Chapter 2. Boosting financial resilience against disasters: Towards better management of disaster-related contingent liabilities

This chapter offers governments a step-by-step guide to boosting financial resilience against disasters through better management of their disaster-related contingent liabilities. For each proposed step, this chapter presents and reflects upon the findings of practices obtained through the nine case studies. It shows how governments can assess disaster-related contingent liabilities and benefit from including the results in fiscal planning and fiscal risk assessment processes. The chapter highlights good practices and provides a discussion of the persisting challenges governments face in increasing their financial resilience to disasters.

    

Introduction

While chapter 1 focusses on the economic and fiscal impacts of natural disasters, this chapter turns to the question of how governments can improve their management of disaster-related contingent liabilities with a view to increasing their financial resilience to natural disasters. The framework presented offers governments a step by step guide to better identifying, quantifying and managing their disaster-related contingent liabilities. Each step includes a set of findings on governments’ achievements to date and highlights good practices with a view to informing governments’ financial management strategies going forward.

The major steps in managing disaster-related contingent liabilities are listed below and illustrated in Figure 2.1:

  • Identify contingent liabilities. Identifying of contingent liabilities establishes a baseline for the sources of financial commitments that may arise for a government when a disaster occurs. This step should not only make clear the sources of the government’s explicit commitments to assume disaster-related costs, but also analyse potential implicit commitments, i.e. implicit liabilities.

  • Quantify contingent liabilities. Once the sources have been established, it is desirable to estimate the level of these commitments under given disaster scenarios, based on disaster risk models that estimate return periods. Estimating the size of a government’s liabilities can provide an informed basis for policy making before disasters occur.

  • Integrate contingent liabilities within a fiscal context. In the case of a severe disaster, it may be important to know whether the public financial impacts could cause fiscal distress for the government. For this reason, disaster-related contingent liabilities should be considered in overall fiscal forecasting and fiscal risk analysis.

  • Evaluate mitigation efforts. Evaluating a government’s efforts to mitigate disaster-related liabilities and finance residual fiscal risk is indispensable as part of an integrative assessment of a government’s contingent liabilities.

  • Disclose contingent liabilities. In a final step, governments should disclose their disaster-related contingent liabilities as well as the way their approach to managing them. Such disclosure can build confidence that these liabilities are well managed.

Figure 2.1. Managing disaster-related contingent liabilities within public finance frameworks
Figure 2.1. Managing disaster-related contingent liabilities within public finance frameworks

Source: Authors

The sections that follow elaborate these steps and document and compare governments’ practices.

Identification of disaster-related contingent liabilities

Distinguishing explicit from implicit contingent liabilities

Assessing a government’s disaster-related contingent liabilities requires the identification of both explicit and, to the extent possible, implicit liabilities. Figure 2.2 provides an overview of the sources of both kinds of liabilities.

Figure 2.2. Sources of a government's disaster-related contingent liabilities
Figure 2.2. Sources of a government's disaster-related contingent liabilities

Source: Authors.

Explicit contingent liabilities, as indicated above, are payment obligations based on government contracts, laws or clear policy commitments that fall due in the event of a disaster. To identify explicit contingent liabilities arising from disasters, it is necessary to understand the legal and policy frameworks that determine a government’s obligations to shoulder the costs caused by disasters. Explicit liabilities can arise from both central and subnational government commitments.

The legal framework, which includes laws, regulations and contractual obligations1 with external entities, provides the basis for identifying explicit commitments for recovery payments that have been made by governments prior to any disaster. As an example, Box 2.1 shows the diversity of such commitments in Japan. Policies that were announced prior to a disaster but are not yet reflected in legal obligations should also be considered as part of this assessment. The laws, regulations, policies and contracts that determine explicit contingent liabilities:

  • Define the central government’s legal responsibility to finance post-disaster response and recovery, including cost-sharing arrangements between the central and subnational governments. More specifically:

    • They define the central government’s legal responsibility to pay for public asset reconstruction and maintenance as well as its explicit liability with respect to providing recovery financing for damaged or destroyed private assets. Where legal requirements related to the insurance of public assets exist, this assessment should analyse the government’s exposure to uninsured losses, including co-insurance or deductibles, and consider the level of compliance with the legal requirements.

    • They define the legal liability of the central government to pay for the recovery of assets owned by regional, local or municipal governments, as well as historical experience in this regard.

    • They establish government guarantees for disaster losses incurred by public corporations and public-private partnerships (PPPs).

Box 2.1. Legal frameworks guiding Japanese government commitments in disaster assistance

In Japan, a number of laws recognise the government’s legal or explicit commitment to support disaster response, reconstruction of public and certain private assets, and social and economic restoration:

  • The Disaster Relief Act (1947) provides for disaster relief and welfare support (including repair of private housing, cash transfers and/or loans); it also establishes subsidisation of local government measures.

  • The Disaster Countermeasures Act (1961) sets out central and local governments’ responsibilities for disaster risk management and defines fiscal mechanisms for disaster response, e.g. subsidy, tax and debt measures.

  • Other laws, such as the Act on Special Financial Support to Deal with Extremely Severe Disasters (1962) and the Act on Support for Livelihood Recovery of Disaster Victims (1998), further extend the scope of the government’s financial responsibility. Following the Great East Japan Earthquake in 2011, the latter law was amended to increase the central government’s responsibility for disaster relief – shared with local governments – from 50% to 80%.

A series of laws provides for government support to certain lines of insurance (earthquake, agricultural, fisheries, fishing boat and forest) and establishes a central government contingent liability to pay a portion of reinsurance payouts under these schemes. For example, in the case of earthquake insurance, the Japanese government is responsible for a specific share of the losses covered by the Japan Earthquake Reinsurance. The share of losses borne by the government increases with the amount of overall losses and is reassessed periodically based on the capacity of the insurance sector to cover earthquake losses.

Source: Sato and Boudreau, 2012.

