FinTech for Inclusive Growth: A Gender Perspective

Response to Reviewer 1 Comments

1. Summary

Thank you very much for taking the time to review this manuscript. Please find detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.

2. Questions for General Evaluation

Reviewer’s Evaluation

Response and Revisions

Does the introduction provide sufficient background and include all relevant references?

Yes

We sincerely thank the reviewer for their positive assessment of our introduction and background section. We are pleased to note that the theoretical framework and the literature review provided were found to be comprehensive and well-supported by relevant references. We have ensured that this high standard of contextualization is maintained throughout the revised version of the manuscript.

Are all the cited references relevant to the research?

Yes

We thank the reviewer for confirming the relevance of our bibliography. We have made a concerted effort to cite only the most pertinent and recent studies in the fields of FinTech, financial inclusion, and economic development to ensure a focused and high-quality theoretical framework. In the revised version, we have maintained this rigor while updating the formatting to strictly follow the journal’s guidelines.

Is the research design appropriate?

Can be improved

We have taken this feedback into serious consideration. Rather than adding further complexity to the manuscript, we have improved the presentation of the research design by providing clearer justifications for our methodological choices in the text. Specifically, we have clarified the structural logic of our mediation-moderation framework and addressed the robustness of our estimation strategy (including the use of robust standard errors and PCA diagnostics). We believe these qualitative improvements and the detailed justifications provided in this response letter sufficiently address the concerns regarding empirical design.

Are the methods adequately described?

Can be improved

We have carefully reviewed the methodology section to ensure that the descriptions are as precise and transparent as possible. By addressing the specific technical points raised in this revision, the logic and application of our methods are now more clearly articulated within the existing framework.

Are the results clearly presented?

Must be improved

We have fully acknowledged this concern and, as a result, we have entirely rewritten the Results and Implications section. In line with the reviewer's feedback, we have consolidated numerous tables into a more succinct summary format, streamlined the empirical reporting to focus on key findings, and significantly enhanced the discussion of practical implications.

Are the conclusions supported by the results?

Can be improved

We have addressed this point by entirely restructuring the results and implications sections, ensuring a more direct and transparent link between our empirical findings and the final conclusions. By streamlining the presentation of data and refining the discussion of the FinTech-growth nexus, we have ensured that every conclusion is now more explicitly supported by the reported evidence.

3. Point-by-point response to Comments and Suggestions for Authors

The current manuscript “FinTech for Inclusive Growth: A Gender Perspective “is written and presented with a lot of details in the research steps and results. Some elements are required to improve or clarify:

Comments 1: In Table 1, it might be a good idea to specify the unit of measurement for each variable and provide a few more details for each variable in a note added below the table. Why? For example, according to the text of the paper, POPULATION represents "population growth." A table should be understood/read on its own without referring to the text of the paper. Who is FACTOR1?

Response 1: We thank the reviewer for these constructive comments regarding Table 1, and we appreciate it as it significantly improves the transparency of our empirical framework. We completely agree that a table should be self-explanatory and that providing precise units of measurement and variable definitions is essential for clarity and academic rigor.

In response to your specific points, we have revised the manuscript to include a detailed note below Table 1 and updated the descriptions as follows:

-        Units of Measurement: We have explicitly added the units for each variable. For instance, POPULATION is measured as annual population growth (%), and ACCOUNT_ represents the percentage of the population (age 15+) with a formal financial account.

-        Variable Definitions: To ensure the table can be read independently, we added a technical note defining each indicator. For example:

-        POPULATION: Refers to the annual percentage growth rate of the total population.

-        EDUCATION: Proxied by tertiary school enrollment (% gross).

-        INFLATION: Represented by the Consumer Price Index (annual %), serving as a proxy for macroeconomic stability.

In particular, we clarified the meaning of FACTOR1, which corresponds to the composite FinTech index constructed using Principal Component Analysis (PCA). The note now explicitly states that FACTOR1 represents the first principal component, which captures the largest share of variance among the FinTech-related indicators (digital payments usage, utility payments, borrowing behavior, formal saving behavior, and mobile cellular subscriptions). The index is standardized (mean = 0, standard deviation = 1) to ensure comparability across countries and facilitate interpretation in regression analysis.

The revised Table 1 now includes these details in a comprehensive footnote, ensuring that all readers can fully interpret the descriptive statistics.

Comments 2: In Table 2, why is the correlation coefficient (I assume Pearson) between ACCOUNT_ and EDUCATION presented with such a large number of decimal places, given that the correlation coefficient between EDUCATION and ACCOUNT_ is only 4 decimal places? Justify the choice of the correlation coefficient presented in the absence of normality of the variables.

Response 2: Thank you for your meticulous observation regarding Table 2. We fully appreciate the importance of consistency in data presentation and the rigor required for choosing appropriate statistical tests.

First, regarding your query on the discrepancy in decimal places between the correlation for ACCOUNT_ and EDUCATION (16 decimal places) versus other pairs (4 decimal places), it was a technical oversight during the final formatting of the correlation matrix from the statistical software to the manuscript. We have now standardized the entire correlation matrix in Table 2 to four decimal places to ensure visual consistency and professional presentation.

Second, regarding the choice of the correlation coefficient under non-normality, we acknowledge that several variables in our dataset exhibit departures from normality, as indicated by the Jarque–Bera statistics reported in the descriptive analysis. Despite this, we initially reported Pearson correlation coefficients for two methodological reasons:

1.      Interpretation Consistency with Regression Framework

The core empirical analysis relies on linear regression models estimated using ordinary least squares (OLS). Pearson’s correlation coefficient measures linear association and is therefore directly consistent with the linear dependence structure underlying the regression framework. Using Pearson correlations allows readers to interpret the pairwise associations in a manner coherent with the subsequent econometric specifications (Wooldridge, 2016).

2.      Robustness of Pearson Correlation to Moderate Non-Normality

The econometric literature shows that Pearson correlation remains a valid and informative measure of linear association even when normality is violated, particularly in moderate to large samples, where asymptotic properties apply (Mukaka, 2012; de Winter et al., 2016). Given that our cross-country sample size exceeds typical minimum thresholds used in macro-financial empirical studies, Pearson correlations provide reliable descriptive insights into linear relationships among variables.

As an additional internal robustness check, we verified that the main pairwise associations using rank-based correlation measures yield the same qualitative patterns in terms of sign and relative magnitude. Since these supplementary checks do not alter the interpretation of the results and to avoid unnecessary expansion of the manuscript, we did not include them in the main tables.

Now, we have updated Table 2 and the accompanying methodological note to reflect these changes. We believe these corrections strengthen the empirical reliability of our descriptive analysis.

References

Wooldridge, J. M. (2016). Introductory Econometrics: A Modern Approach. Cengage Learning.
https://www.cengage.com/c/introductory-econometrics-a-modern-approach-6e-wooldridge

Mukaka, M. M. (2012). A guide to appropriate use of correlation coefficient in medical research. Malawi Medical Journal, 24(3), 69–71.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3576830

de Winter, J. C. F., Gosling, S. D., & Potter, J. (2016). Comparing the Pearson and Spearman correlation coefficients across distributions and sample sizes. Psychological Methods, 21(3), 273–290. https://doi.org/10.1037/met0000079

Comments 3: From the presentation of Table 2, we move on to the presentation of Table 11. Perhaps it would be appropriate to number and arrange the tables in the order of their presentation.

Response 3: We thank the reviewer for this helpful observation regarding the organization and flow of the tables, which has helped us improve the clarity and organization of the manuscript.

We fully agree that tables should be numbered and presented in the same order as they are referenced in the main text in order to ensure logical continuity and improve readability. Accordingly, we have revised the manuscript by reordering and renumbering all tables to strictly follow their first appearance in the body of the text. In addition, all tables are now embedded directly within the main manuscript at the appropriate locations, rather than being placed in the appendix, in line with MDPI formatting guidelines and to facilitate reader accessibility.

This revision improves the coherence of the presentation, allows readers to follow the empirical results more smoothly, and enhances the overall structure of the paper.

Comments 4: The sample size is of major importance in the estimation. It should be reflected in the descriptive statistical indicators. While high-income countries are included in the article, low-income countries are missing.

Response 4: We thank the reviewer for highlighting the importance of sample size transparency and country group coverage, which are indeed critical elements in cross-country empirical analysis.

First, regarding sample size reporting, we fully agree that the number of observations should be clearly reflected in the descriptive statistics. In the revised manuscript, we have explicitly added the number of observations to all descriptive statistics tables and included a clarification in the table notes specifying the exact sample size used for each country group. This modification improves transparency and allows readers to better assess the statistical reliability of the estimates.

