring the importance of the paper. Also, conclusion should highlight the contribution.  

Response 10: We thank the reviewer for this constructive suggestion. In response, we have restructured relevant sections to better articulate the study's limitations and future research directions. Additionally, we have strengthened the conclusion by elaborating on policy implications and highlighting the study's contributions.

 

Comments 11: Extending references with more recent studies from 2024, 2025, 2026.

Response 11: We thank the reviewer for this valuable suggestion. Following the recommendation, we have now incorporated several recent references into the manuscript to enhance the timeliness and relevance of the literature support.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

All critiques are satisfied with the revisions. I am happy to see this improved version.

 

There is one very very minor issue, a typo: in Eq. (8), the if condition is stated with smaller and larger than signs (<,>). This should be smaller and equal or larger (<, <=).