Explicit government commitments for disaster assistance: A range of policies and practices

Governments’ formal conditions for explicit disaster-related contingent liabilities include a mix of practices (Table 2.1). Central governments may have a wide range of explicit commitments to finance post-disaster costs, such as in Japan, or only a limited set of explicit commitments, such as in Colombia.

Japan makes explicit commitments to finance the recovery and rehabilitation of nationally and subnationally owned public infrastructure. It also provides explicit assistance to affected households for housing rehabilitation and temporary tax relief. Credit guarantees are provided for small and medium-sized enterprises (SMEs), complemented by safety net loans to overcome temporary cash-flow problems. The government also retains a portion of the liability of the Japan Earthquake Reinsurance (see Box 1.4).

Table 2.1. Governments’ explicit commitments to provide post-disaster financial assistance

Legal responsibility of central govt. to finance disaster response & recovery

Cost-sharing arrangements between central and subnational govts. to finance disaster response & recovery

Legal responsibility of central govt. to reconstruct/ maintain central government–owned public assets

Explicit liability of central govt. to finance rehabilitation & reconstruction of private assets

Legal liability of central govt. for other expenses incurred by subnational govts. (e.g. payments to businesses or individuals)

Govt. guarantees for disaster losses incurred by public corporations & PPPs

Australia

X

Canada

X

X

X

X

Colombia

X

X

X

X

X

Costa Rica

X

X

X

France

X

X

X

X

X

Japan

Mexico

X

New Zealand

Partially

Partially

Peru

X

X

X

X

Source: Case study reports.

In contrast to the Japanese example stands Colombia, where the legal framework provides for only some explicit public commitments to finance post-disaster assistance. Colombia’s practice is interesting: the government has a legal responsibility to finance post-disaster response and recovery, but no detailed commitments are specified. In contrast to both (and indeed to most economies) is Peru, which does not even establish a general, legally stipulated responsibility for financing post-disaster response and recovery (shown in column 2 of Table 2.1). In actual practice, however, the Peruvian government has regularly paid for the rehabilitation and reconstruction of public infrastructure, as well as compensated the losses incurred by the poorest population groups, hinting at the existence of quasi-explicit commitments or liabilities. France is another interesting example in this regard. While the authorities point to only a very limited set of explicit disaster-related contingent liabilities, across ministries and levels of government various disaster assistance schemes are in place. When activated, assistance through these schemes can be quite substantial, as was for instance the case following Hurricanes Maria and Irma, which struck the French Antilles in late 2017.

Commitments for recovering and replacing central versus subnational public infrastructure assets

The case studies reveal that all governments, explicitly or implicitly, provide finance for the recovery and reconstruction of assets owned by the central government. Many governments have made an effort to reduce this liability by conducting risk and vulnerability assessments, and by incentivising disaster risk reduction. For example, the Fund for Natural Disasters (FONDEN) provides financial assistance for public assets damaged by a disaster, but it reduces the level of assistance when the same asset is damaged more than once and has not been insured. Other governments assume the liability to finance recovery of damaged public assets on more of an ad hoc basis, with no specific financing provisions for meeting these contingent liabilities in place. For subnationally owned infrastructure assets some governments have detailed prescriptions for sharing costs arising from damages across levels of government. Other governments negotiate the process on a case-by-case basis and have few ex ante provisions in place for financing any assistance that is given (Note: more on cost-sharing mechanisms can be found in Chapter 3).

Limited government assistance for state-owned enterprises or PPPs

Most governments have strived to limit their assistance for damages incurred by state-owned enterprises. In New Zealand, for example, government assistance is not usually available for state-owned enterprises, as they are expected to maintain sufficient insurance cover and emergency reserves to manage their risks. Only exceptional hardship may be considered for government financial support. In Colombia and Costa Rica the central governments do not provide guarantees for disaster losses incurred by public corporations and public-private partnerships. In Peru, obligations for financing disaster recovery of state-owned enterprises are delegated to subnational government levels. In contrast to these practices, the Mexican government supports the rebuilding of federal infrastructure, including that owned by the state productive enterprises.

Varying government commitment to support household recovery

Government support for the damages suffered by individual households can influence the degree to which households invest in disaster prevention measures, creating a potential moral hazard. A wide-ranging commitment to support households after a disaster may discourage individual households from investing in preventative action, while clear guidelines for and better communication about possible post-disaster assistance could increase households’ preventative contributions.

Practices in providing post-disaster aid to households could not be more different. While most of the studied governments provide some level of immediate relief assistance (e.g. temporary shelter and food), practices differ significantly when it comes to providing financial support for the rehabilitation and reconstruction of private assets. For example:

  • In Australia, support for private households affected by a disaster goes beyond emergency response needs and includes compensation for the cost of demolishing and rebuilding houses.

  • In New Zealand, the central government makes a commitment to compensate private asset losses through its public earthquake insurance scheme (although only for those who have purchased the government insurance).

  • Without having a clear legal obligation to do so, the government of Colombia has regularly compensated the affected population for loss of private houses.

  • The government of Peru has tended to compensate the poorest households for costs incurred by the destruction of private assets, but this action has no explicit legal basis.

  • In Mexico, assistance is provided for low-income households through FONDEN, which pays compensation of up to USD 6 500 per household for completely destroyed homes, up to USD 1 600 for partially destroyed homes, and around USD 300 for homes suffering minor damage. The payment is based on the extent of damage and is independent of any preventative measures households may have taken.

  • In Japan, a post-disaster subsidy for housing (up to USD 30 000) is paid by the central and subnational governments to affected households, independent of their income. Japan’s government is also among the few governments that make an explicit commitment to pay compensation to families of disaster victims (up to USD 50 000); for low-income households, compensation can be as high as USD 127 000. Japan’s explicit individual household support also includes a temporary tax relief option.

  • In France, basic relief items may be financed from an emergency relief fund, as well as from the relief fund for overseas territories (FSOM). The FSOM may also be used to assist uninsured households with the costs of disaster reconstruction; the exact eligibility criteria are decided following each disaster.

