Cross-Layer Influence of Multiple Network Embedding on Venture Capital Networks in China: An ERGM-Based Analysis

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This article discusses an important and timely research issue related to the emergence of venture capital networks and the determinants of this process in the contemporary economy. Although this issue is analyzed using China as an example, the research results and conclusions are universal enough to be considered for other countries as well. They can be used to study this phenomenon in developing or developed countries, but facing significant challenges in the development of venture capital funds and their involvement in technology enterprises.

For the purposes of this research, the authors adopted a network perspective and created a multilevel model of network embeddedness. They systematically analyzed the factors behind venture capital network development across three dimensions: endogenous network structures, multidimensional relational networks between VC firms, and informal networks of venture capital investors. The approach to the research problem contributes new knowledge and helps narrow the research gap in this area. The authors' results confirm the hypotheses and allow for the development of recommendations for venture capital fund managers regarding the development and management of their networks.

Despite the overall high rating for this work, I believe that with minor additions and corrections, its quality could be further improved. Below are some suggestions in this regard.

1. I believe it's worthwhile to outline the structure of the article at the end of the Introduction—describe the content of each section.
2. The bibliography should be expanded to include more recent literature (5 years or less). Currently, of the 58 references, only 29.31% of the entire bibliography (17 sources) is 5 years or less old. Also relatively recent are the 2019 references—12.07% (7 items). However, nearly 60% of the sources are at least a dozen years old, and some are over 20 years old. The authors should supplement the bibliography with more recent items.
3. I think it is worth emphasizing that informal relationships supporting connections between VC investors should be of a legal nature, rejecting the influence of mafia-like social capital on the development of high-risk funds.
4. In line 203, the authors cite source no. 30, citing: Agnieszka N et al. [30]. They made two errors here. The aforementioned source has an error in the author's surname. Moreover, it is not a collective work. The bibliographic description should be as follows: Niwińska A. Ships and relationships: Competition, geographical proximity, and relations in the shipping industry. J. Bus. Res. 2019, 21, 161-170. https://doi.org/10.1016/j.jbusres.2019.04.021. This should be corrected in the Bibliography and in line 203.
5. The paper contains other editorial errors, such as missing spaces between words and the reference number, and missing periods in some sentences. In the bibliography, the authors have lowercased the names of some countries (China, Korea).

Incorporating the suggested changes and additions will enhance the article's value. The article may then be considered for publication.

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Congratulations for the very interesting research work concepted by Distinguished Authors: Gao, Y, Xie, Y, Yang, Y. It is very well documented, argued and clear written with explicit scientific article presentation.

However, I will make a general and sequentially analyse concerning the followings items of the research:

1. The title called Cross Layer Influence of Multiple Network Embedding on Venture Capital Network: An ERGM - Based Analysis is well chosen for the research article.
2. The abstract presents, on short, the research novelty idea, and the method ERGM based analysis, used to argue it. It underlines the novelty of the subject. It analyses the definitory influence factors from three dimension perspective:
   1. Endogenous networks structures
   2. Multi dimensional relational networks among Venture Capital (VC) companies,
   3. Informal networks of venture capital
3. The introduction argues the utility of the research, scope and objectives. It underlines that the research sustain and brings scientific contribution to the Strategy of Scientific and Technological Development of China according to Chinese Political, Social and Economical objectives. The introduction identify the research motive , related with the fact that the high risk, high return ventures, investments institution, face significant asymetry and uncertainity during the decision making process. The observations of the Distinguished Authors that the most of research works regarding venture capital relations in business / investment networks are analysed from a dyadic perspective, omitting a trydic perspective, is a good argument and novelty subject for the present research. The formation of VC networks ties based on dyadic relationship, that from the perspective of distinguished authors is incomplete. The authors underlines that, in the real world conditions, it should be taken into consideration, the potential influence of the third party members, on bilateral relationships, by this helping mitigate endogeneity concerns.
4. The Literature Review and Research hypothesis are very well documented and argued. The concepts of Endogeneus Networks , Network Structures and Venture Capital Networks are sustained with other scholars research works. The research Hypothesis H1, H2 (H2a, H2b), H3(H3a, H2b) are well formulated, focusing the research to reach at its objectives.
   1. The research Hypothesis, H1 concerns the probability of Venture capital network formation based on closed triad structures that enhances the probability of their formation
   2. The research Hypothesis H2 (H2a) shows that the geographic distance reduce the formation of Venture Capital Networks, 
   3. The research Hypothesis H2b shows that knowledge proximity increase the formation of Venture Capital Networks
   4. The research Hypothesis H3a shows that alumni networks increase the formation of Venture Capital Networks,
   5. The research Hypothesis H3 b shows that Shared Employment Experience Networks, increase also the formation of Venture Capital Networks.
   6. An easy to visualize and read the Conceptual Model of research, is also presented in the Literature Review and Research Hypothesis is presented.  
5. The Research Design, is very well presented, the reader could easy understand all process from data gathering (CVS, data base, BESTLA), large sample of investments 3986 , and 272 companies, to 166 (that have more then 3 investments events) and finally after selection remains 157 companies and 1317 investments events. The accuracy of the presentation is also about the model selection, and chosen, mentioning also both mathematical and scientific adequate software support (ERGM build it with STATNET package in R software). Very accurate description of models, including graphical model presentation to underline personal original scientific contribution besides other research works.
6. Results and discussions, follows the methodology and its appliance in the research.
7. The Research Conclusions are well done and fits with the research hypothesis (H1..H3) and sub hypohesis (H2a, H2b, H3a, H3b)
8. The List of the Reference is enough to support the information that argues the research

