Natural disasters are caused by the exposure and vulnerabilities to natural hazards of people, infrastructure and economic activities. Analysis of these factors has permitted identification of countries and areas within them where disaster-related mortality and economic losses are likely in the future. These high-risk areas are candidates for increased attention to, and investment in, disaster risk identification, reduction and transfer.
Plans are underway to further identify disaster risk levels and factors on national and subnational scales in high-risk countries to create evidence for improved risk management decision-making. In this paper, I review selected recent global and regional risk analyses to highlight findings, areas for improvement and next steps in the overall process of using disaster risk information for more effective risk management and cost-effective reduction of losses.
One of the key challenges in promoting a shift from disaster to risk management is to make the risk factors that cause disasters more visible. Prior to disasters these causal factors may be hidden. Only after a disaster occurs does it become clear the extent to which latent risk factors were present. Unfortunately by then it is too late to prevent losses.
Identifying where and when conditions of hazardousness and vulnerability are present creates the potential for acting before disasters occur to reduce the risks. Risk information can also be used to support schemes to transfer financial risks away from exposed populations and assets or to reduce risks of future disasters during disaster recovery.
In the past several years a concerted effort has been mounted by international agencies seeking to promote an evidence-based shift from emergency to risk management. In this paper, I briefly review selected results from three recent disaster risk analyses: Reducing Disaster Risk: A Challenge for Development (UNDP 2004)2, Natural Disaster Hotspots: A Global Risk Analysis (Dilley et al. 2005)3 and Indicators of Disaster Risk and Risk Management: Program for Latin America and the Caribbean (Cardona 2005; IDEA 2005)4. Work on these analyses was coordinated through a working group on vulnerability, risk and impacts assessment as part of the International Strategy for Disaster Reduction.
These global and regional reports complement additional risk analyses that have been, and are being, undertaken at national and subnational scales. One hopes that this body of evidence may be approaching a critical mass that will begin to allow the factors that cause disasters to be more systematically addressed through the development process. Factors that affect the utility and effectiveness of evidence on disaster risk for risk management applications include the scientific quality of the analysis and the match between the analysis and decision-making alternatives. Since many of the most meaningful risk management decisions are made at the national-to-local levels, it is particularly important to strengthen the evidence base for disaster risk management within high-risk countries.
Following a review of recent risk analysis results, this paper concludes with two additional sections. One suggests ways for improving how loss and risk information are generated. The other identifies decision contexts in support of which further evidence creation on disaster risks and risk factors could be oriented. The intent is to suggest that developing greater capacity for cooperation among scientists, risk assessment practitioners and decision-makers, particularly in high-risk areas, would be a worthwhile contribution to reducing disaster losses.
2. Disaster risk identification: a review of recent global and regional results
Disasters are caused by the exposure of vulnerable communities, infrastructure and economic activities to natural hazards. Three recent global and regional disaster risk analyses use this theory of disaster causality to identify particularly high-risk countries and geographic areas. These reports have provided the institutions that sponsored them—including the United Nations Development Program (UNDP), the World Bank and the Inter-American Development Bank—with an objective basis for drawing attention to countries where disaster risk management is a priority.
In policy circles, evidence seldom if ever has a deterministic effect on decisions. Rather it may become one consideration among many affecting how decisions are taken. Although the reports reviewed below are imperfect, their imperfections do not undermine the usefulness they have demonstrated as a basis for negotiating the appropriate consideration that should be given to disaster risk reduction in the overall context of seeking to achieve important development goals. At the same time, there is no one-size-fits-all evidence base for risk management decision-making. Decisions on specific measures to undertake to reduce and transfer risks in high-risk areas will require qualitatively different, and significantly more accurate, risk analyses.
The following brief summary selectively highlights recently available disaster and risk information and its applications, as well as some of its shortcomings. Readers are encouraged to read the full reports and consult other risk and loss information resources before drawing conclusions on specific results summarized below.
All three reports have many limitations, further documented in the reports themselves. It should surprise no one that available data is inadequate to fully characterize risks, neither for any one country nor even less worldwide. Data shortcomings include spatial resolution, the length of available historical hazard and loss time-series and the selection of available variables.