The role of post-disaster financial support to businesses in limiting a disaster’s economic impact

Governments’ post-disaster financial support to private businesses varies widely. For SMEs that have experienced significant damage to their assets or that have foregone a significant amount of income, many governments (such as those of Japan, Australia and Canada) provide loans at concessional interest rates or interest rate subsidies. In Japan, for example, this explicit assistance includes credit guarantees under a central government safety-net programme, which provides loans for SMEs facing temporary cash-flow problems. Canada has established the Small Business Financing Program, a loan loss-sharing programme between the government of Canada and private sector financial institutions that is designed to increase the availability of financing to small businesses. Each year, the programme helps small businesses access close to CAD 1 billion (7.6 billion USD) in financing for post-disaster recovery. In other economies, such as France, the government’s post-disaster assistance to businesses focuses on preventing wage disruptions. Through a French Labour Ministry programme, for example, employers whose operations have been disrupted due to disasters or other exceptional circumstances can request public funding to help pay salaries.

Governments’ limited effort to encourage disaster prevention by businesses

Governments consider post-disaster support to businesses an important way to limit the scope and duration of the disaster’s negative economic impacts. However, the case studies revealed that most governments have done little to encourage businesses to adopt preventative measures prior to a disaster. In most economies, commercial property and business interruption insurance coverage rates are assumed to be very low, but there appear to be few or no systematic studies by governments designed to understand why businesses tend not to implement measures to reduce disaster risk, such as business continuity plans or physical asset protection measures. Such studies could help governments in promoting disaster prevention efforts among businesses thereby effectively limiting eventual government liabilities.

Tailored support to the agricultural sector

In many economies, tailored post-disaster assistance for the agricultural sector is available. In Mexico, Canada and France, for example, special funding frameworks have been established to support agricultural producers in bearing the costs of major disasters. In Mexico, such support is limited to low-income farmers without insurance who are affected by climate-related hazards, while in Canada support is more broadly available: under Canada’s AgriRecovery Framework, federal and subnational governments have established an initiative that provides farmers with targeted assistance to help cover the extraordinary costs of recovery from natural disasters. The French national guarantee fund for agricultural disasters has been set up to provide compensation to agricultural businesses that suffered uninsurable losses due to natural hazards or disease outbreak. While such assistance programmes make up only a limited share of a government’s overall assistance payments, they should nevertheless be structured with a view to incentivising loss prevention to avoid repetitive or even increasing damage payments.

A rising trend in implicit government commitments due to extreme events

Implicit contingent liabilities, as stated earlier, refer to post-disaster expenditures the government makes due to a perceived moral obligation or political pressure – that is, in the absence of any previous formal commitment to pay for them. Implicit government liabilities are therefore much more challenging than explicit liabilities to identify and to define. Broadly speaking, sources for identifying implicit liabilities can be one or more of the following:

  • Any established practice that may be considered de facto policy.

  • Any informal or implied indications of fiscal support in government statements.

  • Any informal or implied indications of fiscal support in interactions between individual ministers, ministries or agencies and third parties, such as “letters of comfort” to third parties with respect to a public corporation, a subnational government or a PPP partner that suggest the government is well-disposed towards, generally supports or is in favour of an activity, project or entity. These might be implicit contingent liabilities if, for example, such an entity suffered losses in a disaster that impaired its ability to service its debt or deliver important services.

  • Strong media coverage in the immediate aftermath of a disaster. This coverage may increase implicit contingent liabilities by increasing moral and political pressure for governments to provide assistance beyond the explicit commitments previously made. This pressure likely increases as knowledge about the magnitude of the disaster is spread and reiterated (Kunreuther and Michel-Kerjan, 2013).

  • Any elements of the above that limit or offset the gross fiscal impact of the exposures. These include ceilings on coverage or compulsory insurance of public assets such as those of government-owned public corporations or subnational governments.

As these broad categories indicate, the types of assistance that have been provided under more implicit commitments by governments are wide-ranging. Nevertheless, one strong commonality in these implicit liabilities involves their trigger: in many economies, exceptional disaster events are the circumstances that give governments the grounds for going beyond their initial and explicit commitments.

In Japan, the already wide-ranging explicit commitments have been even further expanded during “exceptional circumstances,” such as the Great East Japan Earthquake. This disaster – admittedly a major one – led the central government to shoulder a much greater share of the fiscal burden than it was legally obliged to. For example, all the disaster relief and recovery costs, which normally would have been shared with subnational governments, were in most cases shouldered by the central government. All other cost-sharing arrangements saw an increase in favour of the central government’s share. In addition, tax reductions were broadened, and assistance for SMEs was expanded beyond the legal stipulations.

A similar practice can be seen in Mexico, where the government has responded to exceptional disasters by going beyond its explicit commitments to provide post-disaster financial assistance. During past disasters, for example, it has issued zero coupon loans for subnational governments unable to meet the minimum funding requirement for subnationally owned public assets.

In response to the Canterbury earthquakes, the New Zealand government bailed out a private insurance company, AMI, and provided several welfare benefits to the affected population that were not based on any prior explicit commitments. In addition to a number of temporary tax relief programmes after the earthquake, the government has offered homeowners in high-risk earthquake zones buyouts of their homes at near-market value. Between 2011 and 2015, 95% of eligible property owners participated in that initiative (Mitchell, 2015).

In Australia, an explicit category under the Natural Disaster Relief and Recovery Arrangements (NDRRA) - category D - has been established for exceptional disaster events. This category gives government the green light to go beyond its explicit commitments spelled out in categories A to C and provide unlimited support for any kind of assistance deemed necessary under the exceptional circumstances. Such additional commitments were triggered, for example, for the dredging of a port after the 2010/11 Queensland floods. Measures that are financed under such implicit commitments have to undergo a special approval process by the prime minister.

