A growing body of evidence shows that the world's poor suffer disproportionately from natural disasters, but do not benefit from risk hedging markets. This reality underpins the World Bank's work on natural disaster funding. It is intended to ensure that sufficient liquidity exists very soon after a disaster, that modern risk transfer mechanisms can help to speed economic recovery, and that governments' fiscal exposures to natural disasters are reduced. In addition, catastrophe risk management can assist countries in the optimal allocation of risk in their economy, contributing toward higher economic, and more effective poverty alleviation.

Historically, most governments have taken little interest in hedging against natural disasters (until a major event occurs) because of low perceived vulnerability levels and the fact that most severe natural hazards manifest themselves only infrequently. In addition, there has been a willingness on the part of rich countries and other donors to provide post disaster funding for those vulnerable countries which are subject to frequent events. The World Bank alone has provided well in excess of $7bn of post-disaster funding in the last 20 years.

This attitude is changing however, with the wealth disparities between country and city attracting people to huge, newly-formed industrial and service conglomerations. Changes in the environment are also having an effect. The 1999 World Disasters Report indicates that declining soil fertility, drought, flooding and deforestation drove 25 million ‘environmental refugees' from their land, into often vulnerable squatter communities in fast-growing cities.

The great new population and economic centres are almost always found in locations subject to windstorm, flood and/or earthquake, with South Asia clearly one of the prime examples. The region contains some of the most disaster-prone and vulnerable countries in the world. Of the 40 most deadly natural disasters in the last 30 years of the 20th century, South Asia (excluding Iran) accounted for 15. The Commonwealth Disasters Index, despite being developed to support a case for small states, still includes Bangladesh and India in the five countries most vulnerable to natural disasters.

However, in terms of direct insurance losses, South Asia accounted for none of the top 40 disasters. In 1999, one of the worst years on record for disaster-related insurance losses, South Asia did not rate a mention in the top 20 insurance losses, although in terms of lives lost it accounted for five of the worst events.

Effectively, South Asia is not participating in global disaster loss hedging markets. Despite having close to a fifth of the world's population, the region accounts for only 0.3% of the world's non-life insurance premiums. The region instead has been very dependent on donors. Of 102 natural disaster reconstruction projects approved by the Bank in the 20 years to 2000, 16 were for South Asian countries. Bangladesh and India accounted for 50% of all countries in the world with at least five post-disaster World Bank projects approved in that period.

This absence of adequate catastrophe insurance coverage, combined with rapid increases in the concentrations of urban populations and economic activities in disaster prone areas, makes disaster-related economic losses increasingly relevant in terms of fiscal management. While the countries of South Asia can certainly continue to count on some external funding from donors and international development lenders, the size of such assistance will increasingly fall short of the liquidity needs in the aftermath of major natural disasters. This became evident following the recent experience with the Gujarat earthquake, which has lead to a recasting of India's fiscal strategy and, despite some initial sanguine forecasts, appears likely to have had a measurable negative impact on the country's economic growth.

One caveat is that funding and hedging instruments have a practical limitation imposed by event frequency. Where the return period for an event which would have national consequences is thousands of years there is probably no constituency for risk management. At the other end of the spectrum, the frequency and severity may both be so high that funding and risk transfer are simply not an option. In this situation, mitigation and donor support are likely to be the only answers. Funding and hedging arrangements tend to work in the middle range of frequencies, somewhere in the annual probability range between 0.5 and 6.0%, although there is generally a maximum loss size that can be contemplated in practice, given fiscal reality.

Disaster funding models
A large, well diversified economy funds losses internally through the domestic direct insurance system (and its international reinsurers) and government transfer payments. For instance, most of the direct property losses arising from Hurricane Andrew in Florida were borne by US-based insurers. However, some industrialised country jurisdictions do not have sufficient diversification opportunities given the nature and potential severity of losses, and a number of national or regional/state level funding approaches have developed. These include post-funding, such as use of contingent lines of credit (for example, the Californian Earthquake Authority scheme) and pre-funding (for example, the New Zealand Earthquake Commission), usually in combination with reinsurance. Other variants can be found in France, Norway, Spain and Japan. Reinsurers, in turn, have been making increasing use of capital market instruments such as catastrophe bonds. In virtually all cases, the industrialised countries' schemes rest on the foundation of an established insurance sector with strong penetration of the insurable property base.

In the last five years, the international markets for risk hedging and loss funding have witnessed important positive changes that may further encourage the development of catastrophe risk management at national and sub-national level. First, catastrophic risk reinsurance capacity has increased and prices have, until recently, been declining. Second, insurance companies and investment houses have created new instruments to spread catastrophic risk directly to the capital markets. Third, and despite the first point, public and private initiatives to develop catastrophe risk transfer mechanisms beyond the traditional insurance and reinsurance vehicles have continued, driven by the memory of the three-to four-fold increase in prices that occurred in the global catastrophe reinsurance market following Hurricane Andrew in 1992.

