The political economy of premium subsidies: Searching for better impact and design. Insights for Sovereign Climate and Disaster Risk Finance and Insurance

Scott, Zoe; Panwar, Vikrant; Weingärtner, Lena; Wilkinson, Emily Research Report The political economy of premium subsidies: Searching for better impact and design. Insights for Sovereign Climate and Disaster Risk Finance and Insurance ODI Report Provided in Cooperation with: ODI Global, London Suggested Citation: Scott, Zoe; Panwar, Vikrant; Weingärtner, Lena; Wilkinson, Emily (2022) : The political economy of premium subsidies: Searching for better impact and design. Insights for Sovereign Climate and Disaster Risk Finance and Insurance, ODI Report, Overseas Development Institute (ODI), London This Version is available at: https://hdl.handle.net/10419/280297 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ Report The political economy of premium subsidies: searching for better impact and design Insights for Sovereign Climate and Disaster Risk Finance and Insurance Zoe Scott, Vikrant Panwar, Lena Weingärtner and Emily Wilkinson December 2022 Disclaimer: This advisory report has received financial support from Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH through the InsuResilience Global Partnership (IGP) secretariat. The views expressed do not necessarily reflect the official policies of GIZ or IGP. Readers are encouraged to reproduce material for their own publications, as long as they are not being sold commercially. ODI requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views presented in this paper are those of the authors and do not necessarily represent the views of ODI or our partners. This work is licensed under CC BY-NC-ND 4.0. How to cite: Scott, Z., Panwar, V., Weingärtner, L. and Wilkinson, E. (2022) The political economy of premium subsidies: searching for better impact and design. Report. ODI and InsuResilience Global Partnership. London: ODI (www.odi.org) Acknowledgements The authors would like to thank members of the advisory working group, Annette Detken, Daniel Clarke, Nicola Ranger, Olivier Mahul and colleagues at the IGP secretariat particularly Daniel Stadtmüller, Janek Töpper and Kay Tuschen for their insights, inputs and guidance. We are grateful to the interviewees who participated in the key informant interviews (KIIs) on the political economy of premium subsidies and provided critical inputs for the development this advisory report. At ODI, we are grateful to Silvia Harvey. This study has received financial support from Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH on behalf of the InsuResilience Global Partnership (IGP) secretariat. The views and findings presented here are of the authors’ and do not represent the views of GIZ or IGP. About this publication This advisory report is an output of the Global Risks and Resilience Programme (GRR) at ODI. GRR provides rigorous analysis of multiple interconnected risks, interrogates narratives and risk perceptions, and uses this evidence to recommend tailored solutions for the management of systemic risks in development, humanitarian, climate adaptation and disaster risk management policies and actions. About the authors Zoe Scott An independent consultant, specialising in disaster risk finance and management. She has over 20 years’ experience and has worked on disaster risk in many low and middle-income countries, providing advice and technical services to governments, donor agencies, development banks and international organisations. Her work aims to challenge and reform how the world plans and pays for disasters, promoting evidence-based policy-making, robust evaluation and insightful research. Vikrant Panwar Senior Researcher with ODI’s Global Risks and Resilience programme. An economist with specialisation of macro-fiscal disaster and climate risks and impacts. He conducts research around disaster and climate risk financing, macro-fiscal climate and disaster impacts at the sovereign and sub-sovereign levels. Email: v[email protected].uk ORCID: https://orcid.org/0000-0003-1259-9789 Lena Weingärtner Research Associate with ODI’s Global Risks and Resilience programme, working primarily on disaster risk management and financing. She has experience in researching and informing financial instruments and delivery mechanisms at micro, meso and sovereign levels. This includes approaches – such as parametric insurance, adaptive social protection, or anticipatory action. Emily Wilkinson Senior Research Fellow with ODI’s Global Risks and Resilience Programme, specialising in resilience in Small-Island Developing States. Emily has 25 years’ experience as a researcher, analyst, journalist, lecturer and adviser, identifying critical entry points and opportunities for governments and communities to mange risks, access finance and build resilience to external shocks. She is Chief Scientific Adviser to the Climate Resilience Agency for Dominica (CREAD); and is Director, Resilient and Sustainable Islands (RESI). Contents Acronyms / iv Executive summary / 1 1 Introduction / 3 1.1 Background / 3 1.2 Methodology / 3 2 Factors that influence insurance uptake / 5 2.1 Affordability of premiums and fiscal space / 7 2.2 Understanding and technical capacity / 7 2.3 Availability of alternatives / 8 2.4 Perceptions of reliability / 8 2.5 Relevance of products / 9 2.6 Government processes and bureaucracy / 9 2.7 Political disincentives / 10 2.8 Desire to effectively finance risks and build resilience / 10 2.9 Regional dynamics / 11 2.10 Lesser factors / 11 3 Premium support versus capital support / 13 4 The impact of subsidies / 16 5 Allocating subsidies / 19 6 Improving the design of subsidies / 24 7 Conclusions / 28 References / 30 Acronyms ADB Asian Development Bank ADF African Development Fund, AfDB ADRiFi African Disaster Risk Financing Programme AfDB African Development Bank ARC African Risk Capacity CATDDO Catastrophe Deferred Drawdown Option CCRIF Caribbean Catastrophe Risk Insurance Facility CDRFI Climate and Disaster Risk Finance and Insurance CEO Chief Executive Officer DRF disaster risk finance FCDO Foreign, Commonwealth and Development Office, UK IDA International Development Association, World Bank IGP InsuResilience Global Partnership KII key informant interview MoF Ministry of Finance M&E monitoring and evaluation PCRIC Pacific Catastrophe Risk Insurance Company PCS premium and capital support PEA political economy analysis PIC Pacific Island country TWG technical working group UK United Kingdom 1ODI Report Executive summary Risk pools offering climate-related insurance have been operating for several years in Africa, the Caribbean and the Pacific. All have benefited from donor capitalisation and subsidisation of premiums in the past. With growing climate risks across all these regions, limited fiscal space in lowand middle-income countries, and an over-burdened humanitarian caseload, there is increasing interest in using donor subsidies to grow the risk pools and offer more reliable, more cost-effective and faster support to disasteraffected communities. This report investigates the political economy of country decision-making in relation to sovereignlevel climate