Tendency of implicit contingent liabilities to be greater where explicit commitments ex ante are limited

In economies like Peru and Colombia, where there are limited explicit commitments by the government to provide post-disaster financial assistance, implicit liabilities that are assumed tend to make up a larger share of disaster-related contingent liabilities. Indeed, the government of Peru has previously provided substantial and systematic financial support, including support for the affected population, for the rehabilitation and reconstruction of public assets, and for the reconstruction of private assets for the poorest households. Additional ad hoc actions by the government included cash transfers to each local government in emergency areas, the expansion of grants to protect vulnerable households and an authorisation for the housing ministry to deliver temporary housing solutions to affected citizens. All emergency decrees on which these interventions were based were valid for a limited time, typically for less than a year after approval.

Quantification of disaster-related contingent liabilities

To effectively manage and plan for disaster-related contingent liabilities within a government’s public financial framework, it is useful to quantify or estimate their potential size. There are two suggested approaches to estimating the size of potential government contingent liabilities arising from disasters: direct estimation and estimation through probabilistic modelling.

Direct estimation of disaster-related contingent liabilities

A first approximation of a government’s expected disaster-related liabilities can be obtained by gathering information from existing resources on past government spending on disasters. This includes, but is not restricted to, the following: historical data on government expenditures (e.g. the resources spent via dedicated disaster risk management funds, disaster-related contingent credit lines or catastrophe bonds and specific budget lines for disaster risk management spending), government-backed insurance claim payments and government guarantees for public (or public-private) corporations that materialised. All of the governments reviewed can use at least a number of these sources to establish a baseline understanding of their eventual financial liabilities in the event of a disaster.

Historical data on direct government expenditures in response to past disasters can be a first step in estimating the probability of future losses and the expected size of fiscal costs for governments. The longer the period covered by the historical data and the larger the volume of coverage, the more reliable the estimate is likely to be. In Japan, for example, public spending in response to disasters has been recorded and publicly disclosed for some time. Table 2.2 shows that a quite comprehensive list of direct spending items gets recorded in Japan. In addition to the detailed ex post spending records, Japan has compiled a solid basis of ex ante risk reduction spending information. Figure 2.3 tracks these figures since 1980.

Table 2.2. Information from past disasters as a basis for quantifying disaster-related contingent liabilities: The case of Japan

Type of disaster-related expenditure

What gets recorded

Relief spending

Temporary housing, medical care, waste disposal, dispatching of Self-Defence Forces, etc.

Spending for the reconstruction of damaged public infrastructure and assets

Recovery/reconstruction of infrastructure assets, public schools, government buildings, etc.

Spending for the reconstruction of damaged private assets

Financial support for livelihood recovery of disaster victims, provision of disaster condolence grants, support for the reconstruction of agricultural facilities, etc.

Spending on increased social transfers due to a post-disaster economic slowdown

Items such as school attendance support, tuition support, expansion of job creation programmes and unemployment assistance

Expenditures due to guarantees issued to public or private entities suffering disaster losses

Earthquake reinsurance claims, disaster risk insurance for agriculture and fishery, credit guarantee for small and medium-sized enterprises

Post-disaster payments to subnational governments

Subsidy to disaster-affected subnational governments

Reduced tax collections

General changes in tax revenue published in the highlights of the general account budget document and in the accompanying documentation on Japan’s fiscal condition

Disrupted operations of public corporations

Not included

Disrupted operations of private corporations

Not included

Deterioration in the terms at which the government can in the short term refinance public debt or raise additional debt

Not included

Source: Case study reports

Figure 2.3. Disaster prevention and recovery expenditure in Japan, 1980–2016
Figure 2.3. Disaster prevention and recovery expenditure in Japan, 1980–2016

Note: The figures for the 2016 fiscal year are preliminary figures reflecting the initial budget.

Source: Cabinet Office Japan, 2016.

When the government maintains a dedicated disaster risk or catastrophe fund, payouts from the contingency appropriation are a useful source of information for estimating the size of potential liabilities. Examples of such funds obtained through the case studies can be found in Table 2.3.

Table 2.3. Sources for identifying and quantifying disaster-related contingent liabilities: governments’ catastrophe funds and cost-sharing programmes

Types of funds1

Australia

- Natural Disaster Relief and Recovery Arrangements (NDRRA) provide financial assistance from the central government, reimbursing up to 75% of eligible expenditure on relief and recovery payments made by subnational governments.

Canada

- Disaster Financial Assistance Arrangements (DFAA) provide financial assistance from central government, reimbursing up to 90% of eligible expenditure on relief and recovery payments made by subnational governments once a minimum expenditure threshold has been met.

Colombia

- National Disaster Risk Management Fund finances knowledge generation about risk, risk reduction, risk management, recovery and financial protection activities.

- National Adaptation Fund is dedicated to financing disaster risk prevention.

Costa Rica

- National Emergency Fund (Fondo Nacional de Emergencia, NEF) provides funding for disaster recovery measures.

France

- Emergency relief fund finances assistance for immediate disaster relief

- Relief fund for overseas territories (FSOM) finances assistance for the reconstruction of uninsured private assets and uninsurable subnational assets, and for immediate disaster relief

- National guarantee fund for agricultural disasters (FNGRA) finances compensation for uninsurable crop losses due to natural hazards or disease outbreak

- The CATNAT insurance scheme is backed by a state-guarantee

Japan

- The Annual Reserve for Disaster Recovery and a non-earmarked contingency reserve in the general account budget are two reserves available to finance the cost of disaster recovery

Mexico

- FONDEN (Fund for Natural Disasters) finances ex post disaster risk management measures.

- The Fund to Support the Rural Population Affected by Climate Hazards provides support to low-income farmers who do not have agricultural insurance and who are affected by climate-related hazards

New Zealand

- The Natural Disaster Fund is an accumulated technical reserve in the earthquake insurance scheme

Peru

- FONDES (Fund for Interventions to Face Natural Disasters) finances both ex ante and ex post disaster risk management measures.

- FONIPREL (Promotion Fund for Regional and Local Public Investment) may also be used for financing disaster risk management measures.

