Review Reports - From Construction to Operation: A Public Service Ecosystem Framework for Urban Rail Transit’s Economic Impact

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The core question addressed in this study is whether urban rail systems promote economic growth through agglomeration effects or, conversely, inadvertently hinder productivity development due to fiscal crowding-out effects. Additionally, the study explores the staged characteristics of the economic impacts of urban rail transit construction and operation, as well as the regional heterogeneity of such impacts between the eastern and central-western regions of China. The specific review recommendations are as follows：

1）Although the overall structure of the paper has shown a relatively complete framework, it is still necessary to provide strong support and explanation with a technical roadmap in order to gain a deep and accurate understanding of the operation steps adopted and the desired results.

2）According to the standard structure of common papers, the method part should have described the research steps in detail, followed by experiments and verification of relevant hypotheses. However, the current paper has the problem of unclear structure, and the method description content appears in the result part, so it is necessary to further sort out and optimize the overall structure of the paper.

3）The current conclusion part has the problem of redundancy in content. In order to improve the accuracy and effectiveness of the conclusion, it should be deeply streamlined, focusing on the core conclusions revealed by the calculation results, and ensuring that the conclusion part is concise and logical.
4）In order to further optimize the structure of the paper and enhance the logic and coherence of the discussion, it is recommended to adjust the relevant content of "limitations and future research work" to the "discussion" section.

Author Response

Referee: 1 (Please note that the line and page numbers indicated in the responses may change in the main text word file. Please use the uploaded main text pdf file for reviewing in that case.)
The core question addressed in this study is whether urban rail systems promote economic growth through agglomeration effects or, conversely, inadvertently hinder productivity development due to fiscal crowding-out effects. Additionally, the study explores the staged characteristics of the economic impacts of urban rail transit construction and operation, as well as the regional heterogeneity of such impacts between the eastern and central-western regions of China. The specific review recommendations are as follows：

Although the overall structure of the paper has shown a relatively complete framework, it is still necessary to provide strong support and explanation with a technical roadmap in order to gain a deep and accurate understanding of the operation steps adopted and the desired results.

We sincerely thank the reviewer for the constructive comment regarding the need for stronger support and explanation with a clear technical roadmap. In response, we have added Section 3.1 “Analytical and Technical Framework” in Chapter 3 “Materials and Methods”, where we provide a detailed explanation of how the PSE theoretical framework is integrated with the threshold model and the individual fixed-effects model (line 271-329, page 6-7). We have also included a comprehensive technical roadmap (line 330-332, page 8) to clearly illustrate the operational steps and expected outcomes of our study.

Specifically, we now emphasize the joint role of producers and consumers in value creation under the service-dominant logic of the PSE framework. The construction-phase model is developed from the provider’s perspective, while the operation-phase model is based on the user’s perspective. Value creation and value destruction are conceptualized as a multi-level dynamic process. This “multi-level” encompasses macro-level fiscal policy, meso-level technological capacity, micro-level human resources, and sub-micro-level consumer demand. To further control for externalities, we also consider cross-level environmental factors, which can be represented by an air quality index. Within this framework, transportation infrastructure and services jointly shape urban economic outcomes.

By controlling for related factors such as government expenditure, technology investment, and human resources, we are better able to observe the effects of other key factors, such as capital investment during construction and operational efficiency, on regional economic outcomes. During the construction phase, we posit that URT investment stimulates industrial development, employment growth, and land appreciation, thereby creating value and promoting economic growth. Next, we also highlight potential risks that misaligned investment or excessive debt may lead to value destruction. To capture these nonlinear effects, we employ a threshold model based on URT density, distinguishing between investment levels that generate net value creation and those that pose economic risks. During the operational phase, URT passenger flow intensity reflects system utilization efficiency, linking service provision to economic outcomes. High passenger flow facilitates labor mobility, supports agglomeration economies, and improves productivity, creating additional value. Conversely, operational inefficiencies, such as underutilized routes or distorted pricing, may result in value destruction and hinder economic development.

Technically, we follow a dual-model approach. The FET model captures nonlinear threshold effects during the construction phase, while the IFE model controls for city-specific unobserved heterogeneity and time-varying factors in the operational phase. This stepwise modeling strategy constitutes a clear technical roadmap: first, integration of multi-level control variables; second, application of the PSE framework with lifecycle perspective to emphasize operational-phase effects; third, quantification of nonlinear relationships and regional variations, linking mechanisms to observable outcomes.

Overall, we thank the reviewer for the valuable comments. These revisions provide a comprehensive and transparent roadmap of our methodology, thereby enhancing the clarity and persuasiveness of our study.

According to the standard structure of common papers, the method part should have described the research steps in detail, followed by experiments and verification of relevant hypotheses. However, the current paper has the problem of unclear structure, and the method description content appears in the result part, so it is necessary to further sort out and optimize the overall structure of the paper.

We are grateful to the reviewer for the insightful suggestions concerning the clarity and organization of the Methods and Results sections.

In response, we have thoroughly revised and systematically optimized the overall structure of the paper. Specifically, as mentioned in Comment 1, we have added Section 3.1 “Analytical and Technical Framework” in the Methods part to more clearly elaborate the research logic and technical roadmap. Subsequently, the research hypotheses are presented separately in Section 3.2 “Hypotheses Development”. Thereafter, we have sequentially arranged Section 3.3 Data Collection and Sample Selection, Section 3.4 Variable Selection and Definition, and Section 3.5 Model Design, thereby forming a structure that is logically clearer and more coherent.

In particular, to resolve the problem in the original manuscript where methods and results were mixed together, we have made the following changes: (1) The Descriptive Statistics originally placed in the “Methodology” section have been moved to Section 4.1 of Chapter 4 Results (lines 647- 660, page 16), ensuring that all results-related tables and figures are consistently presented in the Results section. (2) In Section 3.5 Model Design, we have added a new subsection, 3.5.3 Linearity Testing and Robustness Analysis, to consolidate methodological content previously scattered across the Results section. This new subsection covers the procedures for testing model nonlinearity, introducing lag terms in robustness checks, and applying the GMM method to enhance causal interpretation and address potential endogeneity issues (lines 605-631, pages 15).

Furthermore, we have refined some details. For example, the explanation regarding the constant term, which was originally in the Results section, has now been moved to the model specification in Section 3.5.1 “Construction Effect Model (Model I)” (lines 585-591, pages 14).