All three studies involve the use of historical loss data in the calculation of various risk indexes. This type of data, contained in databases maintained by the Center for Research on the Epidemiology of Disasters (CRED), the Munich Reinsurance company, the Latin American network La Red and others, has only been systematically collected in the past several decades.5 Hazard frequency data sets also often only cover recent decades, a shortcoming that particularly affects assessment of geophysical risks and interpretation of the results. Other socio-economic statistics used in the calculation of vulnerability and risk management capacity are similarly limited. The result is that most risk assessments are heavily dependent on what has happened in the recent past and therefore do not provide adequate guidance concerning unprecedented events in the future.
Finally, it is possible to assess risks of many potential outcomes. The reports reviewed below focus on a limited number, including disaster-related mortality, economic losses and the potential gap between disaster losses and available financing. Other outcomes of interest could include impacts on poverty, on specific sectors, risks of failure of key structures and so on. Therefore, it is necessary to understand the types of decisions that a risk analysis is intended to support when assessing the adequacy of the data, methods and results for particular purposes.
(a) Reducing disaster risk: a challenge for development
This report by UNDP, released in 2004, assesses disaster-related mortality risks. Its primary purpose is advocacy. Countries are the unit of analysis. It identifies the relative contributions of hazard exposure and socio-economic vulnerability factors to mortality risk and analyses how these can be reduced or exacerbated through the development process.
Risks are characterized by a Disaster Risk Index (DRI), developed by the United Nations Environment Programme, that measures relative vulnerability to cyclones, floods and earthquakes. Relative vulnerability is measured as the deaths in each country associated with a particular natural hazard compared to the number of people exposed. Countries that have suffered higher loss of life in recent decades than others with similar exposure are considered to have a higher relative vulnerability to each hazard.
The value of the DRI is not so much to precisely rank countries by relative vulnerability, but rather the aggregate point it makes that countries with similar levels of hazard exposure nevertheless may experience quite different mortality outcomes. The report examines potential explanations for these differences in terms of how development can create or reduce risk.
Over the past 21 years, countries that have demonstrated relatively high vulnerability to earthquakes, for example, include Armenia, the Islamic Republic of Iran, Turkey, India, Italy, Algeria and Mexico. Countries whose mortality in relation to exposure suggest relatively low vulnerability include Chile, the United States of America, Argentina and Germany.
It is important to note that, in the case of earthquakes the analysis is based on observed events since 1980. Thus, the results reflect only earthquakes that have occurred in the recent past. During this period, however, it is clear that some countries have fared considerably better in terms of earthquake mortality than others with similar exposure levels. Additional indexes in the other two reports reviewed below use earthquake probabilities and calculated maximum probable losses rather than historical events to complement the DRI.
Tropical cyclones occur annually and their higher frequency allows the calculation of relative vulnerability to be based on a larger sample. Countries assessed as being relatively highly vulnerable to cyclones include Honduras, Nicaragua and Bangladesh. Relatively low-vulnerability countries include Australia, Japan and Cuba. Small island states are special cases with respect to cyclone vulnerability, due to the potential for covariate losses across wide swaths of the population and economy during cyclone landfalls. There are clear differences among small island states, however, in terms of mortality in relation to hazard exposure. Cuba and Mauritius have relatively low vulnerability, for example, compared with that of Haiti, Fiji and the Solomon Islands.
Countries assessed as relatively highly vulnerable to flooding include Venezuela, Morocco, Somalia and Botswana. Argentina and Germany on the other hand have historically experienced relatively low mortality during flood events.
By focusing on relative differences in disaster-related mortality at various levels of hazard exposure and by examining potential causal factors, the Reducing Disaster Risk report seeks to draw attention to the role that development decisions play in creating or reducing risk. It also suggests particular countries in which greater attention to risk management is warranted.
(b) Disaster risk hotspots
The Disaster Risk Hotspots project was a collaboration between Columbia University, the World Bank and a host of partner organizations. For cyclones, drought, earthquakes, floods, landslides and volcanoes, the report assesses global risks of disaster-related mortality and economic losses (Dilley et al. 2005). The intent is to identify relative differences in the risks of these outcomes at subnational scales as a guide for further risk management analysis to inform development policy and program decision-making.