- The Fiscal Stabilisation Fund can be used to finance national emergencies that affect Peru’s fiscal stability.

Note: List of funds and programmes is not exhaustive.

Source: Case study reports.

Another useful source of information could be an insurance programme or a credit guarantee programme for small-scale farmers that includes a disaster cover and that has been operating long enough to generate a solid record of annual claims paid by the government. The JER provides a good example here. The Japanese government shares the liabilities arising from this privately provided disaster risk insurance with the private sector based on specified thresholds of total insurance claims. For example, for total insurance liabilities between USD 2 billion and USD 103 billion, the government will assume 99.5% of the costs. In New Zealand, the EQC provides an unlimited guarantee to compensate private asset losses from natural hazards (including earthquakes, tsunamis, landslides and flood impacts on land) emanating from its public earthquake insurance scheme.

Budget classifications and programmes can also be useful tools for identifying potential costs of disaster-related contingent liabilities (OECD, 2011), although expenditures for disasters in budget categories may not always be made explicit. A good practice in that regard can be found in Peru. Peru introduced the Budgetary Program 0068, a results-based budget that records spending on preparedness and prevention measures for disaster risk management. Although the majority of funding recorded under Budgetary Program 0068 is used for ex ante disaster risk management funding, the budget shows how central and subnational spending can be systematically recorded for disaster risk management and provides a basis for expanding the records towards ex post funding.

Estimation of disaster-related contingent liabilities through probabilistic modelling

In the absence of – or in complement to – existing records, probabilistic modelling can estimate a government’s potential exposure to disaster costs, i.e. contingent liabilities. Probabilistic modelling can also serve to estimate liabilities that could arise during more extreme loss events that are possible but not part of the historical record; governments could find this information helpful when deciding whether to take on new contingent obligations, or when analysing, communicating and managing the potential impact of existing exposures (GFDRR, 2014).

Increasingly, governments are using probabilistic catastrophe risk models, such as Hazus in the United States2 or the CatSim model developed by the International Institute for Applied Systems Analysis (IIASA),3 to gain a better understanding of their potential disaster losses and inform financial decision making. Probabilistic disaster risk assessments allow governments to estimate the total size of contingent liabilities for given disaster scenarios or return periods; they also provide estimates of the potential damage to government-owned buildings and infrastructure as well as to privately owned assets, which can be used to estimate the potential cost of compensation and financial assistance to individuals and businesses facing damages and losses. Where probabilities are known with reasonable confidence, this approach can be incorporated into cost-benefit analyses to determine which financial instruments are best suited to protect against potential fiscal shocks arising from disasters.

Some examples of the use of probabilistic models are provided here:

  • In New Zealand, the government carried out a one-off study to understand the worst-case impact a major disaster could have for the central government. The 2010 study modelled the fiscal impact of a 7.8 earthquake affecting its capital, Wellington, and found an estimated government contingent liability of USD 11 billion to finance response and recovery for three consecutive years following the modelled earthquake scenario. The Canterbury earthquakes proved this study useful, as the actual fiscal costs came close to the estimates established in the study’s model.

  • Australia’s Productivity Commission, the Australian government’s independent research and advisory body, has sought to understand the government’s future disaster-related contingent liabilities (in complement to the exercise of accounting for liabilities on the basis of past commitments through the NDRRA). It projects that the annual costs of disasters will increase from as much as USD 11.1 billion in 2018 to USD 11.5 billion by 2023.

  • As part of the development of a national seismic profile, the Ministry of Economics and Finance in Peru estimated the exposed value of state assets at USD 2.6 billion in order to calculate the probable maximum loss of a 1 000-year return period event.

  • France annually calculates the liability arising from the state guarantee of the CATNAT insurance scheme. To inform the definition of the state guarantee threshold, modelling is used to estimate the cost of disasters that might activate the guarantee.

    An important baseline for estimating the exposure of government-owned assets’ to disasters is the central asset inventory, which provides information on the number and type of assets, their residual value based on the quality of their maintenance and localised information on their natural hazard exposure. Table 2.4 provides an overview of current practices among the cases studied.

Table 2.4. Overview of public asset inventories across governments

 

Public asset inventory?

Australia

Yes, partial; central government–owned buildings are monitored through a central inventory and insured through Comcover

Canada

Not currently established

Colombia

Not currently established

Costa Rica

Yes, but currently considered outdated and underestimated; update is planned that will include information regarding assets’ location, value and insurance coverage

France

Yes, the general government account (Compte général de l’État, CGE) lists all public assets

Japan

Partial: Information is collected, e.g. on water infrastructure by river management authorities, but no national repository exists

Mexico

Yes, created in 2013; includes roads, bridges, water infrastructure, hospitals and schools (among others)

New Zealand

Information available scattered across ministries and agencies; not centrally collected

Peru

There is an information system on state-owned assets (Sistema de Informacion Nacional de Biens Estatales), but it cannot be used to estimate probable losses

Source: Case study reports.

The government of Colombia used the information on the estimated value of exposed public property (and of specific private property) to determine the magnitude of its disaster-related contingent liabilities. Estimated losses associated with public properties were defined as the government’s explicit contingent liabilities, while estimated losses associated with private properties for which the government chose to compensate losses were defined as its implicit contingent liabilities (Colombia Ministry of Finance and Public Credit, 2011).

Governments’ compilation of information to estimate their liabilities

As the discussion above has demonstrated, governments have a wide range of information that can help produce rough estimates of their potential disaster-related contingent liabilities. However, the case studies indicate that this information is not fully exploited in overall public financial planning for managing contingent liabilities:

  • In New Zealand, the government records information on the central and subnational governments’ past response and recovery spending, on spending for the reparation or replacement of damaged public infrastructure, on spending to increase welfare benefits following a natural disaster, on additional public resources allocated to recovery through special policies and on expenditures emanating from guarantees issued to the Earthquake Commission. However, this information is not compiled in a systematic manner for quantifying New Zealand’s overall fiscal exposure to disaster-related contingent liabilities.