This adjustment improves the coherence of the Methodology section. After these revisions, the logical flow of the paper has been improved, and the description of the methodology is now separated from the results.

The current conclusion part has the problem of redundancy in content. In order to improve the accuracy and effectiveness of the conclusion, it should be deeply streamlined, focusing on the core conclusions revealed by the calculation results, and ensuring that the conclusion part is concise and logical.

We appreciate the reviewer’s careful suggestion regarding the redundancy in the conclusion section. In response, we have thoroughly reviewed and revised both the conclusion and the full manuscript.

First, we moved the original “Limitations and Future Work” from 6. “Conclusions” to 5. “Discussions”, making the conclusion more concise and focused on the core findings. (lines 865–882, page 25)

Second, we simplified the first key point in the conclusion. The original version was lengthy, providing a detailed description of the principles such as the nonlinear threshold effect, which had already been extensively discussed in Section 3.1 “Analytical and Technical Framework”. The revised version condenses this to: URT construction investment exhibits a nonlinear threshold effect on economic growth, and operational efficiency is positively correlated with economic development. Therefore, investment strategies should be dynamically adjusted according to development stages, focusing on construction investment in the early stage and on improving operational efficiency in the mature stage, including enhancing transport performance, service quality, and urban connectivity to sustain economic growth (lines 889–896, page 25-26).

Additionally, we carefully reviewed the entire manuscript for redundancy and removed repeated content. For example, the overly detailed “Limitations and Future Work” section was simplified to make the text clearer and easier to follow (lines 865–882, page 25).

We believe these revisions significantly enhance the conciseness and effectiveness of the conclusion, better highlighting the study’s core contributions.

In order to further optimize the structure of the paper and enhance the logic and coherence of the discussion, it is recommended to adjust the relevant content of "limitations and future research work" to the "discussion" section.

We sincerely appreciate your valuable suggestion regarding the placement of the “limitations and future research work” content. In response, we carefully reviewed the manuscript and moved all related content to the 5. “Discussion” section, and, as noted in Comment 3, this part has been simplified (lines 865–882, page 25).

In addition, we added comparisons with international studies in the Discussion section, which echoes the review of domestic and international research on the economic impacts of URT in the Literature Review, enhancing the logical flow and coherence of both the discussion and the entire manuscript. We are grateful for your insightful comment, which has improved the overall structure and readability of the paper, making the discussion more focused and meaningful (lines 834–864, page 24-25).

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review the paper “From Construction to Operation: A Public Service Ecosystem Framework for Urban Rail Transit’s Economic Impact.” The paper has clear objectives and is very well structured and written. The authors do a great job of reviewing the relevant literature and use a rigorous methodological framework to address their objectives. I believe the paper should be accepted for publication, but I have some comments that the authors may find helpful in delivering a more concise manuscript.

More precisely, by using the threshold approach, the authors demonstrate the creation of value added by the construction and operation phases of URT. They also show that this is conditioned by the control variables they used and that it differs between regions of the study area. On the other hand, I do not see how some of the findings mentioned in the discussion section are supported by the results. Specifically, in lines 641–643, the authors claim:

“On the other hand, URT construction and operation may cause resource misallocation or environmental pressures, leading to value destruction that limits sustainable economic development.”

I am not sure how this finding was derived from the models implemented in the analysis. I can see that environmental condition acts as a control variable for testing the effect of URT construction and operation on GDP. Thus, how is this relationship reversed to claim that URT construction may lead to environmental pressures? How can an independent variable exert an effect on a control variable?

Truth be told, when I started reading the paper, I had reservations about the inclusion of the environmental index in the models. All other variables somehow affect GDP, but I do not see how environmental conditions do the same. This could be better suited as a dependent variable in a study analyzing the environmental added value of URT, or even as a sub-indicator in a composite indicator measuring sustainability together with other metrics of economy, society, and environment. If reverse causality cannot be demonstrated by the models, then a stronger justification for selecting this variable is essential. Please try to find past studies that explicitly show that environmental conditions have a causal effect on GDP.

Considering the above, I would ask the authors to clearly explain what kinds of relationships are tested by their modelling approach. In this vein, I would like them to define precisely what “value creation” and “value destruction” mean and how the estimated effects of each model demonstrate the influence of each variable on them. In summary, I want to clearly see how the authors justify claims that independent variables affect control variables. Please review and revise all similar claims in the discussion section accordingly.

Additional comments

Lines 216–219: The research hypothesis is highly context-dependent. Considering that this is a research paper read by many scholars outside China, I would advise the authors to make their hypothesis more general. For example, instead of stating:

“And the threshold effect is stronger in the western regions.”
consider something like:
“…is stronger for regions with more advanced institutions or better local socioeconomic and environmental conditions.”
This would facilitate broader generalization of results and encourage future studies in the field. This comment should also be addressed in the conclusions section. Try to formulate more general statements starting from your case study, and avoid referring explicitly to the study area.

Lines 415–418 and others: Please rewrite the equations. The current font is unclear.

Line 664: Please avoid bold statements such as “resolves the academic debate.” Instead, use more cautious phrasing such as “sheds more light on the academic debate.”

Congratulations on this very nice paper.

Author Response

Referee: 2 (Please note that the line and page numbers indicated in the responses may change in the main text word file. Please use the uploaded main text pdf file for reviewing in that case.)
Thank you for the opportunity to review the paper “From Construction to Operation: A Public Service Ecosystem Framework for Urban Rail Transit’s Economic Impact.” The paper has clear objectives and is very well structured and written. The authors do a great job of reviewing the relevant literature and use a rigorous methodological framework to address their objectives. I believe the paper should be accepted for publication, but I have some comments that the authors may find helpful in delivering a more concise manuscript.

1.More precisely, by using the threshold approach, the authors demonstrate the creation of value added by the construction and operation phases of URT. They also show that this is conditioned by the control variables they used and that it differs between regions of the study area. On the other hand, I do not see how some of the findings mentioned in the discussion section are supported by the results. Specifically, in lines 641–643, the authors claim:

“On the other hand, URT construction and operation may cause resource misallocation or environmental pressures, leading to value destruction that limits sustainable economic development.”

We sincerely thank the reviewer for the valuable comments. The reviewer is correct in pointing out that our previous statement suggesting that URT construction and operation “cause environmental pressures” was misleading. We greatly appreciate the opportunity to clarify this point.