Mortality and economic loss risks for each hazard are assessed for each 5×5 km cell on the earth's surface, excluding those with small populations and/or low agricultural production. The project identifies relative risk levels for each hazard and all hazards together. Risks are calculated based on the estimated hazard exposure of people and Gross Domestic Product (GDP) in each grid cell. Vulnerability is represented by historical loss rates for each hazard stratified by region and four country wealth status categories. Probabilities of occurrence, rather than event frequencies, are used for low-frequency hazards such as earthquakes and landslides. The report emphasizes the theory, data, methods, results and applications of the analysis.
Disaster-related mortality risks associated with hydro-meteorological hazards is found to be highest across the subtropical zones, with drought-related mortality risks being highest in semi-arid regions of Africa. Mortality risks associated with geophysical hazards are highest along plate boundaries, around the Pacific rim and across southern Asia. Some countries such as the Philippines and Indonesia are at high risk from all types of hazards. In general, mortality risks tend to be highest in developing or middle income countries.
The Hotspots project also evaluated total economic loss risks which, compared with mortality risks, are more concentrated in wealthier countries. Significant areas of Europe, North America and Southern and Eastern Asia are total economic loss risk hotspots. Large areas of the caucuses region are at relatively high risk from the full range of geophysical and hydrometeorological hazards.
When risks are evaluated for economic losses in proportion to GDP per unit area, the largest risk burden shifts back to the developing world. Coastal zones and large areas of Africa and Central Asia are at relatively highest risk by this measure. These areas and areas of relatively high risks of disaster-related mortality, will be hard pressed to develop economically due to recurrent losses unless disaster risks are reduced and/or transferred.
The Hotspots report focuses on risk patterns and levels more than on how risk is created or reduced through development. It also includes an analysis of the implications of the results for practical applications such as the development of country assistance strategies, development projects and recovery plans.
(c) Indicators of disaster risk and risk management for the Americas
Additional information about risk patterns is available for a set of countries in Latin America and the Caribbean, through an analysis by the Inter-American Development Bank and the National University of Colombia in Manizales (Cardona 2005; IDEA 2005). This project assessed different aspects of risk and risk management capacity with composite indicators incorporating a variety of variables selected by experts. Each indicator is a complex combination of sometimes dozens of constituent variables. Countries are the unit of analysis.
Similarly to the DRI in Reducing Disaster Risk, the indicators are designed to generate knowledge and awareness within the Inter-American Development Bank and governments concerning the importance of disaster risk management for development and to provide tools for risk management decision making. The report discusses methods, results and interpretation of the composite indexes.
One indicator, the Disaster Deficit Index, concerns the financial capacity of countries to absorb the economic losses and recovery costs of a major disaster. Values greater than one indicate that the country would not be able to meet the financial demands generated by a maximum probable loss. The maximum probable loss over the next century exceeds current financial resources in several countries—including Peru, Chile and the Dominican Republic—several times over.
Another indicator, the Risk Management Index (RMI), evaluates risk management performance in the areas of risk identification, risk reduction, disaster management and governance and financial protection. Countries with higher values are judged to be more able to manage disaster risks. These include Chile, Costa Rica, Jamaica and Mexico. Countries with low-RMI values include the Dominican Republic, Ecuador and Argentina.
A third indicator measures vulnerability based on factors related to exposure and susceptibility, social fragility and financial protection. A fourth measures the degree of spatial concentration of risks within each country.
The indicators and the variables they comprise offer decision makers the ability to diagnose specific areas in which risk management policies and practices can be improved. They also allow comparison among countries. Over time, attention to the factors on which the indicators are based could result in measurable improvement in the indicator values.
(d) Complementarities and areas for improvement
Notable differences between Hotspots and the DRI include the use of earthquake probabilities rather than event frequencies in Hotspots for calculating earthquake-related risks and the use of historical losses in Hotspots to calculate vulnerability coefficients rather than for selecting socio-economic dependent variables in multiple regression equations fitted to maximize the correlation between predicted and observed losses across countries. The lengthy examination of the relationship between disasters and development in Reducing Disaster Risk, on the other hand, provides insights into the role of development in reducing or creating risk.
The Indicators of Disaster Risk and Risk Management for Latin America and the Caribbean focus on specific aspects of risk—such as countries' abilities to respond financially to large losses and their risk management capacity—and demonstrate that a much richer data set can be employed when focusing on a relatively small number of countries. The results provide benchmarks for specific areas of risk management performance.