  • Peru is a similar case. It has information on historical government expenditures for disaster risk management, as well as information from annual budget contingency appropriations and spending by the emergency management authority, but each set of information is separately maintained rather than combined to provide global estimates of the government’s potential financial outlay.

  • In Australia, central and subnational governments recognise the need to assess disaster-related contingent liabilities as part of budget planning (and, as will be shown in the following section, in their consideration of fiscal risk). Regular reviews are carried out based on past disaster-related expenditures and expected future expenditures resulting from past incidents. The NDRRA reimbursement requests are an important basis of information, and are complemented by an examination of subnational governments’ public accounts.

While it may be challenging for governments to quantify their disaster-related contingent liabilities, and in some contexts neither desirable nor feasible, governments are encouraged to have at least a qualitative sense of the level of these liabilities and their potential impact on government finances. To gain this sense, a simple classification method can be used that defines liabilities as probable, possible or remote, based on judgments about their likelihood (IMF, 2016). The method suggests combining such a classification with broad estimates about the potential significance of the contingency, which is the probability multiplied by the estimated disaster exposure, to categorise the arising fiscal risks as minimal, small, moderate or major. Given the level of uncertainty involved, this approach is inevitably crude, but it may nonetheless help to focus attention on areas of greater potential risk. In Australia, for example, the annual Statement of Risks published with the budget each year contains a category of contingent liabilities described as “significant but remote”, which receives a qualitative discussion when the budget plans are drawn up4.

Estimation of fiscal impacts of disaster-related contingent liabilities and their integration in overall fiscal forecasting and fiscal risk assessments

Once disaster-related contingent liabilities are identified, and to the extent possible quantified, this information can be used to design and inform adequate financial planning and instruments as well as regulatory mechanisms. In a subsequent step, the potential comprehensive impact of disaster-related contingent liabilities on an economy’s fiscal prospects can be evaluated as part of fiscal forecasts and fiscal risk assessments. Whether an economy chooses to engage in all or some of these steps might depend on the extent of the possible impact of disasters. Hence the information and practices described here are offered as guidance mechanisms rather than prescriptive policies to be applied across all governments.

A number of methods can be used to estimate the impact of contingent liabilities on the fiscal balance of an economy. The simplest method is to analyse the sensitivity of the current forecast path for public debt or the fiscal deficit to a disaster shock, for example, an assumed shock to the public finances of a certain percentage of GDP from a disaster (OECD, 2014). A standard sensitivity analysis of public debt sustainability can be conducted over a five-year period based on selected shocks. The size of a disaster-related shock could be varied to assess the sensitivity of public finances to shocks of different magnitudes, such as 1%, 5% or 10% of GDP. For example, the Philippine 2013 fiscal risk statement (Republic of the Philippines 2013) included a debt sustainability analysis that incorporated scenario analyses such as the occurrence of large disasters (see Figure 2.4). The 2015–17 Multiannual Macroeconomic Framework in Peru considered the macroeconomic and fiscal consequences arising from a severe El Niño episode; the analysis suggested that the debt and deficit would increase slightly compared to the baseline (IMF, 2015).

Figure 2.4. Philippine national government debt-to-GDP scenario analysis
Figure 2.4. Philippine national government debt-to-GDP scenario analysis

Source: Republic of the Philippines, 2013.

Sensitivity analysis has the advantages of methodological simplicity, limited data requirements and easy communication of results through tables or fan charts. It is commonly used to illustrate the sensitivity of public finances to small changes in macroeconomic parameters (GDP, interest rates, etc.), taken one at a time, or, when combined in plausible combinations, in alternative macroeconomic scenarios.

However, sensitivity and scenario analyses also have a number of shortcomings. In the context of exposure to the risk of a disaster shock, one of the more serious shortcomings is the lack of data on the probability of the shock. Another is the failure to account for interaction between shocks; much like other fiscal risks, a major disaster can cause an economic slowdown that exacerbates the initial shock to public finances, which can in turn potentially trigger further explicit and implicit contingent liabilities.

In a recent detailed survey and analysis of contingent liability realisations of all types, the International Monetary Fund (IMF, 2016) concluded that most fiscal risk scenario and sensitivity analyses tend to explore only modest-sized shocks and do not consider complexities in the impact on revenues and expenditures. In response it has put forward a new fiscal stress test methodology intended to analyse the impact of very large shocks. The stress scenario is forward looking rather than based only on past experience. One of its key elements is the range and likelihood of both explicit and implicit contingent liabilities, and their interaction with large macroeconomic shocks. Disasters are one source in the contingent liability component of the stress test but, especially for advanced economies, not the largest. The framework, however, can incorporate the scenario of a very large disaster occurring during a period of an economic shock already playing out, in turn triggering both explicit and implicit contingent liabilities. An example of such a scenario is the Great East Japan Earthquake, which according to the Japan Cabinet Office was a “crisis in the midst of a crisis” for the Japanese economy and its public finances (Ranghieri and Ishiwatari, 2014, citing Cabinet Office Japan, 2011). This very large disaster occurred during a period of prolonged economic underperformance and when public debt levels were already very high, forcing the government to incur additional debts and increase taxation levels in an already tense fiscal environment (Ranghieri and Ishiwatari, 2014; Sato and Boudreau, 2012).

Practices showing strong recognition of major disasters in overall fiscal management

In several of the case study governments, fiscal management practices take major disasters into account. For example:

New Zealand’s fiscal strategy aims at attaining a high level of fiscal resilience, taking into account a wide range of risks, including natural hazards, that the government may be exposed to. In addition to producing a detailed official account of government expenditures following major disaster events, the Treasury integrates the impact of major disasters in its fiscal forecasting and fiscal risk analysis as part of its annual budget reporting process, in which it sets desired fiscal buffers to withstand economic shocks that have been identified. The Treasury regularly measures the impacts of a number of key stress events, including the fiscal impact of a major disaster such as a major earthquake affecting Wellington (estimated to cause a USD 11 billion liability for the government). In this exercise, the Treasury evaluates the impact on net worth and net debt to GDP by modelling the combined worst-case outcomes stemming from the occurrence of a natural disaster during a financial crisis. While the specific costs of the scenarios are not quantified for the purposes of budget estimates, the impacts are considered in the development of the overall fiscal strategy.

