Finance as the Glue
Finance as the Glue
Abstract and Keywords
Financial and budgetary instruments are the glue that holds credible plans together and makes them strong enough to withstand the whirlwind of highly charged post-disaster politics. When designing and implementing disaster risk financing strategies, details matter. It is important to pay for financial advice and build in-house expertise. Discipline, cost, accuracy, and speed all matter when structuring a disaster risk financing strategy. Speed matters but not all resources are needed at once. The triggers in the financial strategy should match the triggers in the plan. Traditional reinsurance can be particularly useful for locking in plans for reconstruction, and indexed reinsurance can play the same role for financing indexed early actions. Financial instruments can enable multiple parties to contribute to the same plan. Subsidized financial instruments can be used to encourage others to contribute towards the cost of well-defined plans.
In the early morning of 19 September 1985, an earthquake that registered 8.0 on the Richter scale struck Mexico City, bringing down tens of thousands of buildings and seriously damaging tens of thousands more. It killed at least 5,000 people, and left some 250,000 homeless. The city suffered major damage because of the large magnitude of the quake and the ancient lake bed on which Mexico City sits.
That Mexico should be hit by such an extreme event like this earthquake should not have come as a surprise. With its diverse geography, the country is in fact exposed to a wide variety of geological and hydro-meteorological hazards: earthquakes, volcanoes, tsunamis, hurricanes, wildfires, floods, landslides, and droughts. However, despite being a reasonably well-off middle-income country, it was underprepared to deal with such a calamity, and serious political fallout followed.1
The 1985 quake triggered a national dialogue on disaster risk management, and the government of Mexico responded by making it a national priority and by investing to make the country more resilient. One component of this effort was the establishment in 1996 of the Natural Disaster Fund (Fondo Nacional de Desastres Naturales), commonly known as FONDEN. Initially, it was intended to finance post-disaster reconstruction of public infrastructure and low-income housing without the need to gouge other parts of the national budget in the event of a disaster.2
Instead of purchasing insurance for each government-owned road, school, hospital, water distribution pipe, electricity line, and piece of (p.76) other infrastructure, FONDEN acts like an insurer. It covers the full cost of the reconstruction of any federally owned infrastructure and 50 per cent of the cost of any state or municipally owned infrastructure. It finances ‘build back better’, allowing the reconstruction of more resilient infrastructure at higher standards and the relocation of public buildings and low-income communities to safer zones. To pay for all this, FONDEN has an annual budget allocation from the federal government, akin to an insurer’s capital base, and purchases a large reinsurance policy and a catastrophe bond to protect its balance sheet against really big disasters, again like an insurer would.3
FONDEN has very clear rules about what reconstruction it will finance and when, similar to the rules that might be found in an insurance policy, but it faces the risk that state governments will inflate their reported losses from disasters and use political pressure to get their way. To protect against this, the government of Mexico has invested in an army of independent loss adjusters who are responsible for objective post-disaster damage assessments and who are overseen jointly by government and its consortium of reinsurers. Reinsurance therefore has a dual purpose: to help pay for reconstruction costs in the worst years and to act as an Odyssean pact, enabling the government of Mexico to credibly commit to objective independent loss adjustment in line with its rules of operation and in turn to insulate FONDEN from potential post-disaster political pressures.4
If done well, as in Mexico, financial planning for disasters can be the glue that holds together all the pieces of a carefully defined recovery plan and makes it credible and strong enough to withstand the whirlwind of highly charged post-disaster politics. As the government of Mexico found when developing FONDEN, financial wizardry is important, but it should be the servant, not the master, of a disaster response and recovery plan.
Just as different types of glue are available for fixing different materials, so too different kinds of budgetary and financial instruments are available to governments, donors or other organizations for (p.77) different kinds of response and recovery plan. Showing the potential use and misuse of these instruments is the purpose of this chapter.
Can Financial Instruments Replace a Disaster Plan?
The amount or type of glue used, or the number or complexity of instruments adopted, is not a good measure of financial planning. Crowing over a big budget for disaster response or preparedness may be misleading. The total budget for disaster response is hardly a measure of effective planning because it can be easily spent poorly. Likewise, it is easy to be seduced by an especially dramatic-sounding financial instrument that promises to solve all problems. Focusing on a particular financial instrument misses the point, and thinking that financial instruments can replace having a plan is misguided. Just because catastrophe reinsurance has been bought or a large contingency fund has been set up does not mean it is the right solution for the circumstances and the organization involved.
Financial instruments will be useful to governments and donors, just as they are to insurers, but for them to give the biggest bang for buck they should form part of the plan, not define the plan. Without something to stick together, playing with glue just makes a mess, and without a clear plan to stick together, whizzy financial instruments are a waste of money. Just as with glue, the usefulness of a financial instrument for financing the risks associated with disasters comes from applying it appropriately.
So, suppose that a government or donor has a plan it wishes to finance. How should it actually develop strategies to bind the plan together and make it credible? Essentially, such an effort has two parts. The first is ensuring that the right amount of money is available quickly when—and only when—it is required by the plan. The second is ensuring that the money is spent on what it is supposed to be spent on and accounted for in a clear, transparent fashion. For example, a ministry of transportation might ensure that US$10 million is available within a week of an earthquake of magnitude 6.0 and that (p.78) this money actually finances the emergency livelihood support and reconstruction of the lifeline infrastructure for which it is intended. This practice, termed disaster risk finance, should bind the various partners to the pre-agreed objectives, decision processes, and implementation modalities. It should give all parties confidence that the plan is credible.
