Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript examines Romania’s sheep production base, live-sheep and sheepmeat trade, and export-destination concentration over 2003–2024. The topic is relevant, and the integration of production and trade data provides a useful descriptive account of the sector. The clear commodity definitions, distinction between World totals and partner-level records, treatment of provisional data, and explicit recognition that gross export values do not measure domestic value added are notable strengths. Overall, the manuscript is potentially suitable for publication after minor revision. The authors should clarify the scope of statistical inference, reconcile several inconsistencies between the narrative and the reported results, provide additional clarification regarding geographical definitions and auxiliary data, and address the presentation and formatting issues detailed below.

1.Clarify the scope of statistical inference. Section 2.7.3 uses unrestricted permutations to reallocate 22 consecutive annual observations across five contiguous periods. Nonparametric testing does not itself establish exchangeability or remove serial dependence, and the HAC adjustment in Section 2.7.5 applies to a separate segmented regression. Please define the inferential target and justify the reference distributions for the period, trend, and hypothesis tests if the corresponding inferential claims are retained. Given the study’s descriptive contribution, an alternative is to present affected comparisons as historical descriptions and remove unsupported claims of statistical confirmation. Any changes in interpretation should be reflected consistently in the abstract, results, and conclusions. This request does not require converting the study into a causal policy evaluation.

2.Reconcile the substantive summaries with the reported data. Page 36, lines 1496–1501 refers to the absence of a proportional increase in meat exports, whereas the results report an 866.10% increase in sheepmeat export volume. The intended distinction appears to concern domestic meat production and the continuing small export-value share of meat. Page 29, lines 1186–1189 also describes export values and unit values as increasing steadily and peaking in 2020–2024, although Table 7 shows the sheepmeat export-value mean peaking in 2014–2019 and a nonmonotonic meat unit-value path. Please correct these statements separately.

3.Clarify geographic definitions and the role of auxiliary data. Please provide the EU-27 and MENA membership rules, explain the treatment of unspecified partner categories and their annual shares, and show more explicitly how the Eurostat series contribute to the results or cross-checks. In discussing 2024, the authors may also consider the within-window PPR outbreak and movement restrictions documented by WOAH as contemporaneous sector-specific context, without attributing the annual export decline to that event in the absence of an identifying analysis.

4.Correct numbering and cross-references. The analytical-periodization table and the hypothesis-operationalization table are both labelled Table 2. Subsection numbering also jumps from 2.7.5 to 2.7.8, and Section 2.8 is followed directly by 2.8.4. Please check the numbering sequence and update all associated cross-references. Supplementary Section S1 refers to Equation (29) in the main text, whereas the supplied manuscript ends its equation sequence at Equation (28). Please also keep each table caption with its table where possible.

5.Repair the placement of mathematical symbols. In Section 2.4, the year index and the symbols for Sheep Stock, Sheep Slaughtered, and Sheep Meat Production are displaced from their definitions, leaving empty parentheses. In Section 2.5, the HHI threshold expressions are grouped at the end of the paragraph rather than appearing beside the corresponding concentration categories. Please restore each symbol or inequality to its intended position and check the final PDF for similar typesetting problems.

6.Check the statistical-method citation numbers. The references cited in Sections 2.7.3–2.7.5 do not consistently match the bibliography. Under the current reference-list numbering, Benjamini–Hochberg should cite [47], Wilcoxon [48], Kendall [49], and Newey–West [50]. Please correct the relevant callouts and check consistency throughout the manuscript.

7.Standardize numerical presentation and table headings. Decimal commas and decimal points are mixed. For example, some H statistics are reported with commas in tables but with decimal points in the text. Please apply one consistent decimal convention and precision, while clearly distinguishing decimal separators from thousands separators. Replace “periodically” or “Periodically” with “Period” in the period-summary tables, replace “Statistical” with “Test statistic” where appropriate, and check the spacing and wrapping of units such as “kg/head”.

8.Complete missing text and correct terminology. Several passages require correction beyond general language polishing, including incomplete sentences and unfinished phrases in the methods section. Please also correct the placement of MVS_t and subject–verb agreement in the value-share discussion, replace “odds” with “shares” in the HHI definition, and use “sheepmeat” or “sheep meat” consistently for the defined commodity group because “mutton” can imply a narrower category than a group that includes lamb.