and disaster risk finance and insurance, and the role of premium and capital support in these decisions. It also analyses the political economy of donor decisions in relation to the provision of premium and capital support. From the analysis, affordability emerged as the main barrier to insurance uptake, but it is one among many factors. The most significant barriers, after affordability, were lack of understanding and technical capacity; availability of alternatives; and perceptions of reliability, among others. The balance of which factors are most important will vary in each country, affecting the impact of subsidies. Experience has demonstrated that subsidies are not always attractive enough to incentivise insurance uptake, as other barriers may be more important to a country than affordability. The design of subsidies has also proved a barrier in the past – particularly that subsidies were required to go to ‘new’ (previously unsubsidised) countries, or cover new hazards, and that they sometimes required multi-year commitments from countries to co-finance premiums. Stakeholders argued strongly that recipient countries should be much more involved in the design of subsidies, so that donor objectives can be carefully aligned with country perspectives and priorities. They also highlighted that information about the availability of subsidies would need to be communicated to countries much earlier in the insurance policy subscription cycle, as a lack of clarity about the extent of available support has previously prevented some countries from effectively negotiating policies and obtaining insurance coverage. Stakeholders were also adamant that donors should focus on providing grant funding for premium subsidies rather than capital support at this stage. Whilst investment loans for capital support are generally more available and in larger quantities than grant finance within key donor agencies, premium subsidies are now the priority in order to ensure the growth and sustainability of the risk pools. Subsidies can help to grow risk pool membership, but there are reasons why existing members could also be considered for premium support. Many countries who receive subsidies state that they would need to reduce coverage or drop out of the risk pool should the subsidy stop. Subsidies to countries who have been loyal risk pool members, paying premiums out of their own budgets, could ‘reward’ strong risk ownership and ‘good performance’. Furthermore, if governments use the subsidy to expand their policies rather than replace their own costs, it could lead to increased coverage. 2ODI Report Premium subsidies are considered to have few negative impacts. In some cases, subsidy design includes exit strategies to address concerns around dependency and sustainability. There is little evidence that premium subsidies contribute to moral hazard or undermine risk reduction and preparedness. Subsidy allocation is complex and has used different criteria in the past, supporting varying objectives for the risk pools. Actors consulted during the study had diverging views on the importance of different factors in allocating premium subsidies. Overall, ‘proportion of vulnerable population in total population’ and ‘climate and disaster risk profile’ were viewed as the most important factors, particularly by representatives from across the risk pools. ‘Country income level’ and ‘prior risk reduction actions/policies’ were also highly valued, though not universally. However, donors seemed to prioritise and value a range of other factors in making decisions about subsidy allocations, particularly that the product should be high quality; that there is a plausible exit strategy; that the subsidy is for a new country or product; and that the country is a priority for them. Donors also had exclusion criteria; most significantly, they would exclude countries that were not ODA-eligible or were subject to sanctions. Subsidy design can support a donor’s objectives, but there is little consensus on the appropriate size and duration of premium subsidies. Most stakeholders consulted felt that it was important, at least after the first year, that recipient countries made some contribution to ensure buy-in. Views on the appropriate duraction of subsidies ranged from two years to very long term – as an alternative to humanitarian aid, and to help address loss and damage in support of climate justice. The majority of interviewees argued that support should be multi-year, although it was acknowledged that this can be unpopular with some governments as it typically requires a commitment for increasing levels of co-financing. 9ODI Report where a country had had expectations of a payout, even though the disaster arose from a different hazard than the one covered by their policy, or the threshold for a payout had evidently not been met. Regardless, these situations where a payout was expected but did not materialise appear to be very damaging to risk pools’ reputations. Conversely, a belief in the reliability of the product can act as a catalyst for insurance uptake. Some interviewees clearly valued this assurance, particularly in contrast to humanitarian aid. As one PCRIC client commented: ‘I think the most comforting thing about a PCRIC payout is, I know a payout is coming.’ Malawi is an example of a country which dropped out of the ARC risk pool following a basis risk event, but which has recently been persuaded to re-join as their faith in the risk model improved. They were able to groundtruth the satellite data the ARC model used against information collected themselves and, having found that the two were well aligned, and having accessed premium subsidies, they started purchasing insurance again. 2.5 Relevance of products Understandably, uptake will be higher if insurance products are available for hazards that are viewed as a priority risk in a country. For example, ARC only have a limited set of products, focused on drought, with tropical cyclone available in a few cases. However, many countries have expressed a desire for flood coverage as well as, or instead of, drought (OPM, forthcoming). There is also evidence that countries value flexibility of products, so that they can be suited to specific country dynamics and concerns. 