Thinking Like an Insurance Company
In many ways, financial planning for disasters is easier now than ever before because of the huge growth in the range of instruments available and the wealth of knowledge in the insurance industry. There is magic in a good insurance policy: it creates certainty when an unexpected event occurs. That means that even though people face lots of risks, the consequences of some of these risks becoming reality are known with certainty. If a kitchen floods but a good home insurance policy is in effect, the result will be lots of inconvenience, but a predetermined share of the material losses will be covered with certainty. If a driver crashes into someone else’s car, the insurance company of the car owner, the damaged party, will settle with the driver’s insurance company to ensure that the car owner is paid for the damage. The key is certainty, which is created by the credibility of the arrangement: a contract with the insurance company, as well as good regulation and legal frameworks, ensures that insurance companies will indeed do what is expected of them in line with the contract.
Suppose there is an agency in charge of disaster response and preparedness. If it were to learn to think like a reputable insurance company, it would get better at credible financial planning for disasters. Suppose the agency and political leaders have agreed who and what will be protected, against what, and how the protection will work. This ‘how’ provides the blueprint for a well-defined response and recovery plan to which the agency is committed, as if there is a political contract with the citizens it is entrusted to protect. As discussed in Chapter 3, this agency would then work with engineers and (p.79) logistical experts to ensure it has a clear idea of how much cash the plan would require in the aftermath of different disasters. It would also work with scientists and risk modellers to understand the likelihood of different types of disaster occurring. These two tracks of technical work would allow the agency to develop a probabilistic assessment of its potential financial liability, its contingent liability. Armed with this, the agency would then use actuaries and other financial experts to piece together different budgetary and financial instruments to form a strategy that would ensure it could meet this contingent liability as cost-effectively as possible.
Our advice to such an agency, or to a ministry, or a local or international non-governmental organization (NGO) involved in disaster response is this: think like insurance companies and responses will be more cost-effective with better outcomes. National governments, donors and international organizations would function a lot better if they worked together pre-disaster to clarify what contingent liabilities they would each take on, and then ensured that these liabilities were financed in the most cost-effective way. It does not help anyone to declare that all needs will be met, without a precise, measurable definition of need, a clear plan for how to address the need in a timely manner, and a credible financing strategy. In other words, like an insurer, develop a rules-based plan and then finance the plan, do not try to finance the need directly. As we discussed in Chapter 4, there will always be the need for a back-up for when plans fail or the unforeseen happens, but for the bulk of disaster response and reconstruction there is an urgent need for a clearer articulation of who owns what risk and a credible financing strategy to increase its effectiveness and the value for money it can offer. And this is where thinking like an insurer can really pay off.5
Preparing Like an Insurance Company
The contingent liability of a disaster response and recovery plan will encompass the full costs of the risks taken on by the government or (p.80) agency that owns the plan. This financial liability will include the value of the transfers in cash or in kind that need to reach particular people, the staffing and other logistics costs, the costs of repairing infrastructure, and any other people, firms, organizations, or assets whose losses the plan aims to cover. So how do planners go about designing the financing strategy to cover this entire contingent liability?
The insurance industry has plenty of experience in how to finance a contingent liability. Although medieval begging bowls may still rule disaster finance in many countries, the relevant insurance methods and practices have evolved during centuries of experimenting with risk financing. In fact, insurance companies, pension funds, and their regulators have long solved the key problems of effective disaster risk financing. It is worth explaining how this works.
In a well-regulated market, insurance companies are not allowed to sell insurance policies against disasters unless they can demonstrate up front that they would be able to pay claims as they fall due with a very high degree of certainty. In insurance terms, insurers need to be able to demonstrate that they are actuarially sound, which means that a certified actuary has deemed their financing plan sufficient to meet the needs of the plan, even after an unusually devastating disaster.
To meet this requirement, an insurer traditionally has had a capital base (such as a bank account with money in) to cover all but the most extreme potential losses and it has taken out a reinsurance policy in case it faces such extreme losses that they might exhaust its capital base. The capital base enables the insurer to retain some of the risk, making a profit most of the time but suffering a big loss if claims are much higher than expected. The reinsurance policy enables the insurer to transfer some of the risk to a reinsurer, so that the reinsurer would share in some of the profit if claims are lower than expected but contribute to the cost of very large claims. By buying reinsurance, an insurer transfers part of its balance-sheet risk to the reinsurer, which can retain most of it by holding a massive capital base designed to soak up a diversified portfolio of global risks. In short, the insurance company engages in risk retention via a capital base and in risk (p.81) transfer via reinsurance. The reinsurance company does the same, although it typically retains more of the risk. Over time, there has been a lot of growth in the variety and type of risk-retention and risk-transfer instruments available to insurers, but the basic principles remain.
These principles should form the basis of a financial strategy for a government or an organization committed to covering particular contingent liabilities. It must decide how much risk it will retain and how much risk it will transfer, and which financial and budgetary instruments to use for this. The same applies to international organizations or donors: they may not see themselves as insurance companies, but could see themselves as reinsurance providers, reinsuring quite extreme risks from countries or organizations, or as insurance brokers, setting up schemes and instruments that allow risk transfer into bigger and better reserve funds and global insurance markets. A combination of risk-retention and risk-transfer instruments can form the basis of a credible financing strategy of response and recovery plans for natural disasters or pandemics, and indeed the various layers of the global humanitarian system.
