Two-Level System for Optimal Flood Risk Coverage in Spain

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper highlights the CCS's role in flood damage compensation and proposes a two-level system combining solidarity and risk-based premiums for better flood risk management. This paper presents a new perspective and has provided many inspirations. Some minor suggestions for revision are as follows.

Lines 23-24: It would be beneficial to first provide an overview of the global economic losses from floods and a general summary of flood insurance before focusing on the specific issues in Spain. This would provide context for the significance of flood risk management and highlight why Spain’s situation warrants attention.

Lines 115-117 “However, this could be considered more of a tax than a premium for the coverage of catastrophic risks, since it does not take into consideration the exposure to the risk”: There could be a potential conceptual shift between taxes and insurance. If the discussion conflates these two concepts, it may create confusion about the mechanisms for funding flood risk management and compensation. It would be important to clearly distinguish between the two—insurance as a risk-sharing tool and taxes as a broader public funding method.

Lines 173-175: “On the other hand, occasional flood events not considered extraordinary would be covered by the private sector.” It needs a balanced risk management system involving both the public and private sectors. Could you please elaborate on that?

Line 220: Why did annotation 2 appear when annotation 1 was absent?

Line 221: Figure 1 is not marked.

Line 230: Could you summarize the several related formulas of extreme events mentioned in the text? Have there been any new improvements to the formulas?

Lines 282-283: Why not include the condition "cost < 1 thousand € "?

Line 325: This should be Figure 2? The font sizes of the horizontal and vertical axes of the graph are too small, making it difficult to read. Please enhance the clarity of the pictures.

Lines 332 and 342: What is “e” of “Cumulative cost (e)” and “Cost (e)”?

Lines 450-451: “Our motivation for writing this paper was the direct relationship that the CCS establishes between flood risk and the risk from extraordinary or catastrophic events.” It seems that the evidence provided for this statement in the text is a bit insufficient.

Lines 538-587: It would be better to add more references to support the argument.

Line 623: Shouldn't you add a conclusion?

Lines 631-632: Please add the journal volume, issue and page numbers.

Author Response

Two-level System for Optimal Flood Risk Coverage in Spain

 

ANSWER TO REVIEWER 1 (1st ROUND)

Thank you very much for all the suggestions received by the reviewer which have helped us improve the paper substantially.

Additionally, we have made other modifications to the paper with the intention of improving understanding of it.

 

Lines 23-24: It would be beneficial to first provide an overview of the global economic losses from floods and a general summary of flood insurance before focusing on the specific issues in Spain. This would provide context for the significance of flood risk management and highlight why Spain’s situation warrants attention.

We thank the reviewer for this valuable suggestion. In response, we have expanded the Literature Review section to include a comparative discussion of flood insurance systems in several countries, including France, Germany, the United Kingdom, the United States, and the Netherlands. This addition serves to contextualize the Spanish model within a broader spectrum of institutional arrangements and highlights its unique features.

We have added the revised paragraph on page 4 (Introduction) of the manuscript and reads as follows:

Comparative research on flood insurance systems reveals a wide diversity of institutional arrangements across countries. France operates a mixed system in which private insurers provide coverage backed by a public reinsurance scheme through the Caisse Centrale de Réassurance (Paudel, 2012). In Germany, flood insurance is voluntary and offered by private companies, but penetration rates remain low due to limited risk awareness and affordability challenges (Schwarze & Wagner, 2007). In the United Kingdom, the recently implemented Flood Re scheme combines public oversight with private provision to ensure the affordability of premiums for high-risk households, particularly in light of increasing flood risks (Surminski, 2014). The United States manages flood risk primarily through the National Flood Insurance Program (NFIP), a federally administered system facing chronic financial deficits and repeated calls for reform (Kousky, 2018). The Netherlands presents a unique case, where private insurance against flood damage is virtually nonexistent, and flood risk is treated as a public responsibility managed through structural mitigation and state intervention (Botzen & van den Bergh, 2008). These diverse models reflect varying national approaches to risk-sharing, insurance affordability, and state intervention. Against this backdrop, the Spanish model exhibits distinctive institutional and financial characteristics that merit closer examination.