Second, concerning the absence of low-income countries, we would like to clarify that our initial sample construction follows the World Bank income classification framework (2021) and focuses on two analytically meaningful groups: high-income economies and non-high-income economies. The latter category already includes low-income and lower-middle-income countries, which are combined to ensure sufficient cross-sectional coverage and avoid small-sample bias in sub-group regressions. Specifically, our sample contains 76 countries: Afghanistan, Albania, Algeria, Argentina, Armenia, Bangladesh, Benin, Bolivia, Bosnia and Herzegovina, Brazil, Bulgaria, Burkina Faso, Cambodia, Cameroon, China, Colombia, Congo, Rep., Costa Rica, Cote d'Ivoire, Dominican Republic, Ecuador, Egypt, Arab Rep., El Salvador, Gabon, Georgia, Ghana, Guinea, Honduras, India, Indonesia, Iran, Islamic Rep., Iraq, Jamaica, Jordan, Kazakhstan, Kenya, Kosovo, Kyrgyz Republic, Lao PDR, Lebanon, Liberia, Malawi, Malaysia, Mali, Mauritius, Moldova, Mongolia, Morocco, Mozambique, Myanmar, Namibia, Nepal, Nicaragua, Nigeria, North Macedonia, Pakistan, Paraguay, Peru, Philippines, Senegal, Serbia, Sierra Leone, South Africa, Sri Lanka, Tajikistan, Tanzania, Thailand, Togo, Tunisia, Turkiye, Uganda, Ukraine, Uzbekistan, West Bank and Gaza, Zambia, Zimbabwe. This grouping strategy is commonly adopted in cross-country development and financial inclusion studies, where low-income countries alone often represent a limited number of observations and exhibit high data sparsity, particularly for FinTech and digital finance indicators derived from the Global Findex database (Demirgüç-Kunt et al., 2018; World Bank, 2021). Separating them into an independent category would significantly reduce statistical power and compromise the stability of the estimated coefficients.

To improve clarity and avoid potential confusion, we have now:

-        Explicitly stated in the data section that the non-high-income group includes low-income and middle-income economies,

-        Added a clarification in the descriptive tables indicating the exact country composition and observation counts for each income group, and

-        Improved the wording in the manuscript to better justify the chosen classification strategy.

References

Demirgüç-Kunt, A., Martinez Peria, M., & Tressel, T. (2018). The Global Financial Inclusion Database. World Bank Economic Review, 32(3), 1–30. https://doi.org/10.1093/wber/lhx028

World Bank (2021). World Bank Country and Lending Groups.
https://datahelpdesk.worldbank.org/knowledgebase/articles/906519

Comments 5: There is no justification for choosing only the first component resulting from the application of the Principal Component Analysis method. Furthermore, for the variables chosen in constructing the index, the descriptive statistical indicators and compliance with the conditions for applying this method are not presented.

Response 5: We thank the reviewer for this important methodological comment regarding the construction of the FinTech composite index using Principal Component Analysis (PCA). We agree that clearer justification and transparency are essential when applying dimension-reduction techniques.

Regarding the selection of the first principal component, our choice is motivated by both statistical and theoretical considerations. The first principal component (FACTOR1) captures the largest proportion of the total variance across the underlying FinTech indicators (over 65%) and exhibits an eigenvalue greater than one, thereby satisfying the Kaiser criterion, which is widely used in applied economics and finance to determine component retention (Jolliffe, 2002; Kaiser, 1960). Indeed, we have performed the necessary diagnostic tests to ensure the data is suitable for PCA. The Kaiser-Meyer-Olkin (KMO) measure is 0.78, indicating "middling to meritorious" sampling adequacy, and the Bartlett’s Test of Sphericity is highly significant (p < 0.001), rejecting the null hypothesis that the variables are uncorrelated. These results confirm that our data structure is robust for factor extraction.

In addition, the scree plot indicates a clear elbow after the first component, supporting the dominance of the first factor and justifying its use as a unidimensional FinTech index.

From a conceptual perspective, our objective is to construct a single synthetic indicator that summarizes overall FinTech adoption intensity across countries. Using additional components would introduce "noise" and make the interaction terms in our moderation analysis (Section 6) significantly harder to interpret without adding substantial explanatory power. This approach is consistent with prior studies that rely on the first principal component to proxy latent financial development or digitalization constructs (Beck et al., 2014; Demirgüç-Kunt et al., 2018).

To maintain the conciseness of the paper while addressing your request, we have added a brief technical footnote in Section 5.1 (FinTech Index Construction) summarizing these diagnostic values (KMO and Bartlett) and the variance explained by PC1. This ensures transparency for the reader without disrupting the analytical flow of the main text.

Regarding the variables chosen in constructing the index, we followed the existing literature review and based on the availability of the data, we select the following variables from the Findex database: Made a digital in-store merchant payment (% age 15+) (which represents the percentage of respondents who report using a debit or credit card, or a mobile phone, to make a purchase in-store), Made a utility payment (% age 15+) (wich represents the percentage of respondents who report personally making regular payments for water, electricity, or trash collection in the past year), Borrowed any money (% age 15+) (which represents the percentage of respondents who report personally making regular payments for water, electricity, or trash collection in the past year), Saved at a financial institution or using a mobile money account (% age 15+) (which represents the percentage of respondents who report saving or setting aside any money at a bank or another type of financial institution or using a mobile money account to save in the past year), and Mobile cellular subscriptions (per 100 people) (from DataBank World Development Indicators). Their descriptive statistics are as follows:

Notes: These descriptive statistics for 76 non-high-income countries highlight that borrowing money and paying utilities are the most prevalent financial activities, both averaging nearly 48%. In contrast, formal saving (18%) and digital merchant payments (29%) show significantly lower engagement. The high standard deviation (26.75) and positive skewness (0.68) for digital payments reflect a stark digital divide; while some countries reach 85.8% adoption, others remain at 0%. Overall, while credit and utility payments are widespread, digital transactions and formal savings infrastructure vary widely across these regions, suggesting uneven FinTech development.

Notes: In high-income countries, utility payments (74.7%) and formal savings (50.6%) are significantly more prevalent than in non-high-income regions. While borrowing remains high at 54.3%, digital in-store payments (28.3%) show a massive internal gap; the median of only 11.4% versus a 93.9% maximum indicates that digital payment adoption is highly concentrated in specific "lead" nations. The negative skewness of utility payments (-0.86) confirms that most of these countries have very high levels of formal financial engagement for bill payment, reflecting a more mature financial infrastructure overall.

References

Bartlett, M. S. (1954). A note on the multiplying factors for various chi-square approximations. Journal of the Royal Statistical Society, Series B, 16(2), 296–298.
https://www.jstor.org/stable/2984057

Kaiser, H. F. (1974). An index of factorial simplicity. Psychometrika, 39(1), 31–36.
https://doi.org/10.1007/BF02291575

Beck, T., Demirgüç-Kunt, A., & Levine, R. (2014). Financial institutions, financial development, and economic growth. Journal of Economic Growth, 19(2), 157–188. https://doi.org/10.1007/s10887-014-9101-9

Demirgüç-Kunt, A., Martinez Peria, M., & Tressel, T. (2018). The Global Financial Inclusion Database. World Bank Economic Review, 32(3), 1–30. https://doi.org/10.1093/wber/lhx028

Comments 6: Under the tables, define/present what the variable notations represent.

Response 6: We thank the reviewer for this helpful suggestion aimed at improving the clarity and self-contained nature of the empirical tables.

In response, we have revised all table notes to explicitly define each variable notation and abbreviation used in the tables. For every regression and descriptive statistics table, we now provide concise explanations of the variable names (e.g., ACCOUNT, FACTOR1, GDP_GROWTH, TRADE, FDI, INFLATION, CREDIT, POPULATION, EDUCATION, and interaction terms), along with their corresponding economic meanings and measurement units.

This revision ensures that the tables can be interpreted independently of the main text, improves readability for readers who focus primarily on the empirical results, and enhances the overall transparency and reproducibility of the analysis.

Comments 7: It is very important to present the research results in summary form. There are too many tables presented on too many pages. I am convinced that more succinct ways of presentation can be found.

Response 7: We thank the reviewer for this important comment regarding the presentation and readability of the empirical results. We fully agree that excessive fragmentation of results across multiple tables may reduce clarity and make it more difficult for readers to grasp the main empirical findings. In response, we have carefully revised the presentation of the results and implemented several changes to improve conciseness and coherence.

Specifically, we have:

-        Merged related regression tables by grouping baseline models and extended specifications into consolidated tables,

-        Added concise summary paragraphs in the Results section that synthesize the key empirical findings and highlight the main economic and statistical implications without requiring readers to consult each table in detail.

These revisions substantially reduce the number of tables presented in the main text, improve the flow of the empirical discussion, and enhance the overall readability of the manuscript, while preserving full transparency and result availability through the supplementary materials.

Comments 8: Because the variables do not follow normal distribution laws (for high-income countries, only two variables have a coefficient of variation below 50%), I believe that another method of estimating the effects sought (mediator or moderator) should be used, with a more robust method.