Comments on the Quality of English Language

The manuscript is generally understandable, but the English would benefit from careful language editing to improve clarity, grammatical consistency, and professional presentation. In particular, please revise incomplete sentences and awkward constructions, standardize terminology such as “sheepmeat/sheep meat” and the names of statistical indicators, and check subject–verb agreement, article use, punctuation, and spacing around mathematical notation. A final proofread of the typeset PDF is also recommended because several apparent language problems are compounded by displaced symbols and line-break or formatting errors.

Author Response

Dear Reviewer 1

We thank you for the careful and constructive reading of our manuscript and for your time. Below we reproduce each of your comments in full, followed by our point-by-point responses. All changes are marked in the revised manuscript. Section, table and equation numbers below refer to the revised manuscript unless marked “S”, which refers to the accompanying Supplementary Materials file.

1. Clarify the scope of statistical inference. Section 2.7.3 uses unrestricted permutations to reallocate 22 consecutive annual observations across five contiguous periods. Nonparametric testing does not itself establish exchangeability or remove serial dependence, and the HAC adjustment in Section 2.7.5 applies to a separate segmented regression. Please define the inferential target and justify the reference distributions for the period, trend, and hypothesis tests if the corresponding inferential claims are retained. Given the study’s descriptive contribution, an alternative is to present affected comparisons as historical descriptions and remove unsupported claims of statistical confirmation. Any changes in interpretation should be reflected consistently in the abstract, results, and conclusions. This request does not require converting the study into a causal policy evaluation.

Response: Thank you for this comment. We agree that exchangeability is not established by the permutation procedure itself. We have extended Section 2.8.1 (“Methodological limitations”) with an explicit statement that the permutation Kruskal–Wallis period comparisons (Section 2.7.3) assume exchangeable observations, an assumption not strictly guaranteed for annual series with potential serial dependence, and that these comparisons are best read as descriptive contrasts between historical sub-periods rather than as tests of a general causal hypothesis. As a complementary check, we introduced a new Section 2.7.5 paragraph applying the autocorrelation-robust Hamed–Rao (1998) modified Mann–Kendall test (new reference [53]) to the six indicators for which trend significance is reported in the Results and Discussion. After Benjamini–Hochberg adjustment, this test agreed with the classical Kendall test on both the direction and the statistical significance of the trend for all six indicators (full results in Table S6 of the Supplementary Materials), which corroborates that serial dependence does not drive the trend conclusions. We did not find it necessary to remove the period-comparison results, since they are now explicitly and consistently framed, throughout the Methods, Results, Discussion and the new limitations paragraph, as descriptive historical contrasts rather than as confirmatory causal tests; the Abstract and Conclusions were reviewed for consistency with this framing and required no further change, since neither uses hypothesis-confirmation language beyond reporting which comparisons were statistically significant.

2. Reconcile the substantive summaries with the reported data. Page 36, lines 1496–1501 refers to the absence of a proportional increase in meat exports, whereas the results report an 866.10% increase in sheepmeat export volume. […] Page 29, lines 1186–1189 also describes export values and unit values as increasing steadily and peaking in 2020–2024, although Table 7 shows the sheepmeat export-value mean peaking in 2014–2019 and a nonmonotonic meat unit-value path. Please correct these statements separately.

Response: Thank you. Both inconsistencies have been corrected. In the Conclusions, the sentence stating that the herd advantage was exploited “not through a proportional increase in meat exports” has been rewritten to distinguish explicitly between the substantial growth in sheepmeat export volume (confirmed) and the much slower growth of slaughter and physical meat production together with meat’s continuing small share of total export value, which is the intended distinction. In Section 3.5, the sentence describing export values and unit values as “increasing steadily and peaking in 2020–2024” has been corrected to state that the sheepmeat export-value mean in fact peaked in the 2014–2019 period, consistent with Table 7, and that the meat unit-value path is nonmonotonic rather than steadily increasing.

3. Clarify geographic definitions and the role of auxiliary data. Please provide the EU-27 and MENA membership rules, explain the treatment of unspecified partner categories and their annual shares, and show more explicitly how the Eurostat series contribute to the results or cross-checks. In discussing 2024, the authors may also consider the within-window PPR outbreak and movement restrictions documented by WOAH as contemporaneous sector-specific context, without attributing the annual export decline to that event in the absence of an identifying analysis.

