Digital Verification Capability, Motor Loss Ratios, and Insurer Performance: An Exploratory Panel Study of Saudi Arabia, 2020–2025

Response to the Managing Editor and Reviewer Manuscript: Digital Verification Capability, Motor Loss Ratios, and Insurer Performance:An Exploratory Panel Study of Saudi Arabia, 2020–2025 Dear Ms Obravyt, Thank you for the detailed assessment. We have radically revised the manuscript and re-estimated every reported result from the accompanying 102-observation analytical workbook. We agree that the previous version overstated data provenance, reproducibility, the role of National Address verification, and the strength of the empirical results. The revised manuscript withdraws those claims, reconstructs DVCI exactly from its five binary components, reports mixed rather than uniformly significant results, shortens and reorganises all sections, removes duplicate tables and formulas from the literature review, and explicitly identifies the remaining source-traceability limitation. Responses are provided point by point below. 1. Real data and the synthetic Zenodo file Comment: The underlying real data were not provided, while the Zenodo file is explicitly synthetic, making the 102 observations and estimates unverifiable. Response and revision: Agreed. The revised Data Availability Statement now states that the Zenodo file is synthetic and is not the empirical source dataset. It is described only as an illustrative code-and-format resource. The revised estimates are calculated from the accompanying author-compiled 102-observation workbook. We do not describe that workbook as independently source-verified because direct links or archived report extracts are missing for every row. The limitation is stated in Methods, Conclusions, and the Data Availability Statement. 2. Contradictory reproducibility claims Comment: The manuscript claimed that all variables could be independently verified and reproduced, whereas the Data Availability Statement indicated otherwise. Response and revision: Agreed. All claims of full or independent source-level reproducibility have been removed. The revised manuscript distinguishes computational reproducibility re-running the calculations from the supplied workbook) from source-level reproducibility (reconstructing each input from the original report. Only the former is currently possible. The results are explicitly labelled exploratory and provisional until the source trail is completed and audited. 3. Internal inconsistency in DVCI values Comment: Reported DVCI values contradicted the stated equal-weight calculation procedure. Response and revision: Resolved. DVCI has been recalculated as (NA + DI + AU + API + DG)/5 for all 102 rows. The maximum discrepancy between the stored and recomputed index is zero. Because the five items are binary and equally weighted, the only admissible values are 0.00, 0.20, 0.40, 0.60, 0.80, and 1.00. The previous minimum of 0.154 and maximum of 0.985 have been deleted. Table 1 now gives the exact construction and Table 2 reports the corrected range of 0.000–1.000. 4. Title and theoretical framing were too specific Comment: National Address integration was only one of five DVCI components, so the manuscript did not isolate its effect. Response and revision: Agreed. The title has been changed to “Digital Verification Capability, Motor Loss Ratios, and Insurer Performance: An Exploratory Panel Study of Saudi Arabia, 2020–2025.” The Introduction, Literature Review, Discussion, and Conclusions now state that National Address integration is one component of a broad disclosure index and that no isolated National Address effect is estimated. Claims specific to National Address performance effects have been removed. 5. Manuscript length Comment: The article exceeded the journal limit by approximately 3,000 words. Response and revision: Resolved through substantive deletion and restructuring. Repetitive theory, duplicated hypothesis explanations, repeated mediation descriptions, unnecessary section previews, duplicate references, two redundant figures, and duplicate/blank tables were removed. The revised manuscript is substantially shorter and uses six focused tables. 6. Abstract length and content Comment: The Abstract exceeded the limit and needed a clearer relevance–purpose–methods–results–conclusion structure. Response and revision: Resolved. The Abstract has been rewritten as a single concise paragraph. It identifies the research relevance and purpose, specifies the 17-firm/102-observation design and 96-observation motor subsample, reports the fixed-effects and bootstrap methods, presents the corrected mixed results, and ends with a limited non-causal conclusion and the source-reproducibility caveat. 7. Abbreviations and research place/time Comment: Abbreviations needed consistent introduction, and the manuscript needed to identify exactly when and where the research was conducted. Response and revision: Resolved. Every abbreviation is defined at first use, an Abbreviations section has been added, and terminology is standardised throughout. The Methods opening now states that the study concerns listed insurers in Saudi Arabia, covers fiscal years 2020–2025, and that the revised empirical analysis was completed in August 2026. 