Modeling the Structure and Dynamics of Regional Entrepreneurial Ecosystems: Evidence from Serbia

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper addresses a relevant gap in the entrepreneurial ecosystem literature by proposing and empirically testing a structural-mechanism model at the regional level in Serbia. The topic is timely, the methodology is broadly appropriate, and the paper is clearly written. However, several conceptual, methodological, and presentational issues warrant attention before publication.

Social factors (SOCF) appear in two distinct roles, as a structural driver (shaping INSF and TECF) and simultaneously as a direct mechanism (predicting PRED and MSB). This dual positioning is acknowledged but insufficiently justified theoretically. Why do economic factors not also serve a direct mechanism role? The author should either provide a stronger theoretical rationale for this asymmetry or reconsider the model architecture. As written, it risks appearing ad hoc.

It is noted that the PLS-SEM was chosen because the model is prediction-oriented, but the PLS-SEM selection should be justified more rigorously. For instance, referencing the distributional properties of the data, sample size considerations, the formative vs. reflective nature of constructs, or complexity of the model. Notably, all constructs are specified as reflective, which some scholars would argue makes CB-SEM a reasonable alternative worth at least discussing. It should be explained why CB-SEM was not considered.

A clear definition of all four size categories should be added, specifying the classification standard used (e.g., EU recommendation, Serbian legal framework, or another criterion) and whether classification was based on employee numbers, annual turnover, balance sheet total, or respondent self-report. This should appear in Section 3.2 as a methodological clarification, ideally with a reference to the relevant standard.

The sample of 401 enterprises is stratified across four regions, but Vojvodina (35.2%) and Belgrade (30.9%) together account for approximately 66% of the sample, while Southern Serbia, the most underdeveloped region and a key finding of the paper, accounts for only 16.2% (n=65). This raises concerns about whether the findings for Southern Serbia are sufficiently robust. It should be discussed whether the smaller Southern Serbia and Central Serbia subsamples affect statistical power in the regional comparisons.

Table 1 lists one example item per construct, but the full set of 45 items (5 per construct × 9 constructs) is not presented. For a quantitative paper relying on PLS-SEM, full reporting of all measurement items is standard practice and important for replication. These should be included, for instance, in the appendix.

Hypotheses are largely theoretically deduced rather than empirically cumulative. There is the absence of prior quantitative studies that have tested specific hypothesized relationships within a regional entrepreneurial ecosystem framework. Several hypotheses cite studies from China, South Africa, Italy, and other contexts quite different from Serbia. Given that the paper explicitly positions itself within transition economy literature, the empirical motivation would be considerably stronger if prior evidence from transition economies were more systematically incorporated, particularly for H3, H4, H8, and H9.

Both H6 and H7 propose positive relationships between technological factors and innovation-related outcomes, and the theoretical justifications are very similar. The paper does not sufficiently explain why technological factors would differentially affect new product development versus innovation capacity, or why these should be treated as conceptually distinct outcomes predicted by the same antecedent. This raises the question of whether RNPU and INKAP are sufficiently distinct constructs, a concern reinforced by the discriminant validity issue.

Practical implications discussed in the conclusion should be directly linked to empirical evidence, because some recommendations are not based on the measurement instruments and research design. For instance, the implication that “support for digital infrastructure, research and development cooperation, technology transfer, and university-industry collaboration may be particularly relevant" are not associated with any particular item or construct (maybe when all 45 items are listed, the connection will be clearer). Moreover, the implication that firms in weaker regions "may need to rely more strongly on internal capabilities, informal networks, and external partnerships" is not supported by data and findings. The study measures perceived ecosystem conditions, not firm-level strategies or their outcomes. Business performance outcomes are explicitly acknowledged as outside the scope of the paper. Thus, this implication exceeds what the evidence can support.

Author Response

Dear Reviewer,

We thank you for your thorough review and useful suggestions. We have managed to improve our manuscript thanks to your insights. 

The document is attached.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This article attempts to answer the question: How do economic and social factors influence innovation and entrepreneurship through institutional and technological mechanisms? The answer to this question, in the context of the rationality of public spending, legal development, and the impact on society, is crucial. Entrepreneurship development, and thus the country's economic development, depends largely on the quality of institutions. This article begins by constructing a theoretical model and then empirically testing it using data from 401 companies in Serbia. A key assumption is made: The entrepreneurial ecosystem is not a collection of elements, but a system of interconnected layers. These layers consist of structure (economic factors, social factors), mechanisms (institutions, technology), and outcomes (innovations, products/services, entrepreneurial activity).

The article is characterized by a well-thought-out methodology. The PLS-SEM (Partial Least Squares Structural Equation Modeling) approach is used, an advanced method for examining the relationships between latent variables. The analysis was based on 401 enterprises from four regions of Serbia. The following tests were conducted: construct reliability and validity, direct and indirect effects, R², Q² (explanatory and predictive ability), and effect size (f²). The model and data analysis identified a number of relationships of varying strength, which are discussed consistently.

The most general conclusion is that the system operates through chains of dependencies, not individual relationships. However, this conclusion is preceded by a detailed discussion of these chains. The results also highlighted regional differences, but this may be country-specific. However, this also suggests that other countries should be analyzed at the regional, rather than national, level. Therefore, the most important advantages of the article include the transition from a "list of factors" to a systems model, a clear distinction between: structure vs. mechanism, the use of PLS-SEM instead of simple correlations, analysis of indirect effects, model validity, and predictive relevance, a rare study for a transition economy (Serbia), real company data.

However, what requires further clarification are:
1. Insufficient explanation of causality. The article emphasizes that the results represent statistical relationships, not causal evidence, which is somewhat at odds with the interpretation, which does show these relationships. Perhaps the problem lies in conducting a single empirical study instead of conducting periodic studies, or at least twice on the same identifiable sample population.
2. Perceptual (subjective) data is a serious problem when the data is based on company opinions. This poses the risk of common method bias.
3. Limited generalization to Serbia. This raises the question of the transferability of the results to other countries. Nevertheless, the research concept itself is transferable, and this is valuable.
4. If we assume that the results can be generalized to other countries undergoing transformation, is there a possibility of distortion due to the Balkan context?
5. The model is very structured, but how does it relate to reality? What is the context of public policy as a separate variable and temporal dynamics?
6. The article contributes three things: conceptual (the ecosystem as a system of layers and mechanisms), methodological (integration of multiple analyses in a single SEM model), and empirical (data on Serbia (rare in the literature)). Perhaps this should be emphasized more clearly in the summary.
7. It is also worth expanding on the practical implications that are recommended, e.g., developing institutions, technology, and social culture. But perhaps it would be worthwhile to also note how and indicate directions.

8. Technical issues, such as tables, need to be improved.

Author Response

Dear Reviewer,

We thank you for your thorough review and useful suggestions. We have managed to improve our manuscript thanks to your insights. 

The document is attached.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I have completed the evaluation of the revised manuscript. I would like to express my appreciation for the author’s diligent efforts in addressing all the comments and suggestions raised during the previous review round. Upon thorough examination of the revised submission, I am satisfied that the author has responded comprehensively and appropriately to all concerns. The manuscript has been significantly improved in terms of clarity, methodological rigour, and overall presentation.