As is perhaps inherent in all risk analyses, the three studies tend to be heavily weighted by past experience. All of them, for example, preceded the Indian Ocean tsunami and do not include analyses of tsunami-related risks. Unprecedented events and the dynamic nature of risk necessitate continual revision of risk analyses to incorporate emerging threats and developments. Climate change as well as other environmental and socio-economic changes will result in new patterns of risk in the future.
The final measure of these studies and others like them, is the extent to which they effectively inform risk management decision making and lead to reduced losses. This goal is advanced if: (i) analyses are tailored to particular decision contexts and (ii) the evidence base for each decision context is based on adequate data and that justifies relevant conclusions.
The following two sections address both of the above points. The first outlines ways in which risk analyses can be technically improved, focusing on national and subnational scales. The second suggests broad decision-making domains that risk analyses at these scales may seek to support.
3. Loss and risk identification: some priorities
This section outlines areas of work for advancing the state-of-the-art with respect to assessment of disaster risks and losses, particularly at national and sub-national scales, with an eye towards applications. Although work by a wide spectrum of organizations is ongoing in these areas, greater synergies could be obtained by linking efforts more systematically.6
(a) Information on disaster losses: data and applications
Information on disaster losses plays a crucial role in disaster impact assessment and risk identification and therefore in disaster and risk management. Organizing the collection and use of this information to serve multiple purposes adds value (figure 1). The first area of use is for disaster management (box ‘I’ in figure 1). Over time, data on accumulated losses can be captured in loss databases containing essential information on each disaster (box ‘II’). This information on accumulated losses—and the costs of relief, recovery and reconstruction—document the impact of disasters on development. As the analyses reviewed above attest, loss data is also necessary (although not sufficient) for assessing risks of future disasters (box ‘III’). Coming full circle, risk information in turn is useful for contingency planning and disaster preparedness as well as for prioritizing measures to reduce losses over time, including after disasters. The following section briefly sketches priorities for continued attention within each area.
When a disaster occurs, the response can be divided into three phases: relief, recovery and reconstruction (figure 1, box I). Each phase involves impact (damage and loss) assessments that supports key decisions. In the relief phase, assessment data includes how many people were killed, injured or displaced and their degree of access to basic requirements—water, food, shelter and medical care. This information is used to estimate needs for relief supplies. Recovery involves the initial restoration of the functioning of life-supporting sectors, the economy, government and society. Damage and loss data, combined with an analysis of political, livelihood and social systems, supports recovery planning. Data on the economic value of damages and losses is used as the basis for reconstruction planning and for estimating the financial requirements for reconstruction.
Methodologies are in place or under development for obtaining the data needed to support decisions each of these phases.7 More could done to integrate data collection, compilation and analysis across phases, however. Currently, damage assessments are often performed multiple times by multiple organizations for multiple uses and yet losses tend to be inadequately captured by any one assessment. A better approach would be to have a comprehensive initial assessment of damages and losses, including their spatial distribution, followed by ongoing monitoring of the relief, recovery and reconstruction process.
A second area for improvement is to seek more systematic capture of event-by-event losses in national or subnational level disaster databases (figure 1, box II). Disaster-prone countries, by definition, experience recurrent disasters. Disaster databases document cumulative disaster histories, facilitating identification of spatial patterns and temporal trends in losses. They document the cost of disasters to development and the allocation of these costs across economic sectors and socio-economic classes. Loss trends provide important indicators for monitoring the successfulness of disaster reduction efforts. Currently, loss assessment data is generally captured in disaster databases in an ad hoc manner. That is, incorporation of loss assessment results in databases is typically not considered an integral part of the disaster assessment reporting process.
More systematic incorporation of data on assessed losses and the costs of responding to disasters into disaster databases, particularly at national scales, would constitute a significant step forward. Greater capacity is needed in disaster-prone countries, however, to collect and maintain this type of data on a sustained basis. Specific areas of work to improve the global database on disaster losses include the more widespread establishment of sustainable national and local disaster databases, the improvement of their contents and the linking of data on each event across disaster and hazard databases to facilitate analysis.
Even existing disaster databases are currently not well integrated. Existing loss databases include CRED's EM-DAT; NatCat, maintained by Munich Reinsurance; Swiss Re's Sigma database; and DesInventar, a national-level database system in widespread use in Latin America and the Caribbean developed by La Red.