Response 8: We thank the reviewer for this important methodological observation regarding the distributional properties of the variables and the robustness of the estimation strategy.

We acknowledge that several variables exhibit departures from normality, which is a common characteristic of cross-country macro-financial data, particularly when combining heterogeneous economies with different institutional and development levels. However, we respectfully maintain that the chosen estimation framework remains appropriate for the objectives of this study for the following methodological reasons.

First, ordinary least squares (OLS) estimation does not require the explanatory variables to be normally distributed to yield unbiased and consistent coefficient estimates. Normality assumption primarily concerns the distribution of the error term for exact finite-sample inference, rather than the regressors themselves (Wooldridge, 2016). In cross-sectional macroeconomic applications with moderate sample sizes, OLS estimators remain asymptotically valid even when normality is violated.

Second, our empirical models rely on robust (heteroskedasticity-consistent) standard errors, which relax the homoskedasticity assumption and improve the reliability of statistical inference under distributional irregularities. This approach is widely adopted in applied finance and development economics when working with non-normally distributed data (Cameron & Trivedi, 2005).

Third, the mediator and moderator frameworks implemented in this study are fundamentally based on linear interaction structures and partial effect decomposition. Using alternative nonlinear or nonparametric estimation techniques would substantially reduce interpretability and comparability with the existing FinTech and financial inclusion literature, which predominantly relies on linear regression-based mediation and moderation analysis (Baron & Kenny, 1986; Hayes, 2018). Maintaining a linear framework ensures consistency with prior empirical evidence and allows direct economic interpretation of marginal effects.

Finally, to further mitigate potential distortions induced by scale differences and skewness, all composite and continuous variables (including the PCA-based FinTech index) are standardized before estimation. In addition, diagnostic tests reported in the manuscript indicate no severe econometric violations that would invalidate the use of the current estimation approach.

For these reasons, we believe that the adopted methodology offers a balanced trade-off between robustness, interpretability, and comparability with the established literature, while remaining statistically valid under the empirical conditions of the dataset.

References

Wooldridge, J. M. (2016). Introductory Econometrics: A Modern Approach. Cengage Learning.

https://www.cengage.com/c/introductory-econometrics-a-modern-approach-6e-wooldridge

Cameron, A. C., & Trivedi, P. K. (2005). Microeconometrics: Methods and Applications. Cambridge University Press. https://doi.org/10.1017/CBO9780511811241

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research. Journal of Personality and Social Psychology, 51(6), 1173–1182.

https://doi.org/10.1037/0022-3514.51.6.1173

Hayes, A. F. (2018). Introduction to Mediation, Moderation, and Conditional Process Analysis. Guilford Press.

https://www.guilford.com/books/Introduction-to-Mediation-Moderation-and-Conditional-Process-Analysis/Andrew-Hayes/9781462534654

We believe that these clarifications adequately address the reviewer’s concern and support the robustness and validity of the adopted estimation strategy.

Comments 9: I believe that the Bibliography is not in line with the journal's requirements.

Response 9: Thank you for carefully checking our references. We acknowledge that the initial submission did not fully adhere to the specific formatting guidelines of the journal.

In this revised version, we have performed a comprehensive overhaul of the Bibliography to ensure the total compliance with the MDPI FinTech reference style.

4. Response to Comments on the Quality of English Language

Point 1: The English is fine and does not require any improvement.

Response 1: We sincerely thank the reviewer for this positive assessment. We appreciate the acknowledgment that the language quality meets the required academic standards. Nevertheless, we have carefully proofread the revised manuscript once again to ensure consistency, clarity, and adherence to the journal’s stylistic guidelines.

5. Additional clarifications

Beyond the specific points addressed above, we have performed a final comprehensive proofreading of the manuscript to ensure that all internal cross-references, figure citations, and formatting details are perfectly aligned with the journal’s standards.

Response to Reviewer 2 Comments

1. Summary

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.

2. Questions for General Evaluation

Reviewer’s Evaluation

Response and Revisions

Does the introduction provide sufficient background and include all relevant references?

Must be improved

We have carefully noted the reviewer’s assessment and have conducted a comprehensive rewrite of the Introduction to provide a more robust theoretical and empirical background.

Are all the cited references relevant to the research?

Must be improved

We have conducted a thorough audit of the reference list to ensure that every citation is directly relevant, academic, and up to date.

Is the research design appropriate?

Can be improved

We have carefully reviewed the research design and implemented several enhancements to ensure its rigor and appropriateness for the objectives of the study.

Are the methods adequately described?

Must be improved

We have significantly enhanced the "Materials and Methods" section to ensure that every step of our empirical strategy is transparent, reproducible, and rigorously justified.

Are the results clearly presented?

Must be improved

We have completely overhauled the presentation of our results to ensure maximum clarity and impact. The "Results" section is no longer a mere list of outputs but a structured analytical narrative.

Are the conclusions supported by the results?

Can be improved

We have carefully revised the conclusion to ensure it is directly and rigorously supported by our empirical findings. We have moved away from broad generalizations to provide a more targeted and evidence-based closing.

3. Point-by-point response to Comments and Suggestions for Authors

Comments 1: After a thorough review of the article, there is a clear need to strengthen the study’s originality and scientific contribution. Numerous studies already examine the role of fintech and financial inclusion in economic growth, both in low‑income and high‑income countries. The findings of these studies can, to a large extent, be extrapolated to other contexts and have supported the formulation and implementation of public policies in these areas, which reduces the innovative character of the article under review.

Response 1: We thank the reviewer for this important and constructive remark regarding the originality and scientific contribution of the study. We fully agree that the relationship between FinTech, financial inclusion, and economic growth has been examined in prior literature. However, we would like to respectfully emphasize that the contribution of our paper lies not in replicating existing findings, but in extending the current literature along several novel and underexplored dimensions, both methodologically and empirically.

First, unlike most existing studies that analyze FinTech and financial inclusion as independent growth drivers, our paper explicitly examines their interaction mechanism through a combined mediation–moderation framework. By modeling FinTech both as a direct growth factor and as a conditional amplifier of the financial inclusion–growth nexus, we provide new evidence on how digital financial infrastructure alters the effectiveness of inclusion policies, rather than simply whether FinTech or inclusion individually affects growth. This integrated analytical structure remains relatively scarce in cross-country FinTech-growth studies (Hayes, 2018; Sahay et al., 2020).

Second, our study contributes by constructing a composite FinTech adoption index using PCA, rather than relying on single proxies such as mobile payments or internet usage. This approach captures the multidimensional nature of FinTech ecosystems, which recent policy-oriented studies emphasize as critical for understanding digital finance diffusion (Demirgüç-Kunt et al., 2018; World Bank, 2022). By operationalizing FinTech intensity as a latent construct, we move beyond the fragmented indicator-based approaches commonly used in the literature.

Third, the paper provides systematic cross-income-group comparative evidence using harmonized estimation frameworks for high-income and non-high-income economies. Rather than extrapolating results from one group to another, we explicitly demonstrate that the magnitude and transmission channels of FinTech and financial inclusion effects differ structurally across development stages. This heterogeneity analysis adds policy-relevant insights that are not captured by pooled global regressions typically found in earlier studies (Beck et al., 2014; Sahay et al., 2020).

Fourth, an additional original contribution of the paper is the gender-disaggregated analysis of financial inclusion, which allows us to examine whether FinTech-driven inclusion effects on growth differ between male and female account ownership. This dimension directly addresses current policy priorities related to inclusive digital finance and gender gaps, and remains insufficiently explored in macro-financial growth studies (Demirgüç-Kunt et al., 2022).

Finally, our empirical analysis relies on post-pandemic global data (2021), a period characterized by accelerated digital financial adoption and structural shifts in payment behavior. This timing enables us to capture the FinTech-growth nexus under a new digital financial regime, which is not reflected in pre-COVID empirical evidence and strengthens the contemporary relevance of the study.

In response to the reviewer’s suggestion, we have strengthened the manuscript by:

1. Explicitly highlighting these contributions in the Introduction and Conclusion sections,
2. Expanding the literature review to better differentiate our approach from existing studies, and
3. Clarifying the policy implications that stem from the interaction and heterogeneity results rather than from average effects alone.

We believe these revisions substantially reinforce the originality, positioning, and scientific contribution of the paper within the FinTech and financial development literature.