The Australian government has likewise recognised the importance of the potential fiscal impact of a major disaster. It explicitly acknowledges disaster-related contingent liabilities, defined as potential costs to the central government arising outside of its control, in its annual Statement of Risks. In complement to this, the government budget estimates provide for expected ongoing payments through the NDRRA, the main source of funding from the central to subnational governments covering response and recovery costs after disasters. To arrive at longer-term projections of the future expected costs of disasters, the government of Australia conducts projection exercises and holds qualitative discussions to evaluate the disasters’ potential fiscal impacts. Although it has no standard procedures to evaluate the macro-fiscal scenario that follows a combination of extreme events, these discussions allow the government to consider worst-case scenarios, such as the major Queensland flood that occurred during an economic downturn.

As part of the federal budgeting process, Mexico assesses (and develops a strategy for managing) the most relevant fiscal risks, namely short- and long-term macroeconomic risks and various contingent liabilities, which include those related to natural disasters. Owing to their potential significant impact on public finances, disasters are one of the long-term risks the government regularly assesses and considers in both medium- to longer-term fiscal forecasts. The results of these exercises, presented in the General Economic Policy Guidelines, are taken into account in the annual budgeting decisions on Mexico’s major disaster fund, FONDEN.

In Japan, the Cabinet Office publishes an annual economic and fiscal outlook for use in preparing the budget of the next fiscal year. This outlook discusses medium- and long-term projections of government revenues and expenditures, including projected spending to continue recovery and reconstruction efforts for major disasters that occurred in previous fiscal years, such as following the 2011 Great East Japan Earthquake. Aside from this, fiscal impacts of potential future disaster-related contingent liabilities are not integrated in fiscal forecasting documents.

In Colombia, the potential impact of major disasters figures prominently in its fiscal impact and risk assessment. Analysis has shown that the potential fiscal impact of a major disaster is the second-largest fiscal risk the government can expect to deal with (legal actions pose the largest fiscal risk). As a result, the government has formally recognised natural disasters as a fiscal risk and has mandated their integration in fiscal risk assessments and in efforts to design a broader fiscal risk management strategy (World Bank, 2012).

Benefits of increasing visibility in the fiscal policy-making process

The examples above demonstrate that there are many ways to consider government disaster-related contingent liabilities in fiscal impact and fiscal risk assessments. While some governments have based their approach on quantitative methods forecasting future potential worst-case scenarios, others have studied and highlighted the impact of past events. Governments also integrate the results of these exercises in annual fiscal statements in many different ways. Some governments require precise quantitative forecasts of the impact of potential major disasters on a number of fiscal indicators, including debt, while others stick to a more qualitative discourse.

Even though fiscal risk assessment methods may differ, the objectives are very similar for all governments. Most aim at increasing the visibility of disaster-related contingent liabilities in the fiscal policy-making process. To further enhance this visibility, existing knowledge on the physical impact of extreme events – often available from disaster risk managers – could be used to inform the public financial analysis of potential worst-case scenarios. Greater internal dialogue between risk managers and financial officials would help to increase synergies. Similarly, although finance officials have recognised the importance of analysing potentially concomitant shocks (e.g. a disaster occurring during an economic downturn), sharing information on the physical impact of extreme disaster events and analysing governments’ contingent liabilities based on these scenarios could give a more complete picture of the worst case governments would potentially have to deal with.

Disclosing government’s disaster-related contingent liabilities

By disclosing disaster-related contingent liabilities – and the strategy for managing them – governments create trust in their capacity to manage the financial impacts of disaster risks. While identification is a prerequisite for disclosure, the scrutiny that comes with disclosure creates pressure to ensure that risks continue to be identified, and once identified are estimated and managed. It also helps to unlock additional information from parties outside the central agencies, and perhaps outside government, that may help identify (and quantify) fiscal risks, including contingent liabilities from disasters (IMF, 2009). Disclosure also promotes earlier and smoother policy response, increases trust among stakeholders in the quality of fiscal management, reduces uncertainty for investors and taxpayers and may as a result improve access to international capital markets (IMF, 2008). In a similar manner, publishing the government’s strategy for managing disaster-related contingent liabilities subjects the strategy to beneficial scrutiny and provides an incentive to maximise the strategy’s use. In specific cases, clearly defined exemptions to disclosure may be required, for example to minimise moral hazard, avoid negative economic side effects or avoid disadvantaging the economy in negotiations. Reporting on implicit contingent liabilities might also be inappropriate if it creates a sense that post-disaster assistance is unconditionally guaranteed.

Disclosure can take place through mechanisms for incorporating the potential fiscal impact of disaster-related contingent liabilities into budgeting and fiscal reporting or the overall fiscal risk management strategy. These mechanisms apply both to individual government entities and levels of government as well as to central government as a whole. The more important mechanisms for disclosure include these:

  • the annual budget call circular, which may require line ministries to provide information on contingent liabilities, for internal monitoring purposes and possibly for publication in annual budget documents;

  • ministry of finance documentation, including the medium-term fiscal framework, fiscal risk reports or stand-alone reports on contingent liabilities;

  • regulations such as requirements to report contingent liabilities to the ministry of finance;

  • a fiscal responsibility law requiring regular public disclosure of contingent liabilities (possibly alongside other fiscal risks).

Earlier findings on OECD governments’ contingent liabilities and commitments in their budgeting and fiscal reporting showed that disaster-related contingent liabilities are not always featured (OECD, 2016). The results of the case studies undertaken for this report are consistent with this result: disaster-related contingent liabilities are not systematically reported; or else the reporting is limited to a qualitative mention or to ongoing recovery payments rather than future expected government outlays:

  • Colombia formally recognises disaster-related contingent liabilities as a major fiscal risk, but the Medium-Term Fiscal Framework, which requires the government to include explicit expected contingent liabilities, does not include them.