We have added the following cited references on the previous paragraph:

• Botzen, W. J. W., & Van den Bergh, J. C. J. M. (2008). Insurance against climate change and flooding in the Netherlands: Present, future, and comparison with other countries. Risk Analysis, 28(2), 413–426. https://doi.org/10.1111/j.1539-6924.2008.01035.x
• Kousky, C. (2018). Financing flood losses: A discussion of the National Flood Insurance Program. Risk Management and Insurance Review, 21(1), 11–32. https://doi.org/10.1111/rmir.12079
• Paudel, Y. (2012). A comparative study of public-private catastrophe insurance systems: Lessons from current practices. The Geneva Papers on Risk and Insurance – Issues and Practice, 37(2), 257–285. https://doi.org/10.1057/gpp.2012.9
• Schwarze, R., & Wagner, G. G. (2007). The political economy of natural disaster insurance: Lessons from the failure of a proposed compulsory insurance scheme in Germany. European Environment, 17(6), 403–415. https://doi.org/10.1002/eet.457
• Surminski, S. (2014). Flood insurance in England – an assessment of the current and newly proposed insurance scheme in the context of rising flood risk. International Journal of Disaster Risk Reduction, 7, 1–10. https://doi.org/10.1016/j.ijdrr.2013.12.005

 

Lines 115-117 “However, this could be considered more of a tax than a premium for the coverage of catastrophic risks, since it does not take into consideration the exposure to the risk”: There could be a potential conceptual shift between taxes and insurance. If the discussion conflates these two concepts, it may create confusion about the mechanisms for funding flood risk management and compensation. It would be important to clearly distinguish between the two—insurance as a risk-sharing tool and taxes as a broader public funding method.

We thank the reviewer for highlighting this important conceptual distinction. We fully agree that insurance and taxation are fundamentally different mechanisms—insurance is a tool for risk pooling and transfer based on actuarial principles, whereas taxation is a public financing instrument generally unrelated to individual risk exposure.

Our statement was not intended to conflate the two concepts but rather to critically highlight a limitation in the current system: the mandatory surcharge collected by the CCS is uniform and does not vary with the actual risk level of the insured asset. As such, it lacks the risk-based structure that characterizes traditional insurance pricing. In this sense, we argued that the uniform and compulsory nature of the surcharge makes it functionally closer to a tax than to a risk-adjusted premium.

Additionally, we have added this reflection in the discussion section (page 22):

The implementation of a two-tier insurance system—comprising private coverage for minor risks and a tax-funded public scheme—could, in theory, incentivize greater risk prevention among policyholders with respect to low-impact flood events. However, it could also give rise to moral hazard, as the assurance that all flood-related losses are ultimately covered by the CCS might reduce incentives for preventive action, particularly in areas prone to non-catastrophic flooding.

Nevertheless, with appropriate system design, the first tier could effectively encourage mitigation efforts for manageable risks, while ensuring that extreme events—those beyond the capacity of private insurance mechanisms—remain covered through general solidarity. This second tier would be financed via taxation and administered by the CCS, thereby preserving social protection without undermining individual responsibility.

 

 

 

Lines 173-175: “On the other hand, occasional flood events not considered extraordinary would be covered by the private sector.” It needs a balanced risk management system involving both the public and private sectors. Could you please elaborate on that?

We thank the reviewer for raising this important point. Indeed, the current Spanish system assigns coverage for all flood-related damages—including both catastrophic and more frequent, lower-intensity events—to the public insurance mechanism operated by the Consorcio de Compensación de Seguros (CCS). As we elaborate in the revised manuscript, this comprehensive public coverage does not differentiate between risk levels and, in our view, leads to an inefficient and inequitable allocation of resources.

In response to the reviewer’s comment, we have clarified our argument that a more balanced risk management model should involve a clearer division of responsibilities:

• Non-catastrophic flood events—which are statistically frequent, geographically localized, and more predictable—should be managed through private insurance mechanisms based on actuarial principles and risk mutualization.
• Catastrophic or extreme flood events, which pose systemic risks and overwhelm private insurance capacity, should remain under the coverage of the CCS, funded through solidarity mechanisms such as taxes or universal surcharges.

This approach would enhance the sustainability of the current system, improve risk-based pricing signals, and align Spain more closely with international best practices that combine private insurance tools with public reinsurance or last-resort coverage.