Response: Thank you for this observation. Section 2.5 now states the fixed, time-invariant EU-27 and MENA membership rules used throughout the regional analysis (including the treatment of the United Kingdom as a post-2020 non-member), lists the MENA destinations included, and reports the annual share of the residual “Other” category (median approximately 1.5% for live-animal exports and 0.6% for sheepmeat exports, occasionally reaching about 11% and 33% respectively). As a further robustness check responding to this point, a region-restricted HHI was calculated separately for the EU-27 and MENA subsets of each flow; the complete results are reported in new Section S5 (Table S5) of the Supplementary Materials. The role of the three Eurostat series (slaughter statistics, producer price index, agricultural output value) as cross-checks on the primary FAOSTAT and UN Comtrade series is now stated explicitly in Section 2.5, alongside a new sentence describing the EUR-denominated deflation robustness check (Section S7) that uses two additional Eurostat series (EUR/USD exchange rate and Romanian HICP). Finally, Results Section 3.1 now includes a carefully worded sentence noting the 2024 PPR outbreak and WOAH-documented movement restrictions as contemporaneous sector-specific context for that year’s export decline, explicitly stating that no causal attribution is intended in the absence of an identifying analysis, with the WOAH notice cited as new reference [56].

4. Correct numbering and cross-references. The analytical-periodization table and the hypothesis-operationalization table are both labelled Table 2. Subsection numbering also jumps from 2.7.5 to 2.7.8, and Section 2.8 is followed directly by 2.8.4. […] Supplementary Section S1 refers to Equation (29) in the main text, whereas the supplied manuscript ends its equation sequence at Equation (28). Please also keep each table caption with its table where possible.

Response: Thank you for this observation. All tables are now sequentially and uniquely numbered (Table 1 through Table 13, with the periodization table as Table 2 and the hypothesis-operationalization table as Table 3), subsections run sequentially from 2.7.1 through 2.7.6 and 2.8.1 through 2.8.2, and no cross-reference in the main text or Supplementary Materials points to Equation (29) or any equation beyond the manuscript’s final Equation (28). We have asked the typesetter to keep each table caption together with its table where pagination allows; we will review the final proof specifically for this issue.

5. Repair the placement of mathematical symbols. In Section 2.4, the year index and the symbols for Sheep Stock, Sheep Slaughtered, and Sheep Meat Production are displaced from their definitions, leaving empty parentheses. In Section 2.5, the HHI threshold expressions are grouped at the end of the paragraph rather than appearing beside the corresponding concentration categories. Please restore each symbol or inequality to its intended position and check the final PDF for similar typesetting problems.

Response: Thank you for this comment. displaced year index and indicator symbols in Section 2.4 have been restored to their intended positions beside their definitions, and the Net Export Position (NEP) symbol in Section 2.5, together with the HHI threshold inequalities, has been repositioned next to the corresponding concentration categories rather than grouped at the end of the paragraph; an incomplete sentence in the same paragraph (“Conversely, a dominant share or high value…”) has also been completed. We reviewed the surrounding sections for similar displacement and will give the final typeset PDF a further dedicated check before publication.

6. Check the statistical-method citation numbers. The references cited in Sections 2.7.3–2.7.5 do not consistently match the bibliography. Under the current reference-list numbering, Benjamini–Hochberg should cite [47], Wilcoxon [48], Kendall [49], and Newey–West [50]. Please correct the relevant callouts and check consistency throughout the manuscript.

Response: Thank you for this observation, too. We have verified that Benjamini–Hochberg, the Wilcoxon signed-rank test, Kendall’s test and Newey–West are cited consistently throughout Sections 2.7.3–2.7.5 and the Results, matching the reference list. Because keeping every reference strictly in order of first appearance (as required by the journal) required renumbering the full reference list once the new references were added, these callouts are now [50], [51], [52] and [54] respectively, rather than the [47]–[50] identified at the time of review; we checked them for consistency throughout the manuscript.