In our study, environmental conditions are included in the models as control variables rather than as core explanatory variables. In the original manuscript, this point was not clearly described. Therefore, in the revised manuscript, we have added a detailed explanation of the specific mechanisms linking the theoretical framework with the models in Section 3.1. “Analytical and Technical Framework”, along with a technical roadmap. We explicitly clarify that existing literature generally regards environmental quality as an important external factor affecting urban economic growth. Deteriorating environmental conditions may indirectly suppress GDP growth through increased public health expenditures and reduced labor productivity. Accordingly, the purpose of including environmental variables in our models is to control for external differences across cities, ensuring that the estimated effects of URT construction and operation on economic growth are more precise and not confounded by environmental factors (line 288-332 page 7-8).

In Section 5, Discussion, we highlight the interesting finding that the coefficient of the environmental control variable is negative and statistically significant, indicating that environmental conditions are indeed related to economic growth in our sample. Controlling for environmental factors is therefore necessary to obtain unbiased estimates of URT effects. However, we do not interpret this negative coefficient as evidence that URT construction or operation generates environmental pressures. Rather, it reflects the statistical relationship between the selected environmental indicator (e.g., days meeting air quality standards) and economic growth, which may be influenced by factors such as industrial restructuring, the closure of polluting firms, or investment restrictions. This observation has prompted further reflection on our part (line 856-864, page 25).

What’s more, in future research, following your suggestion, we plan to treat environmental factors as dependent variables or incorporate them into a composite sustainability index to more clearly explore the relationship between URT construction and operation and environmental outcomes. This has been addressed in point 2 of our future work (line 873-879, page 25).

In addition, we have reviewed and corrected all confusing statements regarding environmental factors throughout the manuscript. It is now explicitly clarified that the model is intended only to identify the effect of URT on economic growth while controlling for environmental conditions, without implying any causal effect of URT on the environment.

Considering the above, I would ask the authors to clearly explain what kinds of relationships are tested by their modelling approach. In this vein, I would like them to define precisely what “value creation” and “value destruction” mean and how the estimated effects of each model demonstrate the influence of each variable on them. In summary, I want to clearly see how the authors justify claims that independent variables affect control variables. Please review and revise all similar claims in the discussion section accordingly.

We sincerely thank the reviewer for the valuable feedback and for highlighting the need to clarify the relationships tested by our modeling approach, as well as the precise definitions of “value creation” and “value destruction.” We have carefully revised the manuscript to address these points.

First, in the Introduction, we critique traditional economic analysis frameworks for studying the economic impacts of transportation infrastructure and introduce the PSE theory, summarizing core concepts such as value creation and value destruction. Value creation refers to positive outcomes from multi-actor collaboration, such as industrial development, job creation, land appreciation, improved labor market efficiency, and business agglomeration. In contrast, value destruction arises from conflicts, coordination failures, or mismanagement, manifesting as fiscal pressures, resource waste, or public dissatisfaction. Urban economic growth is seen as a result of value creation, while economic downturn reflects value destruction. Recent studies emphasize the joint role of producers and consumers in value determination, calling this phenomenon value co-creation or value co-destruction (lines 98-144, pages 3).

Second, in Chapter 3, "Materials and Methods," we added Section 3.1, "Analytical and Technical Framework," detailing how PSE theory and the dual-model approach connect theoretical constructs with empirical mechanisms. We also included a technical roadmap, showing how multi-level control variables, lifecycle perspectives, and threshold effects jointly determine regional economic outcomes during the construction and operation phases of URT. This structured approach demonstrates how value creation and destruction are operationalized. During the construction phase, urban rail transit investment promotes industrial development, job growth, and land value appreciation, while misaligned investments or excessive debt may cause fiscal pressures, representing value destruction. In the operational phase, efficient services enhance labor mobility, support agglomeration economies, and improve productivity, whereas underutilized routes or inefficiencies can lead to value loss. These lifecycle perspectives provide clearer economic insights than traditional linear models (Figure 2, lines 271-332, page 6-8).

Third, we elaborate on the role of control variables in the model. We emphasize that, according to the PSE framework, value creation and destruction are dynamic processes spanning macro, meso, micro, and sub-micro levels, with control variables aligned accordingly. Specifically, macro-level variables reflect broad conditions for value creation, with government investment (GOV) representing fiscal capacity and policy support. Meso-level variables reflect technological capacity and innovation environments, with technology investment (TEC) capturing innovation spillovers from urban rail transit. Micro-level variables, such as human capital (HUM), represent effective labor mobility and its contribution to regional productivity. Sub-micro-level variables, such as market size (TRS), reflect local demand conditions. Additionally, environmental factors (ENV) are included as a composite sustainability control variable to better account for externalities. This approach allows us to better control the influence of control variables, focusing on the effects of urban rail transit investment and operational efficiency on economic outcomes (lines 288-329, page 7).

In summary, these revisions clarify the theoretical grounding, the modeling approach, and the interpretation of our findings. By defining value creation and destruction explicitly, detailing the relationships examined by each model, and carefully positioning control variables, we believe the manuscript now provides a more rigorous and transparent explanation of how URT affects urban economic outcomes.

Additional comments

Lines 216–219: The research hypothesis is highly context-dependent. Considering that this is a research paper read by many scholars outside China, I would advise the authors to make their hypothesis more general. For example, instead of stating:

“And the threshold effect is stronger in the western regions.”
consider something like:
“…is stronger for regions with more advanced institutions or better local socioeconomic and environmental conditions.”
This would facilitate broader generalization of results and encourage future studies in the field. This comment should also be addressed in the conclusions section. Try to formulate more general statements starting from your case study, and avoid referring explicitly to the study area.

We sincerely appreciate the reviewer’s thoughtful and constructive feedback on the context-specific nature of our research hypothesis. We fully recognize the importance of making our findings more broadly applicable, especially given that this research will be read by scholars from various regions and disciplines.

In response to the reviewer’s suggestion, we have revised the phrasing in lines 216-219 of the original manuscript. Instead of stating, "And the threshold effect is stronger in the western regions," we have generalized the hypothesis to: "…is stronger for regions with more advanced institutions or better local socioeconomic and environmental conditions".(lines 358-359, page 8) This revision allows the research to avoid geographical limitations while still capturing the underlying mechanisms that influence the threshold effect. It also helps to better align our hypothesis with broader research trends, making it applicable to regions beyond the study area.