But before turning to these instruments we would like to recognize another option: a government or organization may not want to prepare or act like an insurance company. Instead, it might prefer to pass on its contingent liability entirely to a regulated insurance company by fully or partially buying insurance policies itself for all the people or specific assets insured. It then prefers to ‘buy insurance’ rather than to ‘be the insurer’. In countries with weak insurance regulation, this may require investment to ensure that commercial insurers manage this disaster risk in an actuarially sound way, or in some cases a government may wish to set up a new insurance provider that is subject to stricter rules. This may be a fruitful route for specific large, publicly owned assets—for example, government buildings in Peru are all legally covered via specific building insurance. And, as will be discussed in more detail shortly, partially subsidizing insurance policies for farmers, homeowners, and firms (p.82) may be a sensible part of a government strategy, but it can by no means be all.
Choosing among Financial Instruments
What sorts of financial instrument are available to a government or organization to cover the contingent liabilities it has taken on? The budgetary and financial instruments available to a government, its partners, and its insurers to finance the cost of disasters can be categorized across two dimensions: risk retention versus risk transfer and ex ante versus ex post. Table 5.1 describes the key instruments and offers some definitions.6
Risk-retention instruments do not take risk off the balance sheet—the cost of a disaster must still be repaid. The instrument just offers more flexibility in how and when one would have to pay. Contingency funds, budget allocations, and lines of contingent credit are all risk-retention instruments, as are budget reallocations, tax increases, and post-disaster credit. By comparison, risk-transfer instruments remove volatility from one’s own balance sheet and transfer it to somebody else’s. Insurance is the classic risk-transfer instrument, but for governments and insurers the range of risk-transfer instruments is enormous and growing by the day, including indemnity reinsurance, indexed reinsurance, catastrophe bonds, and catastrophe swaps. Soliciting contributions from the international community is a form of risk transfer for governments, although it is typically undisciplined, unpredictable, and slow.
Ex post instruments do not require advance planning. For example, governments could choose to finance a disaster through donor contributions, a budget reallocation, new loans, or tax increases. None of these requires pre-disaster planning. By comparison, ex ante instruments require proactive advance planning and include reserves or contingency funds, budget contingencies, contingent debt facilities, and a range of insurance or other risk-transfer products.
Table 5.1: Instruments for Financing Disaster Risk
(arranged before a disaster)
(arranged after a disaster)
Contingency fund or budget allocationa
(changing how or when one pays)
Line of contingent creditb
Traditional insurance or reinsurancec
Discretionary post-disaster aid (begging bowl)
(removing risk from the balance sheet)
Indexed insurance, reinsurance, or derivativesd
Capital market instrumentse
(a) A pre-funded pot of money that can be used for specific purposes, such as responding to a large natural event. Resembles a current account in a bank.
(b) A pre-agreed loan that can only be drawn down in specific pre-agreed circumstances, such as the onset of a large natural event or a declaration of national emergency by government. Slightly resembles an overdraft facility, but one with rules for when the loan can be drawn down.
(c) A contract whereby the insurer or reinsurer is paid a premium and the rules for claim payments are based on the losses incurred, as measured by independent loss adjustors.
(d) A contract whereby the rules for the net transfer to the insurer, reinsurer, or counterparty are based on the index. For insurance and reinsurance, payment of the premium is followed by indexed claim payments. For an indexed derivative, the timing of payments may differ.
(e) Capital market instruments such as catastrophe bonds and catastrophe swaps are financial contracts that can be structured to act in the same way as insurance, but investors, not necessarily reinsurers, provide the protection. A catastrophe bond is an insurance-linked security in which payment of interest and/or principal is suspended or cancelled in the event of a specified catastrophe, such as an earthquake. A catastrophe swap is a contract used by investors to exchange (swap) a fixed payment for a certain portion of the difference between insurance premiums and claims.
well-defined plans. Ex post instruments are not particularly useful for credible pre-disaster plans because typically they cannot be relied on in a contractual sense. Contributions from the international community are famously unreliable: between 2010 and 2014 the annual funding (p.84) received through UN-coordinated appeals ranged from a low of 29 per cent in 2010, as donors committed large volumes of funds to the response to the Haiti earthquake early in the year, to a high of 57 per cent in 2013.7 Budget reallocations, borrowing, and new taxes are all certainly possible post-disaster. In practice, however, they all typically give politicians or leaders some degree of discretionary power to write new plans to be financed, or at the very least to rewrite any pre-agreed plan. As argued in Chapter 4, opening the door to such discretion slows everything down and defeats the point of having a well-defined, credible plan. Within such a plan, discretion should be applied only at the technical level, not the policy level. Post-disaster instruments can be useful for financing costs beyond the plan, but not for financing the plan itself.
This is relevant not just for governments funding their own response and recovery plans, but also for humanitarian NGOs, donors, and international organizations committed to providing people and countries with support. They can take responsibility for parts of these plans, or essentially provide reinsurance to local organizations, agencies, and countries for their plans, or broker or fund the emergence of specific risk-transfer instruments that offer governments and organizations new options for financing risk in cost-effective ways. To be able to act in a credible way, as an insurer or a reinsurer, all these organizations will need to decide whether to retain or to transfer these risks. For most, ex post instruments are largely unhelpful: they are hardly going to be able to tax or raise new budgets, and begging bowls and appeals are unreliable. Building up a portfolio of ex ante instruments will be essential to fulfil these organizations’ contingent liabilities.