7. Standardize numerical presentation and table headings. Decimal commas and decimal points are mixed. […] Replace “periodically” or “Periodically” with “Period” in the period-summary tables, replace “Statistical” with “Test statistic” where appropriate, and check the spacing and wrapping of units such as “kg/head”.

Response: Thank you for this comment. We identified and corrected every instance in which a Kruskal–Wallis H statistic or a rank-biserial correlation value was reported with a decimal comma instead of a decimal point (22 cells across Tables 5, 9, 11 and 12), so that all decimal values now use a point consistently throughout the manuscript, distinguishing clearly from thousands separators (e.g., “20,000 rearrangements”, “41,993 t”). The period-summary table headings already read “Period” rather than “Periodically”/“Periodically”, and “Test statistic” rather than “Statistical”, so no further change was required there. We added a missing space before parenthetical units in several table headings (e.g., “Slaughter rate (%)”, “Carcass yield (kg/head)”, “Live Export Unit Value (USD/head)”) and corrected a run-together “Live-to-Meat ExportValue Ratio” to “Live-to-Meat Export Value Ratio” in two table headings, to improve spacing and wrapping consistency.

8. Complete missing text and correct terminology. Several passages require correction beyond general language polishing, including incomplete sentences and unfinished phrases in the methods section. Please also correct the placement of MVS_t and subject–verb agreement in the value-share discussion, replace “odds” with “shares” in the HHI definition, and use “sheepmeat” or “sheep meat” consistently for the defined commodity group because “mutton” can imply a narrower category than a group that includes lamb.

Response: Thank you for this comment. The MVS_t symbol placement and the subject–verb agreement in the value-share discussion (“do not” corrected to “does not”) have been corrected, “odds” has been replaced with “shares” in the HHI definition, and “mutton” has been replaced with “sheepmeat”/“sheep meat” for the defined commodity group throughout the manuscript, including in the Introduction’s H4 discussion. The incomplete sentence in Section 2.5 noted under point 5 above has also been completed.

Thank you again for this detailed and constructive report, which has substantially improved the precision and internal consistency of the manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

1. The introduction is compelling, but it lacks specific literature regarding live sheep exports from the EU and Romania to MENA markets. Also missing is a discussion on the role of seasonal and religious demand (Eid al-Adha) and comparable international experience, for instance, the phase-out of live sheep exports by sea from Australia.
2. Regarding the research design, the condition LVS > 50% (H1) is identical by construction (Equation 10) to the condition LEV > MEV in the same year (H4). Consequently, the sign test and the Wilcoxon signed-rank test assess the same event, and the claim that the hypotheses are "conceptually close, but not identical" is unsubstantiated. I recommend merging the two hypotheses or reformulating H4, for example, as a hypothesis concerning the evolution of the live-to-meat ratio.
3. Regarding the methodology section:
   • A decline in the slaughter rate is interpreted as a decoupling between the livestock herd and domestic processing; however, live exports and animals retained for herd growth are not integrated into a unified balance sheet. I recommend calculating a total utilization rate (slaughter + net live exports) / herd size, to distinguish between the substitution of domestic slaughter with exports and simple herd accumulation.
   • Regarding statistical inference on autocorrelated series: the Kendall test and year-permutation tests across sub-periods assume independent and exchangeable observations, an assumption that does not hold for annual series exhibiting autocorrelation. I recommend applying the modified Mann–Kendall test (Hamed–Rao) or pre-whitening for trend analysis and block permutation, or at least explicitly discussing this limitation. The segmented regression (Equation 28) relies on only four observations prior to 2007; the immediate increase in meat production in 2007 (β = 0.371) is likely influenced by the 2006 low (41,993 t), a point that warrants discussion.
   4. The analysis relies exclusively on data reported by Romania to UN Comtrade; for MENA destinations, I recommend a comparison with mirror data (reported by importers), as these are a frequent source of discrepancies. Deflating using the US CPI-U is of limited economic relevance for an exporter in the EUR/RON zone. I believe it would be preferable to use an export unit value index or to include an additional test using a EUR deflator in the results.
   5. The HHI threshold of 0.25 originates in antitrust policy and is applied arbitrarily to export geographic concentration; however, without a comparison against similar exporters (e.g., Hungary, Bulgaria, Spain) or the structure of global demand, the verdict regarding H3 hinges entirely on this threshold. I recommend adding such a benchmark and calculating the HHI by region (EU vs. MENA), as regional reorientation is more relevant to resilience than the mere number of partners.
   6.  References and numbering:
   • References [47]–[50] in the text are misaligned with the reference list: Benjamini–Hochberg is cited as [50] but appears as [47] in the list; Wilcoxon is cited as [47] but appears as [48]; Kendall is cited as [48] but appears as [49]; Newey–West is cited as [49] but appears as [50].
   • Subsection 2.7.8 immediately follows 2.7.5; Subsection 2.8.4 appears without subsections 2.8.1–2.8.3.
   • Two different tables bear the number 2.
   7. The recommendations regarding regional slaughterhouses, mobile units, and refrigeration capacities, as well as the conclusion that the results "do not support administrative restrictions" do not follow from the analysis. The study contains no data on slaughterhouse capacity, costs, or the effects of restrictions, and the authors acknowledge this. Therefore, I recommend reformulating these recommendations as research directions or policy hypotheses, shortening the Discussion section (which largely repeats the results), and making the harmonized dataset available.