Moreover, we conducted a thorough review of the manuscript, particularly the Discussion and Conclusions sections, to ensure that similar context-dependent statements were revised for broader generalizability. We have now framed our conclusions in a way that reflects the broader applicability of the findings without directly referring to the study area (lines 897-904, page 26).

In addition, we have incorporated an analysis of relevant international literature into the Discussion section to further support the generalization of our findings. This addition not only enriches the theoretical grounding of our study but also situates our research within a global context, offering more insights for future studies in the field.(lines 836-864, page 24-25)

We believe these revisions significantly enhance the clarity and universality of our conclusions, and we sincerely hope that these changes address the reviewer’s concerns. Thank you again for your valuable suggestions, which have greatly contributed to improving the overall quality of the manuscript.

Lines 415–418 and others: Please rewrite the equations. The current font is unclear.

We sincerely thank the reviewer for pointing out the issue with equation formatting. We have carefully rewritten all equations in Sections 4 Methodology and 5 Results, as well as throughout the manuscript, using a clear and consistent font to ensure readability and uniformity of all mathematical expressions (line 429-430, page 10; line 571-574, page 13; line 598-614, page 15).

Line 664: Please avoid bold statements such as “resolves the academic debate.” Instead, use more cautious phrasing such as “sheds more light on the academic debate.”

We sincerely thank the reviewer for the valuable suggestion. In response, we have revised the sentence on line 664 of the original manuscript, replacing the previous absolute expression “resolves the academic debate” with the more cautious phrasing “sheds more light on the academic debate” (line 884-888, page 25).

Furthermore, building on this revision, we conducted a thorough review of the entire manuscript to identify and adjust other overly absolute statements, ensuring that all claims are presented with appropriate caution and scholarly rigor. This meticulous, manuscript-wide revision reflects our commitment to precision and rigor, and we believe it significantly enhances the clarity and credibility of the paper.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Based on the theory of public service ecosystems (PSE), this paper uses panel data from 26 Chinese cities from 2007 to 2020 to analyze the impact of urban rail transit (URT) construction and operation on economic growth using a fixed-effect threshold model (FET) and an individual fixed-effect model (IFE). It also proposes policy recommendations for this period, which are of certain significance in revealing the relationship between URT and economic growth. However, this paper suffers from significant deficiencies in logical coherence, methodological rigor, depth of results, application to practical applications, and writing standards. The authors are advised to revise the manuscript comprehensively to enhance the scientific and practical value of the research. Specific issues are as follows:

1. The core data processing methods in the Abstract are too brief, and key steps in the panel data balancing process, such as the linear interpolation method for filling missing values, are unclear. This, to a certain extent, weakens the replicability of the research results and the credibility of the conclusions. It is recommended that the Abstract include additional information on the relevant data processing methods to enhance the completeness and formality of the content.

2. The criticism of traditional economic analysis frameworks in the Introduction is somewhat vague. It does not specifically explain the inherent limitations of input-output analysis and other methods. It is recommended to supplement relevant examples and literature citations to enhance the pertinence and academic depth of the theoretical critique.

3. The integration of the concept of "value co-creation and value co-destruction" in PSE theory with the economic impact mechanisms during the construction and operation phases of urban rail transit is insufficiently specific, failing to clearly explain how the theory transcends the limitations of traditional economic frameworks. It is recommended to supplement with a detailed analysis of the relevant mechanisms.

4. In this article, 26 cities were selected as a sample, but their representativeness in terms of urban size, economic level, and stage of urban rail transit development was not fully explained. The reasons for excluding other eligible cities (such as Chengdu, Chongqing, and other central and western cities) were also not fully explained. The representativeness of the sample is questionable.

5. In this article, the environmental indicator is measured by the proportion of days with air quality reaching level 2 or higher, but key environmental indicators related to urban rail transit, such as water quality and carbon emissions, are not included. This makes it difficult to fully reflect the impact of environmental factors on the economy. It is recommended to add other environmental indicators to enhance the validity of the research.

6. This article does not conduct an in-depth analysis of the heterogeneity of cities within the region (such as cities of different levels in the eastern region), and the explanation of the reasons for the differences is superficial. It is recommended to further stratify the analysis into "super first-tier cities, new first-tier cities, and second-tier cities," and to explain the reasons for the differences in regional industrial structure.

7. The logic behind the selection of control variables in this paper is unclear. Although the PSE "multi-level structure of value creation" framework is mentioned, the relationship between variables such as "government investment" and "technological investment" and the economic impact of urban rail transit is not clearly explained. Theoretical consistency in variable selection is insufficient, and further revisions and improvements are recommended.

8. Although the paper addresses existing issues in the literature review, the discussion of the research results does not systematically compare the differences and reasons with existing results (such as those on the economic impact of urban rail transit in Europe and the United States), and the application value of the research results is not yet fully demonstrated. It is recommended to supplement relevant analysis.

9. The charts and figures in this paper are significantly inadequate. The footnotes to Table 1 only list the mileage of urban rail transit in 2023 and do not explain the dynamic changes during the sample period, making it difficult to support phased analysis of the "construction phase" and "operation phase." Although Figures 2-4 illustrate the LR statistics of the threshold model, they do not label key threshold points and lack key information such as legends. This chart needs improvement in terms of information completeness and readability. Core elements should be clarified, formatting should be standardized, and the visualization should provide better support for the conclusions.

10. The formatting of the formulas in this article is disorganized and not edited or formatted as required. This requires revision.

11. The references in this article lack timeliness, with a low proportion of literature from the past five years, and insufficient citation of recent research on the economic impact of urban rail transit. It is recommended that representative research since 2020 be supplemented, particularly cutting-edge research on urban rail transit and regional economics, as well as the application of PSE theory, to enhance the timeliness of the theoretical support.