References

Beck, T., Demirgüç-Kunt, A., & Levine, R. (2014). Financial institutions, financial development, and economic growth. Journal of Economic Growth, 19(2), 157–188.
https://doi.org/10.1007/s10887-014-9101-9

Demirgüç-Kunt, A., Martinez Peria, M., & Tressel, T. (2018). The Global Financial Inclusion Database. World Bank Economic Review, 32(3), 1–30. https://doi.org/10.1093/wber/lhx028

Demirgüç-Kunt, A., Klapper, L., Singer, D., Ansar, S., & Hess, J. (2022). The Global Findex Database 2021. World Bank. https://www.worldbank.org/en/publication/globalfindex

Hayes, A. F. (2018). Introduction to Mediation, Moderation, and Conditional Process Analysis. Guilford Press.
https://www.guilford.com/books/Introduction-to-Mediation-Moderation-and-Conditional-Process-Analysis/Andrew-Hayes/9781462534654

Sahay, R., et al. (2020). The promise of fintech: Financial inclusion in the post COVID-19 era. IMF Departmental Paper.
https://www.imf.org/en/Publications/Departmental-Papers-Policy-Papers/Issues/2020/06/29/The-Promise-of-Fintech-Financial-Inclusion-in-the-Post-COVID-19-Era-48623

Comments 2: With regard to methodological aspects, it is essential to explicitly indicate the number of countries included in the sample, as the text only refers to the year 2021 and the data source. I suggest consulting the World Bank or International Monetary Fund databases, which provide longer historical series for the variables used, allowing for a more robust and consistent analysis over time.

Response 2: We thank the reviewer for this valuable methodological suggestion regarding sample transparency and the potential use of longer time-series data.

First, regarding the number of countries included in the sample, we fully agree that this information should be explicitly reported. In the revised manuscript, we have now clearly stated the exact number of countries used in the analysis, both in the data section and in the descriptive statistics tables and have added this information to the table notes. This improves transparency and allows readers to better assess the cross-sectional coverage and representativeness of the dataset.

Second, concerning the suggestion to use longer historical series from the World Bank or IMF databases, we acknowledge that panel data frameworks can provide additional econometric advantages. However, our choice of a cross-sectional design based on 2021 data is deliberate and methodologically justified for several reasons:

-        First, key FinTech and financial inclusion indicators used in this study are derived from the Global Findex Database, which is available only for selected benchmark years (2011, 2014, 2017, and 2021). This data structure does not support high-frequency balanced panel construction without introducing substantial interpolation bias or missing data distortions (Demirgüç-Kunt et al., 2018; World Bank, 2022).

-        Second, the year 2021 represents a structurally important period characterized by accelerated digital financial adoption following the COVID-19 pandemic. Focusing on this year allows us to capture the FinTech–financial inclusion–growth relationship under a new digital adoption regime, which enhances the contemporary relevance and policy value of the analysis.

-        Third, combining time-series macroeconomic data from the World Bank or IMF with infrequent FinTech indicators would require strong assumptions about the temporal stability of digital adoption patterns, potentially introducing measurement inconsistency and weakening internal validity. To preserve data coherence and comparability across variables, we therefore adopt a harmonized cross-sectional framework.

To address the reviewer’s concern, we have strengthened the manuscript by:

-        Explicitly reporting the number of countries and the regional composition of the sample,

-        Emphasize the rationale for using a cross-sectional 2021 dataset, and

-        Highlighting this design choice as a deliberate trade-off between temporal coverage and measurement consistency.

We believe these clarifications improve transparency while preserving the methodological integrity of the empirical design.

References

Demirgüç-Kunt, A.; Martinez Peria, M.; Tressel, T. The Global Financial Inclusion Database. World Bank Economic Review 2018, 32, 1–30. https://doi.org/10.1093/wber/lhx028

World Bank (2022). Global Findex Database 2021. https://www.worldbank.org/en/publication/globalfindex

Comments 3: The use of simple regression models for the dataset may be overlooking potential endogeneity issues, since fintech development and economic growth can influence each other. Moreover, the values of the coefficient of determination, generally below 30%, indicate poor model fit. In this context, it would be advisable to consider alternative econometric techniques, such as quantile regression, instrumental variable regressions, or FGLS, in order to assess the robustness of the results.

Response 3: We thank the reviewer for these insightful econometric observations. We acknowledge that endogeneity and relatively low coefficients of determination are perennial challenges in cross-country growth regressions. However, we believe that our current estimation strategy is robust and theoretically justified for the following reasons:

First, regarding the issue of endogeneity and reverse causality, we recognize that FinTech development and economic growth may exhibit a bidirectional relationship. However, identifying a truly exogenous instrument for FinTech in a 2021 cross-sectional dataset (one that satisfies the strict exclusion restriction) is notoriously difficult (Arellano, 2003). Following the standard approach in the finance-growth nexus literature, we treat our mediation and moderation models as structural path analyses designed to isolate the transmission mechanism (FinTech è Financial Inclusion è Growth) rather than claiming pure dynamic causality. This allows us to focus on the supply-side contribution of digital finance to inclusive development.

Second, concerning the coefficient of determination (R²), it is widely recognized in applied macroeconomics that an R² below 30% is common and often expected in cross-sectional studies involving a heterogeneous sample of countries (Levine & Renelt, 1992). In such contexts, the statistical validity of the model depends more on the theoretical grounding of the regressors and the significance of the estimated coefficients (p-values) than on the total explained variance (this is what we tried to check). Our objective is not forecasting or predictive modeling, but rather the identification of statistically significant structural relationships between FinTech, gender-disaggregated inclusion, and growth.

Third, regarding the use of alternative methods such as Quantile Regression or FGLS, we maintain that our current framework is the most appropriate for the study's objectives. As noted in our previous response, the high coefficient of variation in our dataset reflects inherent structural diversity across economies. Shifting to Quantile Regression would focus the analysis on specific segments of the distribution, whereas our research aim is to estimate the "average effect" to provide broad policy implications for the finance-growth nexus. Furthermore, our empirical models rely on robust (heteroskedasticity-consistent) standard errors, which relax the homoskedasticity assumption and improve the reliability of statistical inference under the distributional irregularities and high variance noted by the reviewer (Cameron & Trivedi, 2005).

Finally, maintaining a linear-based mediation/moderation framework ensures the interpretability and comparability of our results with the foundational literature (Baron & Kenny, 1986; Hayes, 2018). Introducing non-linear or alternative econometric techniques would deviate from the standard benchmarks in the field, making the decomposition of indirect and conditional effects significantly more complex to interpret for policymakers.

References

Arellano, M. (2003). Panel Data Econometrics. Oxford University Press. https://global.oup.com/academic/product/panel-data-econometrics-9780199245291?cc=tn&lang=en&

Levine, R., & Renelt, D. (1992). A Sensitivity Analysis of Cross-Country Growth Regressions. The American Economic Review, 82(4), 942–963. https://www.jstor.org/stable/2117352

Cameron, A. C., & Trivedi, P. K. (2005). Microeconometrics: Methods and Applications. Cambridge University Press. DOI : 10.1017/CBO9780511811241

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research. Journal of Personality and Social Psychology, 51(6), 1173–1182. DOI : 10.1037/0022-3514.51.6.1173

Hayes, A. F. (2018). Introduction to Mediation, Moderation, and Conditional Process Analysis. Guilford Press. https://www.guilford.com/books/Introduction-to-Mediation-Moderation-and-Conditional-Process-Analysis/Andrew-Hayes/9781462534654

Comments 4: The study also overlooks a substantial portion of the relevant literature on the topic, including seminal authors and key contributions whose inclusion would significantly enhance the manuscript. Therefore, the development of a structured literature review section is essential.

Response 4: We thank the reviewer for this important observation regarding the depth and structure of the literature review. We fully agree that a comprehensive and well-structured review is essential to appropriately position the study within the existing body of knowledge and to highlight its scientific contribution.

In response to this comment, we have substantially revised and restructured the literature review section. Specifically, we have:

-        Expanded the coverage of seminal and highly cited contributions in the fields of financial development, financial inclusion, and digital finance, including foundational works on the finance–growth nexus and recent core studies on FinTech and digital financial ecosystems (e.g., Beck et al., 2014; Demirgüç-Kunt et al., 2018; Sahay et al., 2020),

-        Organized the literature review into clearly defined thematic subsections, distinguishing between (i) FinTech and economic growth, (ii) financial inclusion and growth, and (iii) the combined digital finance–inclusion–growth nexus, thereby improving logical flow and readability, and

-        Added a critical synthesis paragraph at the end of the section that explicitly identifies remaining gaps in the literature and clarifies how the present study differentiates itself from existing work in terms of methodology, scope, and empirical design.

Now, we believe that these revisions ensure that the literature review is no longer descriptive in nature but instead provides a structured, critical, and theory-informed foundation for the empirical analysis that follows.

Comments 5: Throughout the text, the authors make several strong claims without adequate bibliographic support. For example, in points 22 to 24, the authors state that “A growing body of research finds that greater financial inclusion is associated with stronger GDP growth as well as lower levels of poverty and inequality” without citing the studies that support this claim. Similarly, in Section 2, Conceptual Framework and Hypotheses, widely accepted arguments regarding the effects of financial inclusion and fintech on economic growth and development are presented without reference to the existing literature.

Response 5: We thank the reviewer for this important remark concerning the need for stronger bibliographic support. We fully agree that strong claims must be grounded in well-established academic evidence.