The Challenges and Opportunities of Ex Ante Instruments
Choosing an ex ante financial strategy is not easy, but choices can be informed by technical analysis.
Anyone choosing among ex ante instruments should consider four key dimensions: the discipline required to use an instrument, (p.85) its cost, its accuracy, and the speed by which it can be converted into cash for spending on a disaster response. For discipline, any government or donor with a large reserve fund, budget allocation, or line of credit set aside for financing a specific plan when required will always come under pressure to spend those resources on other things. If a disaster does not happen, governments, international donors, and development banks may all come under criticism for not disbursing funds that have been provided to support development, and the liquidity that has been arranged to finance a specific plan may be re-routed to discretionary spending. When a disaster does strike, even if a financial strategy is successful at providing the right amount of money at the right time, poor public financial management—that is, making sure the money goes where it is supposed to according to the pre-agreed rules—can stymie the entire endeavour. Ex ante risk transfer instruments that transfer contingent liabilities to the market or other agencies make it harder to game the system for political reasons. They will be triggered only in particular circumstances and are more easily directed to the plans they were meant to finance. Plus, they require less discipline by the user because discipline is built into the rules of the contract and is part of the service provided.
In addition to instilling discipline in claim payments, indemnity-based insurance and reinsurance can also instil discipline in risk-reduction investments. Often the actual risk-transfer contract will have a sensible set of conditions for investments in risk reduction and resilience, which will be checked by the insurer or reinsurer. Moreover, any additional investments in risk reduction over and above this should reduce the premium, providing an immediate financial reason for cost-effective risk-reduction investments.8 The proof of all this? It is no coincidence that Cyclone Patricia, the strongest cyclone ever to make landfall on the Pacific coast, did not cause loss of life when it hit Mexico in October 2015. The entire financing system in Mexico promotes clear risk ownership and appropriate investments in risk reduction and preparedness.
(p.86) The second issue of importance when choosing among ex ante financial instruments is the cost. There are no hard and fast rules for what combination of instruments will be most cost-effective for governments or donors. Relying on principles or rules of thumb for designing a financial strategy is dangerous because the value for money depends so much on the details. Contingency funds can yield low investment returns and be used as a political slush fund, or they can be run in a disciplined manner. Risk-transfer products can be expensive, slow, and unreliable—or they can be cheap, quick, and reliable. Lines of contingent credit can be expensive and discretionary, or cheap and arranged in a way that promotes discipline. Financial expertise is too important to be on the periphery of planning, and, as with any financial product, the risk of being sold a product that is not fit for its purpose is huge. Not for nothing did we argue in Chapter 3 that the men and women of finance must be a core part of the team that develops the plans and then guides the financial strategy through to implementation.
Having said this, there is one rule of thumb that is useful to bear in mind when choosing between risk transfer and risk retention: use risk retention for more frequent losses and risk transfer for the less frequent, larger losses. Typically, a cost-effective financial strategy will use a number of complementary instruments, not unlike ancient cities, which had different layers of defensive walls—the lower, thinner, outermost walls offered cheap but reasonable protection for minor attacks, and the taller, thicker, innermost walls served as a reliable last line of defence for protection against more sizeable, sophisticated, or determined aggressors. Similarly, cost-effective financial strategies for disaster risk financing typically involve risk-retention instruments for smaller, more frequent disasters and risk-transfer instruments for larger, more extreme disasters. This arrangement appears over and over in risk finance, from insurers and reinsurers to mutual insurance societies and the insurance captives9 of large multinational companies.
This principle is relevant for risk financing by governments, even if they usually can handle large losses on their budget.10 The reason is (p.87) the cost of liquidity. When a government chooses to retain disaster risk, it effectively needs to have a large pot of money sitting somewhere, waiting for a potential disaster. That pot might take the form of a contingency fund or a budget allocation, or the government may pay a lender to keep a pot of money available that it can borrow from quickly in the event of a disaster, essentially an overdraft facility. Either way, the government is paying the full cost of the liquidity itself. But this will not be attractive for less frequent losses: the cash held would be rarely used and thus costly. Risk transfer is, then, more attractive because it allows the cost of liquidity be shared with others: a regular payment can be made to have access to substantial liquid means when required.
The final two key issues to consider when choosing among financial instruments are accuracy and speed. Here the intuition is straightforward: wherever possible, the timing and triggers of the financial instruments—in particular the triggers for any line of contingent credit and risk-transfer instruments—should match precisely the triggers in the plan. If the triggers in the plan are parametric, then parametric triggers should be applied to the risk transfer. If the triggers are based on individual loss adjustment, the risk transfer should be based on this loss adjustment. Finally, speed matters, and it must play a central role in deciding among instruments. If the plan requires immediate access to financial resources, then instruments need to be chosen that offer this. In general, if it is difficult to access risk-transfer or even risk-retention instruments that match the rules in the plan in terms of accuracy or speed, it is likely that the rules of the plan are unworkable in a practical sense, and they may need to be rethought.
Acting Like an Insurance Company
Governments, agencies, or organizations that have prepared themselves like an insurance company for the contingent liabilities they have decided to take on can go a step further: they can start acting like an insurance company, offering a contractual relationship to clients (p.88) that guarantees support when natural disasters hit. They can learn from modern insurance market practice about how to behave like an insurer and then emulate an insurer, sometimes at lower cost.