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

1. Manuscript Summary and Key Contributions

This manuscript explores the dynamics of venture capital (VC) network formation in the context of China's underdeveloped formal institutional environment. Using a multilayer Exponential Random Graph Model (ERGM), the study integrates endogenous structural factors (e.g., triadic closure), institutional-level proximity (e.g., geographic and knowledge proximity), and informal interpersonal relationships (e.g., alumni and employment ties) to assess their influence on VC syndication behavior. Based on 1,317 investment events from 157 VC firms, the paper tests six hypotheses and finds significant evidence supporting the roles of triadic closure, knowledge similarity, and informal ties in network formation.

The manuscript contributes by:

1. Offering a robust multilayer analytical framework combining structural, organizational, and individual factors.
2. Providing empirical insights into how informal networks compensate for institutional weaknesses in emerging markets.
3. Bridging VC network research with social capital and institutional theory through an ERGM lens.
4. Detailed Evaluation and Constructive Feedback (Page-by-Page and Line-by-Line):

• Page 1, Lines 1–27 (Title and Abstract):
  • Suggestion: Add the word “China” to the title for contextual clarity: “Cross-Layer Influence of Multiple Network Embedding on Venture Capital Networks in China: An ERGM-Based Analysis”.
  • Example Recommended Revision: Modify final sentence of abstract to: “This study introduces a novel multilevel ERGM framework to demonstrate how informal ties substitute formal institutions in China's venture capital network formation.”
• Page 2, Lines 29–65 (Introduction):
  • Line 30: Suggest moving the policy motivation higher for emphasis.
    • Recommended Revision: “The central role of venture capital in China’s economic strategy—highlighted by the 2024 CPC meeting—frames the urgency to understand how VC networks form despite institutional constraints.”
  • Line 48–58: Merge discussion of syndication and principal-agent issues.

 

• Recommended Revision: “While syndication reduces investment risk, it introduces agency conflicts due to diverse roles across institutions—e.g., lead vs. follow investors—amplifying the importance of governance structures.”

• Page 3, Lines 86–128 (Literature Gap):
  • Lines 90–92: Clarify the problem with generalizing interpersonal factors to the organizational level.
    • Suggestion: Reference Granovetter’s work or other multilevel studies showing misattributions.
    • Example Recommended Revision: “This assumption—common in prior work (e.g., XXX & ZZZ, YYYY)—ignores how nested individual actions shape broader inter-firm ties.”
  • Line 112–122: Add a specific example.
    • Example Recommended Revision: “For instance, previous regression models (XXX et al., YYYY) ignored the indirect influence of third-party institutions on dyadic VC partnerships.”
  • Pages 4–5, Lines 144–247 (Hypotheses):
    • Suggestion: Define each construct before proposing hypotheses.
    • Example: Add before H1: “Triadic closure is a network configuration where if A is connected to B and C, then B and C are also likely to be connected, facilitating trust and cooperation.”
    • Suggestion: Include a table for clarity:

Hypothesis	Variable	Expected Effect
H1	Triadic closure (Gwesp)	Positive
H2a	Geographic proximity	Negative
H2b	Knowledge proximity	Positive
H3a	Alumni network	Positive
H3b	Shared employment	Positive