In response to the reviewer’s suggestions, we have incorporated the following explanation into the manuscript on page 6 (Introduction)

The current Spanish flood insurance model assigns coverage for all flood-related damages—regardless of severity—to the public sector through the Consorcio de Compensación de Seguros (CCS). While this centralized approach ensures universal protection, it overlooks the potential for a more efficient and actuarially sound distribution of responsibilities. We propose a differentiated system in which non-catastrophic flood losses, typically localized and statistically more predictable, would be covered by private insurance policies, priced according to risk-based criteria and mutualization principles. In contrast, catastrophic or extreme flood events, which present systemic challenges and exceed private sector capacity, should remain under the purview of the CCS, funded through mechanisms of general solidarity such as taxes or mandatory levies. This public-private balance would incentivize risk reduction, reinforce the role of insurance as a risk management tool, and safeguard the long-term sustainability of the system. Similar hybrid models have been adopted in other countries, such as Flood Re in the UK and the French system involving the Caisse Centrale de Réassurance, providing relevant points of comparison.

 

 

Line 220: Why did annotation 2 appear when annotation 1 was absent?

We have fixed it

Line 221: Figure 1 is not marked.

We have added the number in the Figure 1 (page 8) the following text:

Figure 1: Events coverage by CCS (1996 - 2020). The first figure represents the cost of compensation paid (current €). The second figure represents the number of claims

 

Line 230: Could you summarize the several related formulas of extreme events mentioned in the text? Have there been any new improvements to the formulas?

We acknowledge that the treatment of extreme events has been the subject of ongoing theoretical development in actuarial science and risk modeling. However, the objective of our paper is not to determine the precise statistical threshold or formal actuarial condition under which flood-related losses should be classified as extreme. Rather, our aim is to propose a conceptual framework for the reform of the Spanish flood insurance system—one that distinguishes between two broad types of flood losses:

1. non-extreme events, which could be handled by private insurers using actuarial principles and risk-based premiums, and
2. extreme or catastrophic events, which would remain under public coverage via the CCS.

To support this distinction, we use a general characterization of extremal events as proposed by Embrechts et al. (2013), which emphasizes the disproportionate influence that large claims can have on total aggregate loss. This definition is sufficient for the purpose of our argument, which is to show—through the analysis of historical CCS data—that a small number of days account for a very large share of the total cost of flood claims.

Therefore, while we recognize the value of more sophisticated actuarial modeling tools (e.g., Generalized Extreme Value distributions, Peaks Over Threshold, or copula-based models), these are beyond the scope of this paper. Our contribution lies in offering an operational distinction grounded in observed loss concentration, which can serve as a basis for designing a more balanced and efficient two-tier system of flood risk management in Spain.

We have inserted this clarification to the paper, page 10, section 2.1:

 

 

While the actuarial literature offers a wide array of technical criteria and statistical tools for identifying extreme events—such as the Generalized Extreme Value (GEV) distribution or Peaks Over Threshold (POT) models—this paper does not aim to determine a formal threshold at which flood-related losses should be classified as catastrophic. Instead, we adopt a general conceptualization of extreme events, as proposed by Embrechts et al. (2013), to support a practical framework for policy reform. According to this perspective, extreme events are characterized by their disproportionate influence on total losses, as a small number of days account for a very high share of aggregate compensation. This conceptual approach aligns with our empirical analysis of CCS data, which shows that a handful of events concentrate most of the compensation payouts. Therefore, the distinction we propose—between ordinary and extraordinary flood events—is operational rather than statistical, and serves as a basis for designing a more efficient and balanced public–private flood insurance model.

 

Lines 282-283: Why not include the condition "cost < 1 thousand € "?

This was an error in the table. Indeed, we have recorded claims starting from an amount of €0. Accordingly, we have revised and modified the table (on page 11) to indicate the first interval as including all claims with a compensation amount between €0 and €10.

 

Line 325: This should be Figure 2? The font sizes of the horizontal and vertical axes of the graph are too small, making it difficult to read. Please enhance the clarity of the pictures.

An error was identified in Figure 1 due to the missing figure number. We have corrected this, and the figure now corresponds properly to Figure 2.

We have added to the paper:

Figure 2: Events coverage by CCS (1996 - 2020). The first figure represents the cost of compensation paid (current €). The second figure represents the number of claims

Figure 3: Days with a cumulative cost of above 35 million euros over the period 1996-2020

 

Lines 332 and 342: What is “e” of “Cumulative cost (e)” and “Cost (e)”?

All monetary amounts (“e”) refer to current euros. We have explicitly indicated this in all relevant tables and figures where appears this expression.