8. Introduction Comment: The Introduction was too long, included literature review, methods, purpose/tasks, expected results, and article structure; the specified methods sentence had to be removed. Response and revision: Resolved. The Introduction has been reduced to four focused paragraphs covering relevance, institutional context, and the scientific problem. The specified sentence beginning “Using an unbalanced panel...” has been removed, as have the purpose, hypotheses, methods preview, contributions list, and article roadmap. Those elements now appear only in their appropriate sections. 9. Literature Review organisation and formula Comment: The review lacked a clear analytical sequence, contained irrelevant material and a formula, and ended incorrectly. Response and revision: Resolved. The review is organised around three connected themes: information quality and insurance contracting; digital verification as organisational capability; and portfolio outcomes and the proposed pathway. The loss-ratio formula and all methods material have been removed from the review. The section ends with a concise synthesis of the scientific landscape, then states the study purpose, followed immediately by the grouped hypotheses. 10. Hypotheses placement and formulation Comment: Hypotheses should appear together after the Literature Review, without explanatory text between them, and should be distinct and unambiguous. Response and revision: Resolved. H1–H4 now appear consecutively after the review synthesis and purpose. They separately address DVCI–ROA, DVCI–LR, LR–ROA, and the indirect pathway. No explanatory paragraphs interrupt the hypotheses. 11. Methods and correctness of the sample Comment: The methods and data description needed to correspond to the actual analysis. Response and revision: Resolved. The sample is now correctly described as a balanced panel of 17 insurers over six years (102 observations), not an unbalanced panel. The LR models use 96 observations because the six Bupa Arabia rows contain no motor premium. The exact index rule, control choices, two-way fixed effects, firm-clustered standard errors, and 5,000 firm-cluster bootstrap are stated. AGE is omitted because it is collinear with firm and year effects; SOLV is kept for robustness because of its −.987 correlation with LEV. 12. Results and the previously “clean” estimates Comment: The extremely clean results were not credible without the source data. Response and revision: Resolved by re-estimating rather than preserving the earlier coefficients. The revised H1 estimate is β = −0.023 (p = .458), so H1 is not supported. H2 is directionally negative but marginal only at 10% (β = −0.055, p = .071). H3 is supported: LR is negatively associated with ROA (β = −0.571, p = .001). The indirect-effect bootstrap interval [−0.003, 0.064] includes zero, so H4 is not supported. The earlier VAF of 59.6% and all claims of uniformly significant effects have been removed. 13. Discussion Comment: The Discussion needed to interpret the study results and compare them with prior research instead of functioning as another literature review. Response and revision: Resolved. The Discussion is organised around the revised findings. It contrasts strong pooled correlations with attenuated within-firm estimates, explains the null direct effect using implementation and measurement considerations, compares the weak LR result with actuarial pricing research, relates the strong LR–ROA association to insurance economics, and explains why mediation and a National Address-specific interpretation are not supported. 14. Section function and subsection use Comment: Each section should serve its proper function; sections should not begin with subsections or contain unnecessary fragmentation. Response and revision: Resolved. Each major section begins with an orienting paragraph before any subsection. The Introduction, Results, Discussion, and Conclusions have no unnecessary subsections. Only the Literature Review and Methods retain a small number of functional subsections needed for analytical clarity. 15. Tables, figures, formulas, and captions Comment: Tables, figures, and formulas required professional presentation; captions should stand alone; “own development/calculations” labels were unnecessary. Response and revision: Resolved. Redundant figures and duplicate/blank tables have been removed. Six concise tables now report index construction, descriptive statistics, fixed-effects estimates, indirect-association results, robustness checks, and hypothesis decisions. Captions and notes define the sample and abbreviations and can be read independently. Formulas appear only in Methods. No table or figure is labelled as the authors’ own work. 16. Overall validity and scope of claims Comment: The manuscript required radical revision affecting transparency, internal consistency, variable validity, title, framing, and empirical claims. Response and revision: Agreed and addressed throughout. The revised manuscript is deliberately more cautious: it uses “association” rather than causal language, calls the study exploratory, distinguishes firm-wide ROA from motor LR, does not equate LR with adverse selection, does not isolate National Address, and makes the incomplete source trail a prominent limitation rather than hiding it. We recognise that full source-level verification remains a prerequisite if the journal requires auditable row-level provenance for further consideration. We appreciate the opportunity to revise the manuscript and have sought to make every empirical and interpretive claim commensurate with the evidence currently available. Sincerely,Hamed Abdallha Hamed Mosa