Author Response

Referee:3 (Please note that the line and page numbers indicated in the responses may change in the main text word file. Please use the uploaded main text pdf file for reviewing in that case.)
Based on the theory of public service ecosystems (PSE), this paper uses panel data from 26 Chinese cities from 2007 to 2020 to analyze the impact of urban rail transit (URT) construction and operation on economic growth using a fixed-effect threshold model (FET) and an individual fixed-effect model (IFE). It also proposes policy recommendations for this period, which are of certain significance in revealing the relationship between URT and economic growth. However, this paper suffers from significant deficiencies in logical coherence, methodological rigor, depth of results, application to practical applications, and writing standards. The authors are advised to revise the manuscript comprehensively to enhance the scientific and practical value of the research. Specific issues are as follows:

The core data processing methods in the Abstract are too brief, and key steps in the panel data balancing process, such as the linear interpolation method for filling missing values, are unclear. This, to a certain extent, weakens the replicability of the research results and the credibility of the conclusions. It is recommended that the Abstract include additional information on the relevant data processing methods to enhance the completeness and formality of the content.

We sincerely thank the reviewer for pointing out the need for clearer and more detailed explanations of the data processing methods in the Materials and Methods (rather than Abstract, if I understand correctly). In response, we have revised Section 3.3 Data Collection and Sample Selection to provide a comprehensive description of the construction of the panel dataset, including sample selection criteria, data sources, and key preprocessing steps.

Specifically, we further clarified the composition of the balanced panel dataset, which covers 26 Chinese cities with operational URT systems during 2007-2020. The 26 cities were listed one by one according to their regional distribution, and we explained both the geographical distribution and the rationale for their inclusion. We described in detail three selection criteria: (1) stable URT operation for no fewer than five years during the study period, i.e., cities with URT systems in operation before 2016; (2) regional representativeness consistent with the nationwide URT distribution, including major western cities such as Chongqing and Chengdu; and (3) adherence to the China Association of Metros (CAMET) classification standard for metro and light rail systems. These criteria ensure that the sample is representative, valid, and capable of capturing the dual impacts of the construction and operation phases (line 403-421, page 9-10)

Furthermore, we explicitly detailed the preprocessing methods employed to enhance data robustness. The paper provides a concise explanation of the linear interpolation method, which estimates missing values by drawing a straight line between the known data points before and after the missing observation. For example, a missing year such as 2016 is estimated based on data from the adjacent years 2015 and 2017 (line 422-429, page 10)

We also describe the winsorization procedure referenced in the text. Winsorization is a statistical technique that mitigates the influence of outliers by replacing the minimum and maximum values in a dataset with specified percentiles. Specifically, we applied winsorization to the top and bottom 1% of the data to reduce the impact of extreme values, followed by standardization of the variables through mean subtraction and division by the standard deviation. Each method is explained with both its rationale and illustrative examples, allowing readers to understand the procedure and the underlying assumptions.(line 430-434, page 10)

We believe that these revisions significantly improve the transparency, reproducibility, and credibility of our study. By explicitly documenting the data processing workflow, including the linear interpolation for missing values and the handling of outliers, we aim to provide a more complete and rigorous methodological foundation for future replication and validation of our results.

The criticism of traditional economic analysis frameworks in the Introduction is somewhat vague. It does not specifically explain the inherent limitations of input-output analysis and other methods. It is recommended to supplement relevant examples and literature citations to enhance the pertinence and academic depth of the theoretical critique.

We greatly thank the reviewer for the valuable comment regarding the lack of clarity in our critique of traditional economic analysis frameworks in the Introduction. In response, we have substantially expanded and refined both the Introduction and Literature Review sections. Specifically, we conducted a systematic review of studies on transportation infrastructure and traditional economic analysis frameworks in the Introduction with additional 16 relevant references and 5 carefully selected case studies. Through this systematic approach, we critically examined commonly used traditional methods, including Input-Output Analysis, the Neoclassical Growth Model, and Cost-Benefit Analysis, and identified their inherent limitations in studying the economic impacts of transportation infrastructure (line 44-97, page 2-3).

For each method, we have now provided detailed explanations, illustrative examples, and literature citations to clearly demonstrate their limitations. For instance, we discuss how Input-Output Analysis assumes linear and fixed inter-industry relationships, often relying on proxy derivations for regional data, which may reduce precision. The Neoclassical Growth Model’s assumptions of perfect competition and rational agents limit its ability to account for regional heterogeneity and differences in industrial structure. Similarly, Cost-Benefit Analysis, while valuable in ex ante evaluations, faces challenges in monetizing intangible benefits and is sensitive to context-specific conversion factors.

Building on these critiques, we summarize four major methodological shortcomings of traditional economic frameworks: (1) a producer-centric focus that neglects consumers as value co-creators; (2) emphasis on construction-phase impacts while often ignoring operational-phase benefits; (3) reliance on static or homogenizing assumptions that overlook inter-city and regional differences; and (4) linearity assumptions that fail to capture nonlinear, threshold, and dynamic effects in infrastructure lifecycle transitions.(line 83-97, page 2-3)

To address these limitations, we explicitly demonstrate the framework based on the Public Service Ecosystem (PSE). For issues 1 and 2, we analyze and advance the discussion from the perspective of PSE theory, establishing a comprehensive theoretical framework. For issues 3 and 4, we combine the Fixed Effects Threshold (FET) model and the Individual Fixed Effect (IFE) model to show how these gaps can be addressed. We provide a detailed discussion of how the FET model captures nonlinear threshold effects during the construction phase and how the IFE model accounts for unobserved heterogeneity in the operational phase. Each issue is analyzed in detail, highlighting the advantages of PSE theory as well as the FET and IFE models in addressing the economic impacts of URT. This lifecycle and multi-actor perspective enables a comprehensive assessment of both value creation and value destruction, integrating the effects of construction and operational phases in a manner that traditional frameworks cannot achieve (line 98-144, page 3).

Overall, the revised sections provide clearer theoretical critique, supported by systematic examples and literature citations, while also explaining how our proposed methodology overcomes the identified limitations. We believe these revisions substantially enhance the academic rigor, relevance, and depth of our critique, providing a more solid foundation for understanding the economic impacts of URT and other public infrastructure investments.

The integration of the concept of "value co-creation and value co-destruction" in PSE theory with the economic impact mechanisms during the construction and operation phases of urban rail transit is insufficiently specific, failing to clearly explain how the theory transcends the limitations of traditional economic frameworks. It is recommended to supplement with a detailed analysis of the relevant mechanisms.

We deeply thank the reviewer for the valuable comment on the need to more clearly integrate “value co-creation and value co-destruction” within PSE theory and to explain how it addresses the limitations of traditional economic analysis. In response, we have organized our revisions around 2 core ideas.