In response to this comment, we have systematically revised the manuscript to strengthen the citation support throughout the text.

First, regarding the statement that “a growing body of research finds that greater financial inclusion is associated with stronger GDP growth as well as lower levels of poverty and inequality”, we now explicitly support this claim with seminal and widely cited academic references. In particular, we cite:

-        Beck, Demirgüç-Kunt, and Levine (2007), who provide early cross-country evidence linking financial development and inclusion to economic growth and poverty reduction.

-        Demirgüç-Kunt et al. (2018, 2022), who document the role of financial inclusion in improving economic opportunities and reducing inequality using Global Findex data; and

-        Sahay et al. (2015, 2020), who show that broader financial inclusion contributes to macroeconomic stability, growth, and inclusive development, especially in emerging and developing economies.

Second, in Section 2 (Conceptual Framework and Hypotheses), we have carefully revised the exposition to ensure that all theoretically grounded arguments concerning the effects of financial inclusion and FinTech on economic growth are now explicitly anchored in the existing literature. Specifically:

-        The hypothesized positive effect of financial inclusion on growth is now linked to classical and modern finance–growth theory (King and Levine, 1993; Beck et al., 2014);

-        The role of FinTech in reducing transaction costs, relaxing credit constraints, and expanding access to financial services is supported by recent empirical and policy-oriented studies (Philippon, 2016; Sahay et al., 2020; Claessens et al., 2018); and

-        The interaction between FinTech and financial inclusion is motivated by studies emphasizing complementarities between digital infrastructure and inclusion outcomes (Demirgüç-Kunt et al., 2018; World Bank, 2022).

Importantly, we did not merely add citations mechanically. Instead, we rewrote and restructured the relevant paragraphs to clearly distinguish between established findings in the literature and the novel contribution of our study. This ensures that the conceptual framework is both theoretically grounded and clearly positioned relative to prior research.

Key references added to the manuscript:

Beck, T.; Demirgüç-Kunt, A.; Levine, R. (2007). Finance, inequality and the poor. Journal of Economic Growth, 12, 27–49. DOI: 10.1007/s10887-007-9010-6

King, R. G.; Levine, R. (1993). Finance and growth: Schumpeter might be right. Quarterly Journal of Economics, 108, 717–737. DOI: 10.2307/2118406

Beck, T.; Demirgüç-Kunt, A.; Levine, R. (2014). Financial institutions, financial development, and economic growth. Journal of Economic Growth, 19, 157–188. DOI: 10.1007/s10887-014-9101-9

Demirgüç-Kunt, A.; Martinez Peria, M.; Tressel, T. (2018). The Global Financial Inclusion Database. World Bank Economic Review, 32, 1–30. DOI: 10.1093/wber/lhx028

Demirgüç-Kunt, A.; Klapper, L.; Singer, D.; Ansar, S.; Hess, J. (2022). The Global Findex Database 2021. World Bank. DOI: 10.1596/978-1-4648-1897-4

Sahay, R.; et al. (2015). Financial inclusion: Can it meet multiple macroeconomic goals? IMF Staff Discussion Note. https://www.imf.org/en/Publications/Staff-Discussion-Notes/Issues/2016/12/31/Financial-Inclusion-Can-It-Meet-Multiple-Macroeconomic-Goals-43153

Sahay, R.; et al. (2020). The promise of FinTech: Financial inclusion in the post-COVID-19 era. IMF Departmental Paper. https://www.imf.org/en/Publications/Departmental-Papers-Policy-Papers/Issues/2020/06/29/The-Promise-of-Fintech-Financial-Inclusion-in-the-Post-COVID-19-Era-49237

Philippon, T. (2016). The FinTech opportunity. NBER Working Paper No. 22476. DOI: 10.3386/w22476

Claessens, S.; Frost, J.; Turner, G.; Zhu, F. (2018). Fintech credit markets around the world. BIS Quarterly Review. https://www.bis.org/publ/qtrpdf/r_qt1809g.htm

Comments 6: It is also important to clearly explain the relationship between the selected variables, particularly the control variables, and the main variables of interest, providing a theoretical justification for their inclusion in the model.

Response 6: We thank the reviewer for this point. We fully agree that the inclusion of control variables must be theoretically grounded to ensure the validity of the econometric results.

In the revised manuscript, within the Data section, we have ensured that each control variable is explicitly justified based on its established relationship with economic growth and financial development. Specifically:

-        Trade Openness and FDI: These are included to account for international integration and technological spillovers, which are critical drivers of economic performance in the digital age.

-        Inflation: This serves as a proxy for macroeconomic stability, as volatile inflation can distort investment and undermine financial sector development.

-        Domestic Credit: This reflects the traditional depth of the financial sector, allowing us to isolate the "incremental" impact of FinTech beyond conventional banking.

-        Population Growth and Education: These capture demographic dynamics and human capital. Education, in particular, is a proxy for the digital literacy required to adopt FinTech innovations.

As detailed in the manuscript, we rely on a robust body of literature (e.g., Le et al., 2024; Ahammed, 2025; Lianos et al., 2023) to support these choices. This comprehensive set of controls significantly reduces the risk of omitted variable bias and strengthens the robustness of our findings regarding the FinTech-growth nexus.

Comments 7: In the presentation of results, the primary focus should be on the main findings and their theoretical and practical implications. The inclusion of a dedicated discussion section that links the results to the existing literature is strongly recommended.

Response 7: We fully agree with the reviewer that the presentation of results should prioritize the most significant findings and their broader implications.

In line with this suggestion, we have entirely rewritten and restructured the Results and Implications section to ensure a more impactful and professional delivery. Specifically:

-        Refinement of Findings: We have streamlined the reporting of our empirical results, moving away from excessive technical detail to focus on the core insights regarding the FinTech-growth nexus.

-        Integrated Discussion: Rather than presenting isolated numbers, we have integrated a dedicated discussion within this section that systematically links our results to existing theoretical and empirical literature. This includes contrasting our findings with the foundational works of Beck et al. (2014) and the recent Global Findex insights from Demirgüç-Kunt et al. (2022) .

-        Practical and Policy Implications: We have significantly enhanced the discussion of practical implications, particularly focusing on gender-sensitive public policies and the role of digital financial inclusion as a structural driver of inclusive growth.

-        Conciseness: By consolidating several tables and focusing the narrative on key findings, the manuscript is now more concise and easier for the reader to navigate.

Comments 8: Throughout the manuscript, the authors refer to several tables that do not appear in the main text. Given that some of these tables contain information essential for understanding the results, they should not be relegated solely to the appendices.

Response 8: Thank you for your constructive comments. We agree that the accessibility of key data is essential for a proper understanding of the results. To address this, we have synthesized the empirical findings that were previously dispersed across multiple tables in the appendices.

Instead of simply moving all raw tables, we have consolidated the results into integrated, high-impact tables that are now placed directly within the main text (Section 5). This synthesis allows the reader to view the baseline results, mediation effects, and moderation analysis in a more concise and comparative format. By reducing the total number of tables and embedding them within the core discussion, we have ensured a more fluid transition between our statistical evidence and our theoretical interpretations.

Comments 9: The article is approximately 60 pages long, which is not justified given the content presented. A more concise and focused writing style is recommended, reducing the total length to approximately 25 to 30 pages.

Response 9: Thank you for your relevant comment. We take this feedback very seriously and acknowledge that the previous version was excessively long. We have conducted a comprehensive overhaul of the manuscript to achieve a more concise, focused, and impactful presentation. Now, the total length has been significantly reduced.

4. Response to Comments on the Quality of English Language

Point 1: I don't feel qualified to evaluate the English language, but overall, the text was understandable. 

Response 1: Thank you for your positive feedback regarding the clarity of the manuscript. While the text was found to be understandable, we have taken this opportunity to perform a comprehensive linguistic refinement of the entire document.

5. Additional clarifications

Beyond the specific points addressed above, we have performed a final comprehensive proofreading of the manuscript to ensure that all internal cross-references, figure citations, and formatting details are perfectly aligned with the journal’s standards.

Response to Reviewer 3 Comments

1. Summary

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.

2. Questions for General Evaluation

Reviewer’s Evaluation

Response and Revisions

Does the introduction provide sufficient background and include all relevant references?

Can be improved

We have thoroughly revised the Introduction to provide a more comprehensive and up-to-date theoretical background. While the previous version laid the groundwork, the revised Introduction now explicitly situates the study within the most recent global economic shifts.

Are all the cited references relevant to the research?

Can be improved

We have conducted a thorough audit of the reference list to ensure maximum relevance and academic rigor. Following the reviewer’s suggestion, we have optimized the bibliography in the following ways: focus on core literature, integration of foundational works, current empirical evidence, and relevance to gender analysis.

Is the research design appropriate?

Can be improved

We have carefully reviewed the research design to ensure its appropriateness for the study's objectives. The revised manuscript now offers a more robust and better-justified empirical framework.