2.6 Government processes and bureaucracy Both the decision-making process and the flow of funding can be slow and complex. Even once a decision to purchase insurance has been made, the actual flow of money through government systems to pay for the premium can be just as arduous as the decision-making process. As mentioned above, many different government actors, spread across departments, are involved in the purchase of an insurance policy. Turnover of personnel is a significant problem in many governments, from the ministerial to technical level, which can also delay progress as new people have to be brought up to speed. Premium subsidies can help with generating political support for insurance, but it will still be necessary to engage with government bureaucracy to get a policy in place. This is particularly the case for ARC, who require a number of preparatory steps to be complete, including issuing a Certificate of Good Standing and the production of a Contingency Plan, which can take years. Several interviewees mentioned timing as a common problem, and emphasised that conversations need to ‘start early’, because of the set budget cycles on which governments operate. Many complained that the process took too long and was made worse when subsidies were involved, as the timelines of donors, risk pools and governments were not always aligned, and it was often not clear until late in the process what subsidy was being offered. This issue came across more strongly in interviews than it is reflected in the literature. One interviewee stated: This year we are not insured because we received the policy information from ARC too late, and the time available to review the 10 ODI Report contractual and technical arrangements and to implement the premium payment was too short. There are often very short timeframes for subscription and delays in the administration…. Usually, the information comes around April, and we have to sign and make the payment by July, but this year we only received the information in June, so the decision to subscribe was taken, but the payment didn’t go through quickly enough. ARC apparently had some delays on their end with the reinsurer. However, as well as creating barriers for insurance uptake, excessive bureaucracy can also create an incentive for insurance which, in return, provides very quick payouts. For example, a recent cost–benefit analysis of ARC identifies this as a major motivation for one African country with a large economy – a large part of the appeal of insurance being that a payout can arrive in a government bank account within hours of a disaster, whereas arranging internal budget reallocations would take much longer and be more onerous (Lee and Rusconi, forthcoming). 2.7 Political disincentives Risk management tends to be a ‘back-room’ activity that does not attract as much media attention for politicians as disaster response does, and which carries a risk of being perceived as ‘wasted’ expenditure if the risk does not materialise. These are serious concerns for a politician, particularly when operating in a resource-constrained environment, and they create incentives to wait and see if a disaster happens, rather than pro-actively purchase insurance. This is exacerbated as insurance is most cost-effective when used for low-frequency events, but politicians have short timeframes – insurance provides no certainty of a benefit within the political timeframe of the leaders making the purchasing decision. Politicians typically prefer high-frequency coverage to increase the likelihood of a payout, but this reduces the overall value proposition of insurance (OPM, forthcoming). Governments have competing priorities, and money for premiums could be spent addressing pressing needs such as health or education, where a new hospital or school can be a useful way of gaining popular support. In addition, government priorities are constantly changing, and a change in leadership can reverse spending priorities and de-prioritise insurance. Elections create a particular moment of vulnerability, as priorities can shift radically and displace funding earmarked for premiums. This has been noted for Mauritania, Senegal, Kenya and Fiji (Martinez Diaz et al., 2019; e-Pact, 2017; interviews). As one interviewee described the situation: given upcoming elections, should the government buy insurance which benefits an insurance company based outside the country, or would it be better to take that money and use it to support local businesses? 2.8 Desire to effectively finance risks and build resilience Governments are looking for ways to better manage their risks and build resilience, increasingly using risk transfer alongside a combination of financial instruments. Again, this is infrequently mentioned in the literature but was communicated in several interviews – possibly, it is a factor that has become more important in recent years as countries’ capacities in relation to disaster risk finance have grown. Notwithstanding the powerful political disincentives against purchasing insurance noted above, several country representatives described a desire to improve their country’s resilience by having robust risk financing instruments in place. 11 ODI Report A government official from an insurancepurchasing PIC stated: ‘There’s nobody [within government] who says we shouldn’t be investing in ourselves; nobody says that’ – adding that their country sought to be a role model in the region for resilience and self-reliance. In addition, one interviewee noted that buying ARC insurance lso brought wider resilience benefits, including ‘effectiveness gains and gains in transparency and accountability’ through the contingency planning process. Some countries viewed insurance as a necessary instrument that complemented other risk financing approaches they were using. For example, one African government representative mentioned how it fitted into their wider Disaster Risk Financing strategy, developed in collaboration with the World Bank, and enabled them to transfer risk in order to better protect the development budget. Similarly, one representative from the Pacific mentioned how insurance provided them with ‘another option’ for post-disaster finance, while another stated that insurance fitted into their layered approach to DRF: ‘We generally look at the products that are available to us and try to build a layered approach to financing. So each financing instrument complements the other. We look at how much money we have got, for example in trust fund or in surplus, over budget which can be invested in building financial resilience by purchasing insurance.’ 2.9 Regional dynamics Some interviewees expressed a desire to support the risk pools because they are regionally-led initiatives. They valued the risk pools as regionally owned, with regionallybased staff, and wanted to see them succeed. For example, studies have shown that African governments value ARC’s status as an initiative of the African Union (e-Pact, 2017). Similarly, multiple interviewees from PICs mentioned the importance of having people from the region as senior staff and Board members. In particular the new PCRIC Chief Executive Officer (CEO) is an Islander and is credited with having driven new levels of country engagement, created higher levels of trust and improved regional understanding. Some interviewees also mentioned a level of ‘regional peer pressure’ encouraging uptake, as countries saw their neighbours buying insurance and, sometimes, benefiting from payouts. Some countries were viewed as being particularly important in this regard; for example, one interviewee argued that if Fiji were to join the PCRIC risk pool it would be particularly influential with other PICs, given their size and strategic importance in the region. However, regional dynamics may also undermine insurance uptake. Some studies have mentioned regional politics being a hindrance; for example, ARC being viewed as primarily focused on West Africa and, therefore, of less interest to countries from other regions (Martinez-Diaz et al., 2019; OPM, 2022). 