 

 

 

Lines 450-451: “Our motivation for writing this paper was the direct relationship that the CCS establishes between flood risk and the risk from extraordinary or catastrophic events.” It seems that the evidence provided for this statement in the text is a bit insufficient.

We agree that the link between flood risk and the notion of catastrophic or extraordinary events deserves further clarification. In the current Spanish system, the CCS includes all flood events under its coverage as extraordinary risks, regardless of the severity, frequency, or geographic scope of the event. In practice, this means that no distinction is made between floods resulting from catastrophic phenomena and those caused by isolated or moderate events.

Rather than relying solely on a legal or definitional argument, our motivation stems from the empirical evidence presented throughout the paper, particularly the disproportionate financial impact of a catastrophic events—such as those in September 2019—on total flood-related compensation over the period 1996–2020. This concentration of losses strongly suggests that a two-tier flood risk management model, as proposed in our paper, offers a promising foundation for reform.

We believe that such a framework would be more aligned with the principles of risk differentiation and actuarial fairness, while also being more sustainable in a context marked by increasing frequency and severity of extreme weather events due to climate change.

 

Lines 538-587: It would be better to add more references to support the argument.

Following the reviewer’s suggestion in point 1 (line 23-24), we have expanded the introduction by including further references related to the functioning of flood insurance systems in different countries, beyond the Spanish context.

We believe that this inclusion in the introduction provides substantial support for the conclusions presented in this study.

Line 623: Shouldn't you add a conclusion?

In the authors’ opinion, a two-tier system as proposed here—differentiating between catastrophic and non-catastrophic events—would be more efficient than the current system employed by the CCS.

In line with the reviewer’s observations and appreciating their valuable comments. Additionally, we have included at the end of the discussion section some of the potential implications of the proposed model (page 23):

A hybrid system could be effective if mechanisms such as cross-subsidies, prevention incentives, and advanced technologies for more accurate risk assessment are properly implemented. These measures can help address affordability challenges without neglecting the need to align premiums with actual risk exposure. In the long term, a transparent and incentive-based system could encourage greater responsibility among policyholders without creating unsustainable economic disparities.

To prevent premiums from becoming excessively high in geographically vulnerable areas, it is proposed that low-risk zones contribute to financing premiums in high-risk areas, thus preserving affordability while maintaining a fair risk-based pricing structure.

The current model is designed to provide universal coverage against extraordinary risks such as floods, earthquakes, or major catastrophes. However, under a new two-tier approach, the CCS budget could be significantly adjusted.

One of the main policy implications of the proposed model lies in the need to reallocate the current CCS budget, concentrating its resources exclusively on the coverage of catastrophic risks and thus freeing financial capacity to respond more effectively to high-impact events. This reorientation would also allow part of the budget to be allocated to funding prevention incentives, such as premium discounts, subsidies for resilient infrastructure, or educational programs in risk management. Moreover, to ensure territorial equity, it would be necessary to establish cross-subsidy mechanisms to prevent highly vulnerable areas from facing unaffordable premiums, thereby preserving equal access to insurance coverage. The financial sustainability of the system would require the establishment of a clear framework for funding and governance, with coordinated participation from both the public and private sectors. Taken together, these measures would help develop a more efficient, equitable, and resilient system in the face of extreme weather events, aligning public policy with objectives of risk prevention and climate adaptation.

 

Lines 631-632: Please add the journal volume, issue and page numbers.

We have added:

Charpentier, A., & Le Maux, B. (2014). Natural catastrophe insurance: When should the government intervene? Journal of Public Economics, 115(C), 1-17.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper effectively introduces the role of the CCS and outlines the proposed two-level system.

1)   Abstract: Please mention about the main findings from the proposed policy implications and avoid general statement like “we analyze the data but no mention about the key findings”

2)   At the end of the introduction (around lines 69–70), add a short paragraph summarizing the intended outcome of the proposed two-level model. In addition, better link the identified problems (moral hazard, lack of incentives) to the proposed solution.

3)   Figure 1: Events coverage by CCS (1996 - 2020) does not indicate what units are used (claims? costs?). Where does the source come from?

4)   Table 1 lacks source citation and explanation of abbreviations or formatting (e.g., what currency is “Indemnities (€)” exactly? Constant or current euros?).

5)   Besides, in the discussion of Table 2 (accumulated daily cost), the text mentions categories (Low, Average, Extreme) but does not explain the logic of the thresholds clearly.

6)   Table 3 & 4: Good quantitative data, but the connection to the discussion is weak.