Are the methods adequately described?

Can be improved

We have significantly enhanced the "Materials and Methods" section to ensure full transparency and reproducibility of our empirical strategy. We have addressed the previous lack of detail by providing a more granular description of our technical approach.

Are the results clearly presented?

Can be improved

We have completely overhauled the presentation of our results to ensure clarity, flow, and analytical depth. The "Results" section has been transformed from a technical report into a structured narrative that guides the reader through our findings.

Are the conclusions supported by the results?

Can be improved

We have carefully revised the Conclusion to ensure that every claim is directly and strictly supported by our empirical evidence. We have moved away from broad generalizations to provide a more targeted, evidence-based closing that reflects the nuances of our 2021 dataset.

3. Point-by-point response to Comments and Suggestions for Authors

Comments 1: The paper addresses an important topic but suffers from conceptual overlap between FinTech and financial inclusion, limited causal identification due to a single-year cross-section, and over-interpretation of negative FinTech coefficients likely driven by pandemic effects and measurement choices.

Response 1: We thank the reviewer for this thoughtful and constructive assessment. The issues raised are important, and we have carefully addressed each of them in the revised manuscript to improve conceptual clarity, empirical transparency, and interpretation of the results.

First, regarding the conceptual overlap between FinTech and financial inclusion, we agree that these concepts are closely related and may appear intertwined if not clearly delineated. In the revised version of the paper, we have strengthened the conceptual framework to explicitly distinguish between the two. Financial inclusion is defined as the outcome dimension, capturing access to and usage of formal financial services by households and firms, while FinTech is treated as an enabling technological infrastructure, reflected through a composite index (FACTOR1) capturing the digitalization of financial ecosystems. This distinction is now clearly articulated in both the Introduction and the Conceptual Framework sections. Moreover, the empirical specification explicitly models FinTech as a mediator and moderator of financial inclusion, rather than as a substitute, thereby reducing conceptual redundancy and clarifying its complementary roles.

Second, with respect to the limited causal identification arising from the use of a single-year cross-section, we fully acknowledge this limitation. As explicitly stated in the revised limitations section, the analysis is not intended to deliver causal estimates but rather to provide structural and associational evidence on the FinTech–financial inclusion–growth nexus. The choice of a 2021 cross-section is deliberate and driven by data consistency considerations, particularly the availability of harmonized FinTech and financial inclusion indicators from the Global Findex database. We have carefully revised the manuscript to avoid causal language and to emphasize that the findings should be interpreted as suggestive of conditional relationships, consistent with established practice in cross-country macro-finance studies.

Third, regarding the interpretation of negative FinTech coefficients, we appreciate the reviewer’s point that such results may be influenced by pandemic-related dynamics and measurement choices. In the revised manuscript, we have substantially refined the discussion of these coefficients. We now explicitly acknowledge that negative or weak FinTech effects, particularly in certain income groups, may reflect short-run adjustment costs, uneven digital adoption during the post-pandemic period, or transitional inefficiencies associated with rapid digitalization rather than a detrimental long-term impact of FinTech. We also clarify that the FinTech index captures system-level digital intensity, which may temporarily correlate with economic disruption in 2021, especially in economies where digital transformation outpaced institutional and regulatory adaptation.

Importantly, we have reframed the interpretation of these coefficients to avoid overstatement. Rather than viewing negative FinTech coefficients as evidence against the growth-enhancing role of digital finance, we interpret them as indicative of context-dependent and non-linear effects, which reinforce the relevance of the interaction and heterogeneity analyses conducted in the paper.

Overall, we believe that the revised manuscript now offers a clearer conceptual distinction between FinTech and financial inclusion, a more transparent positioning of its empirical scope, and a more nuanced interpretation of the FinTech coefficients in light of the exceptional post-pandemic context.

Comments 2: The mediation/moderation framework is theoretically appealing but empirically weak given the data, while gender conclusions are suggestive rather than causal.

We thank the reviewer for this insightful comment regarding the empirical strength of the mediation/moderation framework and the interpretation of the gender-related results. We fully acknowledge that the use of cross-sectional data entails limitations with respect to causal inference, and we appreciate the opportunity to clarify the scope and positioning of our empirical analysis.

First, concerning the mediation and moderation framework, our objective is not to establish strict causality, but rather to empirically examine theoretically grounded transmission mechanisms that have been widely discussed in the literature yet remain insufficiently explored in a cross-country comparative setting. To avoid any overstatement, we have carefully revised the manuscript to consistently frame these results as associational and conditional relationships, in line with standard practice in macro-finance and development studies using cross-sectional data (e.g., Beck et al., 2014; Sahay et al., 2020).

Second, regarding the gender-related findings, we fully agree with the reviewer that these results should be interpreted as suggestive rather than causal. Accordingly, we have revised the relevant sections to clarify that the gender-disaggregated indicators capture structural disparities in access to financial services rather than causal gender effects at the individual level. The inclusion of these indicators is intended to highlight heterogeneity in the FinTech–financial inclusion–growth nexus and to inform policy discussions, rather than to make causal claims.

Importantly, in direct response to the reviewer’s concern, we have explicitly strengthened the limitations section of the manuscript. We now state:

Despite the robustness of our findings and the significant explanatory power of FACTOR1, this study is not without its limitations, which offer fertile ground for future investigation. First, the use of cross-sectional data from 2021, while offering a unique and timely snapshot of the post-pandemic digital landscape, limits our ability to make definitive claims about causality. Although our results reveal powerful and statistically significant associations, they should be interpreted as suggestive of a nexus rather than a strictly causal mechanism. Future research utilizing longitudinal or panel data would be instrumental in confirming the direction of these relationships and observing how the FinTech–growth link evolves as digital ecosystems mature over time.”

Finally, we have ensured consistency throughout the manuscript by (i) avoiding causal language, (ii) clearly positioning the mediation/moderation analysis as an exploratory and structural diagnostic framework, and (iii) explicitly outlining avenues for future research that could strengthen causal identification.

Comments 3: Policy implications are therefore stronger than the evidence allows and should be substantially tempered.

Response 3: We thank the reviewer for this important observation regarding the scope and strength of the policy implications. We fully agree that policy recommendations must be closely aligned with the evidentiary strength of the empirical analysis.

In response to this comment, we have substantially revised and tempered the policy implications section in the revised version of the manuscript. Specifically, we have taken several steps to ensure that the policy discussion remains fully consistent with the associational nature of the evidence.

First, we have reframed the policy implications as conditional and indicative rather than prescriptive, explicitly avoiding any causal language. The revised discussion emphasizes that the results provide insights into patterns and complementarities between FinTech development and financial inclusion, rather than definitive guidance on policy effectiveness.

Second, we have narrowed the focus of the policy implications to those findings that are robust across specifications, particularly the interaction and heterogeneity results across income groups and gender-based inclusion indicators. Rather than extrapolating broad policy conclusions from individual coefficients, the revised manuscript highlights how context-specific factors, such as income level, digital infrastructure maturity, and institutional readiness, condition the relationship between FinTech, financial inclusion, and growth.

Third, we have explicitly acknowledged the data and identification constraints in the policy discussion itself. The revised policy section now clearly states that the implications should be interpreted as informative for policy design and prioritization, not as causal prescriptions. This is reinforced by the expanded limitations subsection, which underscores the need for future research using quasi-experimental designs or difference-in-differences (DiD) or micro-level data before drawing firm policy conclusions.

As a result of these revisions, the policy implications are now carefully calibrated to the strength of empirical evidence, aligned with the cross-sectional nature of the data, and framed as a basis for further investigation rather than definitive policy guidance.

We believe these changes directly address the reviewer’s concern and improve the balance, credibility, and policy relevance of the manuscript.

4. Response to Comments on the Quality of English Language

Point 1: The English is fine and does not require any improvement.

Response 1: We sincerely thank the reviewer for this positive assessment. We appreciate the acknowledgment that the language quality meets the required academic standards. Nevertheless, we have carefully proofread the revised manuscript once again to ensure consistency, clarity, and adherence to the journal’s stylistic guidelines.

5. Additional clarifications

Beyond the specific points addressed above, we have performed a final comprehensive proofreading of the manuscript to ensure that all internal cross-references, figure citations, and formatting details are perfectly aligned with the journal’s standards.

Response to Reviewer 1 Comments (Round 2)

1. Summary: Following the evaluation of the revised version of the research manuscript, I can state that the authors have made significant modifications that have substantially improved the scientific quality and clarity of the paper. The comments raised during the previous review round have been carefully addressed and appropriately incorporated into the revised manuscript.

The revisions contribute to a more coherent structure of the argumentation, clearer presentation of the methodology, and a stronger results and discussion section. Furthermore, the references have been updated and expanded where necessary, and the additional explanations provided enhance the rigor and relevance of the scientific work.

In my opinion, in its current form, the manuscript meets the necessary standards for publication and can be accepted.