2.10 Lesser factors There are a wide range of additional factors that appear to influence insurance uptake, albeit to a lesser extent than the factors listed above. All the factors listed above were mentioned by interviewees or in the literature at least five independent times. However, a handful of other motivations were mentioned by two or more independent sources. They are presented below, as the literature in this area is very limited; 12 ODI Report note that a different sample of interviewees may have given greater prominence to a different set of factors.3 • Payouts – experiencing a payout makes a country more likely to buy insurance in future (OPM, forthcoming), just as not getting a payout appears to increase the chance of a country dropping out of a risk pool. • Technical support from the risk pools – a recent evaluation states that ARC capacity building is universally valued by member countries (OPM, forthcoming). Similarly, PCRIC’s country engagement and participation in a number of regional working groups was viewed as beneficial. • Recent experience of a high-impact disaster – for example, one interviewee reflected on how CCRIF was born in the aftermath of a major hurricane, with sixteen countries immediately willing to join. • Climate justice – this was mentioned as a barrier to uptake, particularly in the Pacific where loss and damage debates resonate strongly with governments, making them less inclined to use their own resources to pay premiums (Martinez-Diaz at al., 2019). • Need for quick liquidity – some countries noted that their alternative sources of postdisaster finance were primarily development partner funding or budget reallocations, both of which are very slow. Insurance offers a quick payout that can fill the gap while other resources are being mobilised. Interestingly, speed of payouts did not 3 In particular, this research conducted a very limited number of interviews with people based in the Caribbean, instead focusing on Africa and the Pacific. appear to be a motivating factor to the same extent in the Pacific as in ARC-participating countries, possibly because PICs have smaller governments and are therefore able to mobilise budgetary resources more quickly themselves. 13 ODI Report 3 Premium support versus capital support This section provides analysis of donor decision-making and incentives in relation to providing capital support and premium subsidies, and reflects views from across all types of stakeholder on how this could and should evolve in future. It therefore provides a basis for understanding what drives levels of donor support for premium and capital support and briefly explores stakeholder views around future prioritisation of premium subsidies. Key points: • Donors are predominantly driven by the availability of loan versus grant finance within their institutions when deciding whether to provide capital support or premium subsidies. This access is driven by a range of factors. • Generally, investment loans, which can be used for capital support, are more easily available and in larger quantities than grant funds, which are used for premium subsidies. • Other factors that shape donor decisions about whether to provide capital or premium support include the stage of the risk pool’s development; existing levels of capitalisation; and demand from countries and the risk pools. • Stakeholders expect donor support in this area to grow and unanimously supported more premium subsidies. • There is strong evidence that premium subsidies should be prioritised over capital support at the current time. Donor decisions on whether to support risk pools through the provision of capital support or premium subsidies are complex and multi-faceted. Donor agency staff face numerous constraints in how they provide support to risk pools – unfortunately, they are not always at liberty to choose between using funds for premium subsidies or capital support. The reality is much messier and more complicated, with various operational and political factors driving the decision. This creates a risk that a good balance between capital and premium support is not being achieved and donors would be well advised to review and sense-check their approaches. The primary consideration in whether to give capital or premium support is the availability of loan versus grant funding within the donor institution. Most donors now have an investment instrument as well as grant funding streams. These are not interchangeable – a set amount is available for each in a given spending cycle, and officials have to pitch for it based on a combination of factors (for example, their assessment of need and likely allocation to their department), often using technical as well as political arguments. Although capital support has been provided in the past using grant funds (for example, to both CCRIF and PCRIC), for ARC it was provided as loans, and this trend of using loans for capital support is expected to continue. By contrast, best practice 14 ODI Report suggests that premium subsidies should be paid for using grant funding, not loans, given that premiums are not designed to generate future returns that can be used to service the debt and hence raise questions around debt sustainability (Martinez-Diaz et al., 2019). Grant funding, which typically pays for premium subsidies, is much less available than loans. The main donors supporting the risk pools have been the United Kingdom (UK) and Germany, and grant funding is much harder to access for both. Investment loans, which ultimately have to be paid back, understandably carry more benefits for a donor agency and therefore tend to be more available than grant finance, which does not have to be repaid. In general, therefore, larger amounts of capital investment are available compared to grant funds. This has not always been the case; it is part of a wider trend in development finance. For example, in the UK there was a big shift to capital investment from around 2010, meaning that much larger amounts of finance are now available as capital from the UK than as grant funds. This can be clearly seen in the UK’s capitalisation of the Caribbean Catastrophe Risk Insurance Facility (CCRIF) with £3 million in grant funding, compared to ARC who were later capitalised with a £33 million long-term loan. Access to loan or grant finance is also driven by operational issues and personal connections. There can be windows of opportunity when grant funding or capital becomes available within a donor agency, and officials can try to access it at that point – for example, at year end, if there has been an unexpected underspend. The COVID-19 pandemic provided another such opportunity, when grant finance suddenly became available within KfW Development Bank as emergency support was activated. Working relationships can also shape access to the different types of finance; for example, the department you sit in, or the connections you have, can mean you get easier access to one type of finance over the other. One interviewee also noted that the particular type of finance you have previously accessed is likely to stay more accessible to you in future, as you already have the necessary relationships, have built trust and