Author Response

Dear Reviewer 2

Thank you for a rigorous and detailed report that has led to substantial improvements in the manuscript, most importantly a full reformulation of Hypothesis 4. Below we reproduce each comment in full (bold, numbered as in the original report) followed by our point-by-point response. All changes are marked as tracked changes in the revised manuscript; section, table and equation numbers below refer to the revised manuscript unless marked “S”, which refers to the accompanying Supplementary Materials file.

1. Expand the literature on live sheep exports and comparable international experience. The introduction is compelling, but it lacks specific literature regarding live sheep exports from the EU and Romania to MENA markets. Also missing is a discussion on the role of seasonal and religious demand (Eid al-Adha) and comparable international experience, for instance, the phase-out of live sheep exports by sea from Australia.

Response: Thank you for this comment. We have added three new citations to the Introduction (references [33]–[35]) and a corresponding discussion. The revised text now situates the Romania-to-MENA live sheep trade within the wider EU live-export literature, discusses the strong seasonal concentration of demand around Eid al-Adha as a structural driver of the live-export channel, and references the staged phase-out of live sheep exports by sea from Australia as a comparable international policy precedent, noting the differences in fleet composition, destination markets and domestic processing capacity that make a direct transposition of that experience to Romania uncertain.

2. Reconcile the H1/H4 identity. Regarding the research design, the condition LVS > 50% (H1) is identical by construction (Equation 10) to the condition LEV > MEV in the same year (H4). Consequently, the sign test and the Wilcoxon signed-rank test assess the same event, and the claim that the hypotheses are “conceptually close, but not identical” is unsubstantiated. I recommend merging the two hypotheses or reformulating H4, for example, as a hypothesis concerning the evolution of the live-to-meat ratio.

Response: Thank you. We agree with this comment, which identified a genuine flaw in the original research design: since Live Value Share is algebraically LEV/(LEV + MEV), the condition LVS > 50% and the condition LEV > MEV are mathematically equivalent in every year, so the original H1 (sign test on the annual dominance indicator) and H4 (paired Wilcoxon test on LEV versus MEV) tested the identical event under two different statistical procedures and could not be treated as separately confirmatory. Rather than merge the two hypotheses, we have reformulated H4 as a hypothesis about the evolution of the live-to-meat value relationship across the study’s analytical periods, operationalized as a period-comparison test on the Live Value Share series (Section 2.7.4, Table 3). The original paired Wilcoxon/sign-test evidence is retained in the manuscript, but now presented solely as descriptive corroborating evidence for H1, not as a second, independent test. This reformulation is threaded consistently through the Introduction, Methods (Sections 2.7.4 and 2.7.6), the hypothesis-operationalization table (Table 3), the results tables (Tables 9/10 and 12/13), the Results narrative, the Discussion, the robustness-check section (including the Supplementary Materials’ deflation and leave-one-year-out checks, Section S3), and the Conclusions, so that every mention of H4 throughout the manuscript now refers consistently to the period-comparison test on Live Value Share rather than to the original paired-value test.