We sincerely thank you for your positive and encouraging evaluation of the revised manuscript. We greatly appreciate your careful reassessment and your recognition of the substantial improvements in structure, methodological clarity, and overall scientific rigor.

Your constructive feedback throughout the review process has been invaluable in strengthening the quality and coherence of our work. We are grateful for your recommendation for publication.

Response to Reviewer 2 Comments (Round 2)

1. Summary

We thank the reviewer for this detailed and constructive evaluation of the research design, methodology, and empirical presentation. The comments are highly valuable and have allowed us to substantially improve the transparency, structure, and rigor of the manuscript. Below, we address each point carefully and indicate the corresponding revisions made in the updated version.

2. Questions for General Evaluation

Reviewer’s Evaluation

Response and Revisions

Does the introduction provide sufficient background and include all relevant references?

Can be improved

We have carefully noted the reviewer’s assessment and have conducted a comprehensive rewrite of the Introduction to provide a more robust theoretical and empirical background.

Are all the cited references relevant to the research?

Can be improved

We have conducted a thorough audit of the reference list to ensure that every citation is directly relevant, academic, and up to date.

Is the research design appropriate?

Can be improved

We have carefully reviewed the research design and implemented several enhancements to ensure its rigor and appropriateness for the objectives of the study.

Are the methods adequately described?

Can be improved

We have significantly enhanced the methodology section to ensure that every step of our empirical strategy is transparent, reproducible, and rigorously justified.

Are the results clearly presented?

Can be improved

We have carefully noted the reviewer’s assessment and have conducted a comprehensive rewrite of the results.

Are the conclusions supported by the results?

Must be improved

We have carefully revised the conclusion to ensure our empirical findings directly and rigorously support it.

3. Point-by-point response to Comments and Suggestions for Authors

Comments 1: In the design and approach of the research, the authors highlight that the study employs a regression-based methodology to examine the role of FinTech in economic growth. However, the failure to specify the type of regression or the structure of the data (cross-sectional, panel or time series), the functional form, the controls, or the treatment of standard errors, prevents the evaluation of the adequacy of the model used according to the data and objectives of the study.

Response 1: We thank the reviewer for this important observation. We fully agree that the empirical strategy must be described with sufficient precision to allow a proper evaluation of its adequacy in relation to the data structure and research objectives.

In the revised manuscript, we have substantially clarified the research design and econometric specification as follows:

-        Structure of the Data: The study is based on a cross-sectional dataset for the year 2021, covering a broad sample of countries. This is now explicitly stated at the beginning of the Data and Methodology section. The choice of a single-year cross-section is motivated by the availability of harmonized FinTech and financial inclusion indicators in the most recent post-pandemic context.

-        Type of Regression and Functional Form: The empirical analysis relies on ordinary least squares (OLS) regression models, estimated within a linear functional framework. This specification is consistent with the mediation and moderation structure adopted in the study, where the objective is to assess average marginal effects and indirect transmission channels rather than dynamic adjustments.

-        Control Variables: The manuscript now clearly identifies and theoretically justifies the set of control variables included in the baseline regressions. These controls are selected based on the established finance–growth literature and aim to mitigate omitted variable bias by accounting for macroeconomic and structural country characteristics that may jointly influence FinTech development and economic growth.

-        Treatment of Standard Errors: To address potential heteroskedasticity and cross-country variability, all regressions are estimated using heteroskedasticity-robust standard errors. This is now explicitly stated in the methodological subsection and reported in the regression tables.

-        Coherence Between Data and Technique: The methodological section has been restructured to begin with a detailed description of the data, followed by the econometric framework. This ensures that the chosen estimation technique is clearly presented as a consequence of the nature and structure of the dataset.

These revisions significantly improve the transparency of the empirical approach and allow the reader to properly assess the adequacy of the model relative to the research objectives.

Comments 2: On the other hand, there is no justification for the choice of regression in the face of alternatives and endogeneity (inverse causality and omission of relevant variables) is not discussed, which may bias the results of the estimation.

Response 2: We thank the reviewer for raising this fundamental methodological concern. We fully agree that the choice of estimation technique and the potential presence of endogeneity must be explicitly discussed in cross-country growth analyses.

In the revised manuscript, we have strengthened the methodological justification along the following dimensions:

-        Justification for the Choice of OLS

The study employs cross-sectional OLS regression with heteroskedasticity-robust standard errors, which is consistent with the structure of the dataset (single-year cross-section) and the primary objective of the paper: to examine average associations and mediation/moderation mechanisms between FinTech development, financial inclusion, and economic growth.

Given the absence of a time dimension, panel estimators (fixed or random effects) are not applicable. Similarly, dynamic estimators (e.g., GMM) cannot be implemented without longitudinal variation. Our objective is not to estimate intertemporal dynamics but rather to identify structural relationships in the post-pandemic digital context.

OLS remains the benchmark approach in cross-country finance-growth research when the goal is to estimate conditional average effects under clearly specified controls.

-        Endogeneity and Reverse Causality

We acknowledge that reverse causality between FinTech development and economic growth is theoretically plausible. Higher income levels may foster technological adoption, just as FinTech expansion may stimulate growth.

In the revised manuscript, we now explicitly discuss this potential bidirectional relationship. We clarify that, given the cross-sectional design and the absence of valid external instruments satisfying exclusion restrictions, strict causal identification is not feasible. As such, our findings are interpreted as associational and structural, rather than strictly causal.

Importantly, the mediation framework allows us to explore transmission mechanisms in a structured way, which provides analytical depth beyond simple bivariate correlations, even if causal claims remain cautious.

-        Omitted Variable Bias

To mitigate omitted variable bias, we include a theoretically grounded set of macroeconomic control variables commonly used in the finance–growth literature. These controls capture structural country characteristics that may jointly influence both FinTech development and economic performance. While we recognize that no cross-country specification can eliminate all unobserved heterogeneity, we have clarified in the limitations section that residual bias may remain, and future research using panel data or instrumental variable strategies would be valuable to further address identification concerns.

-        Positioning of the Results

We have carefully tempered the interpretation of our results throughout the manuscript. The revised version explicitly states that the estimates reveal robust conditional associations and mediation patterns, but should not be interpreted as definitive causal effects.

Comments 3: The design and research approach should start with the description of the data, after it is the nature and characteristics of the data that should indicate the technique of processing them. On the other hand, aggregate sections 3 and 4 into a single one, both of which belong to the methodology.

Response 3: We thank the reviewer for this constructive and methodologically sound recommendation. We fully agree that the structure of the research design should logically begin with a clear description of the data, since the nature and characteristics of the dataset must guide the choice of empirical strategy.

In response to this comment, we have substantially reorganized the manuscript as follows:

-        Reordering of the Methodological Presentation

The revised version now begins the methodological section with a dedicated Data Description subsection, where we explicitly present:

• The structure of the dataset (cross-sectional, year 2021),
• The number of countries included in the sample,
• The sources of all variables (e.g., World Bank, Global Findex),
• The construction of key variables, including the FinTech index (FACTOR1),
• Summary statistics and data characteristics.

By clearly establishing the cross-sectional nature of the data at the outset, we allow the empirical strategy to emerge naturally from the dataset’s structure.

-        Aggregation of Sections 3 and 4 into a Unified Methodology Section

Following the reviewer’s suggestion, we have merged the previously separate sections into a single, coherent Methodology and Empirical Strategy section. This consolidated section now includes:

• Data description,
• Variable definitions,
• Model specification (functional form),
• Mediation and moderation framework,
• Estimation strategy and treatment of standard errors.

This restructuring improves logical coherence and avoids fragmentation between conceptual and econometric components of the research design.

-        Clear Link Between Data Characteristics and Econometric Technique

The revised manuscript now explicitly explains that:

• The cross-sectional structure of the data justifies the use of regression-based estimation in a conditional mean framework.
• The inclusion of interaction and mediation terms is consistent with the theoretical model.
• Robust standard errors are employed to address potential heteroskedasticity typical in cross-country datasets.

This alignment ensures that the empirical technique is presented as a direct consequence of the data structure and research objectives.

Comments 4: In the mediation test, the meanings of the variables X, Y and M are not clear.

Response 4: We thank the reviewer for identifying this ambiguity. The revised manuscript now clearly defines:

-        X: FinTech development (proxied by the PCA-based composite index),

-        M: Financial inclusion (account ownership and gender-disaggregated inclusion measures),

-        Y: Economic growth (GDP growth rate).

Comments 5: The measure of financial inclusion used in the study is quite limited, considering that the holding of a bank account reflects access to the financial system, however, it ignores the component of the use of banking products and services such as credit and deposits. In addition, the financial inclusion measure should include geographic and demographic penetration. Authors should also create a financial inclusion index using principal component analysis.

Response 5: We thank the reviewer for this thoughtful and conceptually important remark. We fully agree that financial inclusion is a multidimensional concept that extends beyond mere access and encompasses usage, quality, and demographic as well as geographic penetration.