understanding, and are familiar with the process. Other factors driving donor decision-making, beyond availability, include the following: • The stage of development of the risk pool. Capital support is more likely to be given in the early stages of a risk pool, when capital is needed to establish the initiative. Equally, if there has been a period of rapid growth, or one is anticipated, then more capital may be required to enable more risk to be underwritten and ensure the financial stability of the pool. • Existing levels of capitalisation and subsidisation. One donor representative discussed being mindful of not wanting to overcapitalise a risk pool because of the opportunity costs – this would not be a cost-effective use of funds. For example, the UK’s original business case includes £90 million to capitalise ARC, but this has not all been given as there has not yet been a clear need for the full amount. All donor stakeholders recognised that there was a shortage of premium subsidies available at the moment. • Demand. Donor representatives spoke of seeking to respond to countries and risk pools’ preferences and requests, where these were supported with evidence of need (in the case of additional capitalisation). One interviewee 15 ODI Report noted that climate justice arguments made within global forums like the G7 summits and COP26 have helped create political support and generated increased calls for premium subsidies. Both capital and premium support are expected to increase in the future. Donors anticipate giving more to the risk pools, both as capital and premium support. For example, at COP26, the UK announced that both investment and grant financing would be made available for the risk pools in the coming years. Interviewees unanimously expect funding to the risk pools to increase, and several different donors were mentioned as likely to start support imminently, including non-traditional donors. Several interviewees mentioned a hope and desire that the global climate funds will start to provide premium subsidies – this is seen as a route to larger amounts of funding over the long term. However, alongside the optimism, some actors mentioned that this will all depend on there being sustained good performance amongst the risk pools and noted that a high-profile basis risk event, for example, could reduce donor appetite. 4 Some interviewees and literature discussed the case of ARC, where the capital has been provided as repayable loan that can be recalled with three months’ notice – asserting that this arrangement results in its being protected more than would be expected, leading to higher levels of reinsurance being purchased so that the capital is not put at risk. This therefore generates higher reinsurance costs for the risk pool, which potentially undermines some of the anticipated benefits of the capital. This demonstrates that it is important to get the terms of donor capital investment right, not just the overall amount. There is strong evidence that premium subsidies should be prioritised over capital support at the current time. Capital investment can help sustainability, enable rapid payouts and, indirectly, lower premiums. However, both the literature and many interviewees argued that the risk pools are already well capitalised and were designed not to need more capital unless they grew significantly, which is unlikely to happen without more premium subsidies. This is particularly the case for ARC and PCRIC, both of which have been the subject of recent analytical studies that argue that they are already soundly capitalised and could, in fact, support a good deal more in sales without requiring additional capital (Lee and Rusconi, forthcoming; OPM, 2022).4 Most interviewees argued that donors should focus on providing premium support; some stated that both were important (though capital was often argued for as an indirect way of reducing subsidies) and premium subsidies were generally acknowledged as most needed at this stage. Nobody supported the idea of capital support without premium subsidies. This suggests that donors should proceed with premium subsidies as their priority, and only provide capital support after careful analysis of its likely cost-effectiveness, and with consideration of the lending terms and what incentives and additional costs they may create. 16 ODI Report 4 The impact of subsidies This section provides analysis from across the literature review and the key informant interviews on the actual impacts of subsidies where they have been offered. It also collates views on the potential impacts subsidies could have – both positive and negative. It aims to provide clarity on the likely benefits and risks of providing premium subsidies. Key points: • Premium subsidies can help the initial take-up of insurance, and can dramatically increase the size of a risk pool. However, uptake is not guaranteed and, if subsidies are removed, there is a risk that some countries will discontinue coverage. • Subsidies are used by countries to both reduce government expenditure and to increase coverage. It is unclear what drives the decision about which of these routes to choose, although it seems likely to be affected by region and income level. • Premium subsidies are considered to have few negative impacts. Incorporating exit strategies into design can overcome concerns around dependency and sustainability. There is little evidence that premium subsidies contribute to moral hazard or undermine risk reduction and preparedness. There is broad consensus that premium subsidies support the uptake of insurance by improving affordability, offering a useful initial impetus to countries. All the risk pools have used premium subsidies as a way of growing their membership. Senegal is an example of a country that received subsidies, from Japan, for the first year of their ARC membership, and then integrated premium costs into their national budget for subsequent years (they also received an early payout, which helped to build political support). This was an excellent example of donors’ original expectation for how premium subsidies would catalyse government ownership of risk. Malawi presents a similar example – having dropped out of the ARC risk pool over concerns around basis risk, premium subsidies are thought to have been instrumental (although not the sole motivator) in them re-joining the pool several years later. One report acknowledges the success of premium subsidies as a strategy for growing risk pools: ‘For sovereign insurance, evidence suggests that premium subsidies facilitate or increase uptake particularly for countries that would otherwise be unlikely to take out insurance’ (Vivid Economics et al., 2016: viii). As an official from a country that has not yet taken out insurance phrased it: ‘At the very least, [premiums] will entice the government to take up insurance and try it out.’ Many view premium subsidies as essential for much-needed growth across the risk pools, particularly at this time. Risk pools need broad and diverse membership to function effectively. Yet fiscal space in many countries remains very tight – following COVID-19, and as governments struggle to address cost-of-living increases – which reduces the likelihood of new countries joining. PCRIC’s membership is very small, and needs to 17 ODI Report grow for the company to be sustainable into the future (OPM, forthcoming). Experience with ARC demonstrates the dramatic effect that premium subsidies can have on a risk pool’s membership: a recent cost–benefit analysis of ARC assessed its historical levels of business at around $5 million without subsidies, $15 million with some countries accessing subsidies and $25 million if subsidies to humanitarian agencies (via the Replica product) are also included (Lee and Rusconi, forthcoming). One key informant underscored this point, commenting: ‘Either you put premium subsidies in or you don’t have a risk pool.’ Countries use their subsides in different ways, both to displace government costs and to enable increased coverage, and this appears to differ by region. CCRIF offers countries the option to use subsidies to reduce their own contribution or to increase their coverage, and reports that both approaches get taken up. For ARC, it seems that countries are more likely to use premiums to reduce the contribution from their national budget for that fiscal year. Given the relatively small sample size, and the complexity of decision-making around insurance purchase, more empirical research is needed to understand under what circumstances subsidies are most likely to lead to increased coverage rather than displacing government spending. However, it seems likely that the country’s income level and fiscal situation play a part – if a country is extremely fiscally constrained then they are more likely to want to use a subsidy to reduce pressure on their budget. As an example, the Cook Islands – currently a high-income country, though their tourism-based economy was badly impacted by COVID-19 – stated that they would like subsidies for PCRIC insurance and would use them to increase coverage rather than reduce their own contributions. Countries do not always accept premium subsidies, even of 100%. For example, Mozambique was offered 100% premium subsidy for an ARC product and chose not to proceed. One involved stakeholder believed this was because there were concerns about the level of development of the risk model. This is perhaps surprising, but demonstrates the complexity of decisions around insurance purchase and that affordability is only one of the barriers to insurance uptake. The impact of premium subsidies will therefore be lessened in situations where the key barriers to uptake are issues such as lack of data or technical capacity, rather than affordability. Many countries state that they would have to drop out of risk pools without subsidies. For example, a CCRIF survey discovered that 61% of countries would likely discontinue coverage if their fiscal position changed. ‘As of 2021/22, the EU, the World Bank (MTDF), and Canada all continue to provide funds to significantly reduce the costs of premiums to Caribbean countries’ (Lee and Rusconi, forthcoming: 27–28). The situation is similar with PCRIC: ‘In consultations conducted after the pilot program, four countries suggested that they would not have been able to participate without premium subsidies. They indicated that they would “seriously evaluate their ongoing participation if the premium ceases to be subsidised”’ (Martinez-Diaz et al., 2019). However, this is not always the case. As stated above, there are examples, within ARC and CCRIF, of countries graduating successfully from premium subsidies. Less is understood about the impact of new subsidies to existing member countries. While it is clear that subsidies can support expansion to new members, it is less clear what the impact of subsidies for existing members could be. A recent study on ARC proposes a theory that ‘subsidies 18 ODI Report for more countries would produce more demand, whereas a higher level of subsidy for existing participants would not increase demand much at all’ (Lee and Rusconi, forthcoming). During this research, representatives of countries with existing ARC policies stated that they would be more likely to use subsidies to expand coverage to different hazards or try out new products (particularly for flooding) rather than expand their current policy. However, this was not unanimous – one country representative said they would use additional resources to expand geographic coverage, while another said they would extend coverage to more frequent droughts. More empirical research is needed in this area. Premium subsidies are viewed extremely positively, with few concerns over negative impacts or risks. Interviewees were overwhelmingly positive about, and supportive of, premium subsidies. An official from one African country said: ‘I don’t see risks to premium subsidy provision. Insurance is only one component of a larger package of interventions … it’s part of the puzzle, and we’re preparing government in different areas, because disasters will always be there, and insurance helps cover the high impact event layer.’ Some interviewees noted that subsidies have to be designed well so that they do not undermine sustainability. When asked specifically about the negative impacts of premium subsidies, some interviewees noted a risk that they could create dependency and undermine sustainability. However, all then went on to note that this could be overcome by designing subsidies so that there was a clear and agreed exit strategy (see section 6). There is little evidence that premium subsidies for climate insurance contribute to moral hazard by reducing incentives to invest in disaster risk reduction (DRR). The theory that climate insurance disincentivises government to prepare for, or invest in, the management of climate risks – because they know they will receive a payout – was roundly debunked by interviewees. There was a strong consensus that this theory has not played out in practice, with one interviewee describing it as a ‘false narrative’. The amounts of money for both subsidies and payouts were viewed as simply too small to drive a significant change in attitudes around risk reduction. Vulnerable countries are painfully aware of the risks they face and are very strongly incentivised to manage, rather than ignore these. Some also suggested that the theory credited governments with a more joined-up approach to disaster risk management than is usually the case in reality, noting that DRR was typically dealt with by one department, with insurance being organised and purchased by a completely different ministry. Interestingly, only one interviewee felt they had seen cases of moral hazard; this was in the Pacific, where government is notably much smaller and potentially more integrated than in larger countries. 25 ODI Report criteria, but when it comes down to the country level, it has to be custom-designed really to respond to the specific needs of the country.’ This speaks to the need for detailed country-level work in order to agree terms of the subsidy depending on the specific financial and economic situation of the country. Clearly then, government actors need to be fully engaged collaborators in the design of the subsidies. On all sides, there needs to be an appreciation of the longer-term objectives of the subsidy, clarity around the country’s needs, and honest discussion around the necessary subsidy amount and duration given the country’s financial position and economic outlook. One interviewee noted that, in some countries, prior capacity building will be required to arrive at this point. The complex process of agreeing a subsidy is best conceived as a negotiation between actors with differing incentives, and little is reported publicly. These discussions are (and will be) sensitive and private: often not taking place directly between donors and countries, but mediated through third parties, such as the risk pools themselves, who have their own set of incentives. This introduces complexity and mixed incentives between the different publicand private-sector actors. Some interviewees emphasised the element of negotiation that comes in purchasing an insurance product – for example, negotiation around agreeing the level of coverage. Agreeing a subsidised policy is therefore best characterised as a process of intense negotiation and managing trade-offs, involving different actors driven by potentially competing interests. There needs to be better understanding of what the incentives of these actors actually are, and how they interrelate to shape the design and uptake of subsidies. However, gaining clarity is difficult, as the donors, countries and intermediaries interviewed were often unable (or reticent) to share the exact details of subsidies, or otherwise requested that the information remained confidential. The design of subsidies will likely be shaped by donor objectives. As with the allocation of subsidies, subsidy design can support a donor’s longer-term objectives. For example, if a donor is concerned with increasing uptake or coverage, they may be willing to pay 100% subsidy for several years. However, if their primary concern is building government ownership, they will likely want to see some funding coming from the national budget as soon as possible. Most interviewees were of the view that 100% subsidy was not appropriate, except perhaps in the first year of a subsidy. People spoke of needing to ensure that there was some government awareness and ownership that would best come by making a small contribution to premium costs. Respondents also mentioned that very cheap goods are often not properly valued by those who receive them and that making a contribution ensures that governments have some ‘skin in the game’. One study also notes that ‘countries should continue to cover some portion of the premium, even if minimal, as allocating budgetary funds to pay premiums generates a regular process through which finance and other ministries must review national risk exposure. It also prompts a regular dialogue between ministries and legislatures – which must approve the budget – about disaster risk insurance and disaster risk finance more generally’ (MartinezDiaz et al., 2019). There is no clear consensus on what percentage of premium costs should be covered, over what period of time. One respondent noted that it is ‘tricky to get right and will change over time’ as the economy changes 26 ODI Report and different actors move in and out of decisionmaking roles in government. A recent cost–benefit analysis of ARC considered the impact of 50% subsidies, finding that ‘evidence from ARC’s experience shows that this level of subsidy does seem to make a substantial difference to demand, although this is not a robust statistical result, just an impression from the data. Probably this subsidy would not increase demand from countries that were already purchasing unsubsidised insurance’ (Lee and Rusconi, forthcoming: 25–26). There is some consensus that premium subsidies should be multi-year rather than one-off annual offers, although this is unpopular with some countries because of the requirement for co-financing. Donors, in particular, are keen for multi-year subsidies with increasing levels of cost covered by the government in order to offset the risks of dependency and lack of sustainability. Multi-year arrangements also appeal to donors and risk pools given ‘churn’ within government, bearing in mind that a new Minister of Finance may not be as supportive of insurance as the last. However, countries appear to find multi-year arrangements less appealing and find it difficult to commit to funding increasing amounts of subsidy into an uncertain future. There exists also the added challenge of how to actually enforce such arrangements, especially for vulnerable countries who may experience a shock that radically impacts on their fiscal position. Some respondents argued that because multi-year arrangements are unpopular with countries, they have not been strictly applied to countries receiving subsidies. There are a wide range of views on the appropriate duration of subsidies. Interviewees’ suggestions ranged from two years to ‘indefinitely’. One respondent suggested that around five years would be an appropriate duration, on the basis that most risk pools are insuring risks that are expected to occur every 4–10 years, so in a five-year period there is a fairly good chance of receiving a payout and experiencing the benefits of insurance. An alternative view was voiced by a few interviewees who viewed premium subsidies as an alternative to humanitarian aid, making the argument that they should continue indefinitely, over a very long time horizon (for example, 30 years) or until a country graduates from ODA eligibility. These respondents recognised this was contrary to donors’ concerns about dependency and sustainability, but felt it was a more realistic future for CDRFI, particularly in light of climate justice and ‘loss and damage’ debates. There is some support for embedding conditionalities into the design of subsidies to ensure best practice, although this needs to be done with consideration of how it may affect uptake. As noted in section 5 on allocating subsidies, donors are keen to use premium subsidies to incentivise best practice. One way of doing this is to only allocate premium subsidies to risk pools or products that incorporate elements of best practice. Another option is to incorporate specific practices into the design of individual subsidies as required conditions. For example, this could include requirements to develop payout contingency plans that prioritise vulnerable people; not use payouts to fund food transfers if cash is a viable option; or insist that specific risk reduction activities take place as pre-conditions for the subsidy. Some interviewees noted that that subsidies should carry the ‘normal requirements of aid’; for example, that there is an impact on poor communities, that monitoring and reporting is routinely done, or that gender dimensions are considered. These are all ‘hard-won’ areas of best practice in development programming that 27 ODI Report have not always been reflected in how subsidies have been designed. As one respondent said, ‘If it’s going to be called “development insurance”, then you need to hardwire-in developmental outcomes.’ However, there may well be a tradeoff to be made, where some conditionalities could reduce demand. This should be carefully considered in consultation with recipient countries. Donors will also need to consider how adherence to conditionalities could be monitored and reported, and would need a credible mechanism for withdrawing subsidies if necessary. Subsidies can be packaged as part of broader programming, in order to raise their profile and support sustainability and capacity building. Some interviewees noted that subsidies are likely to work best if they are part of a broader benefits package; for example, if they are linked with a broader technical assistance or capacitybuilding programme. This kind of ‘bundling’ improves the visibility of the subsidy and links it with a greater endeavour in the country. Linked capacity building would also have the added benefit of ensuring that there are people embedded in government who really understand the model and the policy, and who are therefore better placed to negotiate coverage in future. 28 ODI Report 7 Conclusions This political economy analysis is part of a wider study seeking to investigate how premium and capital support can best be provided to countries. The overall aim of the study is to further thinking on what the basis should be for prioritising countries for, or excluding them from, subsidy support. There are many relevant findings in this report. Stakeholders argued strongly that donors should focus on providing grant funding for premium subsidies rather than capital support. Whilst finance for capital support is generally more accessible within key donor agencies, premium subsidies are now the priority in order to ensure the growth and sustainability of the risk pools. Subsidies can help to grow risk pool membership, but there are also reasons why existing members could also be considered for premium support. Evidence demonstrates that subsidies help encourage new countries to purchase insurance. However, many countries who receive subsidies state that should the subsidy stop, they would need to reduce coverage or drop out of the risk pool. There is also a difficult question as to whether countries who have been loyal risk pool members, paying premium out of their own budgets, should be excluded from subsidies or not. Including them could send a positive message to other countries about the availability of long-term support; it is a good way of ‘rewarding’ strong risk ownership and ‘good performance’; and it could lead to increased coverage if governments use the subsidy to expand their policies rather than replace their own costs. Affordability emerged as the main barrier to uptake, but it is only one among many factors that influence the uptake of insurance. The most significant barriers, after affordability, were lack of understanding and technical capacity; availability of alternatives; and perceptions of reliability. However, the report provides a much longer list of influential factors, the balance of which will vary in each country, affecting the impact of subsidies. This is complex, and not well understood due to the lack of research in this area. For example, a country lacking technical capacity and understanding of insurance is less likely to purchase a policy. Even though providing a subsidy to that country may help to overcome that barrier, the barrier may prove too significant and the country may decline to accept the subsidy. Experience has demonstrated that subsidies are not always attractive enough for them to be taken up. Previously, certain elements of the design of subsidies have acted as barriers to uptake – particularly that they had to go to new countries or cover new hazards, or where countries have to commit to a multi-year arrangement in which they increasingly co-finance premiums. Stakeholders argued strongly that recipient countries should be much more involved in the design of subsidies, so that donor objectives can be carefully aligned with country perspectives and priorities. 29 ODI Report Finally, actors have diverging views on the importance of different factors in allocating premium subsidies. Overall, ‘proportion of vulnerable population in total population’ and ‘climate and disaster risk profile’ were viewed as important factors for allocating subsidies, particularly by representatives from across the risk pools. ‘Country income level’ and ‘prior risk reduction actions/policies’ were also highly valued, though not universally. However, donors seemed to prioritise a range of other factors, placing particularly value on the product being of high quality; the inclusion of a plausible exit strategy; the subsidy being for a new country or product; and the country being a priority for them. Donors also had exclusion criteria; most significantly, they would exclude countries that are not ODA-eligible or which are subject to sanctions. References Clarke, D. and Hill, R. (2013) ‘Cost-Benefit Analysis of the African Risk Capacity Facility’, IFPRI Discussion Paper 01292 (https://ebrary.ifpri.org/utils/getfile/collection/p15738coll2/id/127813/ filename/128024.pdf). e-Pact (2017) ‘Independent Evaluation of the African Risk Capacity (ARC)’, Oxford Policy Management (www.opml.co.uk/files/Publications/a0603-independent-evaluation-african-riskcapacity/arc-evaluation-report.pdf?noredirect=1). IGP (2021) ‘Background Note on targets and indicators for Vision 2025: A refined Monitoring & Evaluation framework for the InsuResilience Global Partnership’. Bonn: InsuResilience Secretariat (www.insuresilience.org/publication/background-note-on-targets-and-indicators-for-vision-2025a-refined-monitoring-evaluation-framework-for-the-insuresilience-global-partnership/). Lee, S. and Rusconi, R. (forthcoming) ‘Cost–Benefit Analysis of the African Risk Capacity’. Draft for KfW. Oxford Policy Management. Martinez-Diaz, L., Sidner, L., and McClamrock, J. (2019) ‘The Future of Disaster Risk Pooling for Developing Countries: Where Do We Go from Here?’ Working Paper. Washington, DC: World Resources Institute (www.wri.org/research/future-disaster-risk-pooling-developing-countrieswhere-do-we-go-here). Kramer, B., Rusconi, R., and Glauber, J. (2020) ‘Five years of risk pooling: An updated cost–benefit analysis of the African Risk Capacity’. IFPRI Discussion Paper 1965 (https://doi.org/10.2499/ p15738coll2.134046). OPM (forthcoming) ‘Second Formative Evaluation of the African Risk Capacity’. Oxford and London: OPM and UK FCDO. OPM (2022) ‘Review of the Pacific Catastrophe Risk Insurance Company: Technical Proposal’. Oxford Policy Management (unpublished document). Töpper, J. and Stadtmüller, D. (2022) ‘Smart Premium and Capital Support: Enhancing Climate and Disaster Risk Finance Effectiveness Through Greater Affordability and Sustainability’. Bonn: InsuResilience Secretariat (www.insuresilience.org/publication/smart-premium-and-capitalsupport-enhancing-climate-and-disaster-risk-finance-effectiveness-through-greater-af-fordabilityand-sustainability/). Vivid Economics, Surminski and Callund (2016) Understanding the role of publicly funded premium subsidies in disaster risk insurance in developing countries. Evidence on Demand, UK (www.vivideconomics.com/casestudy/the-role-of-publicly-funded-premium-subsidies-indeveloping-countries/).