Programmes in Mexico and Kenya have done just that with good results.11 As we described in Chapter 1 and elsewhere, in Mexico FONDEN is liable for the cost of reconstructing public infrastructure and low-income housing after natural disasters, and it operates akin to insurance, with clear rules on what it will pay for, a financial strategy that combines a contingency fund with risk-transfer instruments, independent loss adjustment, and strong incentives for risk reduction and resilience.
Also as described earlier, in Kenya about 400,000 pastoralist households are registered in the Hunger Safety Net Programme, and the 100,000 of those deemed to be the most in need receive regular cash transfers. In times of drought, however, the programme temporarily scales up by making quick cash transfers to some or all of the remaining 300,000 pastoralist households, providing them with the means to protect their families and their animals. These additional pastoralists are in effect covered by an insurance policy—a social safety net that pays a cash transfer to a predefined group of pastoralists when the rains fail, without delay and without questions, so they can afford to buy food and fodder even though the harvest is bad. Above and beyond what is being given to them for free, they also are given the option of buying affordable insurance for more peace of mind. The programme has simple parametric triggers for payments, in this case satellite data on the ‘greenness’ of ground vegetation, which is a good proxy for drought-related stress in these areas. The government contributes a regular budget for normal years, and an international donor, the United Kingdom’s Department for International Development, acts like a reinsurer, providing additional funds as needed to cover the pre-agreed rules for scaling up in extreme years.
Although Ethiopia’s Productive Safety Net Programme covers millions of people and has positive impacts on people’s lives and livelihoods (see Chapter 1),12 it does not operate with defined triggers. (p.89) However, since its inception in 2005 it has had a facility to scale up the provision of payments to additional beneficiaries and include new beneficiaries in response to droughts. To secure its funding, a risk-financing mechanism was introduced in 2009 to allow the rapid mobilization of additional resources in the event of an emergency. For example, in 2011 the mechanism was triggered, providing support to an additional 3.1 million beneficiaries for three months and extending the duration of transfers for 6.5 million existing beneficiaries for the same length of time. It now serves as the backbone of current responses to drought because it covers some of the most vulnerable in the population. However, further improvements could be made: it lacks early and defined triggers that may cause delays, and there is no a priori registration of those who may be covered when droughts or other extreme events take place.
Some governments may recognize the virtue of insurance-like schemes, but they are reluctant to be the insurer themselves. A few countries such as Panama, Peru, and Colombia legally require all government buildings to be insured.13 Where farmers, homeowners, and businesses may be able to bear some of these risks themselves, many governments do not want to take on the full contingent liabilities for losses to assets and livelihoods in the event of a disaster. Governments may provide incentives or subsidies for people to buy insurance, either using existing commercial insurance companies or at times setting up national insurance companies to implement these schemes. For example, the governments of France, Japan, New Zealand, Spain, Turkey, the state of California, and the Republic of China (Taiwan), have all set up institutions that offer property catastrophe insurance to homeowners and businesses, and the governments of Canada, Cyprus, Greece, India, the Islamic Republic of Iran, and the Philippines have all established government-owned insurers to provide agricultural insurance.14 These schemes, and schemes in which government partners with private insurers, typically offer subsidized insurance whereby government contributes towards the cost of insurance. In some countries, these (p.90) schemes are perceived to have offered a reasonable balance, restricting the contingent liability of the government to make a sensible plan feasible, by expecting individuals to take some responsibility by paying part of the cost of financial protection against disasters themselves.
Although it is often successful, this approach of governments supporting disaster insurance markets has not always been so. Sometimes, these schemes are not run on an actuarially sound basis and end up having to be bailed out by the government after a disaster.15 Even when such schemes are actually sound, a government often succumbs to temptation to structure its subsidies to undermine the fundamental principle of risk-based pricing, whereby the premium reflects the underlying risk—for example, by capping the premium paid by the policyholder. Such approaches may be attractive politically, because subsidies target those at highest risk, but by structuring subsidies in this way governments encourage people to take on more risk, or at the very least to underinvest in sound risk management, because by doing so the value of the subsidy is kept as high as possible.16 Making insurance compulsory is often regarded as a useful strategy for increasing the number of people protected, thereby reducing the potential for the presentation of begging bowls post-disaster. However, such an approach places a very high burden of care on government to ensure that products offer value, because consumers typically have no power to hold insurance companies to account.
Efforts by governments and donors to develop unsubsidized markets for disaster insurance have not generally been too successful, particularly in poorer countries.17 Over the last decade, international donors have supported over a hundred pilot programmes aimed at providing unsubsidized weather-index insurance markets to protect poor, vulnerable farmers against the vagaries of the weather. A typical goal of these programmes is for farmers to pay the full cost of their financial protection against shocks, so that after a large disaster they reliably receive fast cash and have no need to call on government or other benefactors for help. The policies are parametric: payments are triggered when low rainfall or another easily measured indicator is reached.
(p.91) Although noble in their intentions, only a few of these programmes have survived beyond the donor support, and for those that have, there are unanswered questions about how good the protection would actually be in a disaster. Many poor farmers seem unwilling to pay the full cost of their financial protection themselves, and so most agricultural insurance programmes for poor farmers that are still in force are subsidized. No doubt, there is room to continue to experiment with such schemes, because they could offer some certainty, not least when governments or agencies fail to act as insurers instead. But they are probably not quite ready to be a means of protecting large vulnerable populations during disasters. For that, governments and other benefactors will continue to play a role, and are likely to have to retain these contingent liabilities, acting as insurers themselves by covering either fully or partially the costs of providing financial protection.