3. Methodology: total utilization rate, autocorrelation-robust inference, and the pre-2007 segment. A decline in the slaughter rate is interpreted as a decoupling between the livestock herd and domestic processing; however, live exports and animals retained for herd growth are not integrated into a unified balance sheet. I recommend calculating a total utilization rate (slaughter + net live exports) / herd size […] Regarding statistical inference on autocorrelated series: the Kendall test and year-permutation tests across sub-periods assume independent and exchangeable observations […] I recommend applying the modified Mann–Kendall test (Hamed–Rao) or pre-whitening for trend analysis and block permutation, or at least explicitly discussing this limitation. The segmented regression (Equation 28) relies on only four observations prior to 2007; the immediate increase in meat production in 2007 (β = 0.371) is likely influenced by the 2006 low (41,993 t), a point that warrants discussion.

Response: We thank you for this three-part comment and have addressed each element. First, we now calculate a total utilization rate, defined as (animals slaughtered + net live exports) / sheep stock, for each year 2003–2024; the full annual series and period averages are reported in new Section S4 (Tables S4 and S4b) of the Supplementary Materials and referenced from Methods Section 2.4 and Discussion Section 4.1, allowing the decline in the slaughter rate to be read against a unified balance-sheet measure that distinguishes substitution of domestic slaughter by live exports from simple herd accumulation. Second, we have applied the autocorrelation-robust Hamed–Rao (1998) modified Mann–Kendall test (new reference [53]) to the six indicators for which trend significance is reported; after Benjamini–Hochberg adjustment this test agreed with the classical Kendall test on direction and significance for all six indicators (Section 2.7.5, full results in new Section S6, Table S6), and we have added an explicit statement in the new Methodological Limitations subsection (Section 2.8.1) that the Kruskal–Wallis period-comparison tests assume exchangeability, an assumption not strictly guaranteed for annual series with potential serial dependence, so that these comparisons are best read as descriptive historical contrasts rather than confirmatory causal tests. We did not additionally implement pre-whitening or block permutation, since the Hamed–Rao check and the explicit limitation statement together address the reviewer’s underlying concern without further lengthening the Methods section; we note this as a direction for future work in Section 2.8.1. Third, we have added a sentence to Section 2.7.6 explicitly discussing that the segmented regression’s pre-2007 segment rests on only four annual observations and that the estimated post-break increase in meat production is likely influenced in part by the unusually low 2006 base value (41,993 t), and we now present the segmented-regression result with this caveat rather than as an unqualified structural break.

4. Mirror data for MENA destinations and the choice of deflator. The analysis relies exclusively on data reported by Romania to UN Comtrade; for MENA destinations, I recommend a comparison with mirror data (reported by importers), as these are a frequent source of discrepancies. Deflating using the US CPI-U is of limited economic relevance for an exporter in the EUR/RON zone. I believe it would be preferable to use an export unit value index or to include an additional test using a EUR deflator in the results.

Response: Thank you for this comment. We have added an explicit statement to the Methodological Limitations subsection (Section 2.8.1) acknowledging that the analysis relies on Romania’s reporter-side UN Comtrade data without a systematic mirror-data comparison for MENA destinations, noting that reporter/mirror discrepancies are a documented concern for these trade flows and flagging this as a direction for future work; a full mirror-data reconciliation was judged beyond the scope of the present revision given the number of partner countries and years involved. On the deflator point, we agree that the US CPI-U is of limited direct relevance for a Romanian exporter, and we have added a EUR-denominated robustness check (new Section S7 of the Supplementary Materials, Tables S7 and S8) that converts the USD trade values to constant 2024 EUR using the EUR/USD exchange rate and the Romanian Harmonized Index of Consumer Prices (both Eurostat series) in place of the US CPI-U; this check corroborates the main text’s Kruskal–Wallis period-comparison and trend-test conclusions under a EUR/RON-relevant deflator, as now stated in Section 2.5 and referenced from the relevant Results and Discussion passages (Section S7 cross-reference). We did not additionally construct an export unit value index, since the EUR-deflator check directly targets the reviewer’s stated concern (the currency/price-zone relevance of the deflator) without introducing a second, differently defined price measure.

5. HHI threshold and regional benchmarking. The HHI threshold of 0.25 originates in antitrust policy and is applied arbitrarily to export geographic concentration; however, without a comparison against similar exporters (e.g., Hungary, Bulgaria, Spain) or the structure of global demand, the verdict regarding H3 hinges entirely on this threshold. I recommend adding such a benchmark and calculating the HHI by region (EU vs. MENA), as regional reorientation is more relevant to resilience than the mere number of partners.

