Insurance, Developing Countries and Climate Change

More than three-quarters of recent economic losses caused by natural hazards can be attributed to windstorms, floods, droughts and other climate-related hazards, which appear to be increasing at a greater rate than geophysical disasters.

This trend can be largely attributed to changes in land use and increasing concentration of people and capital in vulnerable areas, for example, in coastal regions exposed to windstorms and in fertile river basins exposed to floods.

As indicated by the greater increase in weather-related disasters compared to geophysical disasters, climate change also appears to be a factor in increased disaster losses.

The Intergovernmental Panel on Climate Change has predicted that climate change will increase weather variability as well as the intensity and frequency of weather-related extremes. There is also mounting evidence that climate change is contributing to increasing current risks.

In the past quarter-century, over 95 per cent of deaths from natural disasters occurred in developing countries, and direct economic losses (averaging US$100 billion per annum in the last decade) in relation to national income were more than twice as high in low-income as opposed to high-income countries.

These disaster statistics do not (for the most part) reflect long-term indirect losses, which can be very significant, particularly in countries with little capacity to respond and recover.

Not only are there considerable differences in the human and economic burden of disasters in developed versus developing countries, but also in insurance cover.

In the richest countries about 30 per cent of losses in the period 1980–2004 (totalling about 3.7 per cent of Gross National Product (GNP) were insured; in low-income countries, only about 1 per cent of losses (amounting to 12.9 per cent of GNP) were insured.

Owing to the lack of insurance, combined with exhausted tax bases, high levels of indebtedness and limited donor assistance, many highly exposed developing countries cannot raise sufficient capital to replace or repair damaged assets and restore livelihoods following major disasters, exacerbating the impacts of disaster shocks on poverty and development.

Developing countries with very low catastrophe insurance penetration represent a challenging and under-served market for the private insurance sector.

Entrepreneurs are beginning to find ways to provide insurance for the lower end of the market, particularly through micro-insurance products that are made accessible by support from civil society and the public sector.

This market is only feasible if premiums are affordable to the poor, which opens an opportunity for negotiators seeking opportunities for helping the most vulnerable adapt to climate change.

The case for including insurance and other risk-transfer instruments in a climate adaptation strategy builds on a growing recognition that the developed world, because of its emissions of greenhouse gases, is contributing to weather-related losses in the developing world.

The so-called Copenhagen Agreed Outcome, which will determine the new post-2012 international climate regime, is expected to include both targets and action plans for the reduction of greenhouse gases (mitigation) as well as a framework to facilitate adaptation to the negative effects of climate change that can no longer be mitigated.

It is under the adaptation agenda that insurance solutions are now under serious consideration. The MCII has proposed a climate risk management module to be included in the adaptation agenda.

Fully funded by an adaptation fund or other financial mechanism emerging from climate negotiations in Copenhagen in 2009, this module would provide support for weather-related disaster prevention and insurance in vulnerable countries.

This paper presents the insurance ‘‘pillar’’ of MCII’s proposed risk management module. As background, we begin in the next section by discussing the status of micro- and sovereign-catastrophe insurance programmes currently in place for serving developing country households, businesses and governments.

In the third section, we examine their benefits, risks, costs and affordability. We argue that the private sector, acting alone, cannot provide adequate security to low-income clients, an argument that forms the rationale for support from a climate adaptation strategy.

The rationale for including insurance in a climate adaptation regime rests not only on the failure of the market to serve the most vulnerable, but as we argue in the fourth section, it is also based on the prospect that insurance mechanisms can help reduce the impacts of these events.

In the fifth section, we describe the MCII proposal for a twotiered Insurance Pillar, financed by a Copenhagen Agreed Outcome financial mechanism, to (1) absorb a part of the high-level risks in vulnerable countries; and (2) enable micro- and sovereign-insurance systems to absorb middle layer risks.

This Insurance Pillar would be part of a broader climate risk management strategy, which includes an interlinked prevention pillar.