7)   Figure 1 & 2 is poorly addressed. Perhaps putting a better caption including from which year to which year might help.

8)   L.R: I don’t see L.R on similar or opposing cases. Expand slightly on contrasting models in other countries (e.g., France, USA) to reinforce the novelty and relevance of the Spanish case.

9)   Methodology: “extreme” events remains somewhat subjective. Define the threshold criteria more precisely (e.g., >€20M in daily claims).

10)   Result: Add a histogram or cumulative distribution plot of daily flood claim costs over the 1996–2020 period. This visual would clearly show the heavy-tailed nature of the loss distribution and strengthen the argument for a two-level system.

11)  Overall: Generally good, but some long and complex sentences reduce clarity. 

12)  Strong in arguing the need for shifting from a pure solidarity-based system to a hybrid model but lacking on clarification on how premiums could be adjusted geographically or risk-based without creating affordability issues.

13)  Provide more detail on how preventive actions could be incentivized at Level 1 (private insurance tier).- maybe a conceptual framework at earlier stage can help to summarize the two-level system.

14)  Policy implications: Discuss how the current CCS budget could be reallocated under the new model.

Author Response

Two-level System for Optimal Flood Risk Coverage in Spain

 

ANSWER TO REVIEWER 2 (1st ROUND)

Thank you very much for all the suggestions received by the reviewer which have helped us improve the paper substantially.

Additionally, we have made other modifications to the paper with the intention of improving understanding of it.

 

1)   Abstract: Please mention about the main findings from the proposed policy implications and avoid general statement like “we analyze the data but no mention about the key findings”

In response to the reviewer’s suggestions, we have rewritten the abstract to clearly highlight the observed findings, the proposed model, and its potential implications. The abstract of the paper would be the following:

This study evaluates the current Spanish insurance framework for catastrophic flood risk, administered by the Consorcio de Compensación de Seguros (CCS), based on nationwide loss data reported by the CCS for the period 1996–2020. The analysis of historical claims data enables a clear differentiation between frequent, low-cost events and infrequent, high-impact catastrophes. While the CCS has fulfilled a critical role in post-disaster compensation, the findings highlight the parallel need for ex-ante risk mitigation strategies. The study proposes a more efficient, two-tier risk coverage model. Events whose impacts can be managed through standard insurance mechanisms should be underwritten by private insurers using actuarially fair premiums. In contrast, events with catastrophic implications—due to their scale or financial impact—should be addressed through general solidarity mechanisms, centrally managed by the CCS. Such a risk segmentation would improve the financial sustainability of the system and create fiscal space for prevention-oriented incentives. The current design of the CCS scheme may generate moral hazard, as flood exposure is not explicitly priced into the premium structure. Empirical findings support a shift towards a more transparent, incentive-aligned model that combines collective risk sharing with individual risk responsibility—an essential balance for effective climate adaptation and long-term resilience.

 

2)   At the end of the introduction (around lines 69–70), add a short paragraph summarizing the intended outcome of the proposed two-level model. In addition, better link the identified problems (moral hazard, lack of incentives) to the proposed solution.

In response to the reviewer’s suggestions, we have incorporated the following explanation into the manuscript on page 4 (Introduction)

Under the proposed model, insured parties would have stronger incentives to implement preventive measures aimed at reducing flood-related losses, as these would no longer be fully covered by the CCS. Regarding moral hazard, if policyholders are aware that the CCS will only compensate for severe (catastrophic) events, they may be more inclined to adopt risk-reducing strategies. This is because, under the current system, the perceived need for prevention diminishes when all flood-related claims are covered by the CCS and the premium is not risk-adjusted.

The premium collected by the CCS is calculated as a fixed percentage of the premium paid for general property risk coverage provided by private insurers. Since the CCS is a public entity, policyholders may perceive it as offering low-cost coverage, further reducing their perception of the actual cost of flood risk and undermining preventive behavior.

In contrast, under the proposed framework, if private insurers were responsible for covering minor flood-related damages, they would have a stronger economic incentive to require or promote preventive measures among policyholders. This could lead to a reduction in both the frequency and severity of small losses and, consequently, the cost of premiums associated with this layer of risk. As noted by Carpentier et al. (2021), this system would represent an intermediate model between the current Spanish coverage scheme and the German system.