 

• Pages 6–9, Lines 326–444 (Methodology):
• Line 337: Justify threshold of three investments.
  • Recommended Revision: “Institutions with fewer than three investments were excluded to ensure statistical reliability and reduce noise from inactive or one-off investors.”
• Line 343: Clarify data availability.
  • Recommended Revision: “Although some data were extracted from proprietary databases (Wind, Bestla), summary statistics and network matrices are available upon request to ensure transparency.”
• Line 359: Clarify distance formula in lay terms.
  • Recommended Revision: “Geographic distance was calculated using the Haversine formula based on the registered city coordinates of each VC firm.”
  • Additional Suggestion: Add a statement on multicollinearity:
  • “We performed a VIF test on explanatory variables. All values were below 5, indicating no critical multicollinearity.”
• Pages 10–14, Lines 447–530 (Results):
• Line 477: Provide marginal effect interpretation.
  • Recommended Revision: “A 1-unit increase in triadic closure increases the likelihood of tie formation by 377% (exp(1.529) = 4.614).”
• Line 505–511: Compare magnitudes.
  • Recommended Revision: “The effect of shared employment experience (β = 0.359) is over three times that of alumni ties (β = 0.111), suggesting professional experience fosters stronger trust than shared education.”
• Page 15–17, Lines 531–571 (Diagnostics and GOF):
• Line 546: Add statistics.
  • Recommended Revision: “All GOF metrics fall within 95% CI (e.g., triad census deviation = 0.17; edgewise shared partners deviation = 0.09).”
• Page 18–19, Lines 572–629 (Substitution Mechanism):
• Line 623: Caution in interpretation.
  • Recommended Revision: “Although the substitution effect is promising, the p-value (>0.05) indicates marginal significance. A larger sample or longitudinal data may clarify this pattern.”
• Page 20–21, Lines 630–648 (Robustness Tests):
• Line 640: Add network density result.
  • Recommended Revision: “The binary networks maintained a comparable density (~0.16), ensuring structural similarity for robustness tests.”
• Page 22–24, Lines 649–758 (Conclusion):
  • Line 751: Add example.
    • Recommended Revision: “Excessive reliance on informal networks may lead to cliques or relational lock-in, reducing exposure to diverse deal flows (e.g., closed circles excluding high-potential newcomers).”
  • Page 25 (References):
  • Recommended Revision: Standardize author names, e.g., “Lee H.S.” vs. “H.S. Lee” and ensure journal names are italicized per MDPI style.

Overall Recommendation:

• Major Concerns:

1. Abstract (Page 1): Needs sharper summary of unique contributions.
2. Data & Replication (Page 6): Transparency on data access is needed.
3. Results Interpretation (Pages 10–12): Clarify exponential effect sizes with applied examples.
4. Method Clarity (Page 7): Justify thresholds and equations better.
5. Theoretical Integration (Page 3): Deeper contextualization of hypotheses.

 

• Minor Concerns:

1. Figures 2 & 3: Expand captions for stand-alone comprehension.
2. Language: Minor grammar edits required throughout.
3. Model Table: Refer to changes in coefficients explicitly in text.
4. Future Research: Suggest specific methods or datasets.
5. Formatting: Recommended Revision reference list inconsistencies.

Final Evaluation:

This paper offers a novel and valuable contribution to understanding VC networks under institutional constraints. With careful revision, including improved methodological transparency, richer contextual grounding, and more precise interpretation of statistical outcomes, the manuscript can reach publishable quality and significantly impact the fields of network theory, institutional economics, and venture capital research.

 

 

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The study used an exponential random graph model based on 1317 investment event data from 157 VC institutions to verify the joint impact of these factors on the formation of VC networks, and pointed out that in the context of imperfect formal institutions in China, informal networks can replace formal institutions to promote cooperation. This research has certain value in theoretical framework construction and model application, but there are certain shortcomings in sample representativeness, mechanism testing, and English writing, and necessary modifications are needed before it is suitable for publication. My suggestion is as follows:

[1] The author emphasizes that "shifting from binary relationships to the network level" is the innovation of this article, but ERGM is not uncommon in organizational network research, and I believe that this innovation needs further explanation.

[2] As an industry-level study, I believe the author needs to further demonstrate that the selected sample institutions have sufficient industry representativeness. Furthermore, considering that the threshold setting for investment events is subjective, further robustness testing is required.