In addition, the World Bank has been developing disaster funding models which are not dependent on the existence of a strong and well-entrenched insurance sector. These are typically based around government sponsored pools, supplemented by various reinsurance and capital market instruments. They are typically underwritten by the ability to call on registered or tax-paying property owners and, ultimately, but as a last resort, the government. The scheme which has generated most interest is the Turkish Catastrophe Insurance Pool (TCIP) under which Turkish-registered private residential property owners are obliged to buy a basic level of cover against earthquake loss. Since its launch on 27 September 2000, almost 2.3m earthquake insurance policies have been issued, making the TCIP the second-largest catastrophe pool in the world.

The TCIP operates primarily as a catastrophe risk transfer and risk financing facility (although it does have a mandate to promote mitigation measures and educate the public about earthquake risk), and in this role it has raised the financial preparedness of Turkey for future disasters, reduced government fiscal exposure to major catastrophic events, and made liquidity readily available to insured homeowners affected by such events.

While the TCIP has elements of the California and New Zealand programmes, which also provide earthquake coverage for homeowners and rely heavily on international reinsurance and capital markets, there are important differences. The TCIP's earthquake insurance cover is compulsory for Turkish homeowners, and, in contrast to the EQC program, a minimum 50% of the pool's surplus is to be invested in foreign assets to avoid a double loss (for example on the asset and liability side) in case of a major disaster. The pool provides a cover up to $20,000 (at the current exchange rate) for each dwelling, for a premium which varies across the country depending upon seismicity, local soil conditions, and the type and quality of construction. Local insurers act as distributors of the TCIP policies and provide additional coverage in excess of that offered by the pool.

This World Bank-supported insurance programme ensures that less advantaged, lower income groups of the Turkish population have access to a reliable catastrophe pooling mechanism. In addition the programme, designed for the riskier end of the property market, frees-up risk capital for private insurers, giving an impetus to the development of the local insurance industry, although consumer education remains a critical issue. The TCIP is backed by commercial reinsurance and a World Bank contingent credit line. A number of earthquake-prone countries have indicated interest in this model, including Greece, Taiwan, Algeria, India and Iran.

The South Asian initiative
The low level of insurance penetration in South Asia can be related to income levels only in part, given that other similarly poor regions have higher levels of insurance penetration. India alone has a significant middle class population (in absolute, if not relative terms), which would normally be sufficient to make that country a leading insurance consumer, and there appear to be significant market inefficiencies.

While the lack of fully competitive markets for insurance in South Asia (and in India's case, until recently a fully nationalised industry) has no doubt contributed to this situation, there may be cultural and historical issues at work as well. With the recent and proposed opening up of regional insurance markets, there may now be an opportunity to consider market or market-linked solutions to the transfer and funding of disaster risk and losses. The prospects of liberalisation and more efficient insurance markets has development significance, because delays in obtaining compensation for loss financing are harmful to prospects of prompt economic recovery, particularly when fiscal resources become severely stretched and critical infrastructure has been affected.

The World Bank Group stands ready to offer technical advice and supplementary financial resources to the countries of the region that are interested in building national disaster risk management programmes. It now offers, through either International Bank for Reconstruction and Development or International Finance Corporation stand-by contingent credit facilities, to finance semi-working or catastrophe layers of reinsurance programmes placed by either government-sponsored pools or private insurers and reinsurers in emerging markets. The Bank is also currently evaluating its potential role in facilitating countries' access to the capital markets to hedge their exposure to natural disasters.

As part of its development mandate in the region, the World Bank has recently initiated a study to assess the scope for the application of funding instruments to natural disasters in India, Pakistan, Sri Lanka and Bangladesh. For this purpose, an analysis will be undertaken to assess the frequency and severity of natural hazards in the region, in particular cyclones, earthquake and rainfall extremes. Information will be collected to estimate risk exposures according to hazard event frequencies as well as overall vulnerability of country assets at risk.

In conjunction with this, the financial exposure and estimated value of such assets will be estimated. This work will yield some preliminary loss exceedance curves and the resultant implications for contingent risk protection requirements. Preliminary inquiries indicate that detailed exposure and vulnerability information will be difficult to obtain and, initially at least, some broad estimation techniques will be required. The Bank has developed such techniques in carrying our recent similar studies for Central America.

The study will then determine to what degree the region's exposure to natural hazards is covered by the current insurance market and predictable donor funding, including the cost of such coverage, both present and historical.

Ultimately, the study will present recommendations as to the selection of hazards and locations which are amenable to modern funding and hedging instruments in South Asia Region. In addition, it will include a financial feasibility assessment of the most appropriate institutional and financing arrangements for national catastrophe risk management programmes in South Asian countries.