Additionally, we have added this reflection in the discussion section (page 22):

The implementation of a two-tier insurance system—comprising private coverage for minor risks and a tax-funded public scheme—could, in theory, incentivize greater risk prevention among policyholders with respect to low-impact flood events. However, it could also give rise to moral hazard, as the assurance that all flood-related losses are ultimately covered by the CCS might reduce incentives for preventive action, particularly in areas prone to non-catastrophic flooding.

Nevertheless, with appropriate system design, the first tier could effectively encourage mitigation efforts for manageable risks, while ensuring that extreme events—those beyond the capacity of private insurance mechanisms—remain covered through general solidarity. This second tier would be financed via taxation and administered by the CCS, thereby preserving social protection without undermining individual responsibility.

 

3)   Figure 1: Events coverage by CCS (1996 - 2020) does not indicate what units are used (claims? costs?). Where does the source come from?

The units of claims and cost appear at the section 2), page 7. We can find this text:

The aim of Figure 1 and Table 1 is to show the relative importance of floods among the natural catastrophic events covered by the CCS, both in terms of the number of claims it generates and the compensation paid in current euros.

Anyway, we have added in the Figure 1 (page 8) the following text:

Figure 1: Events coverage by CCS (1996 - 2020). The first figure represents the cost of compensation paid (current €). The second figure represents the number of claims

 

4)   Table 1 lacks source citation and explanation of abbreviations or formatting (e.g., what currency is “Indemnities (€)” exactly? Constant or current euros?).

The units of claims and cost appear at the section 2), page 7. We can find this text:

The aim of Figure 1 and Table 1 is to show the relative importance of floods among the natural catastrophic events covered by the CCS, both in terms of the number of claims it generates and the compensation paid in current euros.

Anyway, we have added in the Table 1 (page 8) the following text:

Table 1: Claims by type of event (analysis period 1996-2020)

The table shows the relative importance of floods among the natural catastrophic events covered by the CCS, both in terms of the number of claims it generates and the compensation paid (current euros).

 

5)   Besides, in the discussion of Table 2 (accumulated daily cost), the text mentions categories (Low, Average, Extreme) but does not explain the logic of the thresholds clearly.

In relation to the suggestion received at this point, we must specify that the classification applied is based on an intuitive rather than a technical criterion. As we are working with aggregated data rather than individual-level information, no scientific basis has been used to define precise thresholds. Establishing such thresholds would, in our view, require a separate and more detailed study. Our aim is not to determine how these limits should be defined, but to illustrate the variability in loss severity.

Accordingly, we propose a preliminary classification based on the daily cost of compensation:

• Low: less than €0.5 million per day
• Medium: between €0.5 million and €20 million per day
• Extreme: above €20 million per day

As an illustrative example, we classified low-loss days as those falling below the 90th percentile, medium-loss days as those between the 90th and 99th percentiles, and extreme-loss days as those exceeding the 99th percentile. However, it is important to note—as emphasized in the conclusions—that this categorization is intended for analytical purposes only and does not constitute a definitive risk threshold (the graph is shown below)

However, as we have indicated at the end of the conclusions’ section (pag 22), future research could apply a technical actuarial model to establish the cutoff between what may or may not be considered an extraordinary natural event.

 

We have decided to insert this graph and clarification on the paper, page 12, section 2.2:

As an illustrative example, we classified low-loss days as those falling below the 90th percentile, medium-loss days as those between the 90th and 99th percentiles, and extreme-loss days as those exceeding the 99th percentile. However, it is important to note—as emphasized in the conclusions—that this categorization is intended for analytical purposes only and does not constitute a definitive risk threshold.

Graph 1. Daily total nominal losses (current euros)

 

The authors believe that with this explanation, we are also addressing the concerns raised in points 9) and 10) of the reviewer’s comments.

 

6)   Table 3 & 4: Good quantitative data, but the connection to the discussion is weak.

We agree with the reviewer’s suggestion, we have added on page 13.

In Table 3 we clearly see the effect on the total cost of the extreme events. This table presents the cumulative relationship between the number of days since the claim occurrence and the total accumulated cost, as well as the cumulative percentage of claims.

Following Table 4, on page 13, we have added on page 15:

As we have explained before, Table 3 indicates that a large proportion of the total cost and claims are concentrated within a short period following the occurrence of the event (over 35% within 2 days and nearly 50% within 5 days). This is consistent with Table 4, where specific days show a high number of claims and elevated costs—particularly on September 12 and 13—likely reflecting the materialization of damages and the initiation of the claims process.