[3] The test of "informal network replacing formal institutions" only represents institutional similarity through "whether the institutions are located in the same place", which I think is too simplistic. I suggest that the author further consider more detailed institutional dimensions, such as legal environment and regulatory intensity, to further strengthen the conclusion.

[4] The policy recommendations are relatively vague and abstract, lacking specific cases or implementation paths.

[5] There are some grammar errors and inconsistent terminology in the paper, and some charts are not clear enough, which affects readability.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

 General Evaluation

I would like to express my appreciation for the substantial effort and thoughtful revisions made in response to the previous round of review. The revised manuscript now demonstrates firm theoretical grounding, methodological transparency, and empirical rigor. The authors' detailed point-by-point responses clearly indicate a genuine engagement with the reviewers' feedback. The paper has improved significantly in clarity, coherence, and scholarly contribution.

1. Theoretical Development and Conceptual Clarity

The revised manuscript shows a much stronger conceptual foundation. The authors have successfully incorporated multilevel perspectives that link individual, organizational, and institutional layers of venture capital (VC) networks. The inclusion of self-organization, market-oriented, and relational logics provides a nuanced theoretical framing that deepens the study's relevance to institutional theory and network governance.

The addition of examples and clarifications on how interpersonal ties translate into inter-organizational dynamics (as inspired by Granovetter's work and subsequent multilevel studies) strengthens the logical consistency of the arguments. These modifications make the theoretical model more persuasive and contextually grounded in China's venture capital landscape.

2. Methodological Refinement

The methodological section has undergone substantial improvement. The justification for selecting firms with three or more investment events as a threshold for inclusion is now well-articulated and supported by methodological precedent. The clarification of the Haversine formula for geographic distance and the explanation of why VIF tests are unnecessary in ERGM models demonstrate a sophisticated understanding of model-specific diagnostics.

Furthermore, the revised text provides detailed explanations of data sources (Wind and Bestla) and their accessibility, promoting transparency and replicability. The expanded discussion of MCMC convergence diagnostics and goodness-of-fit (GOF) assessments through visual plots adheres to accepted ERGM standards, thereby enhancing methodological credibility.

3. Results Presentation and Interpretation

The results section is now presented with exemplary clarity. The addition of marginal effect interpretations (e.g., the exponential interpretation of coefficients, such as a 1-unit increase in triad closure leading to a 4.6-fold increase in tie formation) greatly aids comprehension. The comparison of coefficient magnitudes between the shared employment experience and alums networks provides meaningful insights into the strength of informal trust mechanisms within the VC ecosystem.

The authors' careful justification for using graphical GOF plots instead of deviation statistics is appropriate and conforms with conventions in ERGM-based research. This methodological reasoning strengthens the presentation's academic soundness.

4. Discussion and Theoretical Implications

The revised discussion section now offers a more nuanced interpretation of how informal networks can serve as substitutes for weak formal institutions. The authors also appropriately tempered their claims by noting the marginal significance of the substitution effect and suggesting that future research employ larger or longitudinal datasets.

This balanced interpretation enhances the study's credibility and aligns the findings with established theories of institutional substitution and relational governance. The integration of recent Chinese policy contexts (e.g., the 2024 CPC meeting) effectively situates the study within a real-world framework, improving its practical relevance.

5. Figures, Tables, and Presentation

Figures 2 and 3 now contain expanded captions that ensure standalone comprehensibility. The formatting of tables, particularly Table 6, which compares network density, is clear and informative. The authors' efforts to standardize references according to MDPI guidelines (e.g., italicized journal names, corrected author initials) have also improved the manuscript's professionalism and readability.

The addition of a “Limitations and Future Research” section (Section 5.4) is a valuable enhancement. The discussion of extending the dataset beyond 2021 and incorporating dynamic network models reflects thoughtful forward-looking research planning.

6. Writing Quality and Readability

The overall quality of English writing has improved substantially. The text is now concise, precise, and logically structured. Minor grammatical corrections and stylistic polishing have refined the flow of arguments, while the use of academic tone and terminology aligns well with the standards of the Systems journal.

7. Final Assessment

In conclusion, the authors have addressed all previous comments comprehensively and effectively. The revised manuscript is now methodologically robust, theoretically rich, and well-written. The improvements to theoretical framing, empirical presentation, and discussion collectively strengthen the contribution of this research to both the fields of network science and venture capital studies.

I am satisfied with the revision and recommend that the manuscript be accepted for publication in Systems.