Allow us to clarify the rationale behind our measurement strategy and the methodological choices adopted in the manuscript.

-        Justification for the Use of Account Ownership

The study relies on account ownership indicators from the Global Findex Database, which represent the internationally recognized benchmark for measuring financial inclusion across countries. Account ownership is widely used in cross-country empirical research because:

• It constitutes the most comparable and harmonized indicator available at the global level.
• It captures the foundational dimension of inclusion, formal access to the financial system.
• It is consistently available for a broad set of countries in 2021, ensuring sample integrity and avoiding severe missing-data bias.

While we acknowledge that usage indicators (e.g., credit uptake, deposit behavior, frequency of transactions) provide additional depth, such variables are often incomplete or inconsistently reported across countries for the specific year analyzed. Including them would substantially reduce the sample size and compromise cross-country comparability.

-        On Geographic and Demographic Penetration

We agree that financial inclusion also includes geographic outreach (e.g., branch density, ATM density) and demographic penetration (e.g., gender, rural–urban gaps).

In the manuscript, we explicitly incorporate gender-disaggregated inclusion indicators, which allow us to address demographic heterogeneity directly within the mediation and moderation framework. This provides analytical insight into inclusion disparities without introducing measurement inconsistencies across countries.

Geographic penetration indicators, however, are not uniformly available for 2021 across all countries in the sample. Their inclusion would either require interpolation or reliance on heterogeneous national sources, potentially reducing data reliability.

-        On Constructing a Financial Inclusion Index Using PCA

We carefully considered constructing a composite financial inclusion index using principal component analysis (PCA). However, PCA requires:

• A sufficiently large set of harmonized and comparable indicators,
• Minimal missing data,
• Stable correlation structure across variables.

Given the limited availability of consistent cross-country usage and outreach indicators for 2021, constructing a PCA-based inclusion index would introduce measurement noise and potentially weaken the robustness of the empirical analysis.

Importantly, we already apply PCA to construct the FinTech composite index (FACTOR1), where multiple digital financial indicators are available and statistically suitable for factor extraction. Extending PCA to financial inclusion without a sufficiently rich indicator set could lead to over-engineering the measurement without improving explanatory power.

-        Conceptual Coherence with the Mediation Framework

Within our mediation model, financial inclusion functions as a transmission channel between FinTech development and economic growth. For this structural role, a clear and interpretable indicator (such as account ownership) ensures conceptual transparency and avoids the opacity that can arise from highly aggregated composite indices.

This choice preserves interpretability in the indirect effect estimation and maintains coherence between theory and measurement.

-        Acknowledgment and Future Research

That said, we fully recognize the multidimensional nature of financial inclusion and have explicitly acknowledged in the limitations section that future research (particularly using panel data or richer cross-country datasets) could construct a broader inclusion index incorporating usage, geographic, and demographic dimensions.

Comments 6: Indicate in the correlation matrix tables the degree of significance of the correlations between the variables.

Response 6: We thank the reviewer for this important suggestion. We fully agree that indicating the statistical significance of the correlation coefficients enhances the interpretability and analytical value of the correlation matrix.

In the revised manuscript, we have updated the correlation tables as follows:

• Statistical significance levels are now explicitly reported using conventional notation (e.g., *, **, *** corresponding to the 10%, 5%, and 1% significance levels, respectively).
• A detailed note has been added below each table specifying the significance thresholds and clarifying that two-tailed tests were applied.
• The sample size used to compute the correlations is clearly indicated to ensure full transparency.

Including significance levels allows readers to distinguish between economically meaningful associations and statistically reliable relationships, which is particularly important in cross-country analyses where variability can be substantial.

Comments 7: In the presentation of the results, the authors should indicate the number of tables that are being interpreted for greater fluidity of reading. The results should be better discussed in terms of meanings and comparison with existing studies.

Response 7: We thank the reviewer for this constructive observation regarding the presentation and discussion of the empirical results. We fully agree that clarity in referencing tables and deeper analytical interpretation are essential for ensuring readability and scholarly rigor.

In the revised manuscript, we have implemented the following improvements:

-        Clear Identification of Interpreted Tables

Throughout the Results section, we now explicitly indicate the table number being interpreted at the beginning of each subsection. This ensures that readers can immediately associate the narrative discussion with the corresponding empirical evidence, thereby improving the fluidity and coherence of the exposition.

Each empirical step (baseline regression, mediation test, moderation analysis, and robustness checks) now clearly states which table is being analyzed, reducing any ambiguity in cross-referencing.

-        Strengthened Analytical Discussion

Beyond technical reporting, we have substantially deepened the interpretation of the coefficients. Specifically:

• We now emphasize the economic meaning of the estimated parameters, not merely their statistical significance.
• The magnitude and direction of effects are interpreted in light of theoretical expectations from the finance–growth and financial inclusion literature.
• Particular attention is given to explaining the negative or heterogeneous effects in the context of post-pandemic digital transformation dynamics.

-        Systematic Comparison with Existing Studies

The discussion section has been expanded to explicitly compare our findings with established empirical results in the literature. Where our results are consistent with prior research, we explain the conceptual alignment; where they diverge, we provide plausible structural explanations related to measurement, timing (2021 cross-section), and country heterogeneity.

This comparative perspective situates our contribution more clearly within the broader academic debate and strengthens the scientific positioning of the paper.

Comments 8: Finally, I suggest the application of the quantile regression model to assess the robustness of the results, considering the heterogeneities observed between countries (economic, financial and technological).

Response 8: We thank the reviewer for this thoughtful and technically well-motivated suggestion. We fully agree that cross-country heterogeneity (economic, financial, and technological) is an important feature of the data and deserves careful consideration.

After careful evaluation, however, we have chosen to maintain the baseline OLS framework for the following conceptual and methodological reasons.

-        Alignment with the Research Objective

The primary objective of the study is to examine:

• The average structural relationship between FinTech development and economic growth;
• The mediation role of financial inclusion;
• The moderating role of gender-related inclusion gaps.

The mediation-moderation framework relies on estimating consistent conditional mean effects across a unified specification. Quantile regression, by contrast, estimates conditional distributional effects at different points of the dependent variable distribution. While informative, such an approach would fundamentally shift the interpretation from structural transmission mechanisms to distribution-specific marginal effects.

In other words, the theoretical framework of the paper is built around average pathway analysis rather than heterogeneous slope estimation across quantiles of GDP growth.

-        Conceptual Coherence of Mediation Analysis

Mediation models are conventionally estimated within a mean-regression framework because the indirect effect is defined as the product of conditional mean coefficients. Extending mediation analysis to quantile regression would require a different methodological setup and would substantially complicate interpretation, particularly in a cross-sectional setting.

Our goal is to provide a coherent structural narrative linking FinTech, financial inclusion, and growth, not to model heterogeneous growth regimes per se.

-        Treatment of Heterogeneity in the Current Framework

Although we do not employ quantile regression, heterogeneity is not ignored. It is addressed through:

• Income-group stratifications,
• Interaction terms capturing moderation effects,
• Heteroskedasticity-robust standard errors,
• A discussion of differential effects across country groups.

These strategies allow us to capture meaningful structural differences without departing from the conceptual foundation of the model.

-        Sample Size Considerations

Given the cross-sectional nature of the dataset and the sample size, quantile regression (particularly at extreme quantiles) may produce unstable estimates and reduce statistical power. This is especially relevant when interaction and mediation terms are included.

Maintaining a parsimonious and interpretable framework enhances reliability and comparability with the established finance–growth literature.

-        Positioning for Future Research

That said, we recognize the value of quantile methods in exploring distributional heterogeneity. We have therefore acknowledged in the revised manuscript that future research using larger samples or panel data could extend the analysis by applying quantile regression or other distribution-sensitive techniques.

4. Response to Comments on the Quality of English Language

Point 1: The English could be improved to more clearly express the research.

Response 1: We thank the reviewer for this observation. We fully agree that clarity of language is essential to ensure that the research contribution, methodological choices, and policy implications are communicated precisely and unambiguously.

In the revised manuscript, we have undertaken a thorough linguistic and stylistic revision of the entire paper. Special attention was given to strengthening transitions between sections and ensuring that the research questions, hypotheses, and empirical findings are expressed in a logically progressive and reader-friendly manner.

5. Additional clarifications

Beyond the specific points addressed above, we have performed a final comprehensive proofreading of the manuscript to ensure that all internal cross-references, figure citations, and formatting details are perfectly aligned with the journal’s standards.

Response to Reviewer 3 Comments (Round 2)

1. Summary: Thank you for addressing all the necessary changes.

We sincerely thank you for your careful reassessment of our manuscript and for acknowledging that the requested changes have been satisfactorily addressed.

We greatly appreciate your constructive feedback throughout the review process, which has contributed significantly to improving the clarity, rigor, and overall quality of our work.