One convention useful for building and maintaining disaster databases is the GLobal disaster IDEntifier (GLIDE).8 The GLIDE was conceived by the Asian Disaster Reduction Center in Kobe, Japan. It consists of a unique number for each disaster, incorporating a two-letter hazard code, a unique sequence number and geo-location codes based in ISO standards. The location codes are hierarchically formatted to allow data to be disaggregated or aggregated regionally, nationally or subnationally. Use of the GLIDE allows historical events to be unambiguously identified. This facilitates verifying data on a single event across multiple sources, linking hazard event data to disaster data and referencing reports, evaluations and other material about the disaster.
The GLIDE creates the potential for a global, multi-tiered, decentralized tracking system for disaster losses, in which data on losses, hazards and other information from multiple sources can be virtually linked. Incorporating the GLIDE into existing and emerging loss data bases, cross-linking entries in these databases and linking disaster to hazard event data would constitute a major but worthwhile undertaking.
Accumulated information on historical losses is crucial for understanding disaster risk patterns. While historical losses may not completely reflect future losses, information on historical losses is necessary for corroborating results of risk analyses based on hazard exposure and vulnerability. Applications of risk information—including the setting of risk reduction priorities, identification of risk reduction measures, the design of risk transfer schemes, disaster preparedness and contingency planning—are discussed further below.
Loss data therefore contributes towards reducing losses in two ways: (i) as an important input into risk analyses and ex ante risk management and (ii) (coming full circle in figure 1) it contributes information on disaster risks that is crucial for managing the early recovery process following disasters. The latter is a particularly important application. Post-disaster recovery is a window of opportunity for better rebuilding using recovery and reconstruction funding. The rush to rebuild following a disaster, however, can result in the reconstruction of the same risks factors that led to the disaster in the first place. Information on historical losses provides evidence that can be used in plans to reduce risks, including from multiple hazards, during reconstruction.
(b) Enhancement and integration of risk information at multiple scales
As was noted earlier, the most crucial risk management decisions are those made at the national and local levels. Improved loss data is but one aspect of an overall effort to begin systematically to develop more in-depth information on risk levels and risk factors at larger (e.g. subnational) scales to inform risk management efforts in high-risk areas.
Adherence to some minimal guidelines would contribute to improving the global body of knowledge about risk assessments. Objectives of such guidelines include (i) a minimum standard of quality and compatibility and (ii) the integration of improved data sets and results across scales as well. Guidelines promoting these objectives include:
(i) The outcome of which the risks are being evaluated is specified. Statements about risk are statements about potential outcomes. The particular outcome of which the risks are being evaluated should be explicitly stated. Examples include the risks of mortality, economic loss, structural damage, livelihood or income losses, agricultural losses and so on. The degree to which the outcome has been experienced should be observable and measurable in the real world.
(ii) The analysis is theoretically sound and identifies causal factors. Disaster risks are a function of two sets of causal factors. The first is the degree to which a set of identified socio-economically valuable assets is exposed to natural hazards. The second is the vulnerability of those assets to the hazards to which they are exposed. Example assets, or ‘elements at risk,’ include but are not limited to people, infrastructure, economic activities and economically important land uses. The analysis should be built on this theory of causality and allow attribution of causality to specific hazard exposure or vulnerability factors. The data and methods should be adequate to support the conclusions, verified through a peer review process.
(iii) There is a clear connection between the analysis and risk management decision making. Evidence on disaster risks and risk factors is fundamental for reducing disaster losses. Such evidence will only contribute to reducing losses, however, if it is connected to risk management policy and decision making. Different decision contexts require different types of evidence. Sectoral policies and projects require evidence regarding potential hazard impacts within specific sectors. National-level development planning has different information requirements than local planning. Decision makers and the decision alternatives to be informed by the analysis constitute key stakeholders and should be explicitly acknowledged in the design of the analysis and its outputs.
(iv) Risk analyses are done by local experts within an appropriate institutional context. Risk assessments by outside experts do not significantly contribute to capacity development and the results are unsustainable. Risk identification should be led by local experts, supported with data, technical and financial assistance as appropriate. Ideally the lead actors will be from an appropriate institution or institutions with a mandate, capacity and credibility in the decision context the analysis seeks to inform. For many reasons it is important that local institutions have the primary responsibility for generating and maintaining risk assessment products. These include the need to incorporate local knowledge, sustainability, ownership, maintenance and cost.