As observed in Table 4, September 13, with the highest number of claims and associated costs, may represent one of the “key” days driving the rapid accumulation seen in Table 3. This suggests that discrete events with high frequency and intensity significantly impact the total cost and initial claims burden. It also indicates that risk is not evenly distributed over time but rather concentrated in specific moments characterized by extreme adverse conditions.

Based on this, it can be concluded that since a large share of claims is concentrated in brief, severe events, the implementation of preventive measures and rapid response during these periods is critical to mitigating damages and costs.

 

7)   Figure 1 & 2 is poorly addressed. Perhaps putting a better caption including from which year to which year might help.

We have added to the figures the analysis period that we used to make them:

Figure 2: Events coverage by CCS (1996 - 2020). The first figure represents the cost of compensation paid (current €). The second figure represents the number of claims

Figure 3: Days with a cumulative cost of above 35 million euros over the period 1996-2020

 

8)   L.R: I don’t see L.R on similar or opposing cases. Expand slightly on contrasting models in other countries (e.g., France, USA) to reinforce the novelty and relevance of the Spanish case.

We thank the reviewer for this valuable suggestion. In response, we have expanded the Literature Review section to include a comparative discussion of flood insurance systems in several countries, including France, Germany, the United Kingdom, the United States, and the Netherlands. This addition serves to contextualize the Spanish model within a broader spectrum of institutional arrangements and highlights its unique features.

We have added the revised paragraph on page 4 (Introduction) of the manuscript and reads as follows:

Comparative research on flood insurance systems reveals a wide diversity of institutional arrangements across countries. France operates a mixed system in which private insurers provide coverage backed by a public reinsurance scheme through the Caisse Centrale de Réassurance (Paudel, 2012). In Germany, flood insurance is voluntary and offered by private companies, but penetration rates remain low due to limited risk awareness and affordability challenges (Schwarze & Wagner, 2007). In the United Kingdom, the recently implemented Flood Re scheme combines public oversight with private provision to ensure the affordability of premiums for high-risk households, particularly in light of increasing flood risks (Surminski, 2014). The United States manages flood risk primarily through the National Flood Insurance Program (NFIP), a federally administered system facing chronic financial deficits and repeated calls for reform (Kousky, 2018). The Netherlands presents a unique case, where private insurance against flood damage is virtually nonexistent, and flood risk is treated as a public responsibility managed through structural mitigation and state intervention (Botzen & van den Bergh, 2008). These diverse models reflect varying national approaches to risk-sharing, insurance affordability, and state intervention. Against this backdrop, the Spanish model exhibits distinctive institutional and financial characteristics that merit closer examination.

We have added the following cited references on the previous paragraph:

• Botzen, W. J. W., & Van den Bergh, J. C. J. M. (2008). Insurance against climate change and flooding in the Netherlands: Present, future, and comparison with other countries. Risk Analysis, 28(2), 413–426. https://doi.org/10.1111/j.1539-6924.2008.01035.x
• Kousky, C. (2018). Financing flood losses: A discussion of the National Flood Insurance Program. Risk Management and Insurance Review, 21(1), 11–32. https://doi.org/10.1111/rmir.12079
• Paudel, Y. (2012). A comparative study of public-private catastrophe insurance systems: Lessons from current practices. The Geneva Papers on Risk and Insurance – Issues and Practice, 37(2), 257–285. https://doi.org/10.1057/gpp.2012.9
• Schwarze, R., & Wagner, G. G. (2007). The political economy of natural disaster insurance: Lessons from the failure of a proposed compulsory insurance scheme in Germany. European Environment, 17(6), 403–415. https://doi.org/10.1002/eet.457
• Surminski, S. (2014). Flood insurance in England – an assessment of the current and newly proposed insurance scheme in the context of rising flood risk. International Journal of Disaster Risk Reduction, 7, 1–10. https://doi.org/10.1016/j.ijdrr.2013.12.005

 

9)   Methodology: “extreme” events remains somewhat subjective. Define the threshold criteria more precisely (e.g., >€20M in daily claims).

This suggestion has been taken into account and is specifically addressed in our response to comment 5 of the review.

 

10)   Result: Add a histogram or cumulative distribution plot of daily flood claim costs over the 1996–2020 period. This visual would clearly show the heavy-tailed nature of the loss distribution and strengthen the argument for a two-level system.

This suggestion has been taken into account and is specifically addressed in our response to comment 5 of the review.