A. Disaster risk financing in developing countries

Insurance instruments are only one of many activities involved in managing risks of natural hazards. The first, and arguably the highest priority in risk management, is investing in preventing or mitigating human and economic losses.

Disaster mitigation and prevention can take many forms: reducing exposure to risks, (e.g. land-use planning); reducing vulnerability (e.g. retrofitting high-risk buildings) or creating institutions for better response (e.g. emergency planning).

The residual risk can then be managed with insurance and other risk-financing strategies for the purpose of providing timely relief and assuring an effective recovery. Importantly, insurance can be designed to reward preventive behaviour, and in this way it can directly contribute to disaster loss reduction.

Most commercial disaster insurance is held by citizens of high-income countries (per capital income greater than US$9,361), although even in these countries less than a third of disaster losses are insured. Not surprisingly, the picture is quite different for countries outside of the high-income bracket.

Insurance density drops from around a third to less than a tenth in emerging economy countries, and it is almost negligible (1–2 per cent) in low-middle and low-income developing countries.

As pictured below, in the U.S., parts of Europe and Australia, the average person pays over US$500 in property insurance premium compared to Africa and parts of Asia with less than US$5 in premium (Figure 1).

Instead of insurance, households and businesses rely extensively on post-disaster public assistance. This is the case even in high-income countries. In the U.S., for instance, the federal government provides extensive assistance to private victims.

Taking the 1994 Northridge earthquake as an example, only about 30 per cent of total direct private and public losses were absorbed by private insurance companies.

In stark contrast, in the U.K., which claims 75 per cent flood insurance penetration, the government gives little assistance to private victims. As noted above, insurance is practically non-existent in least developed countries, and public assistance tends to be far lower.

As a typical case, after the severe flooding in Sudan in 1998, the victims themselves absorbed over 80 per cent of the losses. In addition to relying on public assistance, households, farmers and governments in the developing world have many opportunities for financing their recovery after disasters.

As shown in Table 1, individuals can take out emergency loans from their family, micro-credit institutions or money lenders; sell or mortgage assets and land or rely on public and international aid.

Likewise, governments raise post-disaster capital by diverting funds from other budgeted programmes, borrowing money domestically or taking loans from international financial institutions.

Individuals may also make arrangements before the disaster: setting up mutual arrangements with family, microsavings and food storage. Similarly, governments can spread risks temporally or spatially by putting a reserve fund in place, making contingent credit arrangements or forming regional pools.

These informal mechanisms for financing disasters can be less costly and thus more affordable and accessible to very low-income individuals and governments.

Yet, although informal financing appears to work reasonably well for low-loss events, it is often unreliable and inadequate for catastrophic events. At the public level, it is well known that if governments can spread their post-disaster costs over a large tax base, or other lower-cost financing strategies, they should be risk neutral and not purchase insurance.

However, in the aftermath of heavy devastation in their countries, lowincome developing countries may face exhausted tax bases, little reserves and declining credit ratings making external borrowing difficult.

Finally, external assistance is limited, and with the exception of highly publicised disasters it is usually inadequate to meet post-disaster needs. International support for victims of the 2006 Indian Ocean tsunami was estimated at about $7,000 per affected victim, which was exceptional.

On average, international post-disaster assistance has approximated 10 per cent of direct economic losses,13 and it can be much less. For example, support for victims of the devastating floods affecting Bangladesh in 1998 was estimated at about US$3 per affected victim.

Nor can governments rely on postdisaster assistance. As a typical case, 2 years following the 2001 earthquake in Gujarat, assistance from international sources had reached only 20 per cent of original commitments.

The inaccessibility of sufficient and affordable capital to support the recovery process in highly vulnerable countries is the main rationale for donor organisations, and also for a climate adaptation regime, to provide assistance to insurance programmes.

Switching from post-disaster humanitarian assistance to providing predisaster security through insurance instruments also has benefits to donors.16 Because insurance instruments can provide strong incentives for reducing risks, a point we will cover later, ex-ante support of insurance eventually reduces the need for outside assistance.