Adherence to the above guidelines would contribute to a degree of compatability among risk analyses and create the potential for accumulation of results into an overall body of knowledge about risks globally. One means of supporting the process of compiling and enhancing risk information across multiple scales would be to create a global risk information platform into which data at multiple scales can be integrated. Within such a platform, the higher the risk levels the more desirable it is to have data and results at the highest possible resolution.
An integrated risk information system could support the regular release of a global report on the state of disaster risk. The report could regularly update the status of risk globally, built insofar as possible on local-level analyses adhering to the above guidelines, focusing in particular on high-risk areas. Routinely generating a global risk report would be facilitated by having a GIS-based, web-accessible platform into which data and results at different scales could be incorporated. As risk patterns change or new data and results become available, revised results could be generated and interpreted for priority setting and decision making.
A later phase could begin to incorporate dynamic risk assessment and predictions—continuing recalculation of risks based on such things as current storm tracks, cumulative and projected rainfall anomalies, volcanic activity etc. Hydrometeorological hazard probabilities in particular vary seasonally and inter-annually in ways that are increasingly well understood.
(c) Knowledge and tools for capacity development
Systematic progress in both of the areas described previously—improved global loss and risk information—would be supported and enhanced by compiling, reviewing and promoting the adoption of relevant knowledge, tools and capacity through international collaboration and networking. Documenting and updating methodologies and results, good practices and lessons learned would assist practitioners in keeping abreast of, and advancing, the state-of-the-art. This would contribute to the continuing formation of communities of practice in loss and risk analysis and decision-support applications.
4. Risk management decision support
The efficacy of risk and loss information and the utility of developing greater risk identification capacity, depends not only on the quality of the analyses but also on the appropriateness of the resulting evidence base for decision making. This raises the question of what strategy, or combination of strategies, can cost-effectively bring about actual reductions in disaster-related losses in high-risk countries. In other words, what combination of interventions is necessary, and sufficient, to reduce disaster losses? The intent of the following section is not to answer the question of what mix of strategies is optimal, but rather to begin to identify a set of decision-making domains for further consideration.
One set of decision contexts focuses on promoting disaster reduction in advance of disasters. There are many ways in which one could organize pre-disaster strategies; here I focus on five.
(i) The first is to incorporate disaster reduction as part of the overall national development strategy. Vehicles include Poverty Reduction Strategy Papers, Country Assistance Strategies and United Nations Development Assistance Framework, Common Country Assessments, which set development priorities over multiple years by agreement with national governments. These strategies too often do not acknowledge disaster risks, even in high-risk settings. And they are often set back when a disaster happens, leading to losses due to risk factors that were unforeseen and therefore unaddressed by the strategy. Development strategies in high-risk countries should take account of disaster risks and ensure that risk reduction is part of development—not just for the sake of reducing losses but also for the sake of achieving development goals.
(ii) On the often-voiced assumption that ultimately the most local level is the one at which the most significant risk management decisions are made, another complementary strategy is to ‘downscale’ risk analyses to the local level. Localized risk assessments target local planning processes, ranging in scale from urban planning to community-based risk management. Localized risk assessments also offer greater opportunities for participatory processes involving wider cross-sections of society. Such participation is key in the identification and implementation of local risk management solutions.
(iii) A third strategy is to focus on reducing disaster risks in the achievement of specific development goals such as poverty reduction or food security. Work in these two areas, for example, would focus on the impacts of disasters on livelihoods. In areas with frequent and recurrent disasters—which include many areas exposed to climate variability and hydrometeorological hazards—recurrent losses drain away both assets and any incentive to assume additional risks in pursuit of livelihood gains. In high-risk areas, policies and programs are needed to protect livelihoods and provide the poor with both the means and the incentives to invest in their future with less risk that these investments will be lost. This could help spring the poor out of the ‘poverty trap’ in cases where part of the trap is due to recurrent hazard-related losses.