The Building Blocks of a Better System
The virtues of moving towards ex ante financing of disaster risk cannot be emphasized enough. Away goes the primacy of the begging-bowl financing model, leaving space for the emergence of a system that focuses on well-defined, credible plans, with better decision making and clarity. The programmes under way in Kenya, Ethiopia, and Mexico, while different, demonstrate how good pre-disaster planning and preparation by governments can help all those affected by a natural disaster get back on their feet as quickly as possible.
However, governments cannot typically afford to protect everyone fully against every possible disaster. Similarly, if government leaves it to insurance companies to try to persuade people to pay the full cost of insurance against disasters themselves, experience suggests that very few will end up being protected. But there is good experience with governments joining in partnership with private insurers to offer subsidized insurance for agriculture or for property to cope with natural disasters. Subsidies tend to cover things such as investments (p.92) in data, awareness, and education, and often part of the insurance premium as well, and offer a way for government to offer to pay its share if vulnerable people are willing to pay their share as well. For governments that would otherwise be called on to play the role of benefactor after a disaster, subsidized insurance solutions can offer a planned alternative that provides better incentives and faster support.
Other instruments are also available and have been used by various countries. In recent years, countries or organizations acting on behalf of them have bought insurance products on the international market. For example, in 2006 the World Food Programme bought a parametric policy for Ethiopia from AXA that would have paid out US$7 million in case of drought, but it was not renewed. Likewise, beginning in 2008, Malawi experimented with the use of a parametric drought insurance policy, but discontinued that experiment after several years. In other examples, Mexico has been buying catastrophe bonds linked to FONDEN for some years, and the Caribbean Catastrophe Risk Insurance Facility recently purchased one as well.
Development banks and other organizations have also been devising promising instruments. Targeting middle-income countries, the World Bank has developed a Deferred Drawdown Option for Catastrophe Risks (Cat DDO). This loan can be triggered when a natural disaster occurs. It offers immediate post-disaster liquidity, conditional upon having a sound disaster risk management programme in place. Ten countries have made use of this instrument.18
Groups of countries, with donor support and assistance from international organizations, have also set up specialized institutions targeting disaster risk management. The Caribbean Catastrophe Risk Insurance Facility (CCRIF), established by the Caribbean Community in 2007 with support from the World Bank and a range of donors, sells parametric earthquake and tropical cyclone coverage, and more recently excess rainfall coverage. Products pay within fourteen days of an event and are designed to offer quick liquidity to governments for emergency response and reconstruction. For the 2015–16 season, it sold insurance to sixteen Caribbean island states and one Central (p.93) American country, all but two of whom pay the full premium themselves.19 The Pacific Catastrophe Risk Assessment and Financing Initiative (PCRAFI), a partnership between the Pacific Community, development banks, donors, and scientists, has provided five Pacific island states with subsidized insurance against earthquakes and tropical cyclones since 2013.20 The African Risk Capacity, set up as a specialized agency of the African Union, provides rainfall-based parametric drought insurance, initially to four African countries in 2014.21 And, after Ebola, it is likely that some countries and international organizations will establish an insurance facility for pandemic risks, providing insurance based on the incidence of particular highly infectious diseases in developing countries that could pose risks for pandemics.
And there will always be a need for back-up in foreseen circumstances, or when even the most carefully designed plan fails. For such cases, ex post instruments may be necessary, and here too the details matter. Budget reallocations may involve taking money from projects with a high social benefit, thereby making the opportunity cost of funds quite high. Or, at the other extreme, the opportunity costs of funds could be very low, such as in the aftermath of the 2015 Nepal earthquake. Discussions with the government suggest that the emergency procurement processes actually enabled it to invest budget funds that were not disbursed.
Post-disaster fund-raising can also be implemented better or worse, depending on how contributors arrange their finances. Arranging funding beforehand (instead of resorting to unpredictable appeals) will pay here as well in terms of speed and effectiveness. For example, over the last decade the United Nations Central Emergency Response Fund has operated as a contingency fund that can spend relatively quickly in response to appeals to the humanitarian system, targeting underfunded humanitarian crises. Its annual budget is typically US$400 million. This has been an important innovation in the move towards pre-disaster funding, even though the fund is relatively small compared with the total annual humanitarian funding of approximately US$10 (p.94) billion. This sort of approach to funding global appeals could be developed further by, for example, expanding the size or moving further towards an objective, transparent, rules-based approach based on monitoring both natural disasters and the effectiveness of national responses.22
NGOs will also continue to play their role as back-up. And here we also see helpful innovation, moving away from post-disaster begging bowls, fragmentation, and benefactor behaviours. For example, the Start Fund is a pooled funding mechanism managed by nineteen UK-based NGOs providing relatively small-scale funding to NGOs for underfunded emergencies. It is supported by Irish Aid and the UK Department for International Development.23 Its explicit purpose is to ensure fast disbursement based on collective decision making.