B. Disaster insurance for developing countries

Donors and international financial institutions are increasingly supporting insurance systems in the developing world. A number of innovative pilot programmes are providing insurance to farmers, property owners and small businesses, as well as transferring the risks facing governments to the international capital markets.

Examples include index-based crop and livestock insurance systems in Malawi and Mongolia, property insurance in Turkey; a regional catastrophe insurance pool for the Caribbean Island States,21 and the issuance of a catastrophe bond by the Mexican government.

These and other donor-supported insurance systems are for the most part still in a pilot stage, and none has experienced a major and widespread catastrophic event. It is too early, thus, to fully assess their effectiveness in reducing economic insecurity.

In examining this early experience, the broader question arises whether developing countries should, indeed, follow the path of the developed world in insuring against catastrophic events, and which insurance instruments and modifications may be appropriate for better tackling the developmental dimensions of natural disasters?

This question is especially topical given the insurance controversies following Hurricane Katrina’s devastation of poor communities in New Orleans. In what follows, we briefly discuss the benefits, risks, costs and affordability of disaster insurance based on early experience in developing countries.

1. Benefits

By providing low-income households, farmers and businesses with the right to postdisaster liquidity, thus securing their livelihoods, insurance instruments can lessen the burdens from disasters and expedite the recovery process.

For many, an insurance contract is more dignified and secure than dependency on the ad hoc generosity of donors. As insured households and farms are more creditworthy, insurance can also promote investments in productive assets and higher-risk/higher-yield activities.

Insurance instruments, if designed carefully to avoid moral hazard, can also provide incentives to reduce risk, a point we will return to in the next section. For governments, insurance instruments can also have large pay-offs by reducing the risk of a post-disaster financial gap and thus ensuring the timely repair of public infrastructure and provision of relief expenditures.

Just like investments in prevention, timely relief and reconstruction can save lives and livelihoods and prevent disaster-induced poverty traps. With internationally backed risk-transfer programmes, developing country governments will rely less on debt financing and international donations, and assured funds for repairing critical infrastructure will attract foreign investment.

2. Risks

As recent and past experience in developing and developed countries shows, there are also risks to an insurance strategy. Broadly, these risks can be categorised as resulting from:

the potential insolvency and non-sustainability of insurance systems;
moral hazard, adverse selection and basis risk; and
institutional stability, public confidence and trust.

In the absence of strong regulatory frameworks, many micro-insurance systems operating in low-income communities have insufficient backup capital and are thus exposed to high insolvency risks. These risks can be reduced by strengthening market regulation and also by providing outside support to ensure the solvency and stability of local and national systems.

For example, the World Bank has created a contingent credit facility to provide backup for the Turkish Catastrophe Insurance Pool (TCIP), which will cover risks that an earthquake occurs before sufficient premium has been collected.

This support can increase public confidence and trust in the system. Moral hazard can also be effectively addressed by setting up index-based systems.

In the Malawi pilot micro-insurance project, for example, insurance claims by smallholder farmers are triggered by precipitation falling below a prescribed level as measured by local weather stations.

Not only is moral hazard reduced, but also there is no need for expensive individual claims settling, and expedient payments will reduce the need for farmers and herders to sell their assets and livestock to survive the aftermath of a disaster. However, basis risk remains a problem.

Finally, in developing countries institutional stability and trust can be an issue. In Malawi, interviews with the participating bank revealed a strong distrust in the stability of the partner insurance company.

Without World Bank involvement, the bank interviewee revealed that the bank would not participate in the programme. This lack of institutional trust can be a constraining factor in up-scaling these systems beyond the donor-supported pilot phases.

3. Costs

The benefits of insurance make it a potentially integral part of an overall disaster risk management strategy.

However, these benefits can be costly. Insurers offering cover for co-variant risks face large, stochastic losses and thus must hold capital reserves, diversify or purchase reinsurance, all of which ‘‘load’’ or add to the actuarially riskadequate technical premium.