Firstly, as noted in Comment 2, in the Introduction, we conducted an in-depth analysis of traditional economic analysis frameworks used to study the economic impacts of transportation infrastructure and summarized four main limitations. We establish a comprehensive framework based on PSE theory to address the first two issues, and combine the FET and IFE models to demonstrate in detail how to tackle the latter two issues, fully highlighting the advantages of our theoretical framework and methodology. Additionally, in the Introduction, we provide a brief overview of the key mechanisms and clarify central concepts such as value creation and value destruction. Value creation refers to positive outcomes generated through multi-actor collaboration and interaction, such as promoting industrial development, increasing employment, enhancing land value, improving labor market efficiency, and facilitating enterprise agglomeration. In contrast, value destruction refers to reductions or losses in value caused by conflicts among actors, coordination failures, or mismanagement, which can manifest as fiscal pressures, resource wastage, or public dissatisfaction. Urban economic growth is regarded as an outcome of value creation, while economic downturn reflects value destruction. In recent years, an increasing number of scholars have emphasized the joint role of producers and consumers in determining value, referring to this phenomenon as value co-creation or value co-destruction (line 44-144, page 2-3).

Second, in Chapter 3 “Materials and Methods,” we specifically added Section 3.1 “Analytical and Technical Framework,” which details how PSE theory and the dual-model approach explicitly link theoretical constructs to empirical mechanisms. To clarify this integration, we included a technical roadmap (Figure 2, see line 330-332, page 8) illustrating how multi-level control variables, lifecycle perspectives, and threshold effects jointly determine regional economic outcomes during the construction and operation phases of URT. Overall, this structured approach demonstrates how value creation and value destruction are operationalized. During the construction phase, URT investment promotes industrial development, employment growth, and land appreciation, while misaligned investments or excessive debt may cause fiscal pressures or resource inefficiencies, representing value destruction. During the operational phase, efficient URT services enhance labor mobility, support agglomeration economies, and improve productivity, whereas underutilized routes or operational inefficiencies may result in value loss. These lifecycle perspectives, which consider the roles of both producers and consumers, provide a clearer understanding of economic impacts than traditional linear methods and producer-centric logic (line 271-329, page 6-7).

In summary, the revised Introduction, Literature Review, and Discussion sections clarify the theoretical critique, define value creation and destruction, and explicitly explain the modeling approach. This sequence ensures that the PSE framework, dual-model methodology, and URT lifecycle impacts are coherently integrated, providing a comprehensive, empirically grounded explanation of how URT affects regional economic outcomes.

In this article, 26 cities were selected as a sample, but their representativeness in terms of urban size, economic level, and stage of urban rail transit development was not fully explained. The reasons for excluding other eligible cities (such as Chengdu, Chongqing, and other central and western cities) were also not fully explained. The representativeness of the sample is questionable.

We sincerely thank the reviewer for raising the concern regarding the representativeness of our sample cities. We sincerely thank the reviewer for raising the concern regarding the representativeness of our sample cities. Our original sample did include Chengdu and Chongqing; however, in the original manuscript, these cities were only listed in a table and were not explicitly highlighted in the main text, which may led to the reviewer’s misunderstanding. In the revised version, we have now explicitly enumerated all 26 sample cities in Section 4.1 Data Collection and Sample Selection , distinguishing 16 eastern cities (Beijing, Tianjin, Shanghai, Guangzhou, Changchun, Dalian, Shenzhen, Nanjing, Shenyang, Suzhou, Hangzhou, Harbin, Ningbo, Wuxi, Qingdao, Fuzhou) and 10 central-western cities (Wuhan, Chongqing, Chengdu, Xi’an, Kunming, Zhengzhou, Changsha, Nanchang, Nanning, Hefei). In the revised manuscript, we also have highlighted Chengdu and Chongqing in the text to emphasize their important status in the central and western regions (line 403-409, page 9).

We have also supplemented the manuscript to clearly explain the sample selection principles. As mentioned in Comment 1, to ensure robust longitudinal analysis of URT’s economic impacts, we included only cities with URT systems in stable operation for at least five years during the study period 2007–2020, which corresponds to cities that commenced URT operations in 2016 or earlier. All 10 eligible central-western cities meeting this condition are included in our sample, explicitly covering major hubs like Chengdu and Chongqing to reflect regional diversity in urban size, economic development, and URT network maturity (line 410-421, page 9-10).

By clearly listing all the cities and explaining why they were included, and by reflecting the east–central–west distribution of URT in China, we hope the revised manuscript now clearly shows that our sample is representative. We sincerely thank the reviewer for raising this important point.

In this article, the environmental indicator is measured by the proportion of days with air quality reaching level 2 or higher, but key environmental indicators related to urban rail transit, such as water quality and carbon emissions, are not included. This makes it difficult to fully reflect the impact of environmental factors on the economy. It is recommended to add other environmental indicators to enhance the validity of the research.

We sincerely thank the reviewer for the valuable comment regarding the measurement of environmental indicators and acknowledge the importance of including diverse environmental metrics. In response to this suggestion, we collected complete carbon emissions data for all 26 sample cities covering the period 2007-2020. To test the robustness of our results, we tried two independent approaches: first, we directly included carbon emissions as an additional control variable in the models; second, using the principal component analysis (PCA) method, we combined air quality and carbon emissions into a composite environmental index, thereby integrating multiple related indicators into a single factor reflecting overall environmental conditions.

Our analyses show that incorporating carbon emissions—either as a separate control variable or within the composite index—has only a minor effect on the coefficients of other model variables, robustly supporting the main conclusion that depended on the air quality environmental indicator. Since the PCA method can integrate the two related indicators, carbon emissions and air quality, into a single composite index, it simplifies the model, reduces the number of variables, and makes the regression analysis more robust and easier to interpret. We have incorporated this result into the robustness check section of the main text 3.5.3 “Linearity Testing and Robustness Analysis” (line 625-631, page 15), 4.2.3 “Robustness Test for Model I” (line 755-759, page 21), 4.3.3 “Robustness test for model II” (line 800-805, page 23) and provided detailed comparative analyses and related tables in Appendix 1.