These market and non-market based instruments will not guarantee better responses to disasters. Finance or bigger budgets are not in themselves the key problem to be resolved. But the existence and growth of this plethora of pre-disaster instruments for risk transfer do present the possibility of achieving credibly pre-financed, well-defined plans in a more cost-effective and reliable way. And that may then also reduce the need for a large back-up system for natural disasters based on post-disaster fund-raising and other begging bowls. It would also create the space for the humanitarian system to focus its efforts and fund-raising more on the less tractable humanitarian challenges, such as those in conflict settings.
1. Financial and budgetary instruments are the glue that holds credible plans together and makes them strong enough to withstand the whirlwind of highly charged post-disaster politics.
2. When designing and implementing disaster risk financing strategies, details matter. Financial experts add value. It is important to pay for financial advice and build in-house expertise.
4. The triggers in the financial strategy should match the triggers in the plan. Traditional reinsurance can be particularly useful for locking in plans for reconstruction, and indexed reinsurance can play the same role for financing indexed early actions.
5. Partially subsidized financial instruments can be used to encourage others to contribute towards the cost of well-defined plans.
A Snapshot of the Literature
The economics literature focuses on two rationales for governments transferring disaster-associated risk to regulated insurers, reinsurers, or capital markets. The first rationale is that, by swapping uncertain contingent liabilities for a more predictable premium or fee, risk transfer can reduce a government’s budget volatility. This rationale draws on early theoretical work on the use of financial markets to intermediate risk sharing by Debreu (1954), Borch (1962), and Arrow (1971). Although Arrow and Lind (1970) argue that national governments should be able to diversify risk perfectly and therefore should have no need for risk transfer, Ghesquiere and Mahul (2007: 2010) demonstrate that this argument does not hold in two specific cases: where a country is exposed to potential disaster losses that are large relative to the national wealth, and where risk-retention instruments available for immediate post-disaster funding needs are imperfect.
In practice, the former argument does not seem to be very strong, even for the extreme case of vulnerable small island states. Von Peter et al. (2012) use a large panel data set of 203 countries and jurisdictions over fifty-two years to argue that having larger insurance payouts after a disaster helps post-disaster economic recovery, but because they do not consider the cost of insurance they are unable to argue whether, on average, insurance was beneficial to growth. Bevan and Adam (p.96) (2015) address this issue by applying a recent macroeconomic model developed by the International Monetary Fund’s Research Department (Berg et al. 2012) to Jamaica’s tropical cyclone risk. They find that ex post tax-financed reconstruction of public assets would be slightly more cost-effective than insurance, even though it would lead to slower restoration of the public capital stock. Essentially, their finding is that even though Jamaica is a small island state, exposed to large potential shocks, these shocks are not large relative to national income, and smoothing the cost of the shock over time at the national level is slightly more cost-effective than insurance, a finding reminiscent of Gollier (2003). There is empirical support for the argument that risk-retention instruments for immediate post-disaster funding needs are imperfect (World Bank 2014a; 2015). This means that there does seem to be a budget volatility motivation for risk transfer for immediate post-disaster funding needs, but not for the bulk of reconstruction costs, which are not incurred immediately after a disaster.
The second rationale for government to engage in risk transfer is in situations in which the beneficiary pays part of the cost of protection and insurers and reinsurers are better able to implement risk-based pricing than government. Under risk-based pricing, investments in risk reduction would lead to lower-cost protection, thereby increasing the attractiveness to the beneficiary of investments in risk reduction with a positive net present value (Picard 2008; Jaffee and Russell 2013; Charpentier and Le Maux 2014; Kunreuther 2015). Governments do seem to find risk-based pricing challenging to implement in the absence of independent public or private institutions empowered to implement it (Cummins and Mahul 2009).
One further argument for risk transfer, so far inadequately researched, is as a commitment device—that is, a way in which government can commit itself to data systems and rules for how money will flow after a disaster, protect the system against discretion and fraud, and protect public policy from the dilemma of time inconsistency as discussed in Chapter 4. The use of financial instruments as commitment devices is not new and is, for example, the main explanation for (p.97) the success of commitment savings products observed across developing countries (Armendáriz and Morduch 2010). In the case of disaster risk financing, this rationale has been argued by practitioners such as the World Bank (2014a) and Phaup and Kirschner (2010), but as yet there has been no empirical analysis of its merits.
Three arguments are commonly made against government engaging with regulated risk-transfer providers for the joint management of disaster risk: (1) the provider may risk insolvency (Charpentier and Le Maux 2014); (2) for catastrophe risk that is difficult to quantify, regulated risk-transfer providers may not be able to offer attractive prices (Carter 2013; Kunreuther 2015); and (3) government may face a lack of competition (von Ungern-Sternberg 2004).
A variety of arguments have also been made for government providing subsidies or making insurance compulsory. Hill et al. (2014) and Carter et al. (2014) argue that subsidies for agricultural insurance can be designed to correct a clearly stated and well-documented market failure or equity concern. This argument builds on a number of careful impact evaluations of index insurance programmes for farmers, which support, in turn, the argument that credible indexed protection can instil good incentives for climate-smart, risk-sensitive investments. For example, Karlan et al. (2012), Mobarak and Rosenzweig (2012), Janzen and Carter (2013), and Elabed and Carter (2014) all find that index insurance significantly boosts investment or reduces asset decumulation. However, demand for insurance from farmers is still quite low in the absence of subsidies (see, for example, Cole et al. 2013) and quite sensitive to recent experience (Cole et al. 2014).