Response: Thank you for this observation. We have added an explicit acknowledgment in the Methodological Limitations subsection (Section 2.8.1) that the conventional antitrust HHI threshold of 0.25 is used here as a descriptive reference point rather than a validated benchmark for export-market concentration, and that a cross-country comparison against similar exporters (e.g., Hungary, Bulgaria, Spain) was not performed and is noted as a direction for future work. In direct response to the recommendation to calculate HHI by region, we have added a region-restricted HHI, computed separately for the EU-27 and MENA destination subsets of each flow (live-animal and sheepmeat) and for each analytical period; the complete results are reported in new Section S5 (Table S5) of the Supplementary Materials and referenced from Section 2.5 and Section 3.3 (Section S5 cross-reference). This regional analysis shows MENA destinations consistently more concentrated than EU-27 destinations for live exports across all periods, while meat-export concentration within MENA falls sharply over time as the number of MENA meat-importing partners grows, a pattern that speaks directly to the reviewer’s point that regional reorientation, not only partner count, is informative about resilience.

6. Reference and subsection numbering. References [47]–[50] in the text are misaligned with the reference list […] Subsection 2.7.8 immediately follows 2.7.5; Subsection 2.8.4 appears without subsections 2.8.1–2.8.3. Two different tables bear the number 2.

Response: Thank you for this comment. We have verified that, in the revised manuscript, Benjamini–Hochberg, the Wilcoxon signed-rank test, Kendall’s test and Newey–West are cited consistently and correctly throughout Sections 2.7.3–2.7.5 and the Results, matching the reference list. Because keeping every reference strictly in order of first appearance (as required by the journal) required renumbering the full reference list once the new references were added, these callouts are now [50], [51], [52] and [54] respectively, rather than the [47]–[50] the reviewer identified at the time of review; we checked them for consistency throughout the manuscript. Subsections now run sequentially from 2.7.1 through 2.7.6 and from 2.8.1 through 2.8.2, with no gaps or out-of-sequence numbers, and all tables are sequentially and uniquely numbered (Table 1 through Table 13), with the periodization table as Table 2 and the hypothesis-operationalization table as Table 3, so no two tables share a number.

7. Reformulate policy recommendations, shorten the Discussion, and strengthen data availability. The recommendations regarding regional slaughterhouses, mobile units, and refrigeration capacities, as well as the conclusion that the results “do not support administrative restrictions” do not follow from the analysis. The study contains no data on slaughterhouse capacity, costs, or the effects of restrictions, and the authors acknowledge this. Therefore, I recommend reformulating these recommendations as research directions or policy hypotheses, shortening the Discussion section (which largely repeats the results), and making the harmonized dataset available.

Response: Thank you for these recommendations. We agree that these recommendations were stated with more confidence than the data support. The Conclusions’ statement that the results “do not support administrative restrictions” has been reformulated as a hedged research direction, explicitly noting that the study contains no data on slaughterhouse capacity, processing costs, or the effects of past or hypothetical restrictions, and that this limits any inference about the merits of such measures; the recommendations regarding regional slaughterhouses, mobile units, and refrigeration capacity are likewise reframed in the Discussion (Section 4.4) as policy hypotheses warranting dedicated investment-feasibility and cost analysis rather than as findings of the present study, with a cross-reference to the Conclusions (Section 5) to avoid repeating the same material in both places. We have also shortened the Discussion section overall by removing passages that largely restated the Results, so that the Discussion now focuses on interpretation, limitations, and the policy-hypothesis framing rather than re-reporting numerical results already presented in Section 3. Finally, the Data Availability Statement has been strengthened to commit explicitly to making the harmonized annual dataset underlying the analysis (FAOSTAT, UN Comtrade, and Eurostat series, merged and documented) available in the Supplementary Materials, rather than offering it only on request.

Thank you again for this exceptionally thorough and constructive report, which identified a genuine design flaw in the original H1/H4 pair and has materially strengthened both the methodology and the policy framing of the manuscript.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The work has been improved, and I recommend it for publication.