  • In New Zealand, all contingent liabilities valued at more than USD 73 million need to be individually reported annually in the audited Notes of the Financial Statements of the Government and in the Budget Economic and Fiscal Update (BEFU). The guarantee of the Earthquake Commission is included in the BEFU as one of the central government’s contingent liabilities. However, this contingency is considered as unquantifiable, and the BEFU includes it without specific value, offering instead a brief description of its nature and a note on whether it has changed since the previous report.

  • Australia’s government discloses information on its explicit disaster-related contingent liabilities in its Statement of Risks (part of the Budget Papers). Future disasters are recognised as an unquantifiable contingent liability in the budget documents.

  • Mexico requires the disclosure of the most relevant fiscal risks in the annual General Economic Policy Guidelines, which inform the central budget planning process. Fiscal risks that need to be referred to include natural disasters. In addition, FONDEN allocations for post-disaster reconstruction by disaster and sector are publicly disclosed online.

  • Although Peru requires the publication of the government’s explicit contingent liabilities, it has only recently started to mention disaster-related contingent liabilities in its Multiannual Macroeconomic Frameworks.

  • In Japan, the disclosure of disaster-related contingent liabilities is limited to an outlook on projected expenditure for ongoing recovery and reconstruction efforts stemming from disasters that occurred in previous fiscal years, such as following the 2011 Great East Japan Earthquake. Fiscal impacts of potential future disaster-related contingent liabilities are not included.

References

Cabinet Office Japan (2016), “White Paper: Disaster management in Japan 2016”, http://www.bousai.go.jp/kyoiku/panf/pdf/WP2016_DM_Full_Version.pdf.

Cabinet Office Japan (2011), “The guidelines on policy promotion for the revitalization of Japan”, Cabinet Office, Tokyo.

Colombia Ministry of Finance and Public Credit (2011), “Contingent liabilities: The Colombian experience”, Bogotá, http://treasury.worldbank.org/bdm/pdf/Contingent_Liabilities_Colombian_Experience.pdf.

GFDRR (2014), Understanding Risk in an Evolving World: Emerging Best Practices in Natural Disaster Risk Assessment, World Bank, Washington, DC, www.openknowledge.worldbank.org/handle/10986/20682.

IMF (2016), “Analyzing and managing fiscal risks—Best practices”, International Monetary Fund, June, www.imf.org/external/np/pp/eng/2016/050416.pdf.

IMF (2015), “Peru: Fiscal transparency evaluation”, IMF Country Report, No. 15/294, International Monetary Fund, Washington, DC, http://www.imf.org/external/pubs/ft/scr/2015/cr15294.pdf.

IMF (2009), “Fiscal risks: Sources, disclosure, and management”, International Monetary Fund, Fiscal Affairs Department.

IMF (2008), “Good practices in fiscal risks disclosure: International experience”, presentation at the High Level Conference on Fiscal Risks, Paris, 28–29 October, www.imf.org/external/np/seminars/eng/2008/fiscrisk/pdf/velloso.pdf.

Kunreuther, H. and E. Michel-Kerjan (2013), “Managing catastrophic risks through redesigned insurance: Challenges and opportunities”, in Handbook of Insurance, Georges Dionne (ed.), www.opim.wharton.upenn.edu/risk/library/J2013Kunreuther+Michel-Kerjan_Managing-Cat-Risksthrough-Redesigned-Insurance.pdf.

Mahul, O. et al. (2014), “Financial protection against natural disasters: From products to comprehensive strategies”, World Bank, Washington, DC, https://www.gfdrr.org/sites/gfdrr/files/publication/Financial%20Protection%20Against%20Natural%20Disasters.pdf.

Mitchell, M. (2015), “Relocation after disaster: Engaging with insured residential property owners in Greater Christchurch’s land-damaged ‘Residential Red-Zone’”, Brookings-LSE Project on Internal Displacement, Brookings Institution, Washington, DC, https://www.brookings.edu/research/relocation-after-disaster-engaging-with-insured-residential-property-owners-in-greater-christchurchs-land-damaged-residential-red-zone/.

OECD (2016). “Accrual practices and reform experiences in OECD economies: Results of the 2016 OECD Accruals Survey”, OECD Journal on Budgeting, Vol. 2016/1, https://doi.org/10.1787/16812336.

OECD (2014), “Fiscal risks: New approaches to identification, management and mitigation”, GOV/PGC/SBO(2014)4/Final, 12 August, http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=GOV/PGC/SBO(2014)4/FINAL&docLanguage=En.

OECD (2011), “Future global shocks: Improving risk governance”, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264114586-en.

Ranghieri, F. and M. Ishiwatari (2014), Learning from Megadisasters: Lessons from the Great East Japan Earthquake, World Bank, Washington, DC, www.openknowledge.worldbank.org/handle/10986/18864.

Republic of the Philippines (2013), “Fiscal risks report 2013”, http://www.dbm.gov.ph/wp-content/uploads/DBCC_MATTERS/FiscalRiskStatement/FRS_2013.pdf.

Sato, M. and L. Boudreau (2012), “The financial and fiscal impacts”, World Bank, Washington, DC, https://openknowledge.worldbank.org/handle/10986/16165.

Notes

← 1. “Contractual obligations” refers to contracts made between governments and nongovernment organisations providing services to the government, and to contracts with entities providing infrastructure and other services under public-private partnerships.

← 2. See FEMA, “Hazus”, https://www.fema.gov/hazus.

← 3. See IIASA, “CATSIM,” http://www.iiasa.ac.at/web/home/research/researchPrograms/RISK/CATSIM.en.html.

← 4. Government of Australia, “Budget 2016–17. Statement 8: Statement of Risks (continued)”, www.budget.gov.au/2016-17/content/bp1/html/bp1_bs8-05.htm.

End of the section – Back to iLibrary publication page