The low penetration of disaster insurance is, however, not restricted to agricultural insurance in developing countries. In the absence of subsidies, Pomeroy (2010) reports that only 12 per cent of homeowners’ policies in California include earthquake insurance despite the compulsory offer rule whereby any insurer who sells homeowners’ insurance must explicitly give clients the option to include earthquake insurance in their policy. Kunreuther and Pauly (2006) (p.98) point to the fact that only 40 per cent of the residents of Orleans Parish had flood insurance before Hurricane Katrina slammed into New Orleans in 2005. They propose compulsory but subsidized insurance for homeowners as a policy option, motivating the arguments of Jaffee and Russell (2013).
A wide range of practical experience around the world in designing and implementing credible ex ante financial plans for disasters is documented. A rich literature has emerged in the United States, stimulated by the rapid growth in federal government involvement in financial protection against disasters (see, for example, Kunreuther and Michel-Kerjan 2009) and in developing countries (Gurenko et al. 2006; Cummins and Mahul 2009; World Bank 2014a; 2015). McCulloch et al. (2015) consider the potential application of disaster risk financing solutions by humanitarian actors and microfinance institutions, and Mahul and Stutley (2010) and Carter et al. (2014) provide recent analyses of the conceptual and practical opportunities and challenges for government support of agricultural insurance. Private-sector risk carriers regularly produce research on the global market for disaster risk transfer, some of which is publicly available (see, for example, Swiss Re 2015).
(3.) A reinsurance policy is essentially an insurance policy for insurers. A catastrophe bond is an insurance-linked security in which payment of interest or principal is suspended or cancelled in the event of a specified catastrophe, such as an earthquake.
(5.) Beyond insurance, there are other insights from finance that are relevant to financial planning for disasters. For example, if a plan requires delivering grain to drought-stricken individuals it would be important to manage both price and quantity risks. Price risk is typically best implemented through (p.119) dynamic financial management where a portfolio of financial instruments is actively rebalanced as time passes and prices change to hedge the price risk.
(8.) Insurance and reinsurance prices are typically the result of a negotiation. Risk-reduction investments will not automatically reduce the cost of insurance and reinsurance, but, if well explained in these negotiations, they should typically result in premium reductions.
(9.) Captive insurance companies are insurance companies established by a parent company with the specific objective of covering the risks to which the parent is exposed.
(13.) Cabinet Decree No. 17 of June 5, 1991 (Panama), Supreme Decree 007-2008-Housing (Peru), and Article 48, paragraph 63 of Law 734 of 2002 (Colombia).
(14.) Specific examples are France’s Catastrophe Naturelles, the Japanese Earthquake Reinsurance Company, the New Zealand Earthquake Commission, Norway’s Norsk Naturskadepool, Spain’s Consorcio de Compensacion de Seguros, the Florida Hurricane Catastrophe Fund, the Hawaii Hurricane Relief Fund, the California Earthquake Authority, the Taiwan Residential Earthquake Insurance Pool, and the Turkish Catastrophic Insurance Pool (TCIP). See Cummins and Mahul (2009).
(15.) Following the 2011 Christchurch earthquake, which caused about US$17 billion in insured losses, the government of New Zealand injected further capital into the New Zealand Earthquake Commission, an earthquake reinsurer guaranteed by the government, to enable it to pay about half of the total claims (see <http://www.eqc.govt.nz/canterbury-earthquakes/progress-updates/scorecard>). It also bailed out the second-largest insurance company in the country, AMI Insurance (Muir-Wood 2012). Following Hurricane Wilma, which struck Florida in 2005, the Poe Financial Group declared insolvency, and claims had to be paid by the state-sponsored reinsurer, the Florida Hurricane Catastrophe Fund, and by the state-sponsored (p.120) insurer of last resort, the Citizens Property Insurance Company (Linnerooth-Bayer et al. 2011).
(17.) There are some exceptions, however. The Turkish government established a compulsory, unsubsidized earthquake insurance pool for homeowners’ buildings, and the combination of compulsion and clever, well-funded awareness-raising has led to some 6.8 million buildings being insured (TCIP 2013) out of a total of 19.5 billion, with owners paying the full cost of earthquake insurance themselves (Turkish Statistical Institute 2013).
(18.) Cat DDOs are in place in ten countries for a total value of US$1.38 billion: $7 million to the Seychelles in 2014; $102 million to Sri Lanka in 2014; $250 million to Colombia in 2011; $50 million to El Salvador in 2011; $66 million to Panama in 2011; $500 million to the Philippines in 2011; $100 million to Peru in 2010; $85 million to Guatemala in 2009; $150 million to Colombia in 2008; and $65 million to Costa Rica in 2008.
(19.) The premium for Haiti was paid for the 2015–16 season through a grant provided by the Caribbean Development Bank. All other Caribbean countries paid the full premium themselves. Nicaragua paid the premium through a concessionary loan from the International Development Association (IDA), the World Bank fund for the poorest countries. Since its inception, CCRIF has paid thirteen claims totalling US$38 million.
(20.) For 2015–16, the Cook Islands, Republic of the Marshall Islands, Samoa, Tonga, and Vanuatu all purchased parametric catastrophe insurance coverage against major tropical cyclones and earthquakes (including tsunami). Since its inception, the PCRAFI has paid two claims totalling US$3.2 million.
(21.) For 2014–15, Kenya, Mauritania, Niger, and Senegal all purchased parametric insurance coverage against drought. Since its inception, the ARC has paid three claims totalling US$25 million.
(22.) UN OCHA FTS data.