(iv) Another entry point is the targeting of specific vulnerable sectors in hazard-prone areas. Different hazards affect different sectors differently. Identifying the hazards that are present in a particular location and understanding their sectoral impacts provides a means of protecting affected sectors through sector-specific policies and projects. Sectoral analyses, policies and projects are often the main means by which development strategies are implemented. Therefore, sector analytical work should include an assessment of natural hazard-related risks in the sector. Incorporating measures to address sector-specific risk factors through sectoral policies and projects constitutes a major entry point for reducing those risks.
(v) Lastly, risk analysis can be undertaken in support of specific risk reduction measures. Examples include early warning systems, appropriate building codes and risk transfer schemes. The latter transfer financial risks away from at-risk populations and areas as part of a comprehensive risk management strategy. Risk transfer is already well established in the developed world, for example through insurance and commodity hedge markets. Experiments by the World Bank,9 the World Food Program10 and others, are currently continuing to determine if and how risk transfer mechanisms such as weather index insurance can be used to protect poor people from shouldering the entire disaster risk burden themselves. Risk transfer mechanisms offer a potential means of putting a floor under the bottom rung of the development ladder.
(i) Risk reduction during disaster recovery. Taking advantage of the window of opportunity for reducing risks during recovery following disasters benefits substantially from having put in place measures ahead of time that can be rolled out during recovery and reconstruction. Otherwise, the imperative to rapidly restore essential systems may lead to reconstruction of the same patterns of exposure and vulnerability that lead to the disaster in the first place, or worse.
Risk management in the aftermath of disasters becomes more feasible if vulnerability-reducing measures can be pre-identified and prepared prior to disasters to facilitate their adoption from the very initial stages of recovery. This approach is similar to the preventive strategies above, except that the expectation is that implementation will be undertaken rapidly in the period following a major disaster.
Although the logic behind this strategy may seem perverse, it acknowledges the reality that the greatest attention, and resources, to deal with disaster risks are often available in sufficient quantities only after one happens. This strategy, which is essentially one of response preparedness, also works well in combination with relief, recovery and reconstruction contingency planning.
(ii) Contingency planning for improved disaster response. One particular preparedness activity is contingency planning. Identification of high-risk areas creates the possibility of developing scenarios of likely future disasters. Such scenarios can be used to refine response roles and responsibilities, pre-allocate resources and work out logistics in advance of potential disasters. Contingency plans lead to more effective, quicker responses. Risk information allows contingency planning and preparedness efforts to be more systematically prioritized. The potential for such prioritization based on emerging evidence of disaster risk patterns has yet to be fully and systematically explored, neither at the national level in high-risk countries nor within the international humanitarian assistance community.
5. Conclusion: foreseeable risk
The vision presented in this paper—of a systematic project to supply the evidence base for reducing losses and supporting sustainable development—is not just a pipe dream. Although the outline above is necessarily presented in broad and general terms, there are clear signs of an emerging consensus among disaster reduction stakeholders of the importance of risk information.11 Furthermore, continuing networking, dialogue and coordination among loss and risk assessment professionals is contributing to a sense that the vision outlined above could quite conceivably become a reality.
A more systematic global effort to catalogue losses and identify disaster risk levels and causal factors would promote a shift in emphasis from emergency management to risk management. Making disaster risk factors more visible creates incentives for acting on the root causes of disasters rather than the consequences. Reduction of recurrent losses in high-risk areas would increase the potential for secure, sustainable development.
↵1 Thanks to two anonymous reviewers for highlighting the need to underscore the relationship between analysis and application among other helpful suggestions. The views expressed and any errors of fact or interpretation are the author's alone.
One contribution of 20 to a Discussion Meeting Issue ‘Extreme natural hazards’.
↵http://idea.manizales.unal.edu.co/proyectosespeciales/adminidea/centrodocumentacion/docdigitales/documentos/summary%20report%20idb.pdf and http://idea.manizales.unal.edu.co/proyectosespeciales/adminidea/centrodocumentacion/docdigitales/documentos/maintechnicalreportidea1.pdf.
↵A multi-stakeholder Global Risk Identification Program (GRIP) is currently in preparation for supporting this process, organized by UNDP and the ProVention Consortium (see www.proventionconsortium.org/projects.htm).
↵See the Hyogo Framework for Action (http://www.unisdr.org/eng/hfa/hfa.htm), in which risk identification is one of five priority areas for achieving substantial reduction in disaster losses globally.
- © 2006 The Royal Society