 

11)  Overall: Generally good, but some long and complex sentences reduce clarity. 

Following the reviewer's suggestions, we will try to explain some of the issues we face in a more concise way.

 

12)  Strong in arguing the need for shifting from a pure solidarity-based system to a hybrid model but lacking on clarification on how premiums could be adjusted geographically or risk-based without creating affordability issues.

In response to the reviewer’s suggestions, we have incorporated the following explanation into the manuscript on page 20 (section 3.2)

This system aims to distribute risk more equitably across the population and, indirectly, to prevent insurance premiums from becoming prohibitively high for those living in more vulnerable areas. However, this objective may conflict with the need to adjust premiums according to geographical risk, which can result in significantly higher costs for policyholders in high-risk zones, such as flood-prone regions.

To address this, a more balanced risk-sharing mechanism is proposed, whereby part of the revenues generated from premiums in low-risk areas would be allocated to subsidize premiums in high-risk zones. This approach would help mitigate disparities in coverage access, ensuring that insurance remains affordable even in the most vulnerable regions, while still reflecting the underlying risk levels.

Another option would be to implement parametric insurance models, in which the policyholder pays a premium based on predefined parameters—such as rainfall levels or water depth in specific areas. These models can offer a more flexible and equitable solution for high-risk regions, providing timely payouts without significantly increasing costs.

 

13)  Provide more detail on how preventive actions could be incentivized at Level 1 (private insurance tier).- maybe a conceptual framework at earlier stage can help to summarize the two-level system.

Taking into account the reviewer’s suggestions, we have incorporated the following explanation into the manuscript on page 20 ( section 3.2)

A key strategy in the implementation of this system is to reward policyholders who undertake significant preventive measures, such as installing drainage systems, elevating buildings, or developing flood-resilient infrastructure.  Private insurers could offer lower premiums to individuals who implement such measures, while the public reinsurance entity (e.g., the Consorcio) could apply discounts or bonuses for those who demonstrate investment in risk reduction. Private insurers could also establish a certification program for properties that meet specific flood protection standards, signaling that a property has implemented high-quality prevention and mitigation strategies.

Prevention incentives at the first level should combine direct financial rewards (e.g., discounts, bonuses), certification of preventive measures, and continuous education. Private insurers have the opportunity to design tailored programs that reward policyholders for enhancing the resilience of their properties against flooding. This approach not only promotes proactive risk management but also contributes to a reduction in claims, potentially leading to lower overall premiums across the system in the long term.

Additionally, we have added this reflection in the discussion section (page 23):

A hybrid system could be effective if mechanisms such as cross-subsidies, prevention incentives, and advanced technologies for more accurate risk assessment are properly implemented. These measures can help address affordability challenges without neglecting the need to align premiums with actual risk exposure. In the long term, a transparent and incentive-based system could encourage greater responsibility among policyholders without creating unsustainable economic disparities.

To prevent premiums from becoming excessively high in geographically vulnerable areas, it is proposed that low-risk zones contribute to financing premiums in high-risk areas, thus preserving affordability while maintaining a fair risk-based pricing structure.

 

 

14)  Policy implications: Discuss how the current CCS budget could be reallocated under the new model.

In line with the reviewer’s observations and appreciating their valuable comments, we have incorporated the following reflection at the end of the discussion section (page 23):

The current model is designed to provide universal coverage against extraordinary risks such as floods, earthquakes, or major catastrophes. However, under a new two-tier approach, the CCS (Consorcio de Compensación de Seguros) budget could be significantly adjusted.

One of the main policy implications of the proposed model lies in the need to reallocate the current CCS budget, concentrating its resources exclusively on the coverage of catastrophic risks and thus freeing financial capacity to respond more effectively to high-impact events. This reorientation would also allow part of the budget to be allocated to funding prevention incentives, such as premium discounts, subsidies for resilient infrastructure, or educational programs in risk management. Moreover, to ensure territorial equity, it would be necessary to establish cross-subsidy mechanisms to prevent highly vulnerable areas from facing unaffordable premiums, thereby preserving equal access to insurance coverage. The financial sustainability of the system would require the establishment of a clear framework for funding and governance, with coordinated participation from both the public and private sectors. Taken together, these measures would help develop a more efficient, equitable, and resilient system in the face of extreme weather events, aligning public policy with objectives of risk prevention and climate adaptation.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have made all the necessary changes.