In addition, although in the robustness checks we incorporated an additional carbon emissions indicator into the composite environmental index, other environmental indicators, such as water quality, were not included. This omission is primarily due to data availability constraints, as collecting reliable and comparable water quality data within the limited timeframe proved challenging, and including incomplete or inconsistent data could compromise the robustness of our empirical analysis. Therefore, we have discussed this issue in the Discussion section and plan in future research to treat environmental factors as dependent variables, exploring a broader and more comprehensive composite index of environmental indicators to examine the relationship between URT construction and operation and environmental outcomes (line 873-879, page 25).

This article does not conduct an in-depth analysis of the heterogeneity of cities within the region (such as cities of different levels in the eastern region), and the explanation of the reasons for the differences is superficial. It is recommended to further stratify the analysis into "super first-tier cities, new first-tier cities, and second-tier cities," and to explain the reasons for the differences in regional industrial structure.

We thank the reviewer for the constructive suggestion regarding intra-regional heterogeneity. In response, we stratified the sample cities into first-tier and second-tier groups according to the China City Business Attractiveness Ranking (CCBAR) published by the China Urban Competitiveness Research Center. Given that the number of super first-tier cities is limited to four, the small sample size precludes panel threshold analysis. To ensure robustness and sufficient observations, we therefore adopted the broader classification of first-tier versus second-tier cities. Specifically, the first-tier group includes twelve cities: Beijing, Shanghai, Guangzhou, Shenzhen, Tianjin, Chongqing, Chengdu, Wuhan, Hangzhou, Nanjing, Xi’an, and Suzhou. The second-tier group comprises fourteen cities: Changchun, Dalian, Shenyang, Harbin, Ningbo, Wuxi, Qingdao, Fuzhou, Kunming, Zhengzhou, Changsha, Nanchang, Nanning, and Hefei. This classification aligns with CCBAR’s city hierarchy, ensuring sufficient observations for both groups while avoiding the small-sample issue associated with super first-tier cities. The specific methodology is presented in the newly added Section 3.5.4 City Tier Heterogeneity Analysis (line 632-645, page 15-16).

We then re-estimated both the threshold model for the construction effect and the individual fixed-effects (IFE) model for the operational effect within these subgroups. The results consistently support our main conclusions. For the construction effect, first-tier cities exhibit higher marginal returns to urban rail transit construction, reflecting their stronger absorptive capacity, more mature urban ecosystems, and more efficient governance structures. In contrast, second-tier cities show lower thresholds and faster coefficient attenuation, consistent with their relatively underdeveloped infrastructure and institutional capacity. Regarding the operational effect, first-tier cities display a stronger association between intellectual property output and economic growth, while second-tier cities also benefit, albeit to a slightly lesser extent. These differences underscore the role of city-level characteristics, including infrastructure maturity, institutional capacity, and network development, in shaping the economic impacts of urban rail transit. The specific analysis results can be found in the newly added sections 4.2.4 City Tier Heterogeneity Analysis for Construction Effect and 4.3.4 City Tier Heterogeneity Analysis for Operational Effect, as well as in the corresponding Table 9. The Threshold Statistics for First-tier and Second-tier Cities and Table 13. IFEM Statistics for First-tier and Second-tier Cities. (line 759-779, page 21-22; line 808-819, page 23-24)

Finally, we discuss these results in the Discussion and Conclusion section, highlighting that, beyond East and Central-West regional differences, city hierarchy plays a critical role in shaping both the value creation and potential value destruction effects of urban rail transit. This additional analysis not only strengthens the logical chain of the study but also demonstrates that classifying the 26 sample cities into first-tier and second-tier groups is both empirically feasible and theoretically meaningful, thereby better explaining the heterogeneity of cities and their differences. It further refines the policy implications, indicating that urban rail transit optimization strategies should consider both regional location and urban hierarchy, prioritizing service innovation in mature, high-tier cities while expanding coverage in developing, lower-tier cities. (line 825-833, page 24)

We once again thank the reviewer for this excellent suggestion. This classification enables a more thorough substantiation of our core argument and more effectively addresses regional heterogeneity.

The logic behind the selection of control variables in this paper is unclear. Although the PSE "multi-level structure of value creation" framework is mentioned, the relationship between variables such as "government investment" and "technological investment" and the economic impact of urban rail transit is not clearly explained. Theoretical consistency in variable selection is insufficient, and further revisions and improvements are recommended.

We sincerely thank the reviewer for the insightful comment regarding the clarity and theoretical consistency of control variable selection. In response, we have clarified and strengthened the rationale for including each control variable in the revised manuscript.

First, in Section 3.1 “Analytical and Technical Framework”, we explain key concepts of the PSE framework, such as the “multi-level structure of value creation.” We emphasize that, according to the PSE framework, which conceptualizes value creation and value destruction as a dynamic process spanning macro, meso, micro, and sub-micro levels, control variables are explicitly aligned with these hierarchical levels. Specifically, macro-level variables capture broad enabling conditions for value creation, with government investment (GOV) included to reflect fiscal capacity and policy support for urban development. Meso-level variables capture technological capacity and the innovation environment, with technology investment (TEC) incorporated to represent URT-related innovation spillovers that enhance economic impacts. Micro-level variables, such as human capital (HUM), reflect the effective utilization of labor mobility and its contribution to regional productivity. Sub-micro-level variables capture localized demand conditions, including market size (TRS). It is noteworthy that environmental factors are included in line with the PSE lifecycle perspective, moderating value creation and value destruction by accounting for the ecological and social context in which URT operates. Environmental factors (ENV) are positioned as comprehensive sustainability control variables, reflecting systemic constraints on cities’ overall capacity to generate economic value through URT investments, thereby better controlling for externalities (line 288-312, page 7). These updated explanations further clarify the relationship between variables, as suggested by the reviewer.

In addition, we provide a new Technical Framework in Figure2 to better describe the theoretical consistency in variable selection. Specifically, the hierarchical alignment of control variables and their links to URT’s economic impact are detailed in the new Technical Framework (Figure2). This approach ensures that variable selection is theoretically consistent with the PSE framework, captures multi-level influences on urban economic outcomes, and isolates the effects of URT construction and operation from broader socioeconomic and environmental factors (line 272-332, page 6-8).

Overall, these revisions address the reviewer’s concern by providing a transparent and theoretically grounded reasoning for each control variable, strengthening the conceptual consistency of the empirical models and reinforcing the validity of our findings.