Moreover, providing insurance on a small scale involves high administrative costs in reaching clients, estimating and underwriting risks, and handling claims. This is also the case for sovereign risk transfer. The expenses of issuing the above-mentioned Mexican catastrophe bond, for example, amounted to about 2 per cent of the cover amount.

This substantially exceeds this cost for traditional reinsurance, which normally approximates 1 per cent.28 As shown in Figure 2, the costs of catastrophe insurance exceed the annual expected loss by an expense load and contingency load.

The expense load includes the transaction costs in administering the system: costs of starting up, underwriting, etc. The contingency load can be far greater, and includes not only the costs of equity capital and risk transfer, but also frictional costs and an uncertainty load.

According to Cummins and Mahul,29 frictional costs result from informational asymmetries between capital markets and the insurer’s management.

As global capital markets have less information about the insurer’s exposure to catastrophic risk and the adequacy of its loss reserves than do the firm’s managers, the capital market may charge a higher cost of capital to provide a margin for the informational asymmetry.

Adding to the problem of asymmetric information, insurers may load the premium to account for uncertainty and ambiguity in the risk estimates.

Because of the transaction and capitalisation costs, and arguably the extra premium that insurers demand to take on for ambiguous and uncertain risks, catastrophe insurance premiums can be substantially higher than expected losses.

4. Affordability

Many in the developing world cannot afford risk-based premium payments and remain dependent on post-disaster aid and other forms of financing discussed earlier and shown in Table 1.

The inability of very low-income clients to pay insurance premiums sheds doubt on a common view among development and climate adaptation experts, that is, that the private sector, if assured proper market conditions, can act alone to provide sufficient catastrophe insurance coverage throughout the developing world.

Proponents of this argument point to the success of emergent micro-credit and micro-insurance systems serving the poor. These programmes cover independent and often less costly damages, such as health and funeral expenses, and thus are more affordable than catastrophe cover premiums, which as discussed above reflect large uncertainties, ambiguities and capitalisation costs.

A review of catastrophe insurance coverage in Asia, Africa and Latin America shows that, almost without exception, programmes targeting the poor operate with subsidies, capitalisation or other forms of support from the government or international development agencies, or they offer very minimal cover.

The Disaster Preparedness Programme operating in India’s highly hazard-exposed Andhra Pradesh region provides an example. In partnership with a commercial insurer, this programme offers multiple-hazard insurance coverage for property and life risks to groups of women with a minimum size of 250 members.

Since 2000, the Indian regulatory authority has required insurers to service the low-income segment of society, and many insurers offer affordable contracts to low-income communities made possible by cross subsidies from their other lines of business and wealthier clients.

As a second source of subsidy, the U.K.-based donor NGO, Oxfam, paid 50 per cent of the premium in the first year. Furthermore, Oxfam actively convinced the private insurer to offer very low-cost insurance by training disaster management volunteers, who assist in providing insurance services such as helping communities in the claims process.

Other programmes are made affordable by offering limited cover. The nonsubsidised insurance system in Malawi covers only the bank loans for hybrid seeds and does not insure farmers’ livelihoods.

Moreover, start-up costs including the preparation of risk assessments, the business plan and suitable institutional arrangements were supported by the World Bank and World Food Programme. The above-mentioned TCIP is hailed as a non-subsidised national earthquake insurance programme.

Yet, premiums are indirectly reduced by a World Bank con tingent loan facility that absorbs a layer of risk at very low cost. Moreover, the system covers mainly middle-income property owners; very poor households in rural areas are excluded from the pool and continue to rely on post-disaster public relief.

Notably, with the exception of the TCIP, there are no national hazard insurance systems operating in the developing world, such as the public-private systems in the U.S., France, Norway, New Zealand, Japan and several other highly exposed developed countries.

These partnerships do not solve the fundamental problem that the citizens of poor, small and highly exposed countries cannot collectively afford to be paying members of a national risk or solidarity pool for extreme losses. International support is an important prerequisite for serving poorer countries and regions.

Location: Indonesia