Although the paper addresses existing issues in the literature review, the discussion of the research results does not systematically compare the differences and reasons with existing results (such as those on the economic impact of urban rail transit in Europe and the United States), and the application value of the research results is not yet fully demonstrated. It is recommended to supplement relevant analysis.

We sincerely appreciate the reviewer’s constructive comments regarding the discussion of our research findings and their comparison with existing studies of international cities. We fully recognize that the original manuscript did not systematically compare our findings with international data nor adequately highlight the application value of the results. In response to these concerns, we have substantially revised the discussion section to address them in a more detailed and organized manner.

Comparisons with domestic cases indicate that, despite differences in institutional arrangements and financing models, certain patterns identified in international cities are similar. In North America, research on light rail projects in Houston, Minneapolis, and Los Angeles demonstrates that constructing new lines in suburban or underdeveloped urban areas stimulates local economic activity, whereas operational improvements in high-density, high-demand corridors yield greater benefits for employment, property values, and commercial concentration. Credit and colleagues, using firm-level data from 1990 to 2014, found that the opening of the Phoenix light rail system significantly promoted the formation of new firms in knowledge, service, and retail sectors along its corridors, although this effect diminished over time and exhibited pronounced spatial decay. Crampton and colleagues observed that in the United Kingdom and France, light rail construction significantly increased the number of customers in city centers, while office property prices or rents along light rail corridors generally increased faster than in other areas. Additionally, the impact of light rail investment on economic development is most pronounced in non-residential or less mature urban areas. These international cases confirm our findings that urban rail construction tends to drive value creation in underdeveloped or emerging urban systems, while operational optimization generates economic value in mature cities. This consistency suggests that our conclusions based on the Public Service Ecosystem framework and phase-specific analysis represent a more general and even worldwide mechanism, offering valuable insights for international policymakers.

But it should be noted that, although some international studies exhibit patterns similar to those observed in our research, these cases are generally isolated and fragmented, and significant differences exist across countries in terms of institutional arrangements and financing models; therefore, the generalizability of these patterns still requires further in-depth investigation (line 821-865, page 24-25).

Overall, the revised discussion systematically compares our findings with international evidence, elucidates underlying mechanisms, and highlights policy relevance. We believe these revisions comprehensively address the reviewer’s concerns and significantly enhance the clarity, rigor, and applicability of our research.

The charts and figures in this paper are significantly inadequate. The footnotes to Table 1 only list the mileage of urban rail transit in 2023 and do not explain the dynamic changes during the sample period, making it difficult to support phased analysis of the "construction phase" and "operation phase." Although Figures 2-4 illustrate the LR statistics of the threshold model, they do not label key threshold points and lack key information such as legends. This chart needs improvement in terms of information completeness and readability. Core elements should be clarified, formatting should be standardized, and the visualization should provide better support for the conclusions.

We appreciate the reviewer’s careful and constructive comments regarding the charts and figures in our manuscript. In response to these concerns, we have made substantial revisions to all visual and tabular materials to enhance their clarity, completeness, and readability.

First, regarding Table 1(line 469-470, page 11), we have recollected and updated the urban rail transit mileage data to correspond with the study period, adjusting it to reflect the status as of December 31, 2020. This adjustment allows Table 1 to provide a solid empirical foundation for the phased analysis of urban rail transit construction and operation stages. The updated table presents key information for all 26 cities, including system name, region, opening year, mileage, and number of lines.

Second, we have improved Figures 3-5 (original Figures 2-4) (line 689-690, page 18; line 714-715, page 19; line 716-717, page 20) to enhance interpretability and information completeness. Specifically, all key threshold points in the threshold model have been clearly marked in red, allowing readers to visually identify critical transitions in the effects of urban rail transit investment. We also carefully reviewed all legends, axis labels, and formatting to ensure that each figure is fully self-explanatory and visually coherent. In addition, consistency checks were conducted across all other tables and figures to ensure the accuracy and readability of the entire manuscript.

The text also has been reviewed and edited by an expert in the field and a native speaker to further enhance the language expression and logical coherence. We believe that these improvements substantially address the reviewer’s concerns by providing complete, phase-relevant data and clear visualization of threshold effects, thereby strengthening the empirical support for our conclusions. The corresponding revisions are reflected in Table 1 and Figures 2-4 of the revised manuscript. We hope that the enhanced tables and figures now meet high standards of clarity and informativeness, and we are grateful to the reviewer for their guidance in improving the presentation of our research.

The formatting of the formulas in this article is disorganized and not edited or formatted as required. This requires revision.

We sincerely thank the reviewer for pointing out the formatting issues with the formulas. In response, we have carefully revised and standardized the formatting of all equations in the manuscript. Specifically, Equations 1 through 7 have been fully adjusted to ensure clarity, consistency, and adherence to the journal’s formatting requirements. Each formula now follows a uniform style, with proper alignment, numbering, and notation, making them easier to read and interpret. We believe these revisions substantially improve the presentation and professional appearance of the mathematical content, and we are grateful for the reviewer’s constructive guidance (line 429-430, page 10; line 571-574, page 13; line 598-614, page 15).

The references in this article lack timeliness, with a low proportion of literature from the past five years, and insufficient citation of recent research on the economic impact of urban rail transit. It is recommended that representative research since 2020 be supplemented, particularly cutting-edge research on urban rail transit and regional economics, as well as the application of PSE theory, to enhance the timeliness of the theoretical support.

We sincerely appreciate the reviewer’s comment regarding the timeliness of our references. In response, we carefully reviewed and supplemented our literature selection to ensure that recent, relevant studies are appropriately cited. In particular, we paid close attention to research published since 2020 that directly addresses the economic impacts of urban rail transit, regional economic development, and the application of Public Service Ecosystem (PSE) theory.

After careful screening, we cited 138 references, of which 61 were published between 2020 and 2025, accounting for 42.3% of the total. We deliberately prioritized recent studies that are highly relevant to our research topic, ensuring that both empirical findings and theoretical developments are well represented. Although the total number of references is substantial, we focused on the quality and direct relevance of the literature rather than sheer quantity, in order to maintain academic rigor.

We hope that this careful selection and the inclusion of recent cutting-edge research demonstrate our commitment to providing up-to-date theoretical and empirical support, and we believe it substantially enhances the timeliness and credibility of our